Sensory abnormalities in consecutive, unselected patients with central post-stroke pain

Pain, 61 (1995) 177-186 177 © 1995 Elsevier Science B.V. All rights reserved 0304-3959/95/$09.50

PAIN 2664

Research Papers Sensory abnormalities in consecutive, unselected patients

with central post-stroke pain

Karsten Vestergaard a,*, Jesper Nielsen b, Grethe Andersen a, Margrethe Ingeman-Nielsen c, Lars Arendt-Nielsen b and Troels S. Jensen d

Departments of a Neurology and c Radiology, Aalborg Hospital, DK-9100 Aalborg (Denmark) and b Department of Medical lnformatics, Aalborg University Center and d Department of Neurology, Aarhus University Hospital, Aarhus (Denmark)

(Received 22 March 1994, accepted 29 June 1994)

Summary This study examined the sensory abnormalities in an unselected, consecutive group of patients with central

post-stroke pain (CPSP) surviving more than 1 year after stroke. The sensory examination included clinical examination and quantitative measures with detection and pain thresholds to heat and cold stimuli, argon laser, von Frey hair and determination of stimulus-response function in the 10-45°C range. Sensory examination was in 11 identified CPSP patients (5 female, 6 male; aged 43-80 years) carried out in the painful area using the contralateral homologue area as reference. Pain rating was performed using the McGill Pain Questionnaire and a VAS scale.

All patients had ischemic (MRI verified) infarction. Of the 11 patients with supratentorial lesions, 5 had thalamic lesions; in addition, 7 patients had lesions in the brain stem/cerebellum. Median present spontaneous pain intensity on the VAS scale was 3.3 (range: 0-7.7).

All patients had pain in the body part with sensory abnormalities, which in 8 patients extended the area with pain. Warm detection threshold was higher in the pain area in all patients, and all except 1 patient had increased cold detection threshold. Cold and heat pain thresholds were raised as well, but to a lesser degree. Sensibility to touch (von Frey hairs) and pain (argon laser) were changed in only 4 and 3 patients, respectively. A stimulus-response curve in the 10-45°C range showed different pattems compared to the non-affected side. A cold allodynia in the 10-45°C range was present in the painful area in 6 (56%) of the patients.

The results support the theory that damage to the spino-thalamo-cortical pathway is a necessary condition in CPSP. It is proposed that the spontaneous pain in CPSP is linked to hyperexitability or spontaneous discharges in thalamic or cortical neurons that have lost part of their normal input.

Key words: Central post-stroke pain; Quantitative measure; Spinothalamic function; Allodynia; Central hyperex- citability

Introduction

Central post-stroke pain (CPSP) is a neuropathic pain syndrome, where the vascular lesion is assumed to be responsible for the pain (Dejerine and Roussy 1906; Schott et al. 1986; Mauguire and Desmedt 1988; Boivie et al. 1989; Leijon et al. 1989; Lewis-Jones 1990; Boivie

* Corresponding author: Dr. Karsten Vestergaard, Department of Neurology, Aalborg Hospital, P.O. Box 365, DK-9100 Aalborg, Denmark. Tel.: (45) 99-32-23-75; FAX: (45) 99-32-24-21.

1992). CPSP is characterized by pain in the body part corresponding to the entire or part of the cerebral lesion caused by the vascular accident. Although previously considered to be a rare consequence of stroke, a new study (Andersen et al. 1994) has indicated an CPSP incidence of 8% among unselected stroke patients. Since the original description by Dejerine and Roussy (1906) it has become known that the presence of sensory abnormalities is a characteristic feature of thalamic pain (a subgroup of CPSP). While the condition originally described as thalamic pain by Dejerine and Roussy (1906) has previously been associated with

SSDI 0304-3959(94)00140-5

178

lesions involving the thalamus, studies mainly by Boivie et al. (1989) have shown that extrathalamic lesions can give rise to central pain. Subsequent studies (Boivie and Leijon 1989) have indicated that CPSP patients almost invariably show a deficit of temperature and pain sensitivity while touch, vibration and other functions assumed to course mainly in the lemniscal pathways are less severely affected (Boivie and Leijon 1989). These observations have led to the suggestion that lesions along the spino-thalamo-cortical pathway may give rise to an CPSP syndrome. Since previous studies are based on selected CPSP patients, it is not possible to determine whether the presence of somatosensory deficits is a necessary condition for the occurrence of CPSP. Moreover, from selected cases, it is not possible to determine the extent of sensory deficit as measured by quantitative sensory testing. The aim of the presenl study was to study pain characteristics and sensory. abnormalities by quantitative measures in an unselected group of prospectively identified stroke patients in whom CPSP was found present by systematical examination at regular intervals during a l-year follow-up period after stroke onset.

