THE EFFECT OF PERCUSSION OF NERVE - BMJ · Tinel (1917) summarizes the findings ofDejerine andhis...

THE EFFECT OF PERCUSSION OF NERVE

BY

D. DENNY-BROWN and CHARLES BRENNER *

(RECEIVED 8TH AUGUST, 1944)

THAT a physical impact or blow on a peripheral nervemay cause a brief and transient disorder of neuralfunction has long been recognized. The mildestforms of such disturbance, such as occur when theulnar nerve at the elbow is knocked, are within theexperience of all. Tillaux (1866) and other earlywriters distinguished further degrees of commotionand contusion, but did not always distinguishtransient lesions following brief impact from thoseassociated with more prolonged physical force suchas the damage to the brachial plexus resulting fromdislocation of the shoulder, which were also spokenof as " commotion." Weir Mitchell (1872) speaksof contusion as resulting from a blow from a bluntinstrument, and, noting that the immediate effectsare transient, emphasizes the liability to delayedneuralgia, and in some cases to paralysis withdelayed onset. He also noted that, as a result ofmild contusions, the disturbance of function waspartial, and involved sensory function more thanmotor.Tubby (1915) describes " concussion" as varying

from effusion of blood into a nerve to temporaryanemia or hyperaemia resulting from compressionof a nerve against bone by the rapid passage of aprojectile, and cites case histories illustrating thetransient paralysis lasting 5, 7 and 11 weeks beforecomplete recovery (the last an injury of the sciaticnerve at the notch) without reaction of degeneration,and with both motor and sensory changes. Healso described two examples of lesions of mixedseverity where some muscles, delayed in recovery,had lost faradic excitability.

In the World War 1914-18 a proportion of thenerves damaged by gunshot wound presented afusiform enlargement, a " nerve spindle" or"pseudo-neuroma " (Tinel, 1917). Platt (1921)found such a lesion in 7 per cent. of 500operations. The condition is also referred toin the British Medical Research Council Report(1920). Whereas asymmetrical lesion of the nerve,or extensive fibrosis covering a wide extent, maybe attributed to direct laceration of the nerve bythe injury or to involvement in subsequent sepsis,the nerve spindle was supposed to result from aprimary contusion of the nerve by the transientpassage of the projectile. It was no doubt often* From the Neurological Unit, Boston City Hospital, and the

Department of Nervous Diseases, Harvard Medical School.

difficult to decide upon the actual mechanism ofinjury, and Dejerine et al. (1915) remark that thesymptoms of "compression" most often hadappeared at the actual moment of injury. Platt(1921) concludes that war lesion of nerves is almostalways a "primary lesion" and the secondaryinfluence of compression by perineural scarring orcallus " is usually subordinate and often entirelynegligible."

Tinel (1917) summarizes the findings of Dejerineand his collaborators (1915) and describes two typesof nerve lesions in continuity, a neuroma or cicatrixwith axonal proliferation proximal to a denseintraneural scar, and a pseudo-neuroma occurring'" in certain contusions or compressions " of nervewhen the laminated peripheral sheaths were notdestroyed. In the latter condition he describedswelling or fragmentation of the myelin, withfrequent demyelination of the fibres. The axiscylinders were irregular, beaded, barbed, or in theform of a bundle of fine fibrils. There was alwaysan interstitial infiltration of the nerve with " con-nective tissue" and proliferation of the sheath ofSchwann. In the region of the swelling the myelinmay have disappeared without interruption of axiscylinders, and without Wallerian degeneration, inother cases fibres had degenerated. He noted theassociation of incomplete, temporary, dissociatedand irregular paralyses with such lesions. He alsostated that in other cases regeneration was slowand difficult. The " syndrome of nerve irritationaccompanied by trophic pains and disturbances"(causalgia) was common in association with suchlesions as was also the experience of Platt (1921).In two cases of median causalgia under the care ofone of us (D. D-B.) the original disorder in motorfunction of the median nerve following gunshotinjury had been so transient as to lead to theinevitable conclusion that the primary nerve injurywas a contusion. Contusion of nerve therefore hasa possible but unknown part in the pathogenesis ofcausalgia and offers a simple starting point in thisproblem.

Tillaux (1866) subjected the nerves of rabbits toslight blows with a hammer and then examined thecontused portion. Four days after the blow hesaw himorrhages within the epineurium and underthe perineurium, and lessening of the calibre ofnerve fibres, with commencing distal degeneration.

76

copyright. on June 10, 2021 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.7.3-4.76 on 1 July 1944. D

ownloaded from

http://jnnp.bmj.com/

THE EFFECT OF PERCUSSION OF NERVEWeir Mitchell (1872) cites experiments on rabbitsand observed that in lesser degrees of injury manyfibres escaped damage. After a " smart blow"there was immediate complete paralysis and thenerve fibres at the level struck had an irregularlyswollen appearance. Complete functional andstructural recovery could occur in a few days. Intwo rabbits allowed to survive a longer period a" neuroma " developed at the site of injury. Cajal(1928) made a detailed study of the regenerativechanges following a closely similar injury causedby a brief pinch of a nerve by forceps, and describednecrosis of the damaged portion of the axon andrich branching of nerve sprouts within theundamaged sheath of Schwann.

