AophcTpa

M

a1mopet(Fpc(ed

RFAt71©d

7

Allodynia in Relation to Lesion Site in Central Post-Stroke Pain

David Bowsher

Abstract: Seventy-one percent of 122 patients with central post-stroke pain (CPSP) had allodyniathat was tactile-, cold-, or movement-evoked. Site of thalamic (and some infratentorial) lesions asrevealed by magnetic resonance imaging (MRI) was correlated in some cases with allodynia type andsensory perception threshold testing (QST). Notably, patients with cold allodynia tend to have moredorsally placed thalamic lesions than those without, and those with movement allodynia moreanteriorly placed lesions. Suggestions are made for improved correlation.Perspective: Only about half of patients with CPSP have allodynia (pain caused by innocuousstimulation); such stimulation is usually tactile- or cold-evoked or due to activation of stretch receptors(movement). We have found that, in some of our cases, the type of allodynia may depend on lesionlocation within the thalamus.

© 2005 by the American Pain Society

Key words: Allodynia, central post-stroke pain, lesion site.

R

amamfata

(mmigmtophp

aoapoatww

am

previously undescribed form of allodynia, elic-ited by isotonic or isometric muscle contraction,and called “movement allodynia,” was found to

ccur in a substantial number of patients with centralost-stroke pain (CPSP).4 We have also encountered, asave others, CPSP patients with thermal (particularlyold) allodynia4,7; and 1 patient with startle allodynia.ogether, such subjects make up a not-inconsiderableroportion of CPSP patients (39.5% of CPSP patients withllodynia; 29% of all our CPSP patients).

aterial and MethodsThe case notes of 122 CPSP patients who had been seen

nd clinically examined by the author between 1980 and990 were reviewed. Ninety-four patients underwentagnetic resonance imaging (MRI) in Liverpool, and 59f the images were independently analyzed and inter-reted by Dr. K-Å Thuomas2; another 2 had had comput-rized tomographic (CT) scanning only. One hundredwelve patients had also had quantitative sensory testsQSTs), in which perception thresholds for touch (vonrey filaments), vibration, sharpness,9 skinfold pinchain,13 warmth, cold, heat pain, cold pain, and warm-old limen10 were evaluated using a Somedic ThermotestSomedic, Stockholm, Sweden). In the course of clinicalxamination, the absence, or presence and type, of allo-ynia were tested and recorded in 108 patients.

eceived May 3, 2005; Revised June 20, 2005; Accepted June 23, 2005.rom the Pain Research Institute, Liverpool, UK.ddress reprint requests to David Bowsher, MD, ScD, Pain Research Insti-

ute, Clinical Sciences Building, University Hospital Aintree, Liverpool L9AL, UK. E-mail: bowsher@liv.ac.uk.526-5900/$30.002005 by the American Pain Society

ioi:10.1016/j.jpain.2005.06.009

36 The Journal of Pain, Vol 6, No 11 (N

esultsThirty-five of 122 patients (29%) had no discernible

llodynia (Figs 1, 2, and 3B). As previously reported, pureechanical allodynia was the commonest form (Figs 4

nd 5); it occurred in 34 patients (28%). In 9 patients,echanical allodynia occurred in association with some

orm of thermal (usually cold) allodynia; and in 7, it wasssociated with movement allodynia (Fig 6)—so that al-ogether, some 50 patients (41%) displayed mechanicalllodynia.Movement allodynia alone occurred in 10 patients

8%) (Fig 6); in 7, as stated above, it was associated withechanical allodynia. In the 10 patients with pure move-ent allodynia, pain occurred only on movement in 6

nstances, whereas in 5 others, there was moderate back-round pain which was severely exacerbated by move-ent. Three of these patients spontaneously asserted

hat they had no pain if they kept still and that pain onlyccurred when they moved the affected part (actively orassively); a fourth patient (a surgeon) stated that head mild ongoing pain, but this was “as nothing” to theain he experienced on movement.Pure cold allodynia, in the absence of other forms of

llodynia, was found in 10 patients (8%) (Fig 3A); in 11ther subjects, it was associated with some other form ofllodynia (usually mechanical), so that in toto, 17% ofatients displayed some sort of thermal allodynia. Manyf the subjects with cold allodynia averred that they feltn intense burning sensation when the body surface inhe affected area was contacted by a cold object, but nothen the object was at body temperature or slightlyarmer.One subject, with a discrete pontine infarct (Fig 5), had

llodynia to mechanical and acoustic startle (but not toechanical stimulation when she knew it was coming).Five patients with no allodynia, five with pure mechan-

cal allodynia, and two with mechanical plus cold allo-

ovember), 2005: pp 736-740

dno(

1mastcugism

tlla

psti(htclta

tdadTdsn

D

pm

F(fae

Fimtf

Faoc

737ORIGINAL REPORT/Bowsher

ynia (including one whose CPSP followed a subarach-oid haemorrhage), one with pure cold allodynia, andne with movement allodynia all had infratentorialbrainstem) infarcts with crossed symptoms.QSTs were available on 33 patients with no allodynia,

