DOI: 10.1093/brain/awg007
Brain (2003), 126, 57±70
Sensory function in spinal cord injury patients with
and without central pain
N. B. Finnerup,1 I. L. Johannesen,2 A. Fuglsang-Frederiksen,3 F. W. Bach1 and T. S. Jensen1
1Department
of Neurology and Danish Pain Research
Centre, Aarhus University Hospital, 2Department of
Rheumatology, Viborg Hospital and 3Department of
Clinical Neurophysiology, Aarhus University Hospital,
Denmark
Summary
Spinal cord injury (SCI) frequently results in neuropathic pain. However, the pathophysiology underlying
this pain is unclear. In this study, we compared clinical
examination, quantitative sensory testing (QST) and
somatosensory evoked potentials (SEPs) in SCI patients
with and without pain below spinal lesion level, with a
control group of 20 subjects without injury. All patients
had a traumatic SCI with a lesion above T10; 20
patients presented with spontaneous central neuropathic
pain below lesion level, and 20 patients had no neuropathic pain or dysaesthesia. Patients with and without
pain had a similar reduction of mechanical and thermal
detection and pain thresholds, and SEPs. SCI patients
Correspondence to: N. B. Finnerup, Danish Pain Research
Centre, Building 1A, Aarhus University Hospital,
Noerrebrogade 44, DK-8000 Aarhus C, Denmark
E-mail: ®[email protected]
with central pain more frequently had sensory hypersensitivity (brush- or cold-evoked pain, dysaesthesia or
pinprick hyperalgesia) in dermatomes corresponding to
lesion level than SCI patients without pain. There was
no difference in intensity of pain evoked by repetitive
pinprick at lesion level between patient groups. There
was a signi®cant correlation between intensity of brushevoked dysaesthesia at lesion level and spontaneous
pain below lesion level of SCI. These data suggest that
lesion of the spinothalamic pathway alone cannot
account for central pain in SCI patients, and that
neuronal hyperexcitability at injury or higher level may
be an important mechanism for pain below injury level.
Keywords: central pain; allodynia; neuronal hyperexcitability; spinothalamic function; dorsal column function
Abbreviations: ASIA = American Spinal Injury Association; NRS = numeric rating scales; QST = quantitative sensory
testing; SCI = spinal cord injury; SEP = sensory evoked potential; VAS = visual analogue scales
Introduction
Neuropathic pain is a signi®cant problem following spinal
cord injury (SCI). Recent studies suggest that chronic
neuropathic pain occurs in ~50% of SCI patients and that
such pain represent a major burden to the patients (Anke et al.,
1995; Siddall et al., 1999; Finnerup et al., 2001). Despite its
severity, studies on SCI pain and its underlying mechanisms
have been relatively few.
According to the SCI Pain Task Force of the IASP
(International Association for the Study of Pain), there are
two categories of neuropathic pain resulting from SCI: at
level pain (= pain in segments corresponding to the level of
cord injury) and below level pain (= pain in body parts
corresponding to segments below the injury) (Siddall et al.,
2000). While below level neuropathic pain is considered to be
a central pain condition caused by spinal cord damage, at
level pain may have peripheral and central components that
are dif®cult to separate (Siddall et al., 1997). The mechanisms underlying neuropathic SCI pain are still hypothetical.
ã Guarantors of Brain 2003
In particular, it is unclear why some patients develop pain in
body parts that have lost their afferent input while others with
the same degree of deafferentation do not develop such pain.
A dominating clinical feature of central pain conditions is an
abnormal spinothalamic function with altered sensitivity to
temperature and pinprick (Boivie et al., 1989; Vestergaard
et al., 1995; Bowsher, 1996). In post-stroke pain, for
example, it has been suggested that pain only develops in
those patients with partial or complete loss of spinothalamic
functions and signs of increased response to certain stimuli in
the deafferented body parts (Andersen et al., 1995;
Vestergaard et al., 1995).
In spinal cord-damaged patients with at level and below
level pain, some previous studies have determined sensory
function and recorded pain (Eide et al., 1996; Bouhassira
et al., 2000; Defrin et al., 2001). Eide et al. (1996) compared
somatosensory abnormalities in painful and non-painful
denervated areas at or below injury level in 16 SCI pain
58
N. B. Finnerup et al.
patients. They found that allodynia and wind up-like pain
were more common in painful than in non-painful areas, and
suggested hyperexcitability in spinothalamic tract neurons to
be involved in the pathogenesis of pain. Bouhassira et al.
(2000) studied 10 patients with painful and 11 patients with
painless syringomyelia, and observed no signi®cant difference in thermal or mechanical sensory function between
patients with or without pain or between areas of spontaneous
pain and adjacent non-painful areas. Pain was con®ned,
however, to areas with maximal thermal de®cit. In incomplete SCI patients with lesions restricted to the T4±L3
segments, Defrin et al. (2001) found no differences between
thermal and tactile sensations in patients with or without
chronic pain. In the pain patients, however, thermal thresholds were impaired more in body areas with pain than in areas
without pain, which had almost normal thermal sensibility.
Allodynia was only elicited in pain patients. Taken together,
these observations, in SCI patients with at level and below
level pain, suggest that abnormal spinothalamic activity and
signs of hyperexcitability, originating either in the cord itself
or from damaged afferent ®bres at the segmental level, may
play a role in their pain. From the above studies, it is dif®cult
to determine the role of the peripheral and central
components of the pain. As recently pointed out by Wall
(2001), we need to know how much of spinal injury pain is
caused by peripheral as opposed to central lesions. In order to
determine the mechanisms underlying proper central pain
after SCI, we looked only at patients with pain in areas below
spinal injury level, where damaged root or nerve lesions
causing pain are unlikely. To study the underlying mechanisms we (i) characterized the degree of disconnection of
ascending sensory pathways by using quantitative sensory
testing (QST) and somatosensory evoked potentials (SEPs);
and (ii) studied the degree of sensory hypersensitivity at and
below the level of injury. A group of SCI patients with central
neuropathic pain were compared with a group of SCI patients
without neuropathic pain and also with a control group
without SCI. Parts of the study have been presented in
abstract form previously (Finnerup et al., 2002b).
Material and methods
Subjects
Three groups each of 20 subjects participated. Inclusion
criteria for the SCI pain group were: central neuropathic pain
(= pain below level of spinal cord injury) after a traumatic
SCI. Other reasons for the pain below lesion level
(nociceptive, peripheral neuropathic, and psychogenic pain)
should be excluded or considered highly unlikely. To ensure
that pain patients had below level pain, the area of pain should
be at least two dermatome segments below the lesion level,
and only patients with an injury above the conus medullaris
(with a skeletal level above T10) were included. MRI
performed in 14 pain and nine pain-free patients con®rmed
that the lower border of the spinal cord lesion was above the
10th thoracic vertebra (N. B. Finnerup et al., unpublished
results). Spasm-related pain below level, and at level or above
level neuropathic pain were allowed in addition to the central
pain. For the SCI pain-free group, the inclusion criteria were:
a traumatic SCI above T10 with no neuropathic pain and no
spontaneous dysaesthesias (unpleasant abnormal sensations).
