Psychiatry Idiopathic symptomatic trigeminal · s4r2 s5r2 W3 N7 s3r3 P4 P6 WI logy was identified bymeans other than elec-trophysiological trigeminal testing. Seven patients had postherpetic

Journal ofNeurology, Neurosurgery, and Psychiatry 1990;53:1034-1042

Idiopathic and symptomatic trigeminal pain

G Cruccu, M Leandri, M Feliciani, M Manfredi

AbstractThe trigeminal reflexes (corneal reflex,blink reflex, masseter inhibitory periods,jaw-jerk) and far field scalp potentials(nerve, root, brainstem, subcortical)evoked by percutaneous infraorbitalstimulation were recorded in 30 patientswith "idiopathic" trigeminal neuralgia(ITN) and 20 with "symptomatic" tri-geminal pain (STP): seven postherpeticneuralgia, five multiple sclerosis, fourtumour, two vascular malformation, oneTolosa-Hunt syndrome, and one trau-matic fracture. All the patients with STPand two of those with ITN had tri-geminal reflex abnormalities; 80% ofpatients with STP and 30% of those withITN had evoked potential abnormalities.The results indicate that 1) trigeminalreflexes and evoked potentials are bothuseful in the examination of patientswith trigeminal pain, and in cases secon-dary to specific pathologies provide100% sensitivity; 2) in "symptomatic"and "idiopathic" paroxysmal pain theprimary lesion affects the afferent fibresin the proximal portion of the root or theintrinsic portion in the pons; 3) primarysensory neurons of the A-beta fibregroup are involved in both paroxysmaland constant pain, but in the latter thedamage is far more severe.

University ofRome"La Sapienza", RomeDepartment ofNeurosciencesG CruccuM ManfrediNeuroradiologicalUnit IIM Feliciani

University of Genoa,Genoa, Italy,Department ofClinical NeurologyM LeandriCorrespondence to:Dr Cruccu, DipartimentoScienze Neurologiche, VialeUniversita 30, 1-00185Roma, ItalyReceived 6 Novemberl989and in revised form 7 March1990.Accepted 15 March 1990

Trigeminal neuralgia is a highly frustratingcondition for both patient and physician. Des-pite being one of the most painful syndromes,clinical exmination fails to show pain pathwayabnormalities, and the aetiology is unknown.A lesion can be identified occasionally, but a

thorough neuroimaging study usually provesfruitless.Although neurophysiological techniques do

provide objective data on sensory pathways, intrigeminal neuralgia the results are disappoin-ting or contradictory. Electromyographic(EMG) recordings of the blink reflex and jaw-jerk in these patients have been repeatedlyreported to be normal.'2' In contrast, record-ings of scalp potentials evoked by transcutan-eous stimulation of the lip, gum and mentalnerve have shown a significant delay of latecomponents (beyond 10 ms);456 after surgicallesions to the trigeminal ganglion or root,these potentials even appear to improve orbecome normal.67 However, the presynapticwaves (below 3 ms) of the scalp potentialevoked by percutaneous stimulation of the

infraorbital nerve, are rarely abnormal inpatients with trigeminal neuralgia.89Disagreement arises chiefly because com-

parable data are lacking; reports are based onvaried neurophysiological techniques, whichexamine different sets of fibres and trigeminaldivisions. Technical controversies complicatethe issue. The blink reflex provides informa-tion on A-beta afferents of the ophthalmicdivision,'0 and the jaw-jerk on A-alphaafferents of the mandibular division alone.2The scalp potentials evoked by transcutan-eous stimulation are obscured, in the initial 8-10 ms, by artifacts due to direct excitation ofmuscles, and from 9-11 ms onwards by reflexactivation of the same muscles." Scalp poten-tials of uncontaminated neural origin can onlybe provided by stimulation of a trigeminalcutaneous nerve in its bone canal; a techniquehas been developed for the infraorbital nerveof the maxillary division.'2 The potential con-sists of a series of waves from 1 to 7 mis,originating from the maxillary nerve (WI),trigeminal root (W2), presynaptic intratruncalfibres (W3), postsynaptic brainstem and sub-cortical fibres (P4-N5-P6-N7)."1114

Because neurophysiological techniques can-not be applied effectively to the study oftrigeminal pain without solving these ques-tions, we studied both trigeminal reflexes andevoked potentials, concurrently. To ensurethat our results were complete and reliable, werecorded the whole series of trigeminalreflexes, thus testing different sets of fibresfrom all the three divisions,'5 as well as bothpre- and postsynaptic far-field potentialsevoked by percutaneous stimulation of theinfraorbital nerve." Two groups of subjectswere studied comparatively: patients withtrigeminal pains, not necessarily paroxysmal,secondary to a well-documented pathology(symptomatic trigeminal pain, STP), andpatients with typical "tic douloureux" butwithout demonstrable lesions (idiopathictrigeminal neuralgia, ITN).