7 with pure mechanical allodynia, 7 subjects with pureovement allodynia, 6 subjects with pure cold allodynia,

nd the subject with startle allodynia. Tables 1 and 2how the results of these tests. Comparison of QSTs be-ween subjects with pure cold allodynia and those withold plus mechanical allodynia shows that the affected-naffected cold threshold difference is almost twice asreat in patients with cold plus mechanical allodynia as

n those with pure cold allodynia, and almost reachesignificance (P � .06), whereas A-U differences for otherodalities are very similar between the two groups.No radioanatomical differences could be discerned, ei-

her supra- or infratentorially, between patients with al-odynia and those without; whereas all 5 patients withacunes at the posterolateroventral tip of VPL had nollodynia, neither did 9 (Figs 1 and 4) out of 24 other

igure 3. A, Lesion in medulla oblongata. Parasagittal (left) andn right limbs. All mechanical and thermal thresholds raised and

edulla oblongata Parasagittal (left) and coronal (right) views.actile sensation; thermal sensation present in right face but abs

igure 1. Axial view of infarct in posteroventral VP causingching pain in buttock; no allodynia. Raised perception thresh-lds for von Frey, sharpness, and skinfold pinch; only slight in-reases in thermal thresholds.

ace and left limbs. Sweat production increased in right face and lef

atients with VPL lesions; but 10 out of 18 patients withubtentorial lesions had allodynia. MRI revealed no pat-ern to various forms of allodynia in the 18 patients withnfratentorial lesions. Eleven had crossed symptomslower brainstem lesions of Wallenberg type); of these, 5ad no allodynia (Fig 3B), 3 had tactile allodynia (Fig 3A),wo had cold allodynia, and one had both tactile andold allodynia. Of seven patients with upper brainstemesions (contralateral symptoms), two had no allodynia,wo had tactile allodynia, two had movement allodynia,nd one had startle allodynia (Fig 5).The greatest interest attaches to patients with supra-

entorial (thalamic) lesions and different forms of allo-ynia. Compared to lesions causing mechanical (tactile)llodynia (Fig 4), lesions causing pure movement allo-ynia tended to be further rostral and dorsal (Fig 6).here was a tendency also for patients with cold allo-ynia to have lesions in the dorsal part of the somato-ensory relay nucleus, though the number of good mag-etic resonance scans of such cases was very small.

iscussionThe reason(s) why some approximately half of CPSPatients have allodynia whereas the other half do notay be related to the factors that determine whether

igure 2. Pontine lesions: axial (left panel) and parasagittalright panel) views. No allodynia; burning/aching pain in leftace and body half; von Frey, sharpness, skinfold pinch, warmnd cold thresholds all elevated; hot pain threshold very slightlylevated.

nal (right) views. Cold allodynia; aching pain in left face, no painat production reduced in left face and right limbs. B, Lesion inllodynia; aching pain in right face and left body half. No loss of

left limbs; sharpness sensation present but diminished in right

coroswe

No aent in

t limbs.

potsnltn

psdcdocphl

pmfiu

bwad

epfiWsibt

ap(tcnnmmttcistlte

taslficfahet

C

d

F( Frey

Fca

738 Allodynia in Relation to Lesion Site in Central Post-Stroke Pain

atients with lesions in the somatosensory pathways dor do not have CPSP at all3,5—it should be rememberedhat Bogousslavsky et al2 have shown that only 25% ofubjects with lesions in thalamic VPL have CPSP. Exami-ation of 14 MRIs of patients with CPSP with clear VPL

esions reveals no distinguishable differences betweenhe 7 who have some form of allodynia and the 7 who doot have any allodynia.QSTs (Tables 1 and 2) reveal some differences betweenatients with particular kinds of allodynia that may beignificant. For example, patients with mechanical allo-ynia have a greater deficit for intense (small fiber) me-hanical modalities (sharpness and skinfold pinch) thano CPSP subjects with no allodynia, or with other formsf allodynia (cold, movement), whereas subjects withold allodynia have a greater deficit for cold than doatients with mechanical or no allodynia. There does not,owever, seem to be any explanation for the apparent

arge deficit for cold in subjects with movement allodynia.It is also worth emphasising, without seeking to ex-lain, that two forms of allodynia (tactile and move-ent) are elicited by the activation of large peripheralbres, whereas cold allodynia requires the activation ofnmyelinated fibres, as does startle.1