Non-unpleasant sensations (paraesthesias) and spasm-related
pain were allowed. The control group comprised gender- and
age-matched healthy persons with no pain history. Patients
were excluded if they had known or clinical signs of
concomitant cerebral damage including epilepsy or dementia
[total score on the Mini-Mental State Examination (Folstein
et al., 1975) on inclusion below 26]. Patients were recruited
from the Viborg rehabilitation centre for spinal cord injury,
and controls were enrolled after advertisement at the hospital
and the local blood bank. No analgesics were allowed, and
patients were asked to refrain from spasmolytics on the day of
examination.
Methods
All subjects were examined by the same examiner (N.B.F.).
Examinations were done on three occasions. Patients and
controls were ®rst examined with QST including evaluation
of evoked pain below level; SEPs were determined on a
second, and examination of evoked pain at lesion level on a
third occasion. Informed consent was obtained according to
the Declaration of Helsinki. The study was approved by the
local ethical committee (no. 1999/4492) and the Danish Data
Protection Agency (no. 2000-41-0099).
History
A medical history was obtained. The Functional
Independence Measure was used to estimate disability and
dependence (Fuhrer, 1987). Pain patients ®lled out the Danish
version of the McGill Pain Questionnaire (Melzack, 1975;
Drewes et al., 1993) and localized their pain on a body chart
in two dimensions (anterior and posterior). Numeric rating
scales (NRS; 0±10) and visual analogue scales (VAS; 0±10)
were used for assessing pain intensity. SCI patients reported
spasms (considering both intensity and frequency) on an
NRS 0±10.
Clinical examination
A neurological and physical examination was obtained.
Spinal lesions were classi®ed according to the American
Spinal Injury Association's (ASIA) standards for classi®cation of SCI (Maynard et al., 1997). Spasticity was assessed by
the investigator as a combined score of muscle tone using the
Ashworth scale (Ashworth, 1964) and a clinical grading of
tendon re¯exes (Nielsen et al., 1996).
Sensory function and central pain in spinal cord injury
SEPs
SEPs were determined in 14 pain and 12 pain-free patients
with the subject in a comfortable supine position. The
posterior tibial nerve was stimulated at the ankle with
rectangular stimuli of 0.2 ms duration at a stimulus rate of
3 Hz. The stimulus intensity was increased until a de®nite
motor response was seen. Chlorided silver cup recording
electrodes were placed over Cz¢ (2 cm behind Cz) referenced
to Fz, and over C3¢ (2 cm behind C3) referenced to C4¢ (2 cm
behind C4) for left tibial nerve stimulation, and C4¢
referenced to C3¢ for right side stimulation (International
10±20 system). Electrodes were placed over the twelfth
thoracic vertebral spinous process (T12) referenced to a site
over the iliac crest. Lower and upper frequency limits were 10
and 3000 Hz. SEP outcome was scored based on changes in
latency and amplitude and disappearance of SEP waveform
components: normal [compared with a national reference
material with a regression line based on height and age
(Andersen et al., personal communication)], decreased
(abnormal waveform or decreased latency or amplitude) and
absent SEP (no response clearly distinguishable from noise).
QST
QST was carried out 10 cm below the dominant knee on the
anterior-lateral part, except for vibration (tested on the
anterior surface of the tibia) and skinfold pinch (tested
above the knee). Measurements were obtained in a quiet room
at 20±22°C with all patients lying on a couch in a supine
relaxed position. However, ®ve patients were examined in
their wheelchair with the chair in the supine position. In cases
of severe sensory loss, failure to respond to cut-off limits
resulted in assignment of the cut-off limits as the threshold
value. A threshold was considered normal if there was a
modality-speci®c response within the range among the 20
healthy controls.
Tactile sensitivity to single stimuli
This parameter was assessed using von Frey hairs (Semmes±
Weinstein mono®laments, Stoelting, IL, USA). The tactile
detection threshold and pain detection threshold were determined as the force required to bend the ®lament using the up±
down paradigm described by Chaplan et al. (1994). Detection
of a single hair was de®ned as detection within 3 s.
Touch localization
This parameter was assessed by touching the skin with cotton
wool for 2 s.
Identi®cation of a slowly moving stimulus
Cotton was moved slowly 10 cm in a proximal-distal or
transverse direction (Wall and Noordenbos, 1977; Danziger
59
et al., 1996), and the ability to determine whether the stimulus
was moving and the direction of movement was tested.
The sensation of vibration
Vibration sensation was tested using an electronic vibrameter
(Somedic AB, Sweden) (Goldberg and Lindblom, 1979). The
vibration frequency was 120 Hz, the application pressure was
kept at ~650 g, and the amplitude was given as peak to peak
amplitude. The average of the vibration perception threshold
(the lowest amplitude at which the vibration was perceived)
and the vibration disappearance threshold (the amplitude at
which the vibration disappeared) was taken as the vibration
threshold; the average of three consecutive measurements
was used as the ®nal threshold.
Thermal sensation
Thermal thesholds were determined by the Medoc
Thermotest (TSA 2001, Israel) with a thermode of Peltier
elements measuring 32 3 32 mm. From a baseline temperature of 30°C, temperatures changed at a rate of 1°C/s (cut-off
limits at 50.5 and 0°C). Warm and cold detection thresholds,
heat pain thresholds, heat pain tolerance thresholds and cold
pain thresholds (in that order) were recorded using the
method of limits. Thresholds were calculated as the average
of ®ve successive measurements with a random interstimulus
interval of 3±6 s (two measurements for heat tolerance
thresholds).
Pressure pain
Pressure pain threshold and pressure pain tolerance threshold
were assessed over the anterior tibial muscle using a handheld electronic pressure algometer (Somedic AB, Sweden), as
described previously (Brennum et al., 1989), with a circular
probe (1 cm2) and a pressure rate of 30 kPa/s. The thresholds
were calculated as the mean value of three (two for tolerance
threshold) consecutive determinations with an interstimulus
interval of 60 s and expressed in kPa. The cut-off limit was
1200 kPa.
Skinfold pinch pain
Pain and pain tolerance thresholds were estimated by the use
of a forceps coupled to the electronic algometer, and the
thresholds (median of three and two determinations, respectively, as above) were expressed in kPa, with a cut-off of
1200 kPa.