Material and methodsThe study was carried out in 50 patients,referred for one-sided "trigeminal pain". Allthe subjects had a clinical examination andCT scans; some of them were also evaluatedby magnetic resonance imaging (MRI),angiography, cerebrospinal fluid (CSF)examination, visual evoked potentials, orbrainstem auditory evoked potentials.The STP group comprised 20 patients,

aged 25-77 years, in whom a definite patho-

1034

on May 23, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.53.12.1034 on 1 D

ecember 1990. D

ownloaded from

http://jnnp.bmj.com/


Figure 1 Trigeminalreflexes and evokedpotentials in a normalsubject. Electricalstimulation of the cornea(sl), supraorbital nerve(s2), infraorbital nerve(s3), mental nerve (s4);tap to chin (s5).Recordingfrom theorbicularis oculi muscle(rl), masseter muscle(r2), and scalp Cz-Cv7(r3). CR: corneal reflex;Rl and R2: early and lateblink reflex. Two trials aresuperimposed. Calibration10 ms/200 jiV. SPI andSP2: early and latemasseter silent period.Average of 16 rectifiedtrials. Calibration 20 msl200 pV. JJ: Jaw-jerk.Two trials aresuperimposed. Calibration5 ms/l mV. Wl-W2-W3:presynaptic waves(maxillary nerve, root,pons), P4-N5-P6-N7:postsynaptic waves(brainstem andsubcortical) of the evokedpotential. Average of 1000trials. Calibration 1 ms/2p V. Left: responses tostimulation of the left side.Right: responses tostimulation of the rightside.

Right

CR

sirl

s2rl

s3r2

s4r2

s5r2

W3 N7

s3r3

P4 P6

WI

logy was identified by means other than elec-trophysiological trigeminal testing. Sevenpatients had postherpetic neuralgia, five hadmultiple sclerosis, four had benign tumours ofthe cerebello-pontine angle, two had vascularmalformations in the posterior fossa (one abasilar artery ectasia, and the other an arterio-venous malformation), one had a Tolosa-Huntsyndrome (''painful opthalmoplegia": retro-orbital aching pain and deficit of the 3rd, 4th,5th, and 6th cranial nerve), and one had atraumatic fracture of the maxillary bone. TheITN group consisted of 30 patients, aged 38-77 years, reporting pain typical of trigeminalneuralgia, in whom no underlying pathologywas discovered. The fourth column of thetable summarises the sensory symptoms.

All patients had electrophysiological testingof trigeminal function on each side of the face.

The following responses were recorded (fig1): the corneal reflex evoked by electricalstimulation of the cornea (fig 1, sl-rl), theearly (RI) and late (R2) blink reflex evoked byelectrical stimulation of the supraorbital nerve(s2-rl), the early (SP1) and late (SP2) mas-seter silent periods evoked by electricalstimulation of the infraorbital (s3-r2) andmental nerve (s4-r2), and the jaw-jerk to tapsto the chin (s5-r2). Trigeminal evoked poten-tials were obtained by percutaneous stimula-tion of the infraorbital nerve, by means of twoneedle-electrodes inserted into the infraorbitalcanal; responses were recorded from scalpelectrodes (vertex-spinous process of the 7thcervical vertebra); the three pre-synapticwaves (W1-W2-W3) and four post-synapticwaves (P4-N5-P6-N7) of the evoked responsewere examined (fig 1, s3-r3). Stimulation and

1035



j.com/

J Neurol N


ecember 1990. D

ownloaded from


Cruccu, Leandri, Feliciani, Manfredi

Table Clinical and neurophysiologicalfindings in 50 patients with trigeminal pain

Patient number Age, durationand diagnosis of illness Sensory symptoms Reflexes EPs

1 PHN 72 1 Vl constant pain RI, SPl delayed normal2 PHN 70 2 V2-V3 constant pain, hypoaesthesia SPI, SP2, JJ delayed all waves absent3 PHN 77 1 V1-V2 constant pain, dysaesthesia RI, SPI delayed all waves absent4 PHN 67 1 VI constant + paroxysmal pain, hypoaesthesia RI absent, CR delayed/reduced normal5 PHN 62 1 V1-V2 constant pain RI absent, R2, SPI delayed/reduced WI reduced, later waves

absent6 PHN 71 2 VI constant pain RI delayed, CR delayed normal7 PHN 63 1 V1-V2-V3 constant pain, hypoaesthesia RI, SPI delayed/reduced all waves absent8 MS 46 2 V3 paroxysmal pain all reflexes delayed/reduced W3 absent9 MS 25 1 LV2 paroxysmal pain, RV1-V2-V3 hypoaesthesia all reflexes abnormal bilaterally W3 delayed on the left,