Material and methods

Consecutive patients younger than 81 years of age with an acute stroke episode according to established WHO criteria (Aho et al. 1980) referred either to hospital or an outpatient clinic during a 1-year period (1 February. 1991 to 3l January 1992) were examined within 7 days after stroke onset and followed by regular intervals (I. 6 and 12 months after stroke). The details of the clinical examination are described elsewhere (Andersen et al. 1904). Patients with CPSP (defined as "pain after an undisputable stroke episode, where a peripheral neurogenic, nociceptive or psychogenic pain etiology was considered highly unlikely") were identified in the stroke population, All patients with CPSP surviving 1 year after stroke, and who were able to cooperate for additional studies, were invited to participate in the study. Among 267 patients with acute stroke 191 patients survived 1 year after stroke. Fifteen patients fulfilled the criteria fo~ CPSP and 13 of these were alive at the time of examination, which was carried out 17 months (median) after stroke onset (range: 13-23 months). In 2 patients it was not possible to do further reliable testing because of dementia, so 11 patients were examined. None of the patients received any pain relieving treatment at the time of examination. In 1 patient treated with Citalopram. a selective sero~ tonin re-uptake inhibitor, because of post-stroke depression the drug was discontinued 3 days before sensory examination. In 4 patients treated with paracetamol or clomipramin+mianserin because ~f pain, drugs were discontinued 1 day before examination.

Magnetic resonance imaging MRI scans were carried out on a Siemens Impact. 1.0 r in

circular polarized headcoil. T2-weighted double-echo sequences in the axial and saggital planes were obtained. The MRI scans were examined by an experienced neuroradiologist. To exclude small white-matter lesions of unknown significance (Braffman 1988) lesions visible on the proton-weighted sequences with a size greater than 5 mm were considered pathological and drawn on 3 standardized CNS levels: suprathalamic level, thalamic level and, if there were infraten- torial lesions, a drawing of the brain stem or cerebellum was added.

Pain rating Pain was rated using a VAS scale and the Danish version of the

McGill Pain Questionnaire (MPQ) validated on Danish pain patients (Drewes et al. 1993). For calculation of Pain Rating Index (PRI) the rank values were adjusted between 0 and 100% as suggested by Deschamps et al. (1988).

Sensory examinations The following clinical sensory tests were carried out: thermal

testing, sensation to pinprick and sensation to touch. Sensation was graded as either increased, normal or decreased using the supposed non-affected side as reference (Andersen et al. 1994). The area of sensory, change to pinprick and the area of pain distribution was drawn on a figure for each patient.

Sensory testing was done during daytime with patients sitting in a comfortable chair and in a quiet room with a temperature of 22°C. Sensory testing was carried out at the maximum pain site which in all patients involved the thenar eminence of the hand. The corresponding thenar on the opposite non-painful hand served as a control.

fhermal and pain thresholds Warmth, cold and pain detection thresholds to cold and heat

were measured with a computerized version of the Thermotest (Somedic, Sweden). The thermode consisted of series-coupled Peltier elements measuring 25 ×50 ram. Cut-off limits for warm and cold stimulations were 52°C and 10°C, respectively. For determination of warmth and cold detection thresholds, a baseline temperature of 30°C and l°C/sec rate of change was used. For the determination of heat and cold pain detection threshold a 2°C/sec rate of change was used. The subjects indicated when perception changed to warmth, cold or pain by pressing a push button. The thermode automatically returned to baseline temperature when the button was pressed. Both hands were first examined for warm and cold thresholds, and then for heat-pain and cold-pain thresholds. The examination sequence between affected and non-affected sides was changed randomly. The thresholds were calculated as the mean value of 8 determinations. In case of severe sensory loss, the subjects did not press the button before the cut-off limits were reached. To include these trials in the calculations of the thresholds 2 conditions were required. (1) If less than one-half the stimulations (3) reached the cut-off limit, these were included in the calculations as the cut-off limit (52°C or 10°C) (2) If one-half or more stimulations reached the cut-off limit, this temperature level was used as threshold. Accordingly for some threshold determination no standard deviation can be established.

To determine the non-pain (pain free) sensitivity range in which warm and cold was perceived, a sensibility index was used (Jensen et al, lq92).Warm sensibility, index (WSI)=

Heat pain detection threshold - Warm threshold

Heat pain detection threshold - Reference temperature

('old sensibility index (CS1)=

( o l d pain detection threshold - Cold threshold

("old pain detection threshold - Reference temperature

Qualitatit;e sensation The thermode was also used to study (1) the sensation perceived

and (2) the sensory intensity produced by the different thermode stimuli. The thermotester was regulated to a preset temperature and placed on the thenar eminence of the hand for approximately 5 sec. Subjects were then asked to rate both the sensation and intensity of the stimuli. We used in a random order 8 different temperatures ranging from 10°C to 45°C with an interval of 5°C. The modality of

the stimuli was grouped into 9 categories ranging from 'cold pain' to 'heat pain' (see Fig. 3). Pain intensity was marked on a VAS scale for each temperature.