Evoked pain
Evoked pain was studied below the level of injury on the
dominant leg (in pain patients also in the area of maximal
pain) and at injury level. At injury level, the border zone was
mapped for three types of stimuli (brush, pinprick and cold)
60
N. B. Finnerup et al.
Table 1 Subject characteristics
Number
Age (years), mean (SD)
Gender, males/females
Height (cm), mean (SD)
Weight (kg), mean (SD)
SCI pain
patients
SCI pain-free
patients
Controls
P
20
41.5 (11.1)
18/2
179.7 (8.8)
78.0 (19.1)
20
39.6 (9.2)
17/3
179.3 (9.1)
82.6 (18.3)
20
42.8 (9.0)
18/2
178.7 (7.4)
76.7 (10.2)
0.58*
1+
0.93*
0.44²
*One-way ANOVA; +Fisher's exact test; ²Kruskal±Wallis test.
based on methods used for mapping secondary hyperalgesic
areas in experimental pain (LaMotte et al., 1991).
Brush-evoked pain and dysaesthesia. This was assessed by
stroking the skin twice with a small brush at a rate of 1±
2 cm/s. Allodynia was considered to be present if pain was
evoked by the stimulus; the evoked pain was then measured
on a VAS (0 = no pain; 10 = unbearable pain); in the absence
of pain, brush-evoked dysaesthesia was recorded on a VAS
(0 = no unpleasantness; 10 = unbearable unpleasantness).
Cold allodynia. Cold allodynia below the lesion level was
assessed by running a thermo roll (20±22°C) twice across the
skin. At lesion level, it was assessed by applying an acetone
drop to the skin.
Pinprick hyperalgesia. This was considered to be present if
bending a von Frey hair (no. 5.88, bending force 75.9 g/
745 mN) evoked more pain than above injury level, and pain
was assessed on a VAS.
Pain evoked by repetitive pinprick. This was evaluated by
repetitively tapping the skin with a stiff von Frey hair driven
by a computer-controlled solenoid at a rate of 2 Hz for 1 min
or until pain became unbearable, as described previously
(Eide et al., 1994; Gottrup et al., 1998). From a starting point
where the hair is just touching the skin accurately, the hair is
moved 3 mm forward and 3 mm back, contacting the skin for
300 ms. At lesion level, the tapping was performed manually.
Patients quanti®ed evoked pain on a VAS.
Statistics
For each patient, the thresholds for temperature, skinfold
pinch and pressure were considered to be distributed
normally, and mean values were used as ®nal thresholds.
Statistical differences between groups were calculated using
one-way ANOVA (analysis of variance) or the nonparametric Kruskal±Wallis test; and between two sets of
unpaired data using the unpaired t test or the Mann±Whitney
U test. Correlations between non-parametric data were
assessed using Spearman's rank correlation. Fisher's exact
test was used to compare dichotomized data.
Results
Subject characteristics
The three groups were similar in terms of age, gender, height
and weight (Table 1). Eleven SCI pain patients had a
complete SCI (ASIA classi®cation) while 13 SCI patients
without pain had complete lesions. There were no signi®cant
differences in the two patient groups with respect to
ASIA impairment scale, neurological level, Functional
Independence Measure, mechanism of trauma and time
since spinal injury (Table 2). Only four patients used
analgesics (paracetamol, non-steroidal anti-in¯ammatory
drug, and one also a weak opiate). On the last examination,
two additional pain patients were receiving treatment with
lamotrigine. SCI pain patients reported signi®cantly more
spasms than SCI pain-free patients (P = 0.00, Mann±Whitney
U test), but there was no difference in spasticity measured on
the Ashworth scale and grading of tendon re¯exes (Table 2).
Spontaneous pain and sensory abnormalities
All patients in the SCI pain group had spontaneous pain
below lesion level (Fig. 1), and all except one also described
dysaesthesia; ®ve patients had spontaneous pain at lesion
level, and six patients complained of visceral pain. Onset of
pain after SCI occurred within the ®rst 6 months in 13
patients, and later in seven patients. The mean duration of
pain was 14.1 years. The median pain intensity was scored as
5 (NRS 0±10). At the last examination, which was done
~1 year after the two ®rst sessions, three patients had a
decrease in pain intensity, one spontaneously and two
probably attributable to lamotrigine treatment.
In general, the area of pain represented only a fraction of
the body area with sensory abnormality. In 19 patients, pain
was restricted to an area with altered or absent somatic
sensation to pinprick or touch. One patient (no. 5, Fig. 1) had
pain outside the area, with altered sensation to pinprick and
touch examined by bedside tests (ASIA classi®cation). The
most frequent words used spontaneously by the patients to
describe their pain and dysaesthesias were `pricking' and
`tingling', `burning' or `scalding', `freezing' and `taut'. None
in the pain-free group had pain or dysaesthesia at or below
SCI level, but 17 had constant or intermittent non-painful
sensations described as `pricking', `tingling', `taut', `warm'
or `throbbing'.
Sensory function and central pain in spinal cord injury
61
Table 2 Patient characteristics
SCI pain
patients
Time since injury (years), mean (SD)
Mechanism of injury
Vehicle accident
Fall from height
Diving
Other
Neurological level
Cervical
Thoracic
ASIA impairment scale
A
B
C
D
FIM, median (range)
Medication
Spasmolytics
Analgesics
Spasticity, Asworth and re¯ex score, mean (SD)
Spasms (NRS 0±10) median (range)
Spasm-related pain (NRS 0±10) median (range)
Allodynia, n
McGill Pain Questionnaire
PRI (R), mean (SD)
NWC, mean (SD)
15.5 (8.2)
SCI pain-free
patients
16.7 (8.7)
14
2
2
2
9
4
5
2
12
8
12
8
P
0.67*
0.20+
0.66+
0.41+
1+
1+
11
4
2
3
107 (64±124)
13
3
3
1
118 (51±119)
0.75+
1+
1+
0.61+
0.82²
10
4
20.8 (8.9)
4 (1±8)
0 (0±9)
8
9
0
24.1 (10.3)
2 (0±8)
0 (0±5)
0
1+
0.11+
0.29*
0.00²
0.19²
0.00+
25.9 (12.7)
12.5 (4.8)
*t test, +Fisher's exact test, ²Mann±Whitney U test. Neurological level = the last intact neurological
segment. ASIA class: A = complete, no sensory or motor function is preserved in the sacral segments S4±
S5; B = incomplete, preservation of sensory but not motor function; C = incomplete, motor grade <3
(Medical Research Council grade) in more than half of key muscles; D = incomplete, motor grade >3 in
at least half of key muscles. FIM = Functional Independence Measure; NWC = total number of words
chosen; PRI = the pain rating index.
Spinothalamic versus dorsal column function
Most SCI patients had impairment in all sensory modalities,
mainly affecting sensibility for temperature and noxious
stimuli, with no differences between the two SCI groups.