absent on the right10 MS 21 1 VI paraesthesia JJ delayed, CR central defect normal11 MS 21 7 V2 constant pain SPI, SP2 delayed/reduced W3 absent12 MS 40 1 V1-V2-V3 paraesthesia RI delayed W3 delayed13 Tumour 62 1 V2-V3 paraesthesia RI, R2, SPI, SP2 delayed W3 delayed/reduced14 Tumour 57 1 V2 paroxysmal pain RI, SPI delayed W3 delayed15 Tumour 60 1 V1-V2 constant pain, hypoaesthesia RI absent, SPI, SP2 reduced W3 absent16 Tumour 68 1 V1-V2-V3 constant pain RI absent, SP1, SP2 reduced W2 delayed, W3 absent17 VascMalf 71 9 V1-V2-V3 paraesthesia, hypoaesthesia RI, R2, CR, SPI, SP2 delayed/reduced W3 absent18 VascMalf 46 1 V1-V2 paroxysmal pain SPI, SP2 reduced W3 delayed19 Tolosa H. 27 5 V1-V2 constant pain, hypoaesthesia RI absent, R2 reduced WI reduced later waves

absent20 Fracture 26 1 V2 dysaesthesia, hypoaesthesia SPI absent, SP2 delayed/reduced all waves absent21 ITN 77 8 V2 paroxysmal pain Ri, SPI delayed W3 absent22 ITN 76 6 V2-V3 paroxysmal pain SPI, JJ delayed W2 delayed, W3 absent23 ITN 57 25 Vl-V2 paroxysmal pain all responses normal W3 reduced/absent24 ITN 77 3 V2 paroxysmal pain, hypoaesthesia all responses normal W3 delayed25 ITN 66 5 VI paroxysmal pain all responses normal W3 absent26 ITN 38 1 V3 paroxysmal pain all responses normal W3 delayed/reduced27 ITN 75 5 VI-V2 paroxysmal pain, hypoaesthesia all responses normal W2 delayed28 ITN 65 1 V3 paroxysmal pain all responses normal W3 absent29 ITN 67 5 V2-V3 paroxysmal pain all responses normal W3 delayed

PHN: postherpetic neuralgia, MS: multiple sclerosis, ITN: idiopathic trigeminal neuralgia. VI, V2, V3: ophthalmic, maxillary, mandibular divisions of thetrigeminal nerve. RI and R2: early and late blink reflexes, CR: corneal reflex, SPI and SP2: early and later masseter silent period, JJ: jaw-jerk. WI, W2, W3: firstthree waves of the evoked potential. "Reduced" means reduced in magnitude, which may be the amplitude, the duration or the area, see Methods.Twenty-one cases of ITN are not shown in the table: the mean age was 53 years + 11, duration of illness 4-5 years + 5, reflex responses and evoked potentialswere normal in all subjects.

recording techniques for reflex and evokedresponses have been described previously." 15

The responses obtained after stimulation ofthe affected side were compared with thosefrom the healthy side. Latency and size(amplitude of the jaw-jerk, RI and evoked.potentials, duration of the corneal reflex andR2, area of SPI and SP2) were examined. Theupper limits of normality used in this studywere those reported by Cruccu et al 5 forreflex responses, and those of Leandri et al'6for evoked potentials.To study clinical-electrophysiological cor-

relations we grouped the patients into three"electrophysiological" categories (subjectswith all normal responses, with at least one

abnormal response, and with at least one

absent response), and two "pain" categories(subjects with paroxysmal pain, and thosewith constant pain, dysaesthesiae, or annoyingparaesthesiae). Age and duration of illnesswere also considered. Correlations amongparametric data were evaluated by r coeffientof Pearson, correlations among non-para-metric data by R coefficient of Spearman.Frequency differences were assessed by Chi-square test, and differences in parametric databy Student's t test

ResultsSynmptomatic trigeminal pain (STP)The table summarises findings in subjects withSTP (1-20). The patients with postherpeticneuralgia had constant pain, mainly of a burn-ing quaility. All these patients had severe

abnormalities of short-latency cutaneousreflexes (R1, SP1), and four (2, 4, 5, 6) also had

abnormal long-latency reflexes (CR, R2, SP2).In the three patients who had pain in theophthalmic division only, the evoked potentialswere normal (1, 4, 6); in the remaining fourpatients, who had pain extending to the max-

illary division, all the waves of the evokedpotential, including W1, were either absent or