Mechanical threshold Mechanical sensory detection threshold was tested with the use

of Von Frey hair (0.2-10.0 mg). The threshold was defined as the lowest intensity where more than 50% of stimuli were perceived. To examine if temporal summat ion to brush was present, subjects were also tested with von Frey hairs mounted at a rotating axis. The axis revolved at 0.16 and 1.5 Hz. If any of these revolution velocities were perceived as pain or unpleasantness , the actual revolution velocity limit for feeling pain was determined.

Laser threshold Pain threshold for brief argon laser stimuli was determined

according to the stair-case principle as described by Arendt-Nielsen and Bjerring (1988). Pain threshold was defined as distinct pinprick sensation. The argon laser stimuli covered an area of 7 mm 2 and had a duration of 200 msec (Spectra Physics, model 1683). The stimuli were all delivered at the thenar eminence of the hand.

Statistics Calculation of results with the Thermotes t for each patient was

considered distributed in a normal fashion, and mean values and standard deviation were calculated. Elsewhere non-parametr ic statistics have been used: Mann-Whitney 's test for comparison of un- paired data, Wilcoxon's test for paired data and Spearman 's rank correlation was applied for non-parametr ic correlation.

Results

Table I presents a summary of patients with CPSP. Six patients were male and 5 were female with a median age of 74 years (range: 43-80 years). All patients suffered an ischemic infarction. All patients had supratentorial lesions (5 thalamic, 6 solely extrathalamic) and 7 patients in addition had a brain-stem lesion. Fig. 1 and Table II show the lesions found on

TABLE I

S U M M A R Y OF CLINICAL D A T A

179

Patient no.

2 3 4 5 6

7 8 9 10 11

Right Left

Fig. 1. Lesions seen on MRI scans (proton-weighted sequences) from 10 patients with CPSP.

MRI scans. Pain onset ranged from 0 to 9 months (median: 2 months). Five patients had left-sided pain, 4 patients right-sided pain and 2 patients crossed pain distribution.

Patient Sex Age Pain Pain (no.) ( M / F ) (years) onset laterality

(months) ( L / R )

Variability Clinical sensibility examination

Touch Tempera ture Pinprick Allodynia

20°C 40°C Touch Cold

1 F 79 3 L I 1 1 1 1 1 1 2 M 47 0 R C 1 1 1 1 1 1 3 F 74 2 L I 1 1 1 1 0 1 4 M 74 3 R I 1 2 1 2 0 0 5 F 74 3 Cr C 1 1 1 1 0 1 6 F 80 1 R C 1 2 2 2 1 0 7 F 77 9 R C - 2 - - 1 1 8 M 78 1 L C 1 1 1 1 0 0 9 M 66 7 L I 1 2 1 1 1 0

10 M 51 2 Cr C 1 2 1 1 0 0 11 M 43 5 L C 1 1 1 1 1 1 Median 74 2

H, hemorrhage; I, infarction; M, male; F, female; L, left side; R, right side; Cr, crossed symptoms; I, intermittent pain; C, constant pain. Sensibility: 0 = normal, l = hypestesia, 2 = hyperestesia, - = not ratable. Allodynia: 0 = not present, 1 = present.

180 Patient no.

1 2 3 4 5 6

Area of sensibility change

Area of pain

Pat ient no.

7 8 11 9 10

Area of sensibility change

Area of pain

q

Fig. 2. Areas of sensibility change to touch and pinprick compared to areas of pain in 11 patients with CPSP.

The distribution of pain and sensibility changes is shown in Fig. 2 for each of the 11 patients. As can be seen the sensibility change affected one-half of the body in 8 patients while in 3 patients the sensory area was smaller. In 8 patients the area of pain was smaller than the area in which sensibility was changed. In the remaining 3 patients the distribution areas of sensibility and pain were identical. Thus in all cases the pain area was part of the area with sensory disturbance.

Pain rating Pain rating on the VAS scale ranged from 0 to 7.0

(median: 3.3) (Table III). Nine of 11 patients had pain

181

at the time of examination, while 2 patients were pain free.

MPQ could be completely fulfilled in 10 patients. The median number of words chosen (NWC) was 8 (range: 6-15), the median PRI (in %) was 21.5 (range: 12-35) and the median PPI was 2 (range: 0-4). Details of the MPQ rating will be presented elsewhere (Taagholt et al. in preparation).