Tables 3, 4 and 5 shows thresholds for SCI patients compared
with controls. There were no differences in any of the
thresholds between pain and pain-free patients (Table 3).
Three pain patients did not report pain in the area tested (nos
2, 6 and 9). We therefore compared the 17 patients with pain
in the area tested with six patients with paraesthesias and 17
with no abnormal sensations in the area. There were no
differences in any of the thresholds between these groups, or
between the 17 patients with pain compared with 23 patients
without pain in that area (Kruskal±Wallis and Mann±Whitney
U test).
Spinothalamic function
This was de®ned as temperature or pinprick sensation. Three
patients in the pain group and one in the pain-free group had
detectable spinothalamic function (non-signi®cant difference) (Tables 4 and 5). Two patients had preserved function
for both warm and cold; both were in the pain group. One of
these could not feel warm in the normal range but suddenly
felt a strong burning painful heat at the level of tolerance; the
other had warm thresholds within the range of controls, and
heat pain and pain tolerance thresholds below controls (heat
allodynia and heat hyperalgesia); both had decreased cold
thresholds. One patient (pain free) had detectable cold but no
warm sensation within the temperature range. Pinprick
sensation was preserved in four patients (Tables 4 and 5).
All three patients in the pain group with signs of preserved
spinothalamic function had pain in the area tested.
Dorsal column function
Dorsal column function as de®ned by preserved knowledge of
movement or position or a de®nite cerebral response of the
SEP was preserved in four pain patients and ®ve pain-free
patients (non-signi®cant difference) (Tables 4 and 5).
Vibration sense
Vibration sense as a proper dorsal column function has been
questioned (Cook and Browder, 1965; Calne and Pallis,
1966). Seven pain and six pain-free patients had a detectable
vibration threshold. If the vibration sense was included in
dorsal column function, seven pain and seven pain-free
62
N. B. Finnerup et al.
patients had some degree of preserved dorsal column function
(Tables 4 and 5).
Sensory hypersensitivity
History
Eight pain and no pain-free patients reported allodynia
(pain evoked by normally non-painful stimuli) (P = 0.003,
Fisher's exact test) (Table 2). All eight patients had allodynia to touch, and one also reported cold allodynia. Only
one patient described allodynia in the area of central
pain, while seven described allodynia in dermatomes at
lesion level. Additionally, one patient described more widespread allodynia in periods with severe pain, e.g. in these
periods his face is so sensitive that he cannot shave.
Sensory function and central pain in spinal cord injury
Furthermore, several patients reported that light touch
provoked spasms, some of which were painful. Patients
with allodynia (n = 8) had a signi®cantly higher intensity of
63
spontaneous pain (median 7, range 3±8) than patients who did
not report allodynia (median 4, range 3±7) (P = 0.031, Mann±
Whitney U test).
Fig. 1 (A) Area of absent or impaired sensation to pinprick or light touch. The examination was done in two key points in each of 28
dermatomes according to the ASIA classi®cation (Maynard et al., 1997). (B) Area of spontaneous pain in 20 SCI pain patients.
64
N. B. Finnerup et al.
Evoked pain or dysaesthesia below level
In the areas tested, four pain patients had evoked pain (three
had cold allodynia when examined with the thermo roll, three
had pinprick hyperalgesia and three had pain evoked by
repetitive pinprick) (Table 4). No pain-free patients had
evoked pain below injury (Table 5).
Evoked pain or dysaesthesia at injury level
Hyperaesthesia. When screening the border zone, signi®cantly more pain patients than pain-free patients (14 versus
six) reported hyperaesthesia to one of the stimulations applied
(brush, pinprick or cold) (P = 0.026, Fisher's exact test).
Brush-evoked allodynia or dysaesthesia. Nine pain and no
pain-free patients had brush-evoked pain or dysaesthesia
(P = 0.001, Fisher's exact test). The intensity of brush-evoked
pain or dysaesthesia at lesion level was signi®cantly higher in
SCI patients with pain than in patients without pain (P = 0.001,
Mann±Whitney U test) (Fig. 2).
Cold-evoked dysaesthesia. Dysaesthesia to an acetone droplet
was seen in two pain and no pain-free patients (nonsigni®cant difference).
Pinprick hyperalgesia. Nine pain versus three pain-free
patients had pinprick hyperalgesia (non-signi®cant difference). The intensity of pinprick hyperalgesia at lesion level
was signi®cantly higher in SCI patients with pain than
patients without pain (P = 0.028, Mann±Whitney U test)
(Fig. 2).
Pain to repetitive pinprick. Seven patients in each group had
pain evoked by repetitive pinprick (Tables 4 and 5), and no
difference was found in intensity of pain evoked by repetitive
pinprick between the two patient groups (Fig. 2). The evoked
pain was clearly different from several segments above the
border zone. One patient had pain evoked by repetitive
pinprick that did not differ from above injury (face), and he
was not included as having pain evoked by repetitive pinprick
at level. Pain evoked by repetitive pinprick was present
typically at and just above the border zone, and the sensation
of pain started within the ®rst seconds.
Correlation between at level evoked sensations
and below level spontaneous pain
Within the pain group, 10 patients had brush- and/or coldevoked dysaesthesia and/or pinprick hyperalgesia (two
patients with wind up-like pain only not included) compared
with three patients without pain (P = 0.041, Fisher's exact
test) (Tables 4 and 5). The 10 pain patients with evoked pain
or dysaesthesia at lesion level had signi®cantly higher
intensities of spontaneous pain and higher intensities of
spasms than the 10 patients without evoked pain at level
(P = 0.03 and 0.004, respectively, Mann±Whitney U test), and
more patients with evoked pain reported paroxysms (seven
versus one, P = 0.044, Fisher's exact test). The intensity of
evoked sensations to brush at lesion level was signi®cantly
correlated to the intensity of spontaneous central pain below
lesion level (Spearman's r = 0.60, P = 0.005, n = 20), and to
intensity of spasms (Spearman's r = 0.65, P = 0.002, n = 20).
The intensity of pinprick hyperalgesia was correlated to the
intensity of spasms (Spearman's r = 0.61, P = 0.004, n = 20)
but not to the intensity of spontaneous pain (P = 0.08). The
intensity of wind up was not correlated to the intensity of
spontaneous pain or to spasms (P = 0.78 and 0.75, respectively).