severely altered (fig 2).Of the five patients with multiple sclerosis,

two reported the typical paroxysmal pain oftrigeminal neuralgia (8, 9); both short and long-latency reflexes were abnormal, as were W3and later waves of the evoked potentials.Figures 3 and 4 show the MRI scans andneurophysiological recordings in patient 8: onthe neuralgic side, there was a plaque in thedorsolateral pons close to the entry of thetrigeminal root, and W3 of the evoked potentialwas absent. Patient 9 also showed a facialsensory loss, and severe abnormalities of bothreflex and evoked responses, on the side con-tralateral to neuralgia. In patients 10, 11, and12, reporting constant pain or paraesthesiae,different reflexes were abnormal; thispresumably depended on the location of theplaques; the W3 of the evoked potentials wasalso abnormal, except in patient 10, who com-plained of annoying paraesthesiae confined tothe ophthalmic division.Two of the four patients with cerebello-

pontine angle tumour reported constant painand displayed severe neurophysiological altera-tions, including the absence of Rl and of W3(15, 16). The remaining two patients, one

complaining of annoying paraesthesiae and theother who had typical neuralgia, showed a

delay of the short-latency reflexes (RI, SP1)and of the W3 wave of the evoked potential.

1036



j.com/

J Neurol N


ecember 1990. D

ownloaded from



Figure 2 Trigeminalreflexes and evokedpotentials in a patient (no5) with postherpeticneuralgia. Stimulation,recording, calibration as infigure 1. On the affectedside, Rl is absent and R2delayed (s2, rl), SPI tostimulation of theinfraorbital nerve is absent(s3, r2), WI isconsiderably reduced inamplitude and the laterwaves are not reproducible(s3, r3).

Normal side

slrl

s2rl

s3r2

s4r2

s5r2

s3r3

Affected side

One of the two patients (18) with vascularmalformation in the posterior fossa, reportedparoxysmal pain typical of trigeminal neuralgiain the maxillary and ophthalmic division. TheMRI scans showed a basilar artery ectasia,which compressed the pons and distorted andlaterally displaced the trigeminal root near itsentry into the pons, on the neuralgic side (fig 5);the masseter SP1 and SP2 were abnormal, andthe W3 wave of the evoked potential was

delayed. In the other patient, suffering fromhypoaesthesia and dysaesthesia in the threedivisions, all reflexes were severely affected,and W3 and later waves of the evoked poten-tials were absent.

Patient 19, with the Tolosa-Hunt syndrome,and patient 20, with the maxillary bone frac-ture, had sensory deficits and constant pain or

dysaesthesia; the early and late reflexes were

considerably affected and the evoked potentials

absent or clearly reduced in amplitude from thefirst wave (W1) onwards.

Idiopathic trigeminal neuralgia (ITN)Of the 30 patients with ITN, only two (21 and22), showed slight delays of short-latencyreflexes (RI, SP1, JJ). In the other 28 cases, thetrigeminal reflexes were completely normal(see table).Evoked potentials were abnormal in nine

patients (table and fig 6). Abnormalities ofthese responses consisted in delay of W2 andsubsequent waves (22, 27), delay of W3 andsubsequent waves (24, 26, 29), or absence/reduction in amplitude ofW3 and subsequentwaves (21, 23, 25, 28); the W3 wave wastherefore affected in all nine cases. In none ofthe other 21 cases did any waves-from W1 toN7-exceed normal limits.

1037



j.com/

J Neurol N


ecember 1990. D

ownloaded from



Figure 3 MRI scans in apatient (8) with multiplesclerosis and trigeminalneuralgia. Above. SE2000/120 coronal image:shows a demyelinatingplaque (indicated byarrow) in the right lateralportion of the pons; otherfocal areas of high-intensity are visible in theperiventricularhemispheral white matter.Below. IR 1650/450 axialimage: the right pontinedemyelinating lesiondescribed above is locatedat the level of the entry ofthe right trigeminal root(indicated by arrow); bothtrigeminal roots are visiblein the prepontine cistern.

Clinical-neurophysiological correlationsThe correlation between the site of pain andneurophysiological tests was straightforward inthe STP group: almost all the patients withSTP in a certain division also showed abnor-malities of responses mediated by afferentsfrom the same division. Only in patients 12 and18 was neurophysiological testing normal indivisions involved by pain or paraesthesiae, butabnormal in the nearby affected divisions;more often there were abnormalities in div-isions that were clinically unaffected (1, 8, 9,13, 14, 18). Neurophysiological abnormalitieswere far less prominent in the ITN than in theSTP group, and the topographical correlationwas less clear; four patients had abnormalreflex (21) or evoked responses (25, 26, 28) indivisions that were clinically unaffected.

Severe neurophysiological abnormalities,often demonstrated by absence of responses,were seen in STP patients with constant pain.Conversely, in most patients with paroxysmalpain, whether symptomatic or idiopathic, theabnormalities were less severe, and often cons-

isted merely of a delayed response. Thefrequency of normal, abnormal, and absentresponses (see Methods) was significantly dif-ferent in the two groups of patients with or

without paroxysmal pain (p < 0-01).