Quantitative sensibility examination Thermotest (Fig. 3). All 11 patients had higher warm

detection thresholds on the affected extremity; in 9 of these patients the threshold was more than 2-fold

? @

E Ip Im

W a r m T h r e s h o l d C o l d T h r e s h o l d

I I

4 0 ~= 1

I,-

1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11

Subjects Subjects

[ ] Non-affected side

Affected side

Heat Pain Threshold Cold Pain Threshold

60 30 ? ?

• 50 • 20

40 ~ 10

30 0 1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11

Subjects Subjects

1,0

0,8

0 , e

0,4

0,2

0,0

Warm Sensib i l i ty Index

• A • • • • • •

A A

A • ~ . - ' - • A - , , • ,

0 1 2 3 4 5 6

Subjects

• A . A - ,

7 8 9

• Non-affected side

A Affected side Cold Sensib i l i ty Index

1'0 I • ~ ~ • 0,8 • •

o,6 A A A A u A A 0,4

h 0,2 • A

o , o • , -A- -A- i • i i • i • i • i i i • i . i 10 11 0 1 2 3 4 5 6 7 8 9 10 11

Subjects

Fig. 3. Results from Thermotest applied to thenar eminence in 11 CPSP patients. Affected sides are compared to non-affected sides for each patient. Baseline was 30°C. SDs are not shown in cases where no limit was reached in one-half or more of the examinations. Patients 1, 2, 3, 7 and 8: no warm detection threshold was recorded and the value shown is the same as for the heat pain threshold. Patients 2 and 3: no cold detection threshold was recorded and the value shown is the same as for cold pain threshold. WSI and CSI (defined in text) values are 0 in such

cases.

182

9 8

• ~ 6

¢ 4 ~3 2 1

i 1 • 1 i

1 0 2 0 3 0 4 0

T e m p e r a t u r e C

5n

"9-

6 - 5 - 4 - 3 - 2 - 1 -

J i i i i

10 20 30 40

Temperature 'C 50

9

8

• - - 6 -

a 5 -

¢ 4 -

1 -

9 " ~

8 ' I= 7 ' O ~., 6 " I I 5 " el = 4 •

1

9 "

78: == 5 i i ," 4

1

5 6

1 i 1 i i I i i

10 20 30 40 50 0 10 20 30 40

Temperature °C Temperature °C

7 8

u i i u - I

10 20 3o 40 50 Temperature C

9 8

• ,~ 6 5 4

1

_ °!! 0 20 30 0 50

Temperature °c

50

i i ! 1

10 20 30 40 50 Temperature °C

1 0

1 i ! i

10 20 30 40 Temperature °C

50

o

11

i u I !

1 0 2 0 3 0 4 0 5 0

T e m p e r a t u r e '~C

m ,,, Non-affected side

Af fec ted s ide

Sensation l-Painful Cold 4-Slightly Cold 7-Warm 2-Very Cold 5-Neutral 8-Very Warm 3-Cold 6-Slightly Warm 9*Painful Heat

Fig. 4. A stimulus-response curve in the 10-45%" range, applying the Thermotest at the thenar eminence of both hands. The qualitative experience is converted to numbers as indicated in the figure. Patient 7: due to aphasia the patient only said 'pain' in the 10-20°C range; this

statement is rubricated as I = painful cold.

TABLE II

MRI FINDINGS IN 12 PATIENTS W IT H CPSP

Patient ST T H BS

1 * unilateral 2 unilateral 3 unilateral unilateral 4 unilateral bilateral 5 bilateral

6 bilateral

7 unilateral unilateral 8 unilateral 9 unilateral

10 unilateral unilateral 11 unilateral unilateral Total l 1 5

pontine pontine pontine cerebellar bilateral pontine mesencephalic

pontine cerebellar

* Only CT scan available, showing a suprathalamic lesions. BS, brain-stem lesion; ST, suprathalamic lesion; TH, thalamic lesion.

higher. The median warm and detection threshold values on affected and non-affected sides were 43.5°C and 33.4°C, respectively (P = 0.001). Ten patients had lower cold detection thresholds, and in 8 of these patients the threshold was more than 2-fold lower. The median values on affected and non-affected sides were 20.6°C and 28.3°C (P = 0.003).

Temperature pain threshold values were affected as well, but less so. Heat pain threshold was higher on the affected side in 10 patients (P = 0.003). Median values were 48.9°C and 44.4°C. Examination of cold pain threshold detection was less informative, since 4 patients had no threshold on either side above examination limit (10°C). In 5 patients cold pain threshold was lower on the affected extremity and in 2 patients threshold was lower on the non-affected side (n.s.).

