Peripheral nervous system origin of phantom limb pain
, Haim-Moshe Adahan
, Artan Gjika
, Skerdi Zahaj
, Teﬁk Zhurda
, Gentian Vyshka
Pain and Rehabilitation Clinic, National Trauma Center, Trauma University Hospital and Galenus Clinic, Tirana, Albania
Pain Rehabilitation Unit, Chaim Sheba Medical Center, Tel Hashomer 52621, Israel
Biomedical and Experimental Department, Faculty of Medicine, University of Medicine, Tirana, Albania
Department of Cell and Developmental Biology, Institute of Life Sciences and Center for Research on Pain, The Hebrew University of Jerusalem, Jerusalem 91904, Israel
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Received 31 December 2013
Received in revised form 6 April 2014
Accepted 14 April 2014
Phantom limb pain
Nearly all amputees continue to feel their missing limb as if it still existed, and many experience chronic
phantom limb pain (PLP). What is the origin of these sensations? There is currently a broad consensus
among investigators that PLP is a top-down phenomenon, triggered by loss of sensory input and caused
by maladaptive cortical plasticity. We tested the alternative hypothesis that PLP is primarily a bottom-up
process, due not to the loss of input but rather to exaggerated input, generated ectopically in axotomized
primary afferent neurons in the dorsal root ganglia (DRGs) that used to innervate the limb. In 31 amputees,
the local anesthetic lidocaine was applied intrathecally and/or to the DRG surface (intraforaminal epidural
block). This rapidly and reversibly extinguished PLP and also nonpainful phantom limb sensation (npPLS).
Control injections were ineffective. For intraforaminal block, the effect was topographically appropriate.
The suppression of PLP and npPLS could also be demonstrated using dilute lidocaine concentrations that
are sufﬁcient to suppress DRG ectopia but not to block the propagation of impulses generated further dis-
tally in the nerve. PLP is driven primarily by activity generated within the DRG. We recommend the DRG as a
target for treatment of PLP and perhaps also other types of regional neuropathic pain.
Ó2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
The origin of phantom limb pain (PLP) remains uncertain. Reli-
gious and psychiatric interpretations once predominated [54,58],
but these have since been supplanted by neurobiological and cog-
nitive theories. The fact that pressure on amputation stump neuro-
mas provokes PLP (Tinel sign), and the discovery that neuromas
generate ectopic impulse discharge (ectopia), favored the stump
as the pain generator [5,14,29,49,50,55,56,63]. However, PLP fre-
quently persists despite neuroma inﬁltration and nerve/plexus
block [4,27,46]. For this reason most investigators have abandoned
peripheral nervous system (PNS) explanations in favor of the
hypothesis that PLP is a consequence of maladaptive cortical plas-
ticity induced by loss of input from the limb [1,23,28,39,46,48].
The cortical origin of PLP has considerable empirical support.
For example, limb amputation or corresponding nerve injury leads
to conspicuous neuroplastic remapping of somatotopic representa-
tions in the primary somatosensory cortex (S1) [16,21,24,25,31,32,
53,66], with the extent of remapping proportional to the intensity
of the pain . Likewise, distortions in body schema perception
occur when conﬂict is induced experimentally between the
appearance of an individual’s limb and proprioceptive feedback.
In the rubber hand illusion, for example, the perceptual integration
of the rubber hand is so striking that threatening it with injury
evokes anxiety and pain affect–related cortical activations .
Some subjects report unpleasant sensations, perhaps even pain,
due to such sensory–sensory mismatch . Resolving this mis-
match, as implemented in mirror box therapy, can relieve PLP, at
least temporarily [48,53].
However, a second PNS source, outside of the stump, has never
been adequately considered. For decades there has been direct
electrophysiological evidence that afferent somata in the dorsal
root ganglia (DRGs) also generate ectopia [33,37,52,62]. Indeed,
in head-to-head comparisons, the DRG has proved to be a more
robust source of spontaneous ﬁring than neuromas [2,42].
Evidence, if indirect, is even available in humans [38,40,49,50].
For example, Nystrom and Hagbarth  showed that blocking
0304-3959/Ó2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Corresponding author. Address: Department of Cell & Developmental Biology,
Institute of Life Sciences, The Hebrew University of Jerusalem, Jerusalem 91904,
Israel. Tel.: +972 2 6585085; fax: +972 2 6586027.
E-mail address: firstname.lastname@example.org (M. Devor).
155 (2014) 1384–1391
stump neuromas eliminated the percussion-evoked Tinel sign and
associated spike activity, but not the ongoing discharge recorded in
the nerve. This likely originated in the DRG. DRG electrogenesis
could account for the therapeutic failure of neuroma, nerve, and
plexus inﬁltration because these distal blocks do not affect the
Because DRGs share the same cerebrospinal ﬂuid compartment
as the spinal cord, spinal blocks and intraforaminal blocks both
have the potential to arrest all PNS ectopia: stump and DRG. We
are unaware of any systematic reports on effects of either type of
block on PLP. However, spinal block is frequently used in stump
revision surgery, and practitioners we have consulted attest that
this indeed transiently stops PLP (R. Boas and A. Stav, personal
communications). Paradoxically, case studies have reported tran-
sient rekindling of quiescent PLP after spinal block, but this is rare
. A likely explanation is that the injectate used transiently
excited DRG neurons, or the spinal neurons they drive, by a
mechanical, thermal, or chemical mechanism (rapid injection
of large volumes in a restricted space, cold solution, inaccurate
pH/osmolarity, or preservatives). Here we used diagnostic spinal
and intraforaminal blocks in human amputees to determine
whether preventing central nervous system (CNS) access of ectopic
signals generated in the DRG might affect PLP and/or nonpainful
phantom limb sensations (npPLS).
2.1. Subjects, experimental design, and rationale
We report results of 4 related procedures intended to block the
access of nerve impulse discharge originating in the PNS from
reaching the brain. These are represented in 4 experimental
groups. In group 1, our primary focus, we tested effects of blocking
abnormal afferent input by epidural intraforaminal injection. In
group 2, for comparison, we also examined spinal (intrathecal)
block. In a few cases (group 3), local inﬁltration of stump neuromas
or peripheral nerve block was performed. Procedures were carried
out in 2 centers located in regions that have known recent military
conﬂict and that serve relevant patient populations; staff at such
facilities are acutely aware of the limits of current treatment and
encouraged the introduction of better therapeutic options. At
the Trauma University Hospital and the associated Galenus Clinic
(Tirana, Albania), we treated 16 lower limb amputees with ongoing
PLP (11 men, 5 women). These participated in experimental groups
1 to 3, where some of the amputees participated, on separate occa-
sions, in 2 or 3 of the groups. Each of the 16 individuals is identiﬁed
by a unique number to facilitate tracking who underwent which
procedure. Finally, group 4 comprised an additional 15 amputees
(14 men and 1 woman) who were treated with a modiﬁed protocol
of intraforaminal injection at the Pain Rehabilitation Unit, Chaim
Sheba Medical Center (Tel Hashomer [Tel Aviv], Israel).