No significant difference or correlation be-tween neurophysiological abnormalities andduration of illness was found for either group.In the STP patients, the quality of pain depen-ded not on age but on aetiology. The ITNpatients with abnormal neurophysiological res-ponses (mean age 66 years) were significantly(p < 0 01) older than those with normal res-ponses (52 years) or than the STP patients withparoxysmal pain (44 years). This was not duemerely to an age-related slowing ofconduction,since data refer to intraindividual side-differences. Clinical hypoaesthesia was foundin two ITN patients with abnormal electro-physiological responses, but also in three withnormal reponses.

DiscussionTrigeminal neurophysiology and its clinical sig-nificanceIn patients reporting pain in the trigeminalregion, neurophysiological testing of tri-geminal function was clinically useful. Abnor-malities were often disclosed in divisions thatwere clinically unaffected. An objectivedemonstration of dysfunction was provided inall cases with pain secondary to a documentedpathology (STP group), although severalpatients had no clinical sign other than pain.The trigeminal reflexes yielded a 100%

sensitivity, probably because they allowexamination of all the three divisions. The mostsensitive reflexes were the early blink reflex(R1) and the early masseter inhibitory period(SP 1). The scalp potentials evoked by percu-taneous infraorbital stimulation were abnormalin all but four STP cases. Of these, three hadpostherptic neuralgia involving the ophthalmicdivision only; the fourth patient, complainingof poorly-localised painful paraesthesiae overthe face, had multiple sclerosis and showedabnormalities of the corneal reflex and jaw-jerk.

In the majority of patients with ITN, allneurophysiological tests were normal. Onlytwo patients displayed mild reflex abnor-malities. Previous reflex studies by other auth-ors reached the same conclusions'2 but thesewere questioned since the method used did notallow examination of the maxillary division, themost common site of pain. The early waves ofthe evoked potential, and in particular the 3mswave (W3), were abnormal in nine ITNpatients, a proportion slightly higher than thatobserved in a previous study using the sametechnique but not examining the post-synapticwaves;9 in the remaining 21 patients, all thewaves, including both pre- and postsynapticfar-fields (1-7ms), were normal. These find-ings therefore do not support the conclusions ofstudies based on late waves evoked by surfacestimulations; in particular, the reportedrecovery after surgical manoeuvres on theganglion and root67 is hard to explain.

In our diagnostic protocol for patients withtrigeminal pain, we rely primarily on tri-geminal reflexes: the technique is less invasivethan that for evoked potentials and the findingof any abnormality implies an underlying

1038



j.com/

J Neurol N


ecember 1990. D

ownloaded from



Figure 4 Trigeminalreflexes and evokedpotentials in a patient (8)with multiple sclerosis andtrigeminal neuralgia.Stimulation, recording andcalibration as in fig 1. Onthe affected side, CR isdelayed and reduced induration (sl, rl), RlandR2 are delayed (s2,rl), SPI to infraorbitalstimulation is dalayed andSP2 is reduced in area(s3, r2), SPI to mentalstimulation is markedlydelayed and reduced (s4,r2), the jaw-jerk isdelayed (sS, r2), the Wland W2 are normal, thelater waves are notreproducible (s3, r3).

Normal side

CR

sir |

s2rl

s3r2

y0\{ VW{~ s4r2

s5r2

N5 N7

P4\_~~s3r3

P6

WI

Affected side

structural lesion. In cases with paroxysmalpain, the pre-synaptic waves of the evokedpotential are more sensitive, possibly because aslight slowing of conduction or loss of a fewaxons may not be sufficient to produce sig-nificant changes or reflex responses, which areinfluenced by the temporal-spatial summationat each synapse. Although the early waves ofthe evoked potential can indeed be abnormal inpatients with ITN, the finding of any abnor-mality should promote further investigation tosearch for a cause that may require surgicalattention; this holds particularly true in youngsubjects, since our ITN patients with abnormalevoked potentials were significantly older thanthose with normal responses or than STPpatients with paroxysmal pain.

Site of lesionThe recording of trigeminal reflexes may helpto indicate the site of lesion. On the basis of thebehaviour of ipsi- and contralateral responsesto one-sided stimulations, the reflex abnor-malities may be identified as "afferent","efferent" or "mixed";'7 the reflex circuits arelocated at different brainstem levels, midbrain(jaw-jerk), pons (RI, SP1), ponto-medullaryjunction (SP2), and lower medulla (CR, R2).'7In this study, most patients affected by STP,with constant or paroxysmal pain, showed"afferent" abnormalities of Ri and SP1. Allreflexes, however, share the entry of primaryafferents into the pons, and it is impossible totell where, between the nerve outlets on the faceand the brainstem, the lesion lies.