T A B L E 1II

PAIN C H A R A C T E R I S T I C S IN 11 PATIENTS W I T H CPSP

Patient Pain severity (0-5) VAS

(no.) Now W hen worst W hen least (0-10)

1 3 4 3 4.7 2 1 3 1 3.0 3 2 5 1 2.0 4 0 2 0 0 5 4 5 3 7.0 6 3 3 1 3.7 7 3 4 1 - 8 1 2 0 2.9 9 0 4 0 0

10 1 4 1 3.5 11 2 4 1 3.3 Median 2 4 1 3.3

* Could not cooperate.

183

i 1 0

1 2 3 4 5 6 7 8 9 10 11 Subjects

• Non-affected side

[ ] Affected side

Fig. 5. Sensibility to pain (pinprick) measured with argon-laser (W).

WSI was lower on the affected side in all patients, slight in 3 patients and marked in 8 patients. In 5 patients WSI = 0 because no warm detection threshold was found below the heat pain threshold. CSI showed almost equal values in 2 patients, lower values on the affected side in 8 patients, while in 1 patient CSI was lower on the non-affected side. Two patients had CSI values of 0, as no cold thresholds were recorded above the cold pain threshold. WSI and CSI showed no significant correlation to pain intensity measured by the VAS scale or PPI.

The qualitative experience of temperature stimuli in the 10-45°C range was examined fo~ patients 3-11. As can be seen in Fig. 4 there were marked individual differences. In 2 patients (nos. 4 and 6) experience of temperature sensation was almost similar on the 2 sides. Two patients (nos. 5 and 10) had scores closer to neutral on the affected side in the whole range. Three patients (nos. 3, 9 and 11) felt characteristically cold stimuli as burning painful. In 2 patients (nps. 7 and 8) stimuli were felt colder in the entire temperature range on the affected side. No relation could be seen between the different groups of patients and VAS or PPI.

Pain detection threshold (argon laser) (Fig. 5) Pain detection thresholds (sensation to argon laser) were markedly higher on the affected side in 3 patients (nos. 2, 5 and 7), while the other 8 patients showed equal or almost equal values on the 2 sides. No correlation was found between pain detection threshold and VAS or PPI.

E' 8 6

v- 2

0 1 2 3 4 5 6 7 8 9 10 11

Subjects • Non-affected side

[ ] Affected side

Fig. 6. Sensibility to touch measured with yon Frey hairs (mg). A bar height of 10 mg indicates that sensibility was greater than 5 g.

184

Touch detection threshold to con Frey hairs (Fig. 6) Three patients had markedly raised thresholds to yon Frey hairs (nos. 3, 5 and 8) on the affected side, while 1 patient had a higher threshold on the opposite side (no. 2). In 6 patients there was either little or no difference between the 2 sides. Two patients felt noth- ing on either side even with the thickest hair (10 mg). No significant correlation was found between touch detection threshold and VAS or PPI.

Sensibility to rotating Von Frey hairs Only 1 patient (no. 11) felt aggravation of pain on the affected side at a revolution velocity of 5 Hz and above.

Discussion

This study presents somatosensory functions in an unselected group of stroke patients with pain. Most of the patients had more than 1 lesion on MRI scans and some lesions were large, so precautions have to be drawn to conclusions on affected pathways, but in all patients the MRI scans indicated lesions of the rele- vant spino-thalamo-cortical system. This is in accordance with previous studies of CPSP (Schott et al. 1986; Boivie et al. 1989; Leijon et al. 1989; Lewis-Jones et al. 1990), where lesions may be located anywhere along the spino-thalamo-cortical projection pathway. It should be noticed that in 7 of 10 patients, in whom an MRI scan was performed, a brain-stem lesion was found. This is more frequent than in studies in which lesion sites have been based on CT scans and clinical examination (Schott et al. 1986; Leijon et al. 1989).

Although the character, duration, and onset of pain, and the lesions giving rise to these pains, varied consid- erably among patients there were striking similarities in their sensory disturbances. In accordance with previous studies carried out in selected patients with CPSP (Bowsher and Lahuerta 1985; Boivie et al. 1989) we found that patients with CPSP almost always have a somatosensory deficit. This deficit involves a deficit to cold and warm stimuli, and to pinprick, somatosensory functions associated with spino-thalamic pathways in the CNS. In addition, patients often suffer from other sensory deficits including vibration, touch, 2-point dis- crimination sensory functions mainly associated with lemniscal pathways in the CNS. In the present study 100% had increased threshold of thermal detection, 27% of pain threshold (argon-laser) and 18% of touch threshold (von Frey hairs). Boivie et al. (1989) in their study of 27 patients with CPSP reported threshold abnormalities to temperature detection in 100%, to temperature pain in 93% and to touch in 52% of patients. These data do suggest that lesion of the spino-thalamo-cortical somatosensory pathways is a necessary condition for a CPSP syndrome to occur and that CPSP can occur when lemniscal functions are