As stated above, there was no apparent correlationetween site of lesion and type of allodynia in patientsith brainstem lesions. There was also a failure to findny correlation between site of brainstem lesion and mo-ality/ies of sensory deficit.6 However, our patient who

igure 4. Axial (left panel) and parasagittal (right panel) views oUL�LL); tactile allodynia. Raised perception thresholds for von

igure 5. Right pontine lesion: parasagittal (left panel) andoronal views (right panel). Aching pain in left body half; tactile

dnd auditory startle allodynia; deficit for all modalities.

xperienced pain when startled by sudden and unex-ected tactile or auditory stimuli had a single well-de-ned pontine lesion (Fig 5), as had the case described byatson and Colebatch.14 Jankelowitz and Colebatch11

tate that “symptomatic increased startle occurred onlyn pontine lesions.” In the cat, the center for startle haseen shown to be in the pontine reticular formation,1 ando operate through an opioidergic mechanism in the rat.8

The deficits of superficial somatosensory function thatre associated with ventroposterior thalamic lesions de-end to some extent on whether the lesions are lateralventroposterior nucleus) or further medial (medial/in-ralaminar nuclei)6; the former show little deficit for me-hanical pain but considerable deficits for innocuous andoxious thermal stimuli, whereas more medial lesions doot affect (show little deficit for) the transmission ofechanically induced pain (skinfold pinch) or thermalodalities. Pure tactile allodynia is associated with ven-

rolateroposterior thalamic lesions (Fig 4). More ros-rally-situated lesions interfere with thalamic regionsoncerned with the processing of information originat-ng in group I primary afferents, and thus the finding oftretch-receptor-induced allodynia following lesions inhese regions is not unexpected (Fig 1). A number ofesions straddle the ventroposterior and ventrolateralhalamic nuclei, giving rise to mixed mechanical (brush-voked) and movement allodynia.Lenz and Dougherty12 found, by microstimulation in

he course of human stereotactic surgery, that there wastendency for sensations of coolness to be felt following

timulation in the dorsal part of the somatosensory tha-amic relay than in the ventral, which concurs with ourndings. In many instances, they found that tactile andold responses were obtained from the same unit. Weound 9 instances of patients with both tactile and coldllodynia; but only 1 CPSP patient was encountered whoad both cold and movement allodynia, and he had anxtensive lesion involving the SI and SII cortices as well ashe lateral thalamus.

onclusionWithin the thalamus, we have demonstrated a ten-ency for lesions at different sites to be associated with

rct in posteroventral VP causing aching pain in right extremities, skinfold pinch, and slight increases in thermal thresholds.

f infa

ifferent types of allodynia, though in our material, the

Fm

TCS

N

M

C

M

NS

T

N

N

N

M

M

C

N

A

*

739ORIGINAL REPORT/Bowsher

igure 6. Patient with movement allodynia, lesion in VL; coronal and parasagittal views. Some background pain in lower limb;ovement-induced pain like “red-hot poker down shin”; no von Frey deficit; some deficit for sharpness and thermal sensations.

able 1. Means of Somatosensory Perception Threshold Differences (Affected-Unaffected) forPSP With No Allodynia, Mechanical Allodynia, Movement Allodynia, Cold Allodynia, andtartle Allodynia

vs. FREY SHARPNESS SKINFOLD PINCH WARMTH COLD HOT PAIN

o allodynia 31 25 22 32 32 33N 1.2 � 0.9 2.3 � 1.7 0.8 � 1.1 6.5 � 5.3 6.9 � 8.3 2.5 � 4.0Median 1.1 2.15 0.6 5.9 4.8 1.5echanical allodyniaN 17 9 10 17 14 17Mean � SD 1.13 � 1.6 1.3 � 1.2 0.3 � 0.4 6.1 � 4.9 6.6 � 7.4 4.3 � 3.2

old allodynia 1.17 0.5 0.35 7.0 4.0 3.2N 7 6 3 7 7 7Mean � SD 1.0 � 0.8 2.7 � 1.8 2.4 � 4.2 6.1 � 5.5 4.0 � 5.9 2.7 � 5.5Median 0.84 2 0.2 5.5 3.4 1.0ovement allodyniaN 9 5 6 9 9 9Mean � SD 1.05 � 1.3 1.5 � 1.6 0.17 � 0.36 5.1 � 5.1 3.3 � 3.8 2.3 � 2.8Median 0.4 1.6 0.09 4.3 2.0 2.3