1039



j.com/

J Neurol N


ecember 1990. D

ownloaded from



Figure 5 MRI scans in apatient (no 18) withbasilar artery ectasia andtrigeminal neuralgia.Above. SE 450130 coronalimage: shows a windingbasilar artery, with right-sided loop which impingeson the pons. Middle.Gradient-echo 70, 350/15coronal image: the righttrigeminal root (indicatedby arrow) is displacedlaterally. Below.Gradient-echo 50, 350/15axial image: the basilarartery appears as a high-intensity structure, thehorizontal portion of thebasilar artery compressesthe right side of the ponsand laterally displaces theright trigeminal root(indicated by arrow) nearits entry into the pons.

Figure 6 Trigeminal evoked potentials in a patient (no29) with idiopathic trigeminal neuralgia. Stimulation,recording, calibration as in fig 1. On the affected side(below), Wl and W2 are normal, W3, P4 and N5 aredelayed by more than 0 5 ms without amplitudereduction, and the later waves are not reproducible.

The location of the lesion on the afferentbranch can be established more accurately byrecording the early potentials evoked by infra-orbital nerve stimulation. This is particularlytrue for the first three waves (WI, W2, W3),which are extremely constant and whose originhas been demonstrated by direct intraoperativerecordings.'3 1' Wave Wl, originating from themaxillary nerve, was altered in STP patients inwhom damage to that part of the trigeminalpathway was expected (Tolosa-Hunt syn-drome, maxillary bone fracture, postherpeticneuralgia). Of course, in these cases the sub-sequent waves were altered too. Other patho-logies, with more proximal damage to the

afferent pathway (tumours of the cerebello-pontine angle, vascular malformation, multiplesclerosis) caused alterations ofwave W3, whichoriginates from inside the pons. In all patientsthe pre-synaptic waves were in any eventabnormal, and consequently the primaryafferents were damaged.

Patients 8 and 18 are particularly interesting:both suffered from the typical paroxysmal painoftrigeminal neuralgia, the first abnormal wavewas W3, and MRI scans showed a lesionaffecting the root immediately before (fig 5) orafter (fig 3) its entry into the pons. The mostcommonly reported causes of "symptomatic"neuralgia are vascular anomalies in the pos-terior fossa and benign tumours of thecerebello-pontine angle (both impinging on theproximal portion of the trigeminal root) andmultiple sclerosis (with plaques in the rootentry zone).'8

In nine ITN cases, W3 and later waves wereabnormal. In only two cases, with an absentW3, W2 was delayed. None of the other ITNcases showed abnormalities of any later waves(P4-N7). In ITN, like cases ofneuralgia secon-dary to well-documented lesions, the impair-ment of conduction should therefore take placebetween the site of origin of W2 and W3, thatis, in the region of the root entry into the pons.This supports the view of authors who believethat ITN is induced by contact between theroot near its entry into the pons and benignlesions, including vascular anomalies, other-wise asymptomatic and undetected by radio-logical examination but found during surgery,or at necropsy.'120 Whatever the aetiology, theroot entry zone may be a "locus minoris resis-tentiae", since the myelin sheath provided bySchwann cells is replaced by oligodendroglia.As to the complex pathogenetic mechanism

1040



j.com/

J Neurol N


ecember 1990. D

ownloaded from



by which primary damage to trigeminalafferents eventually give rise to tic douloureux,our study by no means refutes a secondarydysfunction in the trigeminal nucleus, a theorywhich is supported by substantial evidence.2122

Implications for the pathophysiology oftrigeminal painMost patients showed abnormalities of the Rland SPI reflex responses or the W3 wave oftheevoked potential. Measurements of the afferentconduction velocity have shown that both RIand SP1 are mediated by intermediately-fastconducting fibres.3 10 23 Furthermore, testingthe effect of controlled lesions to the trigeminalnerve, Cruccu et al'5 found that these tworeflexes were intermediately susceptible to heatand compression, thus suggesting that theafferents belong to the medium-myelinatedfibre group (A-beta). The three pre-synapticwaves (WI, W2, W3) evoked by stimulation ofthe infraorbital nerve-a purely cutaneousnerve-display a similar threshold, onlyslightly above the sensory threshold; in theinfraorbital canal, the conduction velocity issimilar for the three waves, and if meaured at"peak" latencies, it averages some 41 m/s.9Maximum and "peak" conduction velocities ofthe maxillary division are 54 and 42 m/s."4These data indicate that the afferents for theevoked potentials are medium-myelinatedfibres.Of the long-latency reflexes, which were far

less frequently affected, only the corneal reflexis unquestionably mediated by small-myelin-ated afferents (A-delta), while R2 and SP2,though considered by some as "nociceptive"reflexes, are probably mediated by medium-myelinated afferents.'5 However, the cornealreflex was abnormal in some patients with non-paroxysmal pain, and always normal in patientswith paroxysmal pain; four of the latter had anabnormal R1, a finding that proves the in-volvement of first division afferents. Neithercan the apparent sparing of the cornealafferents be attributed to a poor sensitivity ofthe method or to a greater biological resistance,since the corneal reflex is even more vulnerablethan RI to damage from root or brainstemlesions. 17