Five patients (nos 1, 5, 10, 12 and 16, Fig. 1) had at level
spontaneous ongoing pain (de®ned as pain within two
dermatomes below the neurological level) in addition to the
below level pain. However, in all of these patients, the most
severe pain was the below level pain. Four of these ®ve
patients had evoked pain at level compared with six of 15
without at level spontaneous pain (non-signi®cant difference). To study if the co-existence of at level ongoing pain
had any in¯uence on the results, we performed the calculations without these ®ve patients. The six pain patients with
only evoked pain or dysaesthesia at lesion level still had
signi®cantly higher intensities of spontaneous pain and higher
intensities of spasms than the nine patients without evoked or
spontaneous pain at level (P = 0.03 and 0.04), and now
intensity of evoked sensations to both brush and pinprick
were positively correlated to spontaneous pain and spasms
(P = 0.009 and 0.02, respectively, for brush, and P = 0.04 and
0.04, repectively, for pinprick).
Discussion
This study analysed pain, quantitative sensory measures and
somatosensory evoked responses in SCI patients with and
without pain below injury level and compared these with
®ndings in a gender- and age-matched control group without
injury. The main ®nding is a higher intensity of brush-evoked
dysaesthesia and pinprick hyperalgesia at the lesion level in
SCI patients with central pain as opposed to SCI patients
without pain below injury level, despite a similar degree of
lost somatosensory function in the two groups. These ®ndings
suggest that neuronal hyperexcitability expressed as hypersensitivity at the segmental lesion level is an important
mechanism underlying pain below injury level. This notion is
supported further by the demonstration of a signi®cant
correlation between brush-evoked pain intensity at lesion
level and spontaneous pain below lesion level of SCI.
Previous studies (Eide et al., 1996; Defrin et al., 2001),
including SCI patients with pain at and below lesion level,
have also suggested that neuronal hyperexcitability could
play a role in pain in SCI patients. However, since these
studies have included at level pain and lumbar lesions, a clear
distinction between central and peripheral components of the
pain is dif®cult (Siddall et al., 1997).
Sensory function and central pain in spinal cord injury
65
Table 3 Threshold medians (ranges) at the dominant calf in controls, incomplete SCI pain
patients and incomplete SCI pain-free patients
VT (mm)
PPT (kPa)
PPTT (kPa)
SFT (kPa)
SFTT (kPa)
TDT (g)
(log 0.1 mg)
TPT (g)
(log 0.1 mg)
CDT (°C)
CPT (°C)
WDT (°C)
HPT (°C)
HPTT (°C)
Controls
(n = 20)
SCI pain patients
(n = 9)
SCI pain-free patients
(n = 7)
P*
2.27 (0.8±4.04)
501 (273±1200)
944 (489±1200)
347 (100±914)
744 (243±1200)
0.77 (0.41±5.74)
[3.89 (3.61±4.76)]
447 (17.5±447)
[6.65 (5.24±6.65)]
27.7 (23.3±28.6)
0 (0±25.4)
37 (32.5±44.5)
47.3 (41.6±49.4)
50.5 (47.3±50.5)
150 (0.45±150)
1200 (427±1200)
1200 (404±1200)
1200 (380±1200)
1200 (499±1200)
74.7 (1.4±447)
[5.87 (4.15±6.65)]
447 (284±447)
[6.65 (6.45±6.65)]
0 (0±21.7)
0 (0±0)
50.5 (32.5± 50.5)
50.5 (41.2±50.5)
50.5 (42.5±50.5)
18.8 (3.1±150)
1200 (233±1200)
1200 (373±1200)
1200 (378±1200)
1200 (960±1200)
197 (0.82±284)
[5.79 (3.91±6.45)]
447 (447±447)
[6.65 (6.65±6.65)]
0 (0±23.8)
0 (0±0)
50.5 (50.5±50.5)
50.5 (50.5±50.5)
50.5 (50.5±50.5)
0.78
1
1
0.92
0.69
0.93
0.48
1
1
0.48
0.47
0.48
*SCI pain versus SCI pain-free group, Mann±Whitney U test. CDT = cold detection threshold;
CPT = cold pain threshold; HPT = heat pain threshold; HPTT = heat pain tolerance threshold;
PPT = pressure pain threshold; PPTT = pressure pain tolerance threshold; SFT = skinfold pinch pain
threshold; SFTT = skinfold pinch pain tolerance threshold; TDT = tactile detection threshold to single
stimuli; TPT = tactile pain detection threshold; VT = vibration threshold; WDT = warm detection
threshold.
Neuronal hyperexcitability
Brush-evoked pain or dysaesthesia in the border zone at the
lesion level was only seen in pain patients, and in most
patients it was seen in areas with no spontaneous pain present.
The severity of brush-evoked allodynia in the area of
spontaneous pain has been correlated previously to intensity
of ongoing pain in neuropathic pain conditions, e.g. in postherpetic neuralgia (Rowbotham and Fields, 1996). However,
this is the ®rst study, to our knowledge, to show a correlation
between the intensity of evoked symptoms along the edge of
the damaged dermatome at the level of lesion, and intensity of
spontaneous below level pain and intensity of spasms.
Evoked pain and dysaesthesia in pain areas below lesion
level were only present in a few patients, all with incomplete
SCI. However, at level evoked pain and dysaesthesia were
present in both complete and incomplete SCI, which may
indicate that hyperexcitability is also present in complete
SCI. These ®ndings suggest that mechanisms sustaining
spontaneous pain and evoked dysaesthesia and motor excitability share common mechanism(s), and we believe that a
neuronal hyperexcitability due to loss of input to certain
populations of neurons and/or a lack of inhibition could be
responsible for this effect.
Brush-evoked pain (mechanical allodynia) is a classical
feature of peripheral neuropathic pain and is considered to be
a manifestation of central neuronal hyperexcitability driven
from peripheral abnormal afferent ®bres or from sensitized
dorsal root ganglion cells (Jensen et al., 2001). The allodynia
seen in patients with central pain is less clear. Previous
studies in post-stroke pain indicate that the pain is linked to
brush-evoked pain or dysaesthesia and increased thermal and
pinprick sensation in the painful denervated body parts
(Andersen et al., 1995; Vestergaard et al., 1995). These
observations have led to the proposal that post-stroke pain is
caused by the development of neuronal hyperexcitability in
neurons that have lost their normal patterned afferent input. A
similar mechanism may also play a role in the occurrence of
SCI pain in segmental body parts below the cord injury level.
The rostral level at which hyperexcitability responsible for
evoked and spontaneous pain occurs is not known, and may
include higher structures. In SCI pain patients, hyperexcitability in spinal (Loeser et al., 1968) and thalamic neurons
(Lenz et al., 1994) has been found. In animal studies of SCI,
spinal grey matter lesions and hyperexcitability of dorsal horn
neurons adjacent to spinal injuries have been related to at
level pain (Hao et al., 1992; Yezierski and Park, 1993;
Christensen and Hulsebosch, 1997; Drew et al., 2001). The
in¯uence of abnormal activity in the dorsal horn on below
level pain has been unclear. Animal studies point to a
relationships between at level and below level neuropathic
pain, and suggest below level neuropathic pain to be a result
of a combination of long tract deafferentation and an
activation of rostral targets by other sources of spinal activity
(for a review see Vierck et al., 2000).