In view of some doubt concerning the nature of thestructural change underlying the very brief disturb-ance of function in lesser degrees of such lesion asdescribed by Mitchell and others, and the relation-ship of such damage to that described by us inlesions of nerve caused by ischxmia and com-pression (Denny-Brown and Brenner 1944a and b)we have repeated the experiments of Tillaux andMitchell in cats and made observations at varyingintervals after injury.

MethodThe sciatic nerve of the cat was exposed in mid-

thigh uinderf nembutal anisthesia. The nerve wasgently raised on a flat metal surface (at P in thediagram Fig. 1) and the blunt edge of a retractor

1CM.FIG. 1.-Diagram to show cross-section shape of thestriker H, drawn to scale, in relation to the nerve N andunderlying plate P.

handle (cross-section to scale shown in Fig. 1 at H)was laid transversely across the nerve, and struck asharp blow with a light hammer. On removing theinstruments a narrow groove is immediately seen.The depth of the groove depends on the intensity ofthe blow. Our intention was to give just sufficientinjury to leave a visible transverse mark on thenerve. In the next 30 seconds the nerve bordering

the groove became distinctly pale and opaque, evenif no hemorrhage occurred. Within 8 to 15seconds one or two very small streaks of haemorrhagewere often seen alongside the largest vessels on thenerve at the site of injury. Within the next fewminutes the groove filled out, the pallor disappeared,and no further haemorrhages occurred. Theintensity of blow used was not such as to inducehaemorrhage within any nerve bundle except in somecontrol experiments in which haemorrhage withinthe perineurium was especially sought. An instru-ment with more rounded edge had to be used toproduce this effect. The wound was then closedaseptically and the animal observed daily forevidence of impairment of movement or sensation.Before sacrifice the wound was re-opened underlight nembutal anxesthesia and the nerve stimulatedabove and below the damaged region. The animalwas then killed, the nerve fixed in formol-saline andstained by osmic acid, Gros-Bielschowsky method,cresyl-violet, Sudan III, Spielmeyer's myelin stain,iron hlmatoxylin and other methods.

Findings1. Persistent Paralysis

Percussion of nerve sufficient to impair conduc-tion in some degree is followed by the appearanceof a swelling (pseudo-neuroma) of the region ofpercussion (Fig. 2). If the percussion is so lightthat no disturbance of motor function is demons-trable, a few fibres in the nerve may be later foundto have been damaged, but no swelling occurred inthese circumstances in any of five experiments. Itis convenient first to examine the disorder producedby a degree of trauma sufficient to lead to immediateloss of conduction of movement persisting for sixweeks or longer, with Wallerian degeneration belowthe lesion. The changes were identical whethersmall hiemorrhages had occurred at the time ofinjury or not. If rupture of the perineuriumoccurred, additional changes appeared and thesewill be described in a separate section. Motorparalysis below this type of lesion was complete,but some reaction to a pinch of the pads of the toesalways remained and in one experiment to a touch,showing that some pain conduction and sometimestouch conduction also was intact. Detailed observa-tions on sweating and temperature were not made,but no gross change was observed.

Histological examination of the nerve within thefirst 24 hours after a severe degree of injury failedto demonstrate any change in the myelin or axiscylinders. Twenty-four hours after the injury theinjured region of the nerve was swollen and cedema-tous. The myelin sheath of each fibre was greatlyswollen throughout the cedematous region and themyelin irregular. Within three days most of themyelin at the level of percussion and for a distanceof one or two millimetres above and below haddisappeared. Large numbers of phagocytic histio-cytes, containing fat granules, filled the cedematousendoneural spaces. Some vacuolar masses, stain-ing faintly with fat stains, adhered to remnants of

77



j.com/

J Neurol N


ownloaded from


D. DENNY-BROWN AND CHARLES BRENNER

% Ar0p.i! vt..- ;.r

art -

* . bL4.

fr%"* *4

.t47. *.- fle- -- --.A..-

28

.-...---. WA-

¼:- *,

FiG. 2A.-Longitudinal section through the pseudo-neuroma, popliteal nerve, 13 days after injury. Complete motorparalysis, with sensation to pinch and touch persisted from time of injury. Electrical excitability of peripheral motorbranches of the nerve was completely lost. Spielmeyer's stain for myelin. The left of the figure is proximal and rightdistal in all figures of longitudinal section. The ruled line in upper corner is equivalent to 1 0 mm.FIG. 2B.-Popliteal nerve, 13 days after injury. Great weakness of plantar flexion for three days with preservationof sensation to pinch or touch throughout. Spielmeyer's stain. Magnification as in A.PrG. 2c.-As 2B, Gros-Bielschowsky method for axis cylinders.

78

w



j.com/

J Neurol N


ownloaded from


a -ft l--a

I -~ ~ MF&WqqR. P w

a~~~ b--b

aN ~ -.r-r________________ __ _____

eA3 AF - ~ X--,8 *_~~~ -r---- M_,W*f_4,- - -o-* *JL _a_C

n -~ . S-

* aJ X- vr~~~~~~~~~~~~~~0

I.~~~~~~~~~~~~~~~~~~- a.~~~~- - _. _ - -_

-

a.