A-beta, rather than A-delta fibres, thereforeappear to be involved in the pathophysiology ofparoxysmal trigeminal pain.Our comparison of the findings in cases with

constant and paroxysmal pain (table), showedthat A-beta afferents were more extensivelyand considerably affected in patients with con-stant than in those with paroxysmal pain. Thisagrees with the common notion that a dysfunc-tion of few fibres provokes paroxysmal pain,whereas severe damage does not; indeed,neuralgic pain is relieved by surgical deafferen-tation, constant pain is often worsened.

Different mechanisms have been proposedfor the generation of pain through a dysfunc-tion of the large-fibre input.2425 A markedreduction of large-fibre input produces dysin-hibition on nociceptive transmission, accord-ing to the "gating" of the input suggested byMelzack and Wall;25 a conspicuous deafferen-

sation per se, may also be followed by abnormalreinnervation, with new synapses on the wrongsecondary neurons.11 2' Dubner et al 28 recentlyproposed that the loss of few A-beta fibres mayresult in the expansion of the receptive field ofthe wide-dynamic-range neurons of the tri-geminal nucleus, so that bursts of pin-prick orelectric shock-like sensations can be evoked byinnocuous stimuli.28 Finally, neuralgic painmight result from demyelination of A-betaafferents. Mechanical injury to root fibres isthought to produce demyelinating zones whereextra spikes and after-discharges are gen-erated.29 To explain how anomalous bursts ofactivity in A-beta fibres cause pain, severalmechanisms have been proposed, includingephaptic transmission,30 31 excessive primaryafferent depolarisation of brainstem terminalsor unmyelinated axons,29 32 and chronic irrita-tion of wide-dynamic-range neurons.22

It is worth recalling, however, that A-betafibre damage could also be merely the coin-cidental landmark of a nerve lesion, damagethat cannot be demonstrated in other sets offibres. The use of the available methods toquantitatively assess the unmyelinated fibrefunction33 in patients with trigeminal pain,might provide significant information.

This paper is dedicated to Carlo Loeb to mark his retirement.The study was supported by the National Research Council ofItaly, grant 89 03923 CT04.

I Kimura J, Rodnitzky RL, Van Allen WM. Eletrodiagnosticstudy of the trigeminal nerve: orbicularis oculi reflex andmasseter reflex in trigeminal neuralgia, paratrigeminalsyndrome and other lesions of the trigeminal nerve.Neurology 1970;20:574-83.

2 Ongerboer de Visser BW, Goor C. Electromyographic andreflex study in idiopathic and symptomatic trigeminalneuralgias: latency of the jaw and blink reflexes. J NeurolNeurosurg Psychiatry 1974;37:1225-30.

3 Cruccu G, Bowsher D. Intracranial stimulation of thetrigeminal nerve in man. II. Reflex responses. J NeurolNeurosurg Psychiatry 1986;49:419-27.

4 Stoehr M, Petruch F, Schegelmann K. Somatosensoryevoked potentials following trigeminal nerve stimulationin trigeminal neuralgia. Ann Neurol 198 1; 9:63-6.

5 Bennett MH, Jannetta PJ. Evoked potentials in trigeminalneuralgia. Neurosurgery 183;13:242-7.

6 Drechsler F, Neuhauser B. Somatosensory trigeminalevoked potentials in normal subjects and in patients withtrigeminal neuralgia before and after thermocoagulationof the ganglion gasseri. EMG Clin Neurophysiol 1986;26:315-26.

7 Bennett MH, Lundsford LD. Percutaneous retrogasserianglycerol rhizotomy for tic douloureux. Part II. Neuro-surgery 1984;14:431-5.

8 Cruccu G, Leandri M, Favale E, Manfredi M. Trigeminalreflexes and evoked potentials in trigeminal neuralgia.Pain 1987;(suppl 4)S267.

9 Leandri M, Parodi CI, Favale E. Early trigeminal evokedpotentials in tumors ofthe base ofthe skull and trigeminalneuralgia. Electroencephalogr Clin Neurophysiol 1988;71:114-24.

10 Shahani BT. The human blink reflex. J Neurol NeurosurgPsychiatry 1970;33:792-800.

11 Leandri M, Parodi CI, Zattoni J, Favale E. Subcortical andcortical responses following infraorbital nerve stimulationin man. Electroencephalogr Clin Neurophysiol 1987;66:253-62.