Pharmacological studies lend support to the importance of
neuronal hyperexcitability for SCI pain. Lidocaine, which is
thought to inhibit nerve membrane hyperexcitability by
blocking voltage-sensitive sodium channels, and ketamine,
which blocks NMDA (N-methyl-D-aspartate) receptors and
thereby glutamatergic excitation, are effective in SCI at and
below level pain (Loubser and Donovan, 1991; Eide et al.,
1995; Attal et al., 2000). Increasing GABAergic inhibition
66
N. B. Finnerup et al.
Table 4 QST and SEP below the level of injury and evoked sensations below and at the level of injury in SCI patients with pain
SCI patients ASIA Level* Pain
Localize Direction SEP
with pain
class
intensity+ touch
sense
(n = 20)
Vibration Von Frey Pressure
Pinch pain Pinprick Cold
Warm
Evoked
Evoked
Repetitive
<150 gm 0.41±447 g pain
±1200 kPa
30±0°C 30±50.5°C pain below sensations pinprick
±1200 kPa
level²
at level³ at level§
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
±
±
¯
±
­
±
±
±
±
±
±
±
±
¯
®
¯
±
­
±
¯
A
B
C
A
D
B
A
A
A
B
A
A
A
B
D
D
A
C
A
A
C6
C6
C6
T4
C4
C5
T7
T4
C4
C5
T4
C5
C5
C5
T6
C4
C4
T3
T3
T6
3
3
6
7
7 (5)
6
7
8
4
7
4
7
4
3
8 (6)
4 (2)
4
5
4
5
±
±
¯
±
®
±
±
±
±
±
±
±
±
±
®
¯
±
±
±
±
±
±
±
±
®
±
±
±
±
±
±
±
±
±
®
¯
±
±
±
±
±
±
¯
±
¯
NA
±
±
NA
NA
±
±
±
NA
®
¯
NA
NA
±
±
±
±
¯
±
®
±
±
±
±
¯
±
±
±
±
¯
¯
±
¯
±
±
±
NA
NA
±
®
±
±
±
NA
±
±
±
±
±
±
®
±
NA
±
±
±
NA
NA
±
®
±
±
NA
NA
±
±
NA
±
±
±
®
±
±
±
NA
±
±
­
±
­
±
±
±
±
±
±
±
±
±
±
­
±
±
±
±
±
±
±
±
¯
±
±
±
±
±
±
±
±
±
±
¯
±
±
±
±
±
±
±
±
®
±
±
±
±
±
±
±
±
±
±
¯
±
±
±
±
±
+
+
+
±
±
±
±
±
±
±
±
±
±
+
±
±
±
±
±
±
±
+
+
±
+
+
±
+
±
+
±
±
+
+
+
±
±
+
±
±
+
±
±
±
±
+
+
±
±
±
±
±
+
+
+
±
±
+
*Neurological level; +parentheses indicate pain intensity on the last examination (at level hypersensitivity) in patients with a decrease in pain from ®rst to last examination; ²brush or
cold allodynia, pinprick hyperalgesia and/or pain evoked by repetitive pinprick below the injury level; ³evoked dysaesthesia or pain to brush or cold and/or pinprick hyperalgesia in
dermatomes at injury level; §pain evoked by repetitive pinprick at injury level. ASIA class: see footnotes to Table 2.
® = normal, ¯ = decreased, ­ = increased, ± = absent, + = present, NA = not accessible (if patients did not volunteer for SEP, or if pressure and pinch were not possible because of
evoked spasms or autonomic reactions).
Table 5 QST and SEPs below the level of injury and evoked sensations below and at the level of injury in SCI patients without pain
Vibration Von Frey
Pressure
Pinch
Pinprick Cold
Warm
Evoked
Evoked
Repetitive
<150 gm 0.41±447 g pain
pain
30±0°C 30±50.5°C pain below sensations pinprick
±1200 kPa ±1200 kPa
level²
at level³
at level§
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
¯
±
¯
±
±
±
¯
¯
±
NA
±
±
±
¯
±
±
±
®
±
±
C
A
D
A
A
A
B
C
A
A
B
A
A
C
A
A
A
B
A
A
C1
C5
C4
C6
C4
C4
C7
C6
T3
T3
C6
T3
T9
C4
T7
T7
T4
C6
T3
C3
See footnotes to Table 4.
±
±
¯
±
±
±
±
®
±
±
¯
±
±
¯
±
±
±
¯
±
±
±
±
®
±
±
±
±
®
±
±
¯
±
±
±
±
±
±
¯
±
±
±
±
¯
±
±
NA
NA
¯
±
NA
¯
±
±
±
NA
NA
NA
¯
NA
NA
¯
±
¯
±
±
±
®
¯
±
±
¯
±
±
¯
±
±
±
¯
±
±
±
NA
­
NA
±
±
±
±
±
±
±
NA
±
±
±
±
±
±
±
±
±
±
NA
NA
±
±
±
±
±
±
NA
±
±
®
±
±
±
±
±
±
±
±
¯
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
®
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
+
±
±
+
±
+
±
±
±
±
±
+
±
±
±
±
±
±
±
+
+
+
±
+
±
+
±
+
±
±
±
Sensory function and central pain in spinal cord injury
SCI patients ASIA Level* Localize Direction SEP
without pain class
touch
sense
(n = 20)
67
68
N. B. Finnerup et al.
Fig. 2 Intensity of evoked sensations at injury level in pain (n = 20)
and pain-free (n = 20) patients. Boxes represent medians with 25th
and 75th percentiles; error bars are 10th and 90th percentiles.
with either baclofen or propofol is also effective (Herman
et al., 1992; Canavero et al., 1995). Lamotrigine, an
antiepileptic drug affecting sodium channels and excitatory
amino acids involved in neuronal hyperexcitability, is also
suggested to be effective in SCI patients with spontaneous
and evoked neuropathic pain (Finnerup et al., 2002a).
In the present study, not all patients exhibited signs of
hyperexcitability. A similar lack of clinical hyperexcitability
has also been noted in post-stroke pain (Andersen et al., 1995;
Vestergaard et al., 1995), and in central neuropathic pain due
to multiple sclerosis (Svendsen et al., personal communication). While this may indicate that other mechanisms are at
play, another possibility is that the clinical measures are not
sensitive enough to detect hyperexcitability. Alternatively, it
is possible that hyperexcitability had been present at some
earlier time point after injury, since allodynia seems to have
an early onset (Siddall et al., 1999).