FIG. 3A.-Bundles of regenerating fibrils in the centre of the pseudo-neuroma, same experiment as 2A, 13 days afterinjury. Gros-Bielschowsky method, counterstained Sudan III and hrmatoxylin. The ruled line at the top of thefigure is equivalent to 0 1 mm.FIG. 3B.-As 3A, to show detail of branching of parent axis cylinder. The uppermost fibre loses the myelin sheathbetween the left and middle thirds of the figure. The remaining fibres are unmyelinated. Magnification is as in A.FIG. 3c.-Axis cylinders just proximal to the level of percussion which begins at the right of the figure, three daysafter injury. Gros-Bielschowsky method. The ruled line at upper left corner is equivalent to 0-1 mm.FIG. 3D.-As 3c, axis cylinders just distal to the level of percussion. Magnification as in C.FIG. 3E.-Beading of non-medullated fibres 13 days after severe percussion. Gros-Bielschowsky.

...p_B |r

3B

.,y

3D

----

r

d ompollam-

.1

ommiiiiiiiii..

WIN- a.'a,

3E a-.P a.--... _ _ _ .- -.d .--I

v

4 AP--awoo-

I _ .,.

o"

0.

__

JIC.7- - 41111OWam Z .. ...!

--.- ----



j.com/

J Neurol N


ownloaded from


D. DENNY-BROWN AND CHARLES BRENNERthe Schwann tube. The axis cylinders were greatlyswollen, with irregular gaps and distortion in theproximal part of the cedematous area (Fig. 3c) andmore regular expansion in the distal swelling(Fig. 3D). At the point of percussion all the largeraxis cylinders ended in large bulbs or knobs, in themanner clearly described by Cajal (1928) in hisaccount of trauma to nerve by pinch with forceps.For a millimetre above this, the axis cylinder hadsprouted long leashes and whorls of fine fibrils(Fig. 3c), many of which had already forced theirway into the empty Schwann tube for a millimetreor more. Others appeared to splay out into theendoneural spaces. Many of the finer axons weresimply beaded and irregular in calibre as they passedthe percussed region. Distal to the percussion themyelin and axis cylinders showed only the earliestsigns of Wallerian degeneration.

Fourteen days after the injury such a nerve wasinvariably swollen at the site of injury, sometimes toas much as twice the diameter of the trunk aboveand below. The swelling consisted of a localizedcedematous zone in which the nerve fibre strandswere wIdely separated by fluid containing manyhistiocytes loaded with fat granules (Fig. 2A). Themyelin had entirely disappeared in the centre of thisswollen region, and the nerve fibres were repre-sented by chains of Schwann nuclei in almost all ofwhich multiple strands of regnerating nerve fibrils(Fig. 3A), and in the remainder a single axis cylinder,extended completely across the swollen region.These bands of fibrils lay relatively isolated by theendoneural oedema. Nuclear stains such as hema-toxylin or cresyl violet showed the bands as bundlesof nuclei traversing the cedematous region, andLaidlaw's stain revealed that each had a coveringof endoneural reticulum.Traced to the proximal myelin sheath the axon

fibrils were found to spring from the stump of theold axon, either as a spray immediately below thelast myelin or more commonly as the result ofrepeated sprouts from the old axon which was seento run a variable distance as a pale unmyelinatedsingle central core with sparse Schwann cells atintervals (Fig. 3B). Haematoxylin stains revealedan additional outer covering of scattered paleendoneural nuclei. At the point of division theaxon was shrunken and stained darkly with silver.Few old axis cylinders passed the centre of theswelling and these branched before reaching theperipheral zone of Wallerian degeneration. Apuzzling feature of such proliferating fibrils wastheir frequent grouping around a larger centralfibre which at first appeared to be the persistingoriginal axon. When traceable into the distalsegment of the nerve such large axons coursedirregularly around ovoids of degenerated myelinin the old Schwann tube. Close examination con-vinced us that this appearance of persistence of theoriginal axon at the 14-day stage was due to earlymaturation of a main regenerative axonic sprout,and that even at this time the profuse networks andspirals of finer fibrils were in process of resorption,

for though common at the 14-day stage they wererarely present after three weeks. This process ofresorption also explained the frequent occurrenceat 14 days of side branches lying outside the sheathof Schwann and directed proximally (Fig. 3A).Bundles of non-medullated fibres appeared eitherto be intact but with frequent oval swellings, as inFIG. 6F, or to be beaded and irregular where theypassed the axis of the pseudo-neuroma (Fig. 3E).On close examination a number of the thinnestmedullated fibres were found to run far into thecedematous area before losing their myelin whichthey regained after 02 to 0 5 mm. and courseddistally without Wallerian degeneration. Five ofthese were in the section shown in Fig. 2A, but aretoo small to appear in reproduction. Silver stainsshowed corresponding thin single conductors withassociated sheath nuclei, but it was difficult to dis-tinguish these from regeneration except in twoinstances where such an axon was traced into a smalldistal intact myelin sheath.