intact. However, is a lesion of the spino-thalamic system a sufficient condition for development of CPSP? The present data cannot answer this question in detail. From our previous study of stroke patients (Andersen et al. 1994) with somato-sensory disturbance but without pain it was seen that spino-thalamic disturbance as evidenced by changes in thermal sensation does occur in 61% of patients. Thus, these findings indicate that a lesion of the spino-thalamic system is a necessary but not sufficient condition for the development of CPSP syndrome.

The observation that touch sensation is altered in a large proportion of stroke patients without pain (Andersen et al. 1994) and that this modality is intact in some CPSP patients (present study and Boivie et al. 1989) suggest that lesions in the lemniscal system are neither a necessary nor sufficient condition for the development of CPSP.

The exact mechanisms triggering the pain are, however, still unknown.

Sensory abnormalities in CPSP The sensory abnormalities associated with CPSP do

not show a simple pattern. It is important to distin- guish between threshold abnormalities and evoked sensations.

Threshold abnormalities. An increased threshold to warm and cold detection is a consistent feature in this study as well as in the studies by Boivie and Leijon (Boivie et al. 1989; Boivie 1992). On the other hand, an abnormal mechanical detection threshold is a less consistent feature. It is also remarkable that only one- fourth of patients had a clearly elevated threshold for laser pain detection. The explanation for this is not clear, since these stimuli have been linked to activity in small C or A6 type afferent fibres (Arendt Nielsen and Bjerring 1988; Brennum et al. 1992) coursing centrally in the spino-thalamic system. It may be speculated that laser stimuli in addition to activating small fibres also recruit a few large fibres. Heat pain detection thresholds were less markedly affected than warm detection thresholds. The reason for this difference is not clear. One possibility is that noxious input, in addition to a contralateral projection, also has a more diffuse projection via the reticular formation (Willis 1985).

There is a discordance between the results from the clinical examination of sensation to touch and to pain, where all patients had changed sensibility, and the quantitative sensory examinations where 36% and 27%, respectively, had altered detection thresholds for touch (von Frey hairs) and pain sensation (as assessed by argon-laser). Differences in testing procedure may play a role. Although both the clinical and quantitative procedures attempt to assess touch and pain sensation, there are obvious differences in the procedures used. For example, stroking with cotton wool involves a

dynamic stimulation of a large area while sensation to von Frey touch determines sensation in punctate areas. While clinical pain is examined by means of a sharp needle, the quantitative pain threshold is determined by laser stimuli and it is not certain that perception of these two stimulus modes are engaging the same central neurons.

Evoked sensations Allodynia defined as pain evoked by a stimulus which normally does not provoke pain (Merskey et al. 1986) was tested by cotton wool stroking or thermorollers. Touch and cold allodynia were each found in 55% of patients. In the study by Boivie et al. (1989) touch and cold allodynia was seen in 5% and 23%, respectively, while Leijon and Bowsher (1990) found touch and cold allodynia in 28% and 42%, respectively. It is obvious that differences in patient material can play a role, but subtle differences in carrying out the sensory testing may likewise play a role. If evoked dysesthesia is included 88% had allodynia or dysesthesia to cold in the pain group with somatosensory deficits, but only 3% in the non-pain group with somatosensory deficits (Andersen et al. 1994).

We studied the evoked sensations to temperature stimuli in the 10-45°C range in 9 patients. In 78% of the examined patients abnormally evoked sensations were seen. These sensations ranged from an altered evoked cold or warm sensation to a clear paradoxical sensation of the various stimuli. A stimulus response curve was not established outside the 10-45°C range. In this range 56% of patients felt pain for cold stimuli in the affected hand at a temperature level, where pain was not present at the opposite side, indicating that allodynia for cold was present.

The significance of the sensory disturbance in CPSP Given the close association between somatosensory disturbance and pain, what is the significance of these sensory dysfunctions?