Inclusion criteria were age >18 years, good general health, abil-
ity to communicate and understand instructions, and presence of
signiﬁcant PLP with a frequency and intensity that interfered with
daily life. Subjects were excluded if they had signiﬁcant sensory
deﬁcits, major pain complaints other than PLP (including severe
stump pain, which might have distracted from their ability to
report on their PLP), major CNS or PNS neurological disorders other
than diabetic polyneuropathy and trauma associated with the
cause of amputation, major cognitive or psychiatric disorders, or
contraindication to the injection of lidocaine, corticosteroids, or
Subject background and demographic information is provided
in Tables 1 and 2. Experimental protocols were approved by
authorities on human experimentation (Helsinki committees) at
Most subjects had experienced traumatic amputation; in
Tirana, it was frequently from stepping on a land mine. Some
amputations, however, were due to vascular insufﬁciency or other
causes. PLP tends to be similar regardless of the precipitating
pathology . The objectives and risks of the blocks were
explained to the subjects in their language, including the fact that
treatment results may have no effect on PLP, may produce partial
and reversible analgesia, or may yield more prolonged pain relief.
Informed consent was obtained. We then initiated a protocol that
was standardized but subject to minor variations depending on
the individual patient. First, a history was taken, and the present
quality and location of PLP and npPLS was documented by text,
photos, body charts, and sketches. Information on the circum-
stances of amputation, frequency and duration of PLP, changes
over time, and exacerbating and relieving factors was also noted.
Special care was taken to ensure that subjects fully understood
the difference between sensations experienced in the phantom
limb (PLP and npPLS) and those experienced in the stump.
The amputation stump was then systematically examined, and
tender points and points at which a Tinel sign could be evoked by
palpation or percussion were marked on the skin. Finally, subjects
were prepared for injections. No sedation was used so that subjects
Subject demographics, baseline pain, and results of spinal (intrathecal) block.
interval since amputation
Baseline phantom, effect of percussion
over stump neuromas (Tinel ?), (notes)
Level Effect of spinal block on
phantom and Tinel
1 M/61 R AKA, diabetes, 30 y PLP lateral foot (severe), npPLS leg below knee, Tinel ?PLP L3–4 PLP, npPLS and Tinel lost, recovery
after >3 h
2 F/40 AKA bilateral, trauma,
bilateral PLP, bilateral npPLS (numbness, sensation of movement),
Tinel ?stump pain (‘‘electric’’)
L3–4 PLP lost, npPLS and Tinel persists,
3 F/65 BKA, scleroderma, 7 days PLP, npPLS, Tinel ?stump pain L3–4 PLP, npPLS and Tinel lost
4 M/52 L AKA, trauma, 3 y, R AKA,
vascular, 1 y
L PLP (modest ‘‘shooting’’), R PLP (severe, ‘‘pulsing’’), npPLS bilaterally,
Tinel ?stump pain
L3–4 PLP, npPLS and Tinel lost
5 F/24 R hip disarticulation,
trauma, 2 y
PLP (severe), npPLS (knee to foot), Tinel ?PLP L3–4 PLP, npPLS and Tinel lost
6 M/61 R AKA, vascular, 5 d PLP (‘‘electric’’), npPLS, Tinel ?PLP L2–3 PLP, npPLS and Tinel lost
7 M/48 R AKA, trauma, 10 y PLP, npPLS, stump (itch + burning), Tinel ?PLP (lateral toes) L4–5 PLP, npPLS and Tinel lost. Stump
8 M/22 R lateral foot (toes 2–5),
trauma, 9 y
PLP (toe 5), npPLS, Tinel ?stump pain, scar ‘‘cold’’ L4–5 PLP, npPLS and Tinel lost
9 M/24 R BKA, trauma, 10 y PLP (toes 4, 5), npPLS, Tinel ?PLP, ongoing stump pain L4–5 PLP;, npPLP and Tinel lost
10 M/39 R BKA, trauma, 10 y PLP, Tinel ?PLP + stump pain, ongoing stump pain (cold) L4–5 PLP, Tinel and stump pain lost
11 M/51 L foot, trauma, 10 y PLP (sole), npPLS (foot) Tinel ?stump pain L4–5 PLP, npPLS and Tinel lost
R, right; AKA, above knee amputation; PLP, phantom limb pain; npPLS, nonpainful phantom limb sensation; Tinel, evoked Tinel sign; BKA, below knee amputation; L, left.
A. Vaso et al. /PAIN
155 (2014) 1384–1391 1385
would remain fully alert and responsive. Just before beginning the
procedure, we asked the subjects to rate the intensity of their PLP
on a 10-point scale. They were told that this value would serve as a
baseline for assessing changes associated with subsequent injec-
tions. The rating used a numerical or a visual analog scale (VAS),
where 0 indicates no pain and 10 the worst pain imaginable. ‘‘Loss
of PLP’’ indicates a drop from the preinjection value to 0 (Tables 1
and 2). When loss of pain was incomplete, subjects were asked to
gauge the percentage reduction from the baseline value. In all
spinal and intraforaminal procedures, subjects were unaware of
the order of injections; they were blinded to the speciﬁc material
injected on each occasion; and in most instances, they were una-
ware of when exactly the injections were made. After completion
of the spinal or intraforaminal procedure, subjects rested for 1 to
3 h, during which time the status of their phantom and stump sen-
sations was periodically noted. They were then released. Because
of the inaccessibility of many of the subjects, the study focused
on the short-term effects of the blocks. We did not attempt system-
atic long-term follow-up, although sporadic feedback was obtained
from some subjects.
2.2. Spinal block (group 2)
Spinal block was carried out (in Tirana only) in 11 subjects. In 8
of them, it was followed up with an intraforaminal block within a
few days. The procedure, based on a routine protocol described by
Lund , involved intrathecal midline delivery of 2 mL lidocaine
(1% or 2%, Propharma, Tirana, Albania) within the spinal canal via
interspace L2–3, L3–4, or L4–5 (medial approach).
2.3. Intraforaminal block (groups 1 and 4)
The leg is innervated largely by DRG L4, L5, and S1, with les-
ser L3 and S2 contributions. When PLP was uniform across the
phantom foot, we began by targeting the L5 or S1 DRG. How-
ever, if pain predominated in 1 part of the phantom foot and/
or percussing neuromas evoked pain that was easily localized,
we began with the corresponding segment. Because of the sub-
stantial overlap between adjacent spinal dermatomes [26,35],
we anticipated that it would be necessary to block several adja-
cent segments. However, this proved necessary only occasionally.