After two weeks the loss of myelin in generalcoincided with the extent of the swelling bothproximally and distally (Fig. 2A). More proximallythe myelin sheaths were irregular and swollen fora further 5 mm., with corresponding irregularityand swelling of the contained axis cylinders. Thisirregularity and swelling we have found associatedwith simple cedema of nerve (Denny-Brown andBrenner (1944b)) and regard it here as resultingalso from cedema. The distal portions of the nervewere in Wallerian degeneration. The myelin andaxis cylinders were partly broken up into the familiarovoids. The heavy content of fat and large frag-ments of undigested myelin in the whole peripheralsegment was in marked contrast to the fineintracellular droplets in the traumatic swelling,described above.The perineurium was in places lifted from the

nerve bundles by fluid containing fat-laden phago-cytes. This perineural cedema was variable indegree and never intense. There was no evidenceof migration of macrophages or fluid proximally ordistally from the swelling. The swelling of thenerve in such a lesion was associated with a localizedstasis and congestion of all the vessels in the area.In the region of the nerve just distal to the traumaticswelling the regenerating neurofibrils were seen topass directly into the degenerating Schwann sheath,coursing irregularly round the obstacles providedby ovoids and fat droplets (Fig. 3A).

After twelve weeks, the swelling at the site oftrauma still persisted but was slight in degree. Inan experiment where severe weakness, with retainedsensation only to a severe pinch, had persisted for30 days, with full recovery of spreading of the toesdelayed until the eighth week, full regeneration hadoccurred. The regenerated fibres had each aseparate sheath of Schwann, with commencingmyelination, both in the swelling and beyond. NoPerroncito spirals remained in the traumaticswelling. As shown in the two upper fibres inFig. 8A, a few very small spiral branches besides the

80



j.com/

J Neurol N


ownloaded from


THE EFFECT OF PERCUSSION OF NERVEoriginal conductor had become medullated. Thiswas the exception and no other trace remained of therich branching that was seen in the first two weeks.It is assumed that most of the young fibres are sub-sequently resorbed, for the nerve contained fewmore fibres than just above the lesion. The peri-neurium was not thickened and there was noepineural change. There was, however, a per-sistent increase in the oval endoneural cells, whichin a section stained with cresyl violet or othernuclear stain was seen to be limited to the region of;'previous percussion, and to be only very slight inthe previously degenerated peripheral segments ofthe nerve. Associated with this nuclear increasewas an intensification of the endoneural reticulumas revealed by Laidlaw's stain. The whole dis-turbance was much less severe than that caused bycrushing the nerve with a haemostat at the same level,and the onset of recovery is earlier. The com-pletion of final recovery is, however, about the samefor both types of lesion.A complete transverse lesion of nerve caused by

percussion is therefore remarkable in the speed ofdevelopment of a localized swelling at the site oftrauma, the rapid removal of myelin from the regionof swelling and the relative escape of non-medullatedand a few finely medullated fibres. Regenerationis precocious, but once the distal segment of nerveis reached proceeds as after any other interruptionwith a fairly complete ultimate reconstruction of thenerve. An increase of endoneural reticulum andnuclei marked the region of previous trauma.

2. Transient ParalysisAn intensity of blow sufficient to leave an

immediately visible indentation on the nerve trunkand occasionally the appearance of a small pinpoint effusion of blood in the coats of one or moreof the larger vessels in the nerve sheath may never-theless not cause persistent paralysis. We have notattempted to measure the exact intensity of per-cussion causing such transient effects, for intensitydepends on a number of factors which are difficultto determine, such as surface area struck and com-pressibility of the nerve. It was, however, foundthat within a wide range of intensities a light per-cussion would cause an immediate motor paralysisusually complete for dorsiflexion of the foot andspreading, and partial for plantarflexion, lasting threeto four days, followed by rapid recovery in the next48 hours. Spreading of the toes was always themost severely affected motor function in partiallesions and the last to recover in all lesions. Therewas no loss of sensation when the animal was firstexamined on recovery from anmsthetic 18 hoursafter the injury. The foot was even slightlyhypervesthetic to a light touch in the first three days.If pain sense appeared impaired in any degree, therewas invariably also some persistent weakness,lasting more than two weeks. The nerve wasexplored, stimulated, and submitted to histologicalexamination at varying intervals after the causationof this mild- type of injury. Stimulation of the

exposed nerve invariably confirmed the presence orabsence of motor conduction as observed clinically,and, if the anaesthetic were sufficiently light, con-firmed the maintenance of sensory conduction.

Three days after the percussion of the nerve afusiform swelling had formed at the site of per-cussion, increasing the diameter of the nerve at thispoint to as much as 150 per cent. of normal. In verymild lesions the swelling was correspondinglyslighter. This swelling was still remarkable afterthree weeks, and only just discernible after threemonths, in spite of early complete recovery of con-duction. The swelling of " pseudo-neuroma" isdue to interstitial cedema and the presence of largenumbers of histiocytes in the widened endoneuralspaces, in every way identical with that of severepersistent lesions. The cedema lifted the peri-neurium slightly at the height of the swelling. Inthe cedematous area the myelin of all large fibreshad become greatly swollen and pale, and in somelarge fibres was in complete dissolution by the thirdday (Figs. 4A and B), leaving naked axis cylinders(Fig. 4c). The myelin of smaller fibres appearedto be unaffected in osmic preparations. Coarse,darkly staining fatty granules were seen in amoderate number of histiocytes which had appearedin the cedematous endoneural spaces.