In patients with CPSP the pain was restricted to the area with altered somatic sensation to pinprick, which in size was either equal or smaller than the sensory deficit area. It is of interest to note that similar findings can be found in patients with other neuropathic pains. For example, in patients with phantom limb pain following amputation, hyperalgesia may extent several dermatomes proximal to the stump (Jensen and Ras- mussen 1994). In patients with a peripheral nerve lesion, the sensory abnormality may also extend beyond the area of pain (Lindblom and Verillo 1979); in spinal cord injury central dysestesia also involves an area equal to or larger than the body part where pain is located (Beric et al. 1988). These observations raise the possibility that common mechanisms for the sensory disturbances are shared by the various neuropathic pain syndromes. While little is known about the plastic mechanisms of sensory abnormality in CPSP, experi-

185

mental studies following peripheral nerve lesions have given clues about what is going on in some types of peripheral neuropathic pain. It has been shown that nerve damage gives rise to several changes in spinal cord neurons that have lost their afferent input including among others: a central hyperexitability with spontaneous neuronal activity, lowered threshold, a change in modality that drives 'pain producing' cells in the dorsal horn and an expansion of peripheral receptive fields from which these cells can be activated (Dubner 1991; Wall 1991; Bennett and Laird 1992).

We propose that in CPSP the sensory defect repre- sents the loss of spino-thalamic or thalamo-cortical input onto 3rd or 4rd order neurons and that the spontaneous and evoked pain is linked to development of hyperexitability of thalamic or cortical neurons that have lost part of their normal input similar to the hyperexitability observed in spinal cord neurons that have lost part of their normal input (Bennett and Laird 1992). That thalamic neurons may indeed show such hyperexitability has been shown by the elegant studies of Lenz et al. (1988, 1990) with the demonstration of increased neuronal activity in the ventro-posterior thalamus in patients suffering either from spinal cord injury pain or from phantom pain following limb amputation (for review see Lenz 1992; Boivie 1994). It is possible that the increased activity in the contralateral thalamus as determined by single-photon emission to- mography in 4 patients with CPSP may reflect such neuronal hyperactivity (Cesaro et al. 1991).

In conclusion the present study shows that CPSP has many of the essential characteristics of a neuropathic pain syndrome: spontaneous and evoked pain, pain confined to areas with sensory deficits and often allodynia or hyperalgesia. The lesions can be anywhere along the spino-thalamo-cortical projection and a dis- sociated sensory loss is a characteristic feature. Hyper- excitability related to central plasticity of neuronal pools in the CNS may be a possible mechanism for the pain.

Acknowledgements

This work has been supported by Grants from the Danish Medical Research Council (12-0828-1), the NOVO Foundation (1991-12-16), the Danish Cancer Society (78-4024) to T.S.J. and Andreas Marius Hansen (private donation) to K.V.

References

Aho, K., Harmsen, P., Hatano, S., Marquardsen, J., Smirnov, V.E. and Strasser, T., Cerebrovascular disease in the community: results of the WHO collaborative study, Bull. World Hlth Organ., 58 (1980) 113-130.

186

Andersen, G., Vestergaard, K, lngeman-Nielsen, M. and Jensen. T.S., The incidence of central post-stroke pain, Pain, 61 C1995) 187-193.

Arendt-Nielsen, I_.. and Bjerring, P., Sensory and and pain threshold characteristics to laser stimuli, J. Neurol. Neurosurg. Psychiat., 51 (1988) 35-42.

Bennet, G.J. and Laird, J.M.A., Central changes contributing to neuropathic hyperalgesia. In: W.D. Willis (Ed.), Hyperalgesia and Allodynia, Raven Press, New York, 1992, pp. 3115 310.

Berie, A., Dimitrijevic, M.A. and Lindblom, U., Central dysestesia syndrome in spinal cord injury patients, Pain, 34 (1988) 11)9-116.

Boivie, J., Leijon, G. and Johansson, I., Central post-stroke pain: a study of the mechanisms through analysis of the sensory abnormalities, Pain, 36 (1989) 173-185.

Boivie, J., Hyperalgesia and allodynia in patients with CNS lesions. In: W.D. Willis, Jr. (Ed.), Hyperalgesia and AIIodynia, Raven Press, New York, 1992, pp. 363-373.

Boivie, J., Central pain. in: P.D. Wall arid R. Melzack (Eds.). Textbook of Pain, Churchill Livingstone, Edinburgh, 1994. pp. 871-902.

Bowsher, D. and Lahuerta, J., Central pain in 22 patients: clinical features, somatosensory changes and (~f-scan findings, J. Neurol., 232 11985) 297.

Braffman, B.H., Zimmerman, R.A., Trojanowski, J.Q., Gonatas, N.K., Hickey, W.F. and Schlaepfer, W.W., Brain MR: pathologic correlation with gross and histopathology. Part 1. Lacunar infarction and Virchow-Robin spaces. Am. J. Neuroradiol., 9 (1988) 621-628.

Brennum, J., Arendt-Nielsen, L., Secher, NH., Jensen, T.S. and t~,jerring, P., Quantitative sensory examination in human epidur~ anesthesia and analgesia. 1. Effects of lidocaine, Pain, 51 11992) 27-34.