Changes in the response to tapping stump neuromas (Tinel sign)
also provided information on the completeness of blocks. These
checks tend to be painful, however, and thus we used them with
There were minor differences in procedures at the 2 venues (ie,
between groups 1 and 4). In Tirana (group 1), injections were made
under radiographic guidance using a Somatom ARC CT (Siemens
AG, Munich, Germany). Subjects were placed in the prone position
and prepared as for spinal epidural injection. The intended
segmental level was identiﬁed using anatomic landmarks, and a
trajectory for targeting the intervertebral foramen on the side of
Subject demographics, baseline pain, and results of intraforaminal block.
Baseline phantom, effect of
percussion over stump
neuromas (Tinel ?), notes
Level Effect of foraminal block on... Notes
PLP npPLS Tinel ?
1 M/61 R AKA, diabetes, 30 y PLP lateral foot (severe), npPLS leg
below knee, Tinel ?PLP
L3 Lost Lost Lost "PLP provoked during
maintained during 5 d
4 M/52 L AKA, trauma, 3 y, R
AKA, vascular, 1 y
L PLP (modest ‘‘shooting’’), R PLP
(severe, ‘‘pulsing’’), npPLS bilaterally,
Tinel ?stump pain
R–L5 Lost Lost Lost "PLP and npPLS provoked
7 days later L–L5 Lost Lost Not certain
5 F/24 R hip disarticulation,
trauma, 2 y
PLP, npPLS knee to foot,
L4 ;90% ;90% Lost ‘‘Shadow’’ of phantom
7 M/48 R AKA, trauma,10 y PLP, npPLS, stump (itch + burning),
Tinel ?PLP (lateral toes)
L4 Lost No change Lost
8 M/22 R lateral foot (toes
2–5), trauma, 9 y
PLP (severe in toe 5), npPLS,
Tinel ?stump pain, scar ‘‘cold’’
L5 Lost Lost Lost
9 M/24 R BKA, trauma, 10 y PLP (toes 4, 5), npPLS, ongoing
L4 Lost Lost Lost
10 M/39 R BKA, trauma, 10 y PLP (‘‘pinching, like a very tight sock’’),
npPLS, Tinel ?PLP + stump pain,
ongoing stump pain (cold)
L5 Lost Quality
Lost PLP replaced with
11 M/51 L foot, trauma, 10 PLP (sole), npPLS (foot), Tinel ?
L5 Lost No change Not certain
12 F/55 R BKA, trauma, 17 y PLP (foot only), npPLS
(foot only), Tinel ?stump pain
L4 Lost (?
;60% No change Foot telescoped to stump,
can be moved
Next day L5 Not
13 M/55 L BKA, trauma, 11 y PLP, npPLS (‘‘tingling’’), Tinel ?PLP
(in toe 1)
L5 ;60% Lost ;50% Foot telescoped to stump,
toes can be moved.
14 M/57 R foot, trauma, 11 y PLP (toe 1 ‘‘bound’’), npPLS
Tinel ?PLP (all toes, ‘‘electric’’)
L5 Lost Only
Foot telescoped to stump,
can be moved
Soon after L5 L4 Still
Lost To lateral
15 M/52 L at knee, diabetes, 45 d PLP (toe 1 and ankle), npPLS
(whole leg), Tinel ?stump pain
L4 Lost Lost Lost Result maintained during
12 d infusion
16 F/77 L medial toe (toe 1),
diabetes, 17 d
PLP (‘‘sharp’’), npPLS, Tinel ?
L5 Lost Not certain Lost Result maintained during
10 d infusion
R, right; AKA, above knee amputation; PLP, phantom limb pain; npPLS, nonpainful phantom limb sensation; Tinel, evoked Tinel sign; BKA, below knee amputation; L, left.
1386 A. Vaso et al./ PAIN
155 (2014) 1384–1391
the amputation was chosen and inﬁltrated with lidocaine (Pro-
pharma). An 18-gauge Tuohy epidural needle with plastic obtura-
tor (Portex, Smiths Medical International, Ashford, UK) was
inserted and guided into the foramen. The obturator was then
removed and an injection syringe was attached to the needle. Dur-
ing needle insertion, subjects were encouraged to report on sensa-
tions felt in the phantom limb or the stump. We took advantage of
the fact that 3 of the patients (patients 1, 15, and 16), 2 with recent
amputations, were hospitalized for severe pain and stump issues.
In these, a polyethylene catheter was inserted to the tip of the nee-
dle, and the needle was then withdrawn. This permitted repeated
lidocaine injections over subsequent days. Once tip placement
was satisfactory, we slowly injected 2 mL saline followed by
0.5 mL contrast medium (Ultravist300 diluted 1:2 with saline;
Iopromide, preservative-free, Bayer Schering Pharma, Berlin, Ger-
many). This was followed by lidocaine (2 mL, 1% or 2%). The saline
was used to open the intraforaminal space so as to reduce central
spread of lidocaine, and the contrast was used to monitor lidocaine
coverage of the DRG and its spread beyond.
In lieu of carrying out full nontherapeutic dummy procedures,
we controlled for potential effects of the patients’ anticipation,
solicitousness, and placebo response as follows. Just before each
injection, one of a set of messages, varied randomly, was deliv-
ered to the subject. For example, the subject might be told that
a solution (without specifying which) was about to be injected.
Alternatively, the subject was explicitly told that the impending
injection was expected to relieve PLP, and then the bolus was
either injected or withheld. Most often the injection was covert;
the subject was not informed of the actual timing or type of the
injection . To minimize potential experimenter bias, subjects
were encouraged to report, on their own initiative, changes in
phantom sensation. However, if more than about 10 min passed
without a self-initiated report, a prompt was given, especially if
the time for an injection was approaching. We avoided prompt-
ing subjects for a report on sensory changes in the ﬁrst few min-
utes after injections. If a subject expressed uncertainty as to
whether a change had occurred, he or she was encouraged to
decide, but was not pressed.
We made special note of changes, if any, reported within 2 to
5 min after the delivery of a message without actual injection
and after injections of saline, contrast, and lidocaine. The observa-
tion time was extended to at least 10 min when lidocaine was
delivered. In early trials, many subjects reported a transient cold
sensation in the lower back after administration of verum injec-
tions. This alerted us to prewarm all solutions to 37°C, which
eliminated such reports. Finally, a bolus of dexamethasone (4 mg,
1.0 mL; Propharma) was injected in order to enhance the patient’s
chances of obtaining extended pain relief. The needle was
then withdrawn. When a second intraforaminal block was car-
ried out at another level, we proceeded immediately using the
identical protocol. The procedure typically took between 30 and
In the Tel Hashomer (group 4), the technique differed some-
what, as follows. Imaging for intraforaminal needle insertion used
ﬂuoroscopy rather than computed tomography (OEC 9900 Elite
C-arm ﬂuoroscope; GE Healthcare, Hatﬁeld, UK). Tip location was
conﬁrmed using 1 to 2 mL contrast (Iopamiro 370 without preser-
vative; Dexon Pharma, Or-Akiva, Israel). To enhance the therapeu-
tic effect, in most cases, both the L5 and S1 ganglia were injected,
with the initial block usually directed to S1. The injectates were
mixed rather than being provided in separate boluses. Speciﬁcally,
at each level, we injected 3 mL of a mixture containing lidocaine 1%
(Esracaine 1 mL; Rafa Laboratories, Jerusalem, Israel), Iopamiro 370
(1 mL), and methylprednisolone acetate (Depo-Medrol 40 mg,
1 mL, without preservative, Pﬁzer, New York, NY, USA). Results
are presented as a case series.