Three days after the injury a few axis cylinders ineach section showed the frayed appearance we havefound to be general in severe, persistent lesions.The remainder were all greatly swollen throughoutthe pseudo-neuroma reaching diameters of 10 to12,u as in Fig. 3D. At the point struck and for ashort distance proximal to this the larger axons weredistorted in a sinuous fashion, with intermittentpallor or failure of impregnation seen in some fibresin more severe lesions, as in the lower part ofFig. 3c. At either extremity of the pseudo-neuromathese swollen axons abruptly regained a swollenmyelin sheath which became of normal appearancewhen traced a farther 1-2 mm. away from the lesion.Cell stains showed the presence of moderate numbersof histiocytes in the endoneural spaces at this stageand the oval, pale nuclei of the endoneurium wereunduly frequent. The Schwann nuclei were dis-torted, but not otherwise changed. No mitoses ofendoneural or Schwann nuclei were seen.

Transverse section of the swelling on the fifthday revealed myelin sheaths in all stages of dis-appearance. at the level of percussion (Fig. 4D).The axis cylinder, unstained, or but slightly darkenedin osmic preparations, was sometimes completelydenuded of myelin, but more often had a crescent-shaped remnant applied to one side as in many ofthe unstained fibres in Fig. 4D. The residual myelinwas swollen and pale except for small foci whichwere densely blackened by osmic acid. The nervefrom which this figure was taken had shown clinicalevidence of 50 per cent. loss of motor conduction fortwo days, with clinical improvement on the thirdday and no evidence of block of motor conductionon proximal direct stimulation on the fifth (last day).These charges slowly became intensified. The

81



j.com/

J Neurol N


ownloaded from



A.

¶ 4~~~~~~~~~~~~~~~~~~~~~~~~~0

At~~~~~~*o,..,

\j o*,t w(C

--ONt$: ".*

.7 r .:.

0: w

I.:

~~~~~~~~~~~~~~-I I"-&.5._

-i ~~~~--i-r-- TAA-=;~~~-a-.- -

4~~ ~ ~ ~ ~ -.a-e;r.- -

~~~~~~~~~- _ _- *b- -S

5 A . . _~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r a

- -, - -- a",.aa, - .-

0ma

.#!B3 - ---

-~~Ia r C 'i' . * V2 -8

_~~~~a -- 4 t,~ _-~r _ . -c _w_-r4_.% r a -_. O'- - -...