Cesaro, P., Mann, M.W., Moretti, J.L., Dfer, G., Rouldes, B., Nguyen. J.P. and Degos, J.D., Central pain and thalamic hyperactivity: a single photon emission computerized tomographic study, Pain, 47 (1991) 329-336.

Dechamps, M, Band, P.R. and Coldman, A.J., Assessment of adult cancer pain: shortcomings of current methods, Pain, 32 11988) 133-139.

Dejerine, J. and Roussy, G., La syndrome thalamique, Rev. Neurol.. 14 (1906) 521-532.

Drewes, A.M., Helweg-Larsen, S., Petersen, P., Brennum, J., An- dreasen, A., Poulsen, I,.H. and Jensen T.S., McGill Pain Ques- tionnaire translated into danish: experimental and clinical findings, Clin. J. Pain, 9 (1993) 80-87.

Dubner, R., Neuronal plasticity and pain following peripheral tissue inflammation or nerve injury. In: MR. Bond, J.E. Charlton and C.J. Woolf (Eds.), Pain Research and Clinical Management. Vol. 4, Proc. Vlth World Congress on Pain, Elsevier, Amsterdam, 1991, pp. 263-276.

.lensen, T.S., Bach, F.W., Kastrup, J., Dejgaard, A. and Brennum, I , Vibratory and thermal thresholds in diabetics with and without clinical neuropathy, Acta Neurol. Scand., 84 11991)326-333.

Jensen, T.S. and Rasmussen, P., Phantom pain and other phenom- ena after amputation, ln: P.D. Wall and R. Melzack (Eds.), Textbook of Pain, 3rd edn., Churchill Livingstone, Edinburgh, 1994, pp. 651-665.

I.cijon, G., Boivie, J. and Johansson, I., Central post-stroke pain: neurological symptoms and pain characteristics, Pain, 36 (1989) 13-25.

Leijon, G. and Bowsher, D., Somatosensory findings in central post-stroke pain and controls, Pain, Suppl. 5 (1990) 468.

Lenz, FA., Tasker, R.R., Dostrovsky, J.O., Kwan, H.C., Gorecki, J., Hirayama, T. and Murphy, J.T., Abnormal single-unit activity and responses to stimulation in the presumed ventro-caudal nucleus of patients with central pain. In: R. Dubner, G.F. Gebhart and M R Bond (Eds.), Proc. Vthe World Congress on Pain, Elsevier, Amsterdam, 1988, pp. 157-164,

Eenz, F.A, Kwan, H.C., Dostrowsky, J.O. and Tasker, R.R., Charac- teristics of the bursting pattern of action potentials that occurs in the thalamus of patients with central pain, Brain Res., 496 (1990) 357-3611.

Lenz. F.A., Ascending modulation of thalamic function and pain; experimental and clinical data. In: F. Sicuteri (Ed.), Advances in Pain Research and Therapy, Raven Press, New York, 1992, pp. 177- 196.

I.ewis-Jones, H., Smith, T., Bowsher. D~ and Leijon, G., Magnetic resonance imaging in 36 cases of central post-stroke pain (CPSPk Pain, Suppl. 5 (1990) 278~

Lmdblom, U. and Verillo, R,T., Sensory functions in chronic neural- gia, J. Neurol. Neurosurg. Psychiat., 42 11979) 422-435.

Mauguiere, F. and Desmedt. J.E., Thalamic pain syndrome of Dejer- ine-Roussy. Differentiation of four subtypes assisted by somatosensory evoked potentials data, Arch. Neurol., 45 (1988) 1312-1320.

Merskey, H., Lindblom, U., Mumford, J.M., Nathan, P.W., Norden- boos, W. and Sunderland, S., Pain terms. A current list with definitions and notes on usage, Pain. Suppl. 3 11986) 216-221.

Schott, B., Laurent, B. and Maugiere, F., Les douleurs thalamiques: 6rude critique de 43 cas., Rev. Neurol., 142 11986) 308-315.

Wall, P.D., Neuropathic pain and injured nerve: central mechanisms, Br. Med. Bull., 47 (1991) 631--643.

Willis, WD., Jr., Ascending nocieeptive tracts. In: The Pain System, Karger (Ed.), New York, 1985, pp. 145-206.

Woolf, C.J., Excitability changes in central neurons following peripheral damage: role of central sensitization in the pathogenesis of pain. In. W.D. Willis, Jr. (Ed.), Hyperalgesia and Allodynia. Raven Press, New York, 1992, pp. 221-243.

Sensory abnormalities in consecutive, unselected patients with central post-stroke pain

Documents