2.4. Nerve block (experimental group 3)
In 3 amputees in Tirana (patients 1, 12, and 13), we evaluated
the effects of inﬁltrating stump neuromas by administering lido-
caine and/or by carrying out femoral or sciatic nerve blocks. Effects
were followed for at least 30 min. Although these trials were not
systematic, we report results brieﬂy.
2.5. Statistical analysis
The proportion of subjects who reported criterion reduction in
PLP and/or npPLS was evaluated by the
or Fisher exact probabil-
ities tests (SigmaStat v3.1). Means ± SD are given. P6.05 was con-
3.1. Case description, intraforaminal block (amputee 14 in group 1)
The procedures and key outcomes, which were fairly uniform
across subjects, are illustrated by patient 14. Patient 14, an intelli-
gent and articulate 57-year-old man from Kosovo, experienced
traumatic amputation of the right foot above the ankle 11 years
previously when he stepped on a land mine. He had severe stump
pain for the ﬁrst few weeks after the injury and became aware of
his phantom foot only after about 5 weeks. When we saw him,
he described his usual sensation: the feeling of phantom toes
emerging from the end of the stump (‘‘foreshortening,’’ ‘‘telescop-
ing’’ ). Their position was natural but in forced pronation. The
big toe (toe 1) dominated the phantom and felt tightly constricted
(‘‘bound’’), with PLP rated as 5 to 6 on a scale of 0 to 10. The
remaining toes (toes 2 to 5) were also felt, but they were not pain-
ful (npPLS, pain score = 0) and could be voluntarily moved laterally,
separating them from toe 1. Two sensitive stump neuromas were
identiﬁed. Pressing on the medial one evoked an electric shock–
like pain in the phantom toes, especially toe 1, and a noticeable
ﬂinch. Pressing on the lateral one evoked a local stabbing sensation
in the stump (stump pain). There was no obvious tactile allodynia,
but the subject reported that the stump felt cold (it was not objec-
An injection needle was placed in the L5–S1 intervertebral fora-
men under computed tomographic guidance, targeting the L5 DRG.
Then the subject was told, ‘‘We are about to make an injection’’ and
that he should report any changes felt. No injection was actually
delivered (sham injection), and he reported that he felt no change
in the phantom or stump. After 2 to 3 min, 2 mL saline at 37°C was
injected with no alert given, and this was followed by 0.5 mL con-
trast. No sensory change was reported, and both the medial and
lateral Tinel signs produced the usual responses. After an addi-
tional 5 min, 2 mL 1% lidocaine at 37°C was covertly injected, and
within 2 min, he volunteered that the painful constriction of the
big toe was gone except for the edge closest to the small toes,
which still felt pinched. The npPLS of toes 2 to 5 remained, but
the forced pronation relaxed, and he lost the feeling that he could
move the toes. Over the next few minutes all feeling of toe 1 was
lost, and both Tinel signs weakened markedly. The npPLS of toes
2 to 5 remained. At this point, 1.0 mL dexamethasone was injected.
There were no further sensory changes, and the needle was
removed. A second needle was then placed in the L4–5 interverte-
bral foramen near the L4 DRG.
About 30 min after the L5 DRG lidocaine injection, we were pre-
pared to proceed. At this point, the phantom remained as it was
after the L5 procedure (ie, residual npPLS of toes 2 to 5). The ﬁrst
step, a sham injection, evoked no change, nor did subsequent cov-
ert injection of 2 mL saline. Minor adjustment of the needle posi-
tion provoked brief pain in toes 2 to 5 (not in toe 1). Contrast
A. Vaso et al. /PAIN
155 (2014) 1384–1391 1387
and lidocaine were then injected. This rapidly caused loss of the
toes 2 to 5 npPLS. Sensation on the stump was grossly normal, pre-
sumably because of incomplete block. Both Tinel signs were sup-
pressed (patient 14 estimated by >80%), and even strong
percussion of the medial neuroma no longer caused ﬂinching.
Dexamethasone was then administered, and the needle was with-
drawn. When the subject was released about 90 min later, his PLP
and npPLS were still absent. This remained the case when he was
contacted the next day.
3.2. Spinal (intrathecal) block (group 2 results)
Observations were made of 11 amputees in Tirana who were
given spinal blocks. Two were bilateral amputees (patients 2 and
4). Eight of the subjects, including subject 4, went on to have an
intraforaminal block as well. All 11 experienced PLP at the time
of the procedure (baseline pain ratings 7.1 ± 2.1; Table 1), mostly
in the phantom foot or a part of it. Common pain descriptors were
‘‘electric shock–like,’’ ‘‘shooting,’’ ‘‘constricting,’’ and ‘‘pulsating.’’
The remainder of the phantom leg was felt and could sometimes
be moved, but it was not painful. ‘‘Telescoped’’ phantoms were felt
by at least 3 of the subjects, all of whom had undergone amputa-
tion many years previously. One or more obvious stump neuromas
were present in all subjects, and percussion usually evoked pain in
the phantom (or a part of it) and/or in the stump. The sensation
evoked was most often likened to a stab or an electric shock. A
few subjects reported ongoing stump pain, usually burning or cold
in quality, and some had tenderness on stump scars. We did not
track these stump sensations systematically.
The technical adequacy of spinal block was veriﬁed by numb-
ness and paresis of both legs, and in most subjects, loss or major
attenuation of pain was evident upon percussion of stump neuro-
mas. Spinal block drastically obtunded ongoing PLP in all 11 ampu-
tees, usually within 5 to 10 min, to the point that it was no longer
felt (P< .001 compared to preblock; Table 1,Fig. 1). In subjects 2
and 4, PLP was lost bilaterally. Interestingly, npPLS, which was
present in all of the amputees, was also reported to have vanished
shortly after the block in all but one. The exception was patient 2,
in whom the Tinel sign also persisted, suggesting incomplete
spinal block. Ongoing stump pain, when present, was also
suppressed. PLP and npPLS usually began to return by 2 to 3 h after
injection, roughly in parallel with recovery of motor control of the
legs. Two subjects reported that PLP was still mostly absent 24 h
after injection. We are uncertain whether this reﬂects a persistent
effect of the block or spontaneous remission.