adg!T~ ~ ~

*_ _ _ '_ ^ , .ta q - _r -S a r. a .f

* ~~~~ralt f 1

>~ ~ ~ ~ ~ t. S a; - ir z-S- _ ;

- -~~~~~~~~~~A.

5E3v _tS - "-

._~ ~ ~ c~t.J,.-- -,--^' CZ < x ;;5-:2

D. DENNY-BROWN AND CHARLES BRENNERcedema and cellular reaction reached a maximumfourteen days after the original percussion. At thistime axis cylinders appeared more widely separatedand the phagocytic histiocytes more numerous(Fig. 2c). Occasional axis cylinders that had beeninterrupted now showed the plexiform branchingwith network of proliferated Schwann cells wehave described in persistent lesions. These fewbundles communicated directly with distal myelinsheaths that displayed commencing Walleriandegeneration. A few axis cylinders were swollenand irregular throughout the lesion but showed noirregularity in depth of staining with silver. Theirregularity took the form of numerous projections,as in the large fibre near the top of Fig. 6B. Bycounterstaining frozen sections stained by theBielschowsky method with Sudan (Oil Red 0) stainfor fat, these irregular swollen axons were shown topossess an intact though oedematous myelin sheaththroughout the lesion. They were also clearlyobserved in Sudan sections counterstained withhematoxylin (Fig. 4E), and were evidently the" barbed" fibres of Tinel (1917). These intactfibres were present in inverse proportion to thenumber of degenerate (fibrillar) fibre bands. It wasnot found possible to obtain a pseudo-neuroma freefrom both. The bulk of the nerve fibres, however,were strikingly different, for the remaining 70 to90 per cent. of axis cylinders stained unevenly withsilver for a variable extent, and in the combinedstaining method many could be traced from thepoint when proximal myelin abruptly ceased to wherethe distal myelin abruptly recommenced (Figs 6A,B, C, D, and E). The demyelinated gap varied from001 mm. to 1l1mm. It was not possible to be sureof continuity of fibre in longer gaps which are sug-gested by sections such as that shown in Fig. 2B.In none of these was there evidence of distaldegeneration of myelin, though for a distance upto 1 mm. on either side of the lesion the myelinwas swollen and irregular. The peripheral nervestained with osmic acid usually showed degenerationin fewer than 2 per cent. of fibres, and numeroussections of muscle failed to show degenerated endplates.The demyelinated segment of axon was a single

stout structure with limiting membrane and finelygranular contents (Figs. 6c and D). There wasalways a cellular covering which was more distinctafter the seventh day, partly owing to the separationof fibres by increased cedema, partly to collapse of thesheath of Schwann around the naked axon (Figs. 5cand D and Fig. 7). Two weeks after the percussiontwo types of nuclei accompanied the bared axons.At intervals of approximately 01l to 0-2 mm. thenarrow darkly staining Schwann nuclei lay closelyapplied to the axon, usually indenting the surface(Figs. 6c, D and E and 7B). Schwann nuclei,normally distributed so that each one lies in thecentre of a segment of about one millimetre, wereredistributed between the seventh and fourteenthday so that one lay on the centre of each of the new,shorter and lightly staining, internodal segments.

Various steps in this migration were observed,though mitosis was rarely seen. The youngSchwann nuclei were smaller than normal, and had ahigh affinity for silver, so that in a lightly stainedsilver preparation only those proliferating nucleimay be stained (Figs. 6c, D, and E).At similar intervals, but usually more numerous,

oval, pale nuclei, lacking the nucleolus usually seenin the Schwann nucleus, lay wrapped around thefibre, often obliquely (Figs. 5c and 7A). Those wereusually separated from the axon by a distinctinterval, and in places the outline of a membranecould be seen between these pale nuclei and theSchwann nucleus (Fig. 7B). The pale oval nucleiwere therefore interpreted as those of endoneuralcells. They were readily distinguished from themore narrow pointed nuclei of fibroblasts. Theircytoplasm was rarely stainable and then appearedto be the fibrous endoneural reticulum of Laidlaw.With the Bielschowsky method the cytoplasm of theSchwann nucleus sometimes stains, giving a furtherdifferentiation from the endoneural nucleus(Fig. 7c).The naked segment of axis cylinder did not show

fibrillar change and did not branch in any specimenthat could be traced into an intact distal myelinsheath. All axis cylinders that began proximallyas a single core and more distally showed bifurcationor multiple branching (Fig. 3B) were, if traceabledistally, found to enter Schwann tubes containingdegenerated myelin. Lateral sprouts or brancheswere therefore always a sign of interruption of con-tinuity. Proximal to the point of branching such anaxon presented the same appearance as a demye-linated but uninterrupted cylinder.Though increase in number of endoneural cells

was observed on the third day the manner in whichthese cells embrace the demyelinated segment ofaxon is not clear until after the tenth day, when theSchwann sheath becomes more closely applied to thenaked axon. The earlier distension of this sheathappears to be due to cedema and vacuolation. Theendoneural cells did not embrace fibres whichretained a myelin sheath (Fig. 7A). The penetra-tion of the Schwann sheath by the phagocytichistiocytes, so commonly observed in Walleriandegeneration, was very rarely observed, and thenonly when some large myelin droplet remained in theSchwann sheath. The histiocytes remained in theendoneural spaces where they collected in largenumbers (Figs. 5c, 6A, and 7A), their roundedcytoplasm filled with fatty droplets. Occasionallymphocytes and Mast cells were also seen. Therewas no evidence of migration of cells or oedema upor down the nerve. The blood vessels in the pseudo-neuroma were greatly congested (Fig. 7D) but showedno other change.The smallest medullated fibres (2 to 3yt in frozen

sections) appeared to escape demyelination in pas-sage through the pseudo-neuroma. The myelin inthese was often granular and poorly stained (Fig. 5D)by Spielmeyer's stain, but was intact in Sudan andosmic preparations. In the same figure the distal

84



j.com/

J Neurol N


ownloaded from


MP-am~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

__ -'

6 A'

-6,

-o .ir0

S ..JffiYw.1-X___