The spinothalamic and dorsal column pathways
and pain
This study shows that pain is present in those areas with lost
afferent input. This observation is similar to what is seen in
post-stroke pain, another central pain condition (Vestergaard
et al., 1995). The role of the spinothalamic tract in mediating
pain is well known (Willis and Westlund, 1997). Clinical
studies have also shown that lesioning of the spinothalamic
tract or its cortical projection sites plays an important role in
the development of central neuropathic pain (Boivie et al.,
1989; Vestergaard et al., 1995). In agreement with other
reports on SCI pain (Eide et al., 1996; Defrin et al., 2001), the
present study found that spinothalamic tract damage may be a
necessary but not suf®cient condition for the development of
SCI below level pain, since de®cits of spinothalamic
functions (pinprick and temperature sensation) were equally
severely affected in SCI patients without pain. Interestingly,
one previous study compared stroke patients with and without
central pain and found that central post-stroke pain patients
with supratentorial lesions had larger de®cits of Ad-®bre-
mediated sharpness and cold than pain-free stroke patients,
and infratentorial pain patients additionally had a de®cit of
C-®bre-mediated warmth and hot pain (Bowsher et al., 1998).
Their study also showed that allodynic central post-stroke
pain patients had greater de®cits for warmth than patients
without allodynia. A comparison between allodynia and
small ®bre function cannot be done in the present study where
only three patients had any signs of preserved thermal
sensibility.
The present study could not con®rm the imbalance theory
proposed by Beric et al. (1988), which suggests that
dysfunction of the spinothalamic tract and preservation of
the posterior columns are prerequisites for development of
central pain. This may be related to different methodology
used to access spinothalamic and dorsal column function or to
the criteria used for classi®cation. In this study, dorsal column
function was de®ned as preserved knowledge of movement or
position (Wall and Noordenbos, 1977; Nathan et al., 1986;
Danziger et al., 1996) or preserved SEP (Halliday and
Wake®eld, 1963; Dimitrijevic et al., 1983). Vibration sense
was not included in dorsal column functions since studies
have suggested that combined lesions of dorsal and posterolateral columns are necessary to abolish sense of vibration
(Cook and Browder, 1965; Calne and Pallis, 1966).
Furthermore, vibration threshold was dif®cult to determine
correctly because of projection of vibration to areas with
preserved sensation, lability of threshold, persistence of
sensation, and presence of spontaneous or evoked sensations
that were dif®cult to separate from true vibration sense.
Pinprick and temperature sensation were used as predictors of
the spinothalamic tract function (Willis and Westlund, 1997).
Further studies are needed to clarify the role and interaction
of spinothalamic tract and dorsal column in pain.
Conclusions
The present study shows pain in body segments below a
spinal cord lesion to be linked to the presence of abnormal
evoked sensations in segments at the level of injury,
suggesting that neuronal hyperexcitability in second or third
order neurons, which have lost their normal afferent input, is
an important mechanism for pain below spinal injury.
Acknowledgements
We wish to thank Dr Rana Das-Gupta for correction of the
manuscript. The study was supported by grants from Ludvig
og Sara Elsass' Foundation, the Institute of Experimental
Clinical Research Aarhus University, the Danish Medical
Research Council (nos 22000547 and 9700565), the Danish
Society of Polio and Accident Victims (PTU), the Foundation
for Research in Neurology and the Axel Klees Pain Research
Foundation.
Sensory function and central pain in spinal cord injury
References
Andersen G, Vestergaard K, Ingema-Nielsen M, Jensen TS.
Incidence of central post-stroke pain. Pain 1995; 61: 187±93.
Anke AG, Stenehjem AE, Stanghelle JK. Pain and life quality
within 2 years of spinal cord injury. Paraplegia 1995; 33: 555±9.
Ashworth B. Preliminary trial of carisoprodol in multiple sclerosis.
Practitioner 1964; 192: 540±2.
Attal N, Gaude V, Brasseur L, Dupuy M, Guirimand F, Parker F,
et al. Intravenous lidocaine in central pain: a double-blind, placebocontrolled, psychophysical study. Neurology 2000; 54: 564±74.
Beric A, Dimitrijevic MR, Lindblom U. Central dysesthesia
syndrome in spinal cord injury patients. Pain 1988; 34: 109±16.
Boivie J, Leijon G, Johansson I. Central post-stroke painÐa study
of the mechanisms through analyses of the sensory abnormalities.
Pain 1989; 37: 173±85.
Bouhassira D, Attal N, Brasseur L, Parker F. Quantitative sensory
testing in patients with painful or painless syringomyelia. In: Devor
M, Rowbotham MC, Wiesenfeld-Hallin Z, editors. Proceedings of
the 9th World Congress on Pain. Seattle: IASP Press; 2000. p.
401±10.
Bowsher D. Central pain: clinical and physiological characteristics.
J Neurol Neurosurg Psychiatry 1996; 61: 62±9.
Bowsher D, Leijon G, Thuomas KA. Central poststroke pain.
Correlation of MRI with clinical pain characteristics and sensory
abnormalities. Neurology 1998; 51: 1352±1358.
Brennum J, Kjeldsen M, Jensen K, Jensen TS. Measurements of
human pressure-pain thresholds on ®ngers and toes. Pain 1989; 38:
211±7.
Calne DB, Pallis CA. Vibratory sense: a critical review. Brain 1966;
89: 723±46.
Canavero S, Bonicalzi V, Pagni CA, Castellano G, Merante R,
Gentile S, et al. Propofol analgesia in central pain: preliminary
clinical observations. J Neurol 1995; 242: 561±7.
Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL.
Quantitative assessment of tactile allodynia in the rat paw. J
Neurosci Methods 1994; 53: 55±63.
Christensen MD, Hulsebosch CE. Chronic central pain after spinal
cord injury. [Review]. J Neurotrauma 1997; 14: 517±37.
69
allodynia after contusive spinal cord injury. Brain Res 2001; 893:
59±69.
Drewes AM, Helweg-Larsen S, Petersen P, Brennum J, Andreasen
A, Poulsen LH, et al. McGill Pain Questionnaire translated into
Danish: experimental and clinical ®ndings. Clin J Pain 1993; 9: 80±7.
Eide PK, Jorum E, Stubhaug A, Bremnes J, Breivik H. Relief of
post-herpetic neuralgia with the N-methyl-D-aspartic acid receptor
antagonist ketamine: a double-blind, cross-over comparison with
morphine and placebo. Pain 1994; 58: 347±54.
Eide PK, Stubhaug A, Stenehjem AE. Central dysesthesia pain after
traumatic spinal cord injury is dependent on N-methyl-D-aspartate
receptor activation. Neurosurgery 1995; 37: 1080±7.
Eide PK, Jorum E, Stenehjem AE. Somatosensory ®ndings in
patients with spinal cord injury and central dysaesthesia pain. J
Neurol Neurosurg Psychiatry 1996; 60: 411±5.