3.3. Intraforaminal block (group 1)
Intraforaminal blocks were provided to 13 amputees in Tirana
and 15 in Tel Hashomer. Considering the 13 amputees (baseline
pain ratings = 7.3 ± 2.2; Table 2), PLP was eliminated in 11 and
reduced by an estimated 60 and 90% in 2 (patients 5 and 13). Sur-
prisingly, major pain relief was usually achieved after blocking a
single segment. In patient 4, a double amputee, intraforaminal
block on the right eliminated PLP on the right side, with no effect
on PLP on the left side. One week later, a left-sided block was car-
ried out, with the opposite result. PLP on the left side was lost, with
no effect on PLP on the right side. Overall, compared to preinjection
pain, these blocks signiﬁcantly reduced PLP (P< .001, Fisher test,
Complete loss or near-complete attenuation of npPLS occurred
in parallel with the loss of the PLP in the majority of subjects
(P= .005 compared to preinjection; Table 2). This included loss of
npPLS on the right side, and subsequently on the left, in patient
4, who had undergone bilateral amputation. There were 6 excep-
tions, however. In patients 7, 10, 11, 12, and 16, attenuation of
npPLS was modest or uncertain, or no loss was reported at all. Sub-
ject 10 volunteered that his painful phantom was replaced by a
pleasant nonpainful sensation. We conjecture that in these cases,
the afferent drive of the residual npPLS originated in an adjacent
DRG. Because PLP had already been eliminated, we usually did
not inject additional levels to test this possibility. However, in
patient 14 (described above), after a L5 DRG block had eliminated
the PLP but not the npPLS, we administered a subsequent block to
the L4 DRG. This eliminated the npPLS, supporting our conjecture.
Attenuation of the Tinel sign was variable. In cases where both
PLP and npPLS were lost, the Tinel sign also tended to vanish. In
others, the Tinel sign was partly attenuated or largely unaffected
(Table 2). In some subjects, the Tinel sign from one stump neuroma
was attenuated, but not from a second.
During the ﬁnal phase of needle insertion, or during needle
repositioning which was required occasionally, a brief intensiﬁca-
tion of the PLP was frequently provoked. This sensation faded
within seconds or within a minute or two. Deceptive statements
by the physician that an impending (sham) injection would sup-
press phantom sensation, and at least 2 injections per subject of
nonblocking solutions (saline and contrast), were almost never fol-
lowed by a report that PLP or npPLS had changed. An exception
was a subject who reported ‘‘80% reduction’’ of her npPLS after
the saline injection. PLP remained unchanged until lidocaine was
injected, at which time it rapidly disappeared. Overall, lidocaine
signiﬁcantly obtunded both PLP and npPLS compared to nonblock-
ing solutions (P< .001, Fig. 1). Finally, in patients 1, 4, and 15, the
injection needle was replaced with a catheter that was left in place,
permitting repeated bolus injections of lidocaine (1%, 3 mL every 3
to 4 h) for periods of 5, 10, and 12 days. This produced sustained
absence of PLP at least for the full duration of the block.
3.4. Intraforaminal block using dilute lidocaine (group 4)
Intraforaminal block was carried out on a therapeutic basis at
Tel Hashomer in an additional 15 unilateral lower limb amputees.
All had baseline VAS pain scores of 7 to 10. Ten underwent trau-
matic amputation, and 5 were amputated for other reasons (3 vas-
cular, 1 septic, 1 malignancy). Ten had below-the-knee amputation
(BKA; 9 transtibial, 1 Symes [foot]) and 5 had above-the-knee
Fig. 1. Covert intraforaminal block using high and low concentrations of lidocaine,
and similar covert spinal (intrathecal) block, consistently suppressed phantom limb
pain (PLP) and nonpainful phantom limb sensation (npPLS). Control procedures
(sham, saline, contrast injections) did not. Group sizes were as follows: spinal block
(n= 11); intraforaminal block with 1% to 2% lidocaine (n= 13); controls (n= 13);
dilute intraforaminal lidocaine (n= 15).
1388 A. Vaso et al./ PAIN
155 (2014) 1384–1391
amputation (AKA; all transfemoral). Two segmental levels, L5 + S1,
were injected in all but 4 cases. In these, we injected only S1 (n=3)
or only L5 (n= 1). When queried after the completion of the intra-
foraminal blocks, 14 subjects (93%) reported complete (n= 10,
including 2 single-level cases) or substantial (75–95%, n= 4) elim-
ination of their preinjection PLP. The remaining amputee (BKA)
estimated 50% reduction in PLP. Attenuation of npPLS was similar:
9 reported complete loss of npPLS, 3 reported major attenuation
(70–90%), 2 reported no effect, and 1 was uncertain. Compared to
preinjection, the intensity of both PLP and npPLS was signiﬁcantly
reduced (P< .001).
3.5. Nerve block (group 3)
Dense femoral or sciatic nerve block using perineurial lidocaine
2% was carried out in Tirana in 2 patients (patients 1 and 13; base-
line pain ratings: 10, 7.5). Spinal and/or intraforaminal block had
eliminated or obtunded PLP in both of them. Background PLP and
npPLS were unaffected by the nerve blocks, although in the ampu-
tee in whom the femoral nerve was blocked, pain felt in the phan-
tom limb after neuroma percussion was lost. Lidocaine inﬁltration
of stump neuromas in 2 patients (patients 12 and 13; baseline pain
ratings: 10, 7.5) eliminated the Tinel sign from the injected neuro-
mas without markedly affecting phantom sensation. Intraforami-
nal block was effective in both. Patient 12 volunteered that her
sensation of being able to move phantom toes was noticeably
weakened after neuroma inﬁltration.
Spinal and intraforaminal block consistently attenuated, and
often completely eliminated, both PLP and npPLS in lower-limb
amputees. Control injections did not. This was documented in 31
amputees at 2 independent centers. The effect came on rapidly
because ganglionic sheaths are permeant . Sham injections
and intraforaminal injections of nonblocking solutions never
blocked PLP, even when patient were intentionally told to antici-
pate a pain-relieving injection. Blocking solutions, even when
injected covertly, consistently did. Nonetheless, we acknowledge
that our observations fall short of accepted criteria for randomized
blinded placebo controlled trials. For practical reasons, we did not
include an experimental arm randomized to placebo injections
exclusively, and although patients were blinded to the type and
timing of the injections, the medical staff was not. After spinal
block, the effect faded within hours, although after intraforaminal
block it lingered, probably as a result of the anti-inﬂammatory and
membrane-stabilizing (ectopia-silencing) effects of the coinjected
corticosteroids [15,36,41]. Systematic documentation of the dura-
tion of effect awaits further research. The abolition of PLP occurred
in both recent and veteran amputees, and irrespective of telescop-
ing, with no obvious ‘‘pain memory,’’ ‘‘centralization,’’ or ‘‘transi-
tion to chronicity’’ [13,34]. Although conscious perception is
undoubtedly a high CNS function, our data suggest that the raw
feel of a phantom limb is driven by activity originating in the
PNS, which feeds the CNS in a bottom-up manner.