..._f. W-A le,..

.Jow; e_I ,, _.4_~mfimw__~~~~~~~~-A._ _;~~~-_&m_~~ ~ ~~~I___

rF -_.w ~ . .66 _W

_ _ f __Q

i a aok

a- S'eL....

-a

-w .rO-

,-r. .:_m..v.~~~~~~~~~~~~~~~~~~~NNO

6F t ete

FIG. 6A.-Popliteal nerve. Centre of pseudo-neuroma from same nerve as shown in Fig. 2c, 13 days after injury,10 days after recovery from severe weakness. Gros-Bielschowsky method.

FIG. 6B.-As 6A, higher magnification from another section, to show penetration of axis cylinder in cedematousmyelinsheath (second from top of figure), demyelinated and beaded axons.

FIG. 6c.-As 6A, to show short segment of demyelinated axon. Gros-Bielschowsky, Sudan III and hrmatoxylin.The small ruled line is equivalent to 0-01 mm.

FIG. 6D.-To show demyelinated axon joining the distal myelin sheath (above), and a regenerating bundle of fibrils(below).

FIG. 6E.-As 6A, to show junction of demyelinated axon with distal myelin sheath and two other naked axons.FIG. 6P.-As 6A, non-medullated sympathetic fibres with normal Schwann nuclei and areas of oval expansion in the

pseudo-neuroma. Gros-Bielschowsky. The small ruled line is equivalent to 0-01 mm.

Z.i-=NM4r7 tr r ;U

6o - ~ I- s S- I ..j

-- X-OO

NW

W

_~~~~~~I_ -

6D ---- -- I* ~~~~~~~~~~~~~~~~.1

i.. -.. FA. JPWF

.4'T%V %,., . 1 .

mm-

--,T.MMWA26p:-

Amw4, -

- -m

- -- ---

- - --= ---w- ". Ad.

mck.- a-E!blmgb

liw !:*.-.4so-ol

4. ft.- 1)aqw.. "I-,Aw- ."

-I.,6c

mw. .-

6E qw:z.



j.com/

J Neurol N


ownloaded from



a,,%_ M4W~~~~~~~~

-:


end of the demyelinated segment of a moderatelysmall fibre is shown. The demyelinated gap in suchfibres was short, as can be seen in Fig. SA wherethey are seen to run far into the cedematous areabefore losing myelin. Non-medullated axiscylinders appeared to cross the pseudo-neuromaintact, though oval formations were seen in thefibrils, probably an expression of oedema (Fig. 6F),as in severe lesions. Their Schwann nuclei remainedintact.

3. The Formation of Pseudo-neuromaPercussion, whether leading to rupture of axis

cylinder or only to transient disorder of conduction,is associated with a precocious and localized break-down of the myelin sheath. The process of slowingestion of the myelin tube in the form of globulesor " ovoids" by phagocytes, as seen in Walleriandegeneration, does not occur in the percussed region,but is replaced by a rapid dissolution of myelin,complete in three to five days. In that short periodenormous numbers of fat granule cells with kidney-shaped or irregular nucleus appear in the tissuespaces, which are widely distended with fluid. Theobserved swelling of the nerve is directly related tothis cedema. The vessels in the cedematous regionare widely distended, but show no evidence ofthrombosis. There is seldom any haemorrhage intothe endoneurium, and the swelling and other changesare independent of such occurrence. The histolo-gical picture is comparable with acute but localizedinflammation.. Polymorphonuclear leucocytes donot appear unless in relation to some small andadventitious hemorrhage.We found that if at the same exploration at which

the nerve was percussed a tight silk ligature wastied 1 cm. or more proximal to the level of per-cussion all the phenomena of " pseudo-neuroma "could be produced, uncomplicated by the process ofregeneration. Thus, three days after the per-cussion, all the myelin distal to the ligature was stillintact and the axis cylinders swollen and irregular,except at the level of percussion. Here there haddeveloped the usual localized swelling withendoneural cedema, and all myelin had disappeared(Fig. 8D). Numerous phagocytes filled theendoneural spaces. The axis cylinders of all sizeshad broken into weakly staining fragments or hadcompletely disappeared (Fig. 8E). On either sideof the percussed region the axis cylinder was greatlyswollen and frequently ended in a rounded, abruptfashion. The Schwann cells had not proliferated.The appearance was consistent with a great accelera-tion of lysis and phagocytic activity.

In the course of previous work on compressionof nerve (Denny-Brown and Brenner, 1944b) wehave observed that the nerve on either side of acompressed area became oedematous and swollen.This change was related by us to vascular stasis.In moderate degrees of compression with dis-sociated block ofconduction the oedematous reactionwas also accompanied bya rapid dissolution ofmyelinand appearance of histiocytes, though these were

absent in the cedema produced by minimal com-pressions not causing block. Guttmann andMedawar (1942) note the swelling of a nerve trunkabove a ligature and state that such swelling doesnot develop if an additional ligature be tied10-20 mm. proximally at the same time as the first,or if the proximal nerve be interrupted by theinjection of chromic acid. These authors con-clude that in the absence of nerve fibres, a swollencentral end-bulb is not formed. Such experimentsappear to us, however, to overlook other possibilities,for the vascular supply to the nerve between theligatures or other obstruction was also abolished bytheir method. The same objection also appears tous to be inherent in the experiments of Weiss (1943),who, after slipping an arterial cuff over the proximalcut end of nerves, observed a larger swelling of thenerve proximal to the contracted segment of arterythan distal to it, and concluded that the process ofcedema then arose from obstruction of a centrifugalflow of endoneural and axonic fluids.We have endeavoured to test these hypotheses by