Finnerup NB, Johannesen IL, Sindrup SH, Bach FW, Jensen TS.
Pain and dysesthesia in patients with spinal cord injury: a postal
survey. Spinal Cord 2001; 39: 256±62.
Finnerup NB, Sindrup SH, Bach FW, Johannesen IL, Jensen TS.
Lamotrigine in spinal cord injury pain: a randomized controlled
trial. Pain 2002a; 96: 375±83.
Finnerup NB, Johannesen IL, Fuglsang-Frederiksen A, Gyldensted
C, Bach FW, Sindrup EH, et al. Sensory function in patients with
and without central neuropathic pain after spinal cord injury
[abstract]. In: Scandinavian Association for the Study of Pain, 25th
annual meeting. Denmark, 2002b, p. 39.
Folstein MF, Folstein SE, McHugh PR. `Mini-mental state'. A
practical method for grading the cognitive state of patients for the
clinician. J Psychiatr Res 1975; 12: 189±98.
Fuhrer MJ, editor. Rehabilitation outcomes: analysis and measurement. Baltimore: Brookes; 1987.
Goldberg JM, Lindblom U. Standardised method of determining
vibratory perception thresholds for diagnosis and screening in
neurological investigation. J Neurol Neurosurg Psychiatry 1979; 42:
793±803.
Gottrup H, Nielsen J, Arendt-Nielsen L, Jensen TS. The relationship
between sensory thresholds and mechanical hyperalgesia in nerve
injury. Pain 1998; 75: 321±9.
Cook AW, Browder EJ. Function of posterior columns in man. Arch
Neurol 1965; 12: 72±9.
Halliday AM, Wake®eld GS. Cerebral evoked potentials in patients
with dissociated sensory loss. J Neurol Neurosurg Psychiatry 1963;
26: 211±9.
Danziger N, Remy P, Pidoux B, Dormont D, Samson Y, Fournier E,
et al. A clinical and neurophysiological study of a patient with an
extensive transection of the spinal cord sparing only a part of one
anterolateral quadrant. Brain 1996; 119: 1835±48.
Hao JX, Xu XJ, Yu YX, Seiger A, Wiesenfeld-Hallin Z. Transient
spinal cord ischemia induces temporary hypersensitivity of dorsal
horn wide dynamic range neurons to myelinated, but not
unmyelinated, ®ber input. J Neurophysiol 1992; 68: 384±91.
Defrin R, Ohry A, Blumen N, Urca G. Characterization of chronic
pain and somatosensory function in spinal cord injury subjects. Pain
2001; 89: 253±63.
Herman RM, D'Luzansky SC, Ippolito R. Intrathecal baclofen
suppresses central pain in patients with spinal lesions. A pilot study.
Clin J Pain 1992; 8: 338±45.
Dimitrijevic MR, Prevec TS, Sherwood AM. Somatosensory
perception and cortical evoked potentials in established
paraplegia. J Neurol Sci 1983; 60: 253±65.
Jensen TS, Gottrup H, Kasch H, Nikolajsen L, Terkelsen AJ,
Witting N. Has basic research contributed to chronic pain
treatment? [Review]. Acta Anaesthesiol Scand 2001; 45:
1128±35.
Drew GM, Siddall PJ, Duggan AW. Responses of spinal neurones
to cutaneous and dorsal root stimuli in rats with mechanical
LaMotte RH, Shain CN, Simone DA, Tsai EF. Neurogenic
70
N. B. Finnerup et al.
hyperalgesia: psychophysical studies of underlying mechanisms. J
Neurophysiol 1991; 66: 190±211.
Lenz FA, Kwan HC, Martin R, Tasker R, Richardson RT,
Dostrovsky JO. Characteristics of somatotopic organization and
spontaneous neuronal activity in the region of the thalamic principal
sensory nucleus in patients with spinal cord transection. J
Neurophysiol 1994; 72: 1570±87.
Loeser JD, Ward AA Jr, White LE Jr. Chronic deafferentation of
human spinal cord neurons. J Neurosurg 1968; 29: 48±50.
Loubser PG, Donovan WH. Diagnostic spinal anaesthesia in
chronic spinal cord injury pain. Paraplegia 1991; 29: 25±36.
Maynard FM Jr, Bracken MB, Creasey G, Ditunno JF Jr, Donovan
WH, Ducker TB, et al. International Standards for Neurological and
Functional Classi®cation of Spinal Cord Injury. American Spinal
Injury Association. Spinal Cord 1997; 35: 266±74.
Melzack R. The McGill Pain Questionnaire: major properties and
scoring methods. Pain 1975; 1: 277±99.
Nathan PW, Smith MC, Cook AW. Sensory effects in man of
lesions of the posterior columns and of some other afferent
pathways. Brain 1986; 109: 1003±41.
Nielsen JF, Sinkjaer T, Jakobsen J. Treatment of spasticity with
repetitive magnetic stimulation; a double-blind placebo-controlled
study. Mult Scler 1996; 2: 227±32.
Rowbotham MC, Fields HL. The relationship of pain, allodynia and
thermal sensation in post-herpetic neuralgia. Brain 1996; 119: 347±54.
Siddall PJ, Taylor DA, Cousins MJ. Classi®cation of pain following
spinal cord injury. Spinal Cord 1997; 35: 69±75.
Siddall PJ, Taylor DA, McClelland JM, Rutkowski SB, Cousins MJ.
Pain report and the relationship of pain to physical factors in the
®rst 6 months following spinal cord injury. Pain 1999; 81: 187±97.
Siddall PJ, Yezierski RP, Loeser JD. Pain following spinal cord
injury: clinical features, prevalence, and taxonomy. IASP Newslett
2000; 3: 3±7.
Vestergaard K, Nielsen J, Andersen G, Ingeman-Nielsen M, ArendtNielsen L, Jensen TS. Sensory abnormalities in consecutive,
unselected patients with central post-stroke pain. Pain 1995; 61:
177±86.
Vierck CJ Jr, Siddall P, Yezierski RP. Pain following spinal cord
injury: animal models and mechanistic studies. [Review]. Pain
2000; 89: 1±5.
Wall PD. Letter to the editor [letter]. Pain 2001; 93: 197±8.
Wall PD, Noordenbos W. Sensory functions which remain in man
after complete transection of dorsal columns. Brain 1977; 100:
641±53.
Willis WD, Westlund KN. Neuroanatomy of the pain system and of
the pathways that modulate pain. [Review]. J Clin Neurophysiol
1997; 14: 2±31.
Yezierski RP, Park SH. The mechanosensitivity of spinal sensory
neurons following intraspinal injections of quisqualic acid in the rat.
Neurosci Lett 1993; 157: 115±9.
Received May 23, 2002. Revised July 26, 2002.
Accepted July 29, 2002