4.1. PNS location of the ectopic generator
Amputation-induced neuroplastic changes occur in the spinal
cord as well as in the PNS. These include somatotopic remapping
much like that featured in cortical theories of PLP [16,20] and
spontaneous bursting discharge [10,12,43] (although this is mostly
driven from the periphery [13,52]). Central sensitization also
develops, a phenomenon that both ampliﬁes normal and ectopic
nociceptive input and renders low-threshold Abinput painful
. However, the possibility that spinal neuroplasticity drives
PLP can be ruled out for 2 reasons. First, intrathecal injections at
L4–5/L3–4 vertebral interspaces mostly act on primary afferent
axons (‘‘cauda equina’’). Because the spinal gray (lumbar enlarge-
ment) lies within the T10–12 vertebrae, >12 cm further rostrally,
our injections would have blocked spinal access of PNS ectopia
with minimal effect on the spinal cord per se.
Results of intraforaminal block further reinforce the conclusion
that neither the dorsal horn nor the brain are primary generators of
PLP. In 22 amputees, the S1 and/or L5 foramen were injected
>16 cm caudal to the lumbar enlargement. Fluoroscopy, which per-
mitted real-time tracking of the lidocaine–contrast mixture,
showed that spread was largely limited to the vertebral level or
levels injected (Fig. 2). Even if some had reached the lumbar
enlargement, it would have been highly diluted in the cerebrospi-
nal ﬂuid. Finally, attenuation of PLP (and npPLS) was topographi-
cally appropriate. Most notably, in patient 4, a double amputee,
left-sided L5 injection eliminated PLP on the left with no effect
on PLP on the right, and vice versa. Had the intraforaminal lido-
caine acted at the lumbar enlargement (T10–12) or even the cauda
equina (L5), PLP should have ceased bilaterally on both trials. This
observation also rules out a systemic action of the lidocaine.
4.2. DRG vs neuroma as the principal generator of PLP
When applied to nerves or dorsal root axons, pain relief reﬂects
block of spike propagation.This requires a high drug concentra-
tion; 2% lidocaine (100 mM) is typical . In contrast, for sup-
pression of spike electrogenesis(initiation) in the DRG, much
lower concentrations are sufﬁcient (10
impulses generated in both stump neuromas and the DRG would
have been blocked by 1% to 2% lidocaine administered intraforami-
nally, but lower concentrations would selectively block DRG ecto-
pia, sparing through-propagation of impulses generated further
distally in the stump. The therapeutic protocol used at Tel Hasho-
mer incorporated this factor by injecting lidocaine intraforaminally
at a subanesthetic concentration (0.3%). Selectivity (electrogenesis
vs spike propagation) was achieved because even when both L5
and S1 levels were injected (11 amputees), the stump did not
become numb, and amputees were fully mobile on their prosthetic
limbs moments after the block. The fact that nonblocking 0.3%
Fig. 2. Fluoroscopic image illustrating limited degree of spread typical of that
observed in patients injected intraforaminally with 3 mL lidocaine–steroid solution
containing contrast medium (left L5 DRG injection).
A. Vaso et al. /PAIN
155 (2014) 1384–1391 1389
lidocaine was as effective as 1% to 2% lidocaine suggests that ecto-
pia originating in the DRG is the primary generator of spontaneous
PLP. We suspect that electrogenesis in stump neuromas becomes a
more prominent generator of PLP when mechanical forces are
applied to neuromas during weight bearing and walking. This
accounts for the disappointing clinical experience with nerve
blocks and neuroma inﬁltration, including our own observations.
It is noteworthy that intraforaminal block at a single level usu-
ally provided total or near-total relief of PLP despite the fact that at
least 3 DRGs contribute to leg innervation. There are several likely
explanations. First, PLP is generally located in the foot (primarily L5
and S1) rather than the calf or thigh, and often in only part of the
missing foot. Choice of the ﬁrst segment to block was guided by the
location of symptoms; it was not random. Second, even if drivers of
PLP originate in 2 or 3 DRG, pain may not become a complaint until
the sum of the activity crosses some threshold. Thus, in patients
relieved of PLP after L5 block, S1 block might also have worked
had it been tried ﬁrst. Finally, we cannot exclude lidocaine spread-
ing from the injected to DRG to an adjacent one.
4.3. PLP and maladaptive cortical plasticity
Our data are inconsistent with maladaptive cortical plasticity
being the primary driver of PLP and npPLS. However, some of the
secondary peculiarities of phantom limb sensation (eg, telescoping
and reference) may well reﬂect plasticity of cortical processing.
Had the impulses interpreted by a conscious brain as PLP origi-
nated in the cortex, spinal and intraforaminal blocks would have
been ineffective and certainly not topographically appropriate.
Beyond that, the very foundations upon which the cortical plastic-
ity hypotheses rest are equivocal. For example, correlation
between the extent of somatotopic remapping and the degree of
PLP  does not prove causation. This correlation equally sup-
ports a CNS effect of a PNS cause. It is known that PNS activity
can drive CNS remapping [19,30,57,61]. Thus, discharge originating
ectopically in the DRG could well be the cause of both PLP and of
remapping. Indeed, brachial plexus block sometimes reverses both
. This result might be obtained in all amputees using intrafora-
minal rather than plexus block.
PLP models based on multisensory mismatch also fall short.
Perceptual conﬂict is usually attributed to loss of afferent input
from the limb after amputation. However, this ignores abundant
evidence of ectopic electrogenesis in the PNS after nerve section
. Amputation may cause the slow dying back of some of the
axotomized leg afferents, but most survive for decades and remain
capable of signaling pain—thus the Tinel sign. Noninvasive func-
tional recording also challenges the idea of afferent silence. The
cortical representation of adjacent skin ‘‘invades’’ that of the
amputated limb [20,23], a phenomenon that is thought to account
for the frequent reference of sensation from stump and nearby skin
into the phantom limb [9,51]. However, this should not evoke
(phantom) pain because even direct electrical stimulation of the
primary somatosensory cortex is not painful . Importantly,
the ‘‘invasion’’ does not displace input originating in the severed
afferents that used to serve the (amputated) limb. In fact, the cor-
tical representation of the (phantom) limb actually increases, as
one might predict given the ectopia coming off stump neuromas
and the DRG . Moreover, the increase is proportional to the
intensity of the PLP. These are the observations expected if PLP is
driven by a bottom-up process.
We propose that ectopic PNS discharge, primarily that originat-
ing in DRG serving the amputated limb, drives CNS somatic repre-
sentations to generate a conscious percept of the phantom limb.
The quality of the sensation, PLP or npPLS, presumably depends
largely on the types of primary afferent neurons that contribute
to the ectopic barrage . The fact that stimulating adjacent skin
sometimes evokes sensation felt in the phantom [9,53] probably is
due to CNS plasticity and likewise the sense of limb ownership and
distortions of the phantom limb with respect to body schema,
including telescoping, movement, and unnatural orientations of
phantom limbs [47,54]. In this regard, it is noteworthy that several
of our subjects indicated that walking on their prosthesis in the
absence of their usual phantom limb sensation was disconcerting.