two sets of experiments. The effect of percussionof a section of nerve between two ligatures wasexamined with and without persistence of bloodsupply to the isolated segment. The ligatures wereplaced 10 to 15 mm. apart, and arranged so that insome experiments the isolated segment receivedthe lower branch of the- sciatic artery and vein, inothers no blood supply. Without blood supply theisolated segment of nerve becomes shrunken andyellowish in colour in 14 days. There is someadherence to surrounding tissue. Percussion in themiddle of the segment at the time of ligature hadleft no local swelling. Sections revealed only auniform necrosis of all nerve structure. If theisolated segment were allowed to retain a bloodsupply from a branch of sciatic artery and vein itbecame somewhat swollen as a whole with greaterswelling next each ligature. Percussion in the centreof the segment a few moments after the ligatureswere tied had resulted in a small localized swelling atthe percussed level fourteen days later. Within thisswelling there was precocious disappearance of allmyelin and axis cylinder remnants, and a localizedintensity of histiocyte response with oedema.Further, if the proximal ligature was placed so thatthe nerve trunk received direct branches of sciaticvessels on either side the swelling occurred on thetwo sides of the ligature, perhaps more often alittle less in degree on the distal side as Weiss hasobserved. Weiss concludes that the distal swellingis not cedema because the nerve fibres are notseparated by fluid. We also found less endoneuralfluid below the ligature, but we observed that thefibres themselves were more cedematous, being inprocess of degeneration.The vascular supply to nerve trunks involves a

longitudinal anastomosis between widely spacedtributaries, no one of which is essential for meta-bolism at any level. As we have recently described(Denny-Brown and Brenner, 1944a) ligature ofpudic, sciatic and popliteal vessels to the sciatic

87



j.com/

J Neurol N


ownloaded from


D. DENNY-BROWN AND CHARLES BRENNERnerve in cats results only in patchy aedema withoutdisturbance in function. The experiments_describedabove, however, indicate that cedematous swellingsrelated to ligature or other compression are greaterwhen closely related to the entry of a vascular tribu-tary, and we consider this to indicate that they areprimarily derived from vascular factors.The precocious disappearance ofmyelin in associa-

tion with the localized cedema following percussionraises a further problem. The cedema requires anintact vascular supply. Is the cedema part of aninflammatory response to traumatic breakdown ofmyelin, or is the primary damage vascular andincreased phagocytosis and myelin dissolutionrelated to ischemia ?

4. Percussion of Degenerated NerveWe proceeded to degenerate the sciatic nerve by

section between two ligatures, allowing the sciaticartery to give direct supply to the distal segment.After an interval of three weeks for degeneration,the nerves were re-explored and the distal-degenerated segment was percussed. After intervalsof 3 and 14 days, the animals were killed and thenerves examined.The effect of percussion of degenerated nerve was

remarkable. Although no pseudo-neuroma wasobserved after intervals of 3 and 14 days, in frozensections some slight increase of the cedema ofthe percussed region was present in comparisonwith the remainder of the degenerated nerve. Thearrangement of macrophages and their generalappearance was considerably changed in thedamaged area. The change was most evident infrozen sections stained by Sudan III for fat (Figs. 9Aand B). The macrophages were no longer vacuo-lated and compressed into the well-known " ovoids "of Wallerian degeneration, where the borders of onecell are difficult to distinguish from its neighbour.The macrophages in the percussed area instead layfree, with rounded cytoplasm filled with fat granules.The vacuoles had disappeared. The nucleus waslarger and had lost the distortion typical of " foamcells." They had taken up an arragement in parallellines, a re-arrangement that is obvious also in thebands of Bungner, which were also converted intoparallel columns in the affected region (Fig. 9D)instead of the meshwork in which they containWallerian ovoids (Fig. 9c). Since this change wasfully accomplished in three days and was confirmedin H and E, and Mallory phosphotungstic prepara-tions, and since an occasional remnant of ovoidwas rarely found, it was concluded that the macro-phages had in fact been freed by the injury. Allhistologists have agreed that the macrophagespresent in Wallerian ovoids are " so compressedthat the intercalary spaces are often obliterated"(Cajal, 1928). The source of this pressure has notbeen determined. The dramatic release of all themacrophages by a single sudden trauma indicates,however, that the Schwann tube, confining thesephagocytes, was ruptured by the trauma, followingwhich the macrophages came to lie in the endoneural

space. The rapid disappearance of the characteristicvacuoles is also remarkable, and indicates that theyare related to the previous tension. Whether theconfining membrane is the neurilemma (wall of theSchwann tube) or an endoneural sheath is open toconjecture. We are in agreement with Holmes andYoung (1942) that the neurilemma remains distinctin degenerated nerve, and in stained sections foundno evidence that the endoneurium in such nervesconfines the Schwann cell more completely than in anormal nerve.

Experiments on the effect of percussion ofdegenerated nerve therefore show that the effect ofsuch trauma is then traceable to rupture of theSchwann tube. The " foam cell" then becomes anordinary macrophage. The great cedema peculiarto a percussed normal nerve is absent. It is there-fore concluded that the cedema and phagocyticreaction of the pseudo-neuroma is determinedprimarily by traumatic breakdown of myelin. Fromthe appearance of splits in the myelin three daysafter percussion of normal nerve (Fig. 4B), and theremarkable decompression of phagocytes indegenerated nerve, it appears reasonable to assumethat rupture of the neurilemma occurs in bothinstances by longitudinal splits, and that the tissuereaction in normal nerve is due to contact of myelinwith endoneurium. Such rupture would alsoexplain the precocious spread of leashes of neuro-fibrillar strands into the endoneurium and theirlater atrophy, an occurrence which at first we wereat a loss to understand. It even appears possiblethat at the moment of the blow axis cylinder andmyelin flow laterally through ruptures in theneurilemma.

5. Rupture of the PerineuriumIn some experiments, immediately following

percussion, a small white herniation appeared on thenerve at the point struck. This was likely to happenif the surfaces of the striking object and of the plateunderlying the nerve were not parallel, so that themomentary compression of the nerve bundle wasmore to one side. The perineurium had rupturedand allowed nerve fibres to bulge through the open-ing. In such an experiment, where the peronealdivision had developed such a hernia, the nerve wassectioned after 14 days' survival, paralysis of spread-ing and