1 1 — 1 1 1

tartle 0.47 0.85 — 4.4 5.0 2.9

able 2. Comparison of Deficits in Allodynia: Unpaired Pvs. FREY SHARPNESS SKINFOLD PINCH WARMTH COLD HOT PAIN

o allodynia vsmechanical allodynia

0.4 NS 0.053 NS 0.03* 0.4 NS 0.45 NS 0.06 NS

o allodynia vs coldallodynia

0.5 NS 0.3 NS 0.3 NS 0.3 NS 0.2 NS 0.1 NS

o allodynia vsmovement allodynia

0.3 NS 0.2 NS 0.01* 0.2 NS 0.04* 0.4 NS

echanical allodynia vscold allodynia

0.46 NS 0.01* 0.2 NS 0.3 NS 0.2 NS 0.02*

echanical allodynia vsmovement allodynia

0.4 NS 0.4 NS 0.2 NS 0.3 NS 0.1 NS 0.06 NS

old allodynia vsmovement allodynia

0.4 NS 0.09 NS 0.2 NS 0.35 NS 0.1 NS 0.2 NS

o allodynia vs allallodynia

0.4 NS 0.4 NS 0.2 NS 0.4 NS 0.4 NS 0.2 NS

bbreviation: NS, not significant.

Significant to 0.05.

dotffa

A

1c

R

1c

2s1

3

4a

5S2

6ius

7pco

740 Allodynia in Relation to Lesion Site in Central Post-Stroke Pain

istinction is not hard and fast. Modern scanning meth-ds make possible very accurate localisation of lesions inhe brain (most of the MRIs used in this publication daterom before 1990). If associated with QSTs, anatomo-unctional understanding of patients with brain lesions

nd functional deficits will be greatly enhanced. T

gy 51:1352-1358, 1998

8lu

9sb5

1qJ

1i

1sn2

1t3

1f2

cknowledgmentsMost of the MRIs were performed between 1988 and

990 when they were funded by a grant from The Well-ome Trust (19009/1.5); further funding was provided by

he Pain Relief Foundation.

eferences

. Ascher P: La réaction de sursaut du chat anesthésié auhloralose. DSc thesis, Paris, 1965

. Bogousslavsky J, Regli F, Uske A: Thalamic infarcts: clinicalyndromes, etiology, and prognosis. Neurol 38:837-848,988

. Bowsher D: Central pain. Pain Rev 2:175-186, 1995

. Bowsher D: Central pain: clinical and physiological char-cteristics. J Neurol Neurosurg Psychiat 61:62-69, 1996

. Bowsher D: Central neuropathic pain, in Willis RW,chmidt RF (eds): Encyclopaedia of Pain. Berlin, Springer,005, in press

. Bowsher D: Somatosensory association and dissociationn pathways ascending from the spinal anterolateral funic-lus to the thalamus demonstrated by lesions in man. 2006,ubmitted

. Bowsher D, Leijon G, Thuomas K-Å: Central post-strokeain: correlation of magnetic resonance imaging with clini-al pain characteristics and sensory abnormalities. Neurol-

. Bowsher D, David J-M, Pellet J: Naloxone reverses reticu-ar habituation to high-intensity somatic and acoustic stim-li in the rat. J Physiol 398:26P, 1988

. Chan AW, Campbell JA, Bowsher D: Weighted pinprickensory thresholds: A simple test of sensory function in dia-etic peripheral neuropathy. J Neurol Neurosurg Psychiat5:56-59, 1992

0. Fruhstorfer H, Lindblom U, Schmidt WG: Method foruantitative estimation of thermal thresholds in patients.Neurol Neurosurg Psychiat 39:1071-1075, 1976

1. Jankelowitz SK, Colebatch JG: The acoustic startle reflexn ischemic stroke. Neurology 62:114-116, 2004

2. Lenz FA, Dougherty PM: Neurons in the human thalamicomatosensory nucleus (Ventralis caudalis) respond to in-ocuous cool and mechanical stimuli. J Neurophysiol 79:227-2230, 1998

3. Lynn B, Perl ER: A comparison of four tests for assessinghe pain sensitivity of different subjects and test areas. Pain:353-365, 1977

4. Watson SRD, Colebatch JG: Focal pathological startleollowing pontine infarction. Movt Disorders 17:212-218,

002