One patient stated, ‘‘It feels as if I am on a wooden leg, not on my
4.4. Therapeutic implications
Our study, together with earlier work, highlights the DRG as a
critical source of ectopic impulse discharge in amputees with
PLP. The therapeutic potential of targeting the DRG is documented
by our 3 amputees in whom PLP was suppressed for up to 12 con-
tinuous days with sustained intraforaminal lidocaine given
through an indwelling catheter. Because a low concentration of
lidocaine (and other membrane stabilizers) is sufﬁcient , and
undoubtedly less toxic than 1% to 2% lidocaine, current implant-
able pump systems might provide extended pain relief using a sin-
gle reservoir charge and a slow pumping rate. Novel anesthetic
modalities that are selective to small-diameter afferents  might
be a way to attenuate PLP while preserving the beneﬁts of non-
painful phantom limb sensation in the maintenance of body image.
Furthermore, these approaches might well be applicable to other
neuropathic pain conditions in which DRG ectopia is a root cause
Conﬂict of interest
The authors report no conﬂict of interest.
We thank Leonard Grazhdani for assistance with some of the
procedures, Yitzhak Zivner for guidance and encouragement, Bob
Boas (Aukland, New Zealand) and Anatoly Stav (Hadera, Israel)
for inspiration and permission to cite their unpublished observa-
tions, Bernd Borchardt for his steadfast support, and Ze’ev Seltzer
and Tamar Makin for their comments. We also acknowledge the
amputees who volunteered as subjects for this study. Financial
support in Tirana was provided by the German Health Ministry
and in Israel by the Israel Science Foundation (ISF) and the Hebrew
University Center for Research on Pain. The funders played no
direct role in the design or execution of the work.
 Anderson-Barnes VC, McAuliffe C, Swanberg KM, Tsao JW. Phantom limb
pain—a phenomenon of proprioceptive memory? Med Hypotheses
 Babbedge RC, Soper AJ, Gentry CT, Hood VC, Campbell EA, Urban L. In vitro
characterization of a peripheral afferent pathway of the rat after chronic sciatic
nerve section. J Neurophysiol 1996;76:3169–77.
 Binshtok AM, Bean BP, Woolf CJ. Inhibition of nociceptors by TRPV1-mediated
entry of impermeant sodium channel blockers. Nature 2007;449:607–10.
 Birbaumer N, Lutzenberger W, Montoya P, Larbig W, Unertl K, Töpfner S, Grodd
W, Taub E, Flor H. Effects of regional anesthesia on phantom limb pain are
mirrored in changes in cortical reorganization. J Neurosci 1997;17:5503–8.
 Black JA, Nikolajsen L, Kroner K, Jensen TS, Waxman SG. Multiple sodium
channel isoforms and mitogen-activated protein kinases are present in painful
human neuromas. Ann Neurol 2008;64:644–53.
 Byrod G, Rydevik B, Johansson BR, Olmarker K. Transport of epidurally applied
horseradish peroxidase to the endoneurial space of dorsal root ganglia: a light
and electron microscopic study. J Periph Nerv Syst 2000;5:218–26.
 Clark RL, Bowling FL, Jepson F, Rajbhandari S. Phantom limb pain after
amputation in diabetic patients does not differ from that after amputation in
nondiabetic patients. PAIN
 Colloca L, Lopiano L, Lanotte M, Benedetti F. Overt versus covert treatment for
pain, anxiety, and Parkinson’s disease. Lancet Neurol 2004;3:679–84.
1390 A. Vaso et al./ PAIN
155 (2014) 1384–1391
 Cronholm B. Phantom limb in amputees: study of changes in integration of
centripetal impulses with special reference to referred sensations. Acta
Psychiatr Neurol Scand 1951;72:1–85.
 Dalal A, Tata M, Allègre G, Gekiere F, Bons N, Albe-Fessard D. Spontaneous
activity of rat dorsal horn cells in spinal segments of sciatic projection
following transection of sciatic nerve or of corresponding dorsal roots.
 deJong RH. Local anaesthetics. St Louis: Mosby; 1994.
 Devor M. Central changes after nerve injury. In: Willis W, Schmidt R, editors.
Encyclopedia of pain. Berlin: Springer Verlag; 2007.
 Devor M. Centralization, central sensitization and neuropathic pain. Focus on
‘‘sciatic chronic constriction injury produces cell-type-speciﬁc changes in the
electrophysiological properties of rat substantia gelatinosa neurons’’. J
 Devor M. Neuropathic pain: pathophysiological response of nerves to injury.
In: McMahon SL, Koltzenburg M, Tracey I, Turk DT, editors. Wall and Melzack’s
textbook of pain. London: Churchill Livingstone; 2013. p. 861–88.
 Devor M, Govrin-Lippmann R, Raber P. Corticosteroids suppress ectopic
neuronal discharge originating in experimental neuromas. PAIN
 Devor M, Wall PD. Reorganization of the spinal cord sensory map after
peripheral nerve injury. Nature 1978;276:76.
 Devor M, Wall PD, Catalan N. Systemic lidocaine silences ectopic neuroma and
DRG discharge without blocking nerve conduction. PAIN
 Ehrsson H, Wiech K, Weiskopf N, Dolan RJ, Passingham RE. Threatening a
rubber hand that you feel is yours elicits a cortical anxiety response. Proc Natl
Acad Sci U S A 2007;104:9828–33.
 Faggin BM, Nguyen KT, Nicolelis MA. Immediate and simultaneous sensory
reorganization at cortical and subcortical levels of the somatosensory system.
Proc Natl Acad Sci U S A 1997;94:9428–33.
 Finnerup NB, Nikolajsen L, Jensen TS. Are we neglecting spinal reorganization
following nerve damage? PAIN
 Finnerup NB, Norrbrink C, Fuglsang-Frederiksen A, Terkelsen AJ, Hojlund AP,
Jensen TS. Pain, referred sensations, and involuntary muscle movements in
brachial plexus injury. Acta Neurol Scand 2010;121:320–7.
 Flor H, Elbert T, Knecht S, Wienbruch C, Pantev C, Birbaumer N, Larbig W, Taub
E. Phantom-limb pain as a perceptual correlate of cortical reorganization
following arm amputation. Nature 1995;375:482–4.
 Flor H, Nikolajsen L, Staehelin Jensen T. Phantom limb pain: a case of
maladaptive CNS plasticity? Nat Rev Neurosci 2006;7:873–81.
 Florence SL, Hackett TA, Strata F. Thalamic and cortical contributions to neural
plasticity after limb amputation. J Neurophysiol 2000;83:3154–9.
 Florence SL, Kaas JH. Large-scale reorganization at multiple levels of the
somatosensory pathway follows therapeutic amputation of the hand in
monkeys. J Neurosci 1995;15:8083–95.