12 Leandri M, Parodi CI, Favale E. Early evoked potentialsdetected from the scalp ofman following infraorbital nervestimulation. Electroencephalogr Clin Neurophysiol 1985;62:99-107.

13 Leandri M, Campbell JA. Origin of early waves evoked byinfraorbital nerve stimulation in man. ElectroencephalogrClin Neurophysiol 1986;65:13-9.

14 Cruccu G, Inghilleri M, Manfredi M, Meglio M. Intra-cranial stimulation of the trigeminal nerve in man. III.Sensory potentials. J Neurol Neurosurg Psychiatry 1987;50:1323-30.

15 Cruccu G, Inghilleri M, Fraioli B, Guidetti B, ManfrediM. Neurophysiologic assessment of trigeminal functionafter surgery for trigeminal neuralgia. Neurology 1987;37:631-8.

1041



j.com/

J Neurol N


ecember 1990. D

ownloaded from



16 Leandri M, Parodi CI, Favale E. Normative data on scalpresponses evoked by infraorbital nerve stimulation.Electroencephalogr Clin Neurophysiol 1988;71:415-21.

17 Ongerbor de Visser BW. Anatomical and functional organ-ization of reflexes involving the trigeminal system in man:jaw reflex, blink reflex, corneal reflex and exteroceptivesuppression. Adv Neurol 1983;39:729-38.

18 Selby G. Diseases of the fifth cranial nerve. In: Dick PJ,Thomas PK, Lambert EH and Bunge E, eds. PeripheralNeuropathy Vol 2, Philadelphia: Saunders, 1984:1224-65.

19 Jannetta PJ. Arterial compression of the trigeminal nerve atthe pons in patients with trigeminal neuralgia. JNeurosurg1967;26:159-62.

20 Burchiel KJ, Steege TD, Howe JF, Loeser JD. Comparisonof percutaneous radiofrequency gangliolysis andmicrovascular decompression for the surgical man-agement of tic douloureux. Neurosurgery 1981;9:111-9.

21 Kugelberg E, Lindblom U. The mechanism of the pain intrigeminal neuralgia. J Neurol Neurosurg Psychiatry 1959;22:36-43.

22 Fromm GH, Terrence CF, Maroon JC. Trigeminal neural-gia. Current concepts regarding etiology and pathogen-esis. Arch Neurol 1984;41:1204-7.

23 Cruccu G, Agostino R, Inghilleri M, Manfredi M, Onger-boer de Visser BW. The masseter inhibitory reflex isevoked by innocuous stimuli and mediated by A betaafferent fibres. Exp Brain Res 1989;77:447-50.

24 Darian-Smith I. Discussion for "Peripheral versus central

factors in trigeminal neuralgia". In: Hassler R and WalkerR, eds. Trigeminal neuralgia. Pathogenesis and patho-physiology Philadelphia: Saunders, 1970:187-8.

25 Melzack R, Wall PD. Pain mechanisms: a new theory.Science 1965;150:971-9.

26 Liu CN, Chambers WW. Intraspinal sprouting of dorsalroot axons. Arch Neurol Psychiat 1958;79:46-61.

27 Westrum LE. Electron microscopy of deafferentation in thespinal trigeminal nucleus. In: Bonica JJ, ed. Advances inNeurology Vol 4, New York: Raven Press, 1974:53-60.

28 Dubner R, Sharav Y, Gracely RH, Price DD. Idiopathictrigeminal neuralgia: sensory features and pain mechan-isms. Pain 1987;31:23-33.

29 Burchiel KJ. Abnormal impulse generation in focally de-myelinated trigeminal roots. JNeurosurg 1980;53:674-83.

30 Calvin WH, Devor M, Howe JF. Can neuralgias arise fromminor demyelination? Spontaneous firing, mechano-sensitivity, and after-discharge from conducting axons.Exp Neurol 1982;75:755-63.

31 Devor M. The pathophysiology and anatomy of damagednerve. In: Wall PD and Melzack R, eds. Textbook of PainEdinburgh: Churchill Livinstone, 1984:49-64.

32 Calvin WH, Loeser JD, Howe JF. A neurophysiologicaltheory for the pain mechanism of tic douloureux. Pain1977;3:147-54.

33 Parkhouse N, Le Quesne PM. Quantitative objective assess-ment of peripheral nociceptive C fibre function. J NeurolNeurosurg Psychiatry 1988;51:28-34.

1042



j.com/

J Neurol N


ecember 1990. D

ownloaded from


Documents

Psychiatry Idiopathic symptomatic trigeminal · s4r2 s5r2 W3 N7 s3r3 P4 P6 WI logy was identified bymeans other than elec-trophysiological trigeminal testing. Seven patients had postherpetic