 Greenberg SA. The history of dermatome mapping. Arch Neurol
 Halbert J, Crotty M, Cameron ID. Evidence for the optimal management of
acute and chronic phantom pain: a systematic review. Clin J Pain
 Harris AJ. Cortical origin of pathological pain. Lancet 1999;354:1464–6.
 Henderson WR, Smyth GE. Phantom limbs. J Neurol Neurosurg Psychiatry
 Jenkins W, Merzenich M, Ochs M, Allard T, Guíc-Robles E. Functional
reorganization of primary somatosensory cortex in adult owl monkeys after
behaviorally controlled tactile stimulation. J Neurophysiol 1990;63:82–104.
 Jones EG. Cortical and subcortical contributions to activity-dependent
plasticity in primate somatosensory cortex. Annu Rev Neurosci 2000;23:1–37.
 Kaas J, Merzenich M, Killackey H. The reorganization of somatosensory cortex
following peripheral nerve damage in adult and developing mammals. Ann
Rev Neurosci 1983;6:325–56.
 Kajander KC, Wakisaka S, Bennett GJ. Spontaneous discharge originates in the
dorsal root ganglion at the onset of a painful peripheral neuropathy in the rat.
Neurosci Lett 1992;138:225–8.
 Katz J, Melzack R. Pain, ‘‘memories’’ in phantom limbs: review and clinical
 Keegan JJ, Garrett FD. The segmental distribution of the cutaneous nerves in
the limbs of man. Anat Rec 1948;102:409–37.
 Kingery WS, Agashe GS, Sawamura S, Davies MF, Clark JD, Maze M.
Glucocorticoid inhibition of neuropathic hyperalgesia and spinal Fos. Anesth
 Kirk E. Impulses in dorsal spinal nerve rootlets in cats and rabbits arising
from dorsal root ganglia isolated from the periphery. J Comp Neurol
 Kirk EJ, Denny-Brown D. Functional variation in dermatomes in the macaque
monkey following dorsal root lesions. J Comp Neurol 1970;139:307–20.
 Knotkova H, Cruciani RA, Tronnier VM, Rasche D. Current and future options
for the management of phantom-limb pain. J Pain Res 2012;5:39–49.
 Kuslich S, Ulstro C, Michael C. The tissue origin of low back pain and sciatica.
Orth Clin North Am 1991;22:181–7.
 Li JY, Xie W, Strong JA, Guo QL, Zhang JM. Mechanical hypersensitivity,
sympathetic sprouting, and glial activation are attenuated by local injection of
corticosteroid near the lumbar ganglion in a rat model of neuropathic pain.
Reg Anesth Pain Med 2011;36:56–62.
 Liu CN, Wall PD, Ben-Dor E, Michaelis M, Amir R, Devor M. Tactile allodynia in
the absence of C-ﬁber activation: altered ﬁring properties of DRG neurons
following spinal nerve injury. PAIN
 Loeser JD, Ward AA, White LE. Chronic deafferentation of human spinal cord
neurons. J Neurosurg 1968;29:48–50.
 Lund P. Principles and practice of spinal anesthesia. Springﬁeld: Charles C.
 Makin TR, Scholz J, Filippini N, Henderson Slater D, Tracey I, Johansen-Berg H.
Phantom pain is associated with preserved structure and function in the
former hand area. Nat Commun 2013;4:1570.
 Melzack R. Phantom limbs, the self and the brain. Can Psychol 1989;30:1–16.
 Moseley GL, Brugger P. Interdependence of movement and anatomy persists
when amputees learn a physiologically impossible movement of their
phantom limb. Proc Natl Acad Sci U S A 2009;106:18798–802.
 Moseley GL, Gallace A, Spence C. Is mirror therapy all it is cracked up to be?
Current evidence and future directions. PAIN
 Nordin M, Nyström B, Wallin U, Hagbarth KE. Ectopic sensory discharges and
paresthesiae in patients with disorders of peripheral nerves, dorsal roots and
dorsal columns. PAIN
 Nystrom B, Hagbarth K. Microelectrode recordings from transected nerves in
amputees with phantom limb pain. Neurosci Lett 1981;27:211–6.
 Penﬁeld W, Rasmussen T. The cerebral cortex of man. New York: Macmillan;
 Pitcher GM, Henry JL. Governing role of primary afferent drive in increased
excitation of spinal nociceptive neurons in a model of sciatic neuropathy. Exp
 Ramachandran V, Stewart M, Rogers-Ramachandran D. Perceptual correlates
of massive cortical reorganization. Neuroreport 1992;3:583–6.
 Sherman R, Devor M, Jones C, Katz J, Marbach J. Phantom pain. New
York: Plenum; 1996.
 Sherman RA, Arena JG. Phantom limb pain: mechanisms, incidence and
treatment. Rev Phys Med 1992;4:1–26.
 Sherman RA, Grifﬁn VD, Evans CB, Grana AS. Temporal relationships between
changes in phantom limb pain intensity and changes in surface
electromyogram of the residual limb. Int J Psychophysiol 1992;13:71–7.
 Soros P, Knecht S, Bantel C, Imai T, Wüsten R, Pantev C, Lütkenhöner B, Bürkle
H, Henningsen H. Functional reorganization of the human primary
somatosensory cortex after acute pain demonstrated by
magnetoencephalography. Neurosci Lett 2001;298:195–8.
 Szasz TS. Pain and pleasure: a study of bodily feeling. New York: Basic Books;
 Tanelian DL, MacIver MB. Analgesic concentrations of lidocaine suppress tonic
A-delta and C-ﬁber discharges produced by acute injury. Anesthesia
 Tessler MJ, Kleiman SJ. Spinal anaesthesia for patients with previous lower
limb amputations. Anaesthesia 1994;49:439–41.
 Waberski TD, Gobbelé R, Kawohl W, Cordes C, Buchner H. Immediate cortical
reorganization after local anesthetic block of the thumb: source localization of
somatosensory evoked potentials in human subjects. Neurosci Lett
 Wall PD, Devor M. Sensory afferent impulses originate from dorsal root ganglia
as well as from the periphery in normal and nerve injured rats. PAIN
 Wall PD, Gutnick M. Properties of afferent nerve impulses originating from a
neuroma. Nature 1974;248:740–3.
 Wallace MS, Dyck JB, Rossi SS, Yaksh TL. Computer-controlled lidocaine
infusion for the evaluation of neuropathic pain after peripheral nerve injury.
 Woolf CJ, Salter MW. Plasticity and pain: role of the dorsal horn. In: McMahon
SB, Koltzenberg M, editors. Wall and Melzack’s textbook of
pain. London: Elsevier Churchill Livingstone; 2006. p. 91–105.
 Yang TT, Gallen CC, Ramachandran VS, Cobb S, Schwartz BJ, Bloom FE.
Noninvasive detection of cerebral plasticity in adult human somatosensory
cortex. Neuroreport 1994;5:701–4.
A. Vaso et al. /PAIN
155 (2014) 1384–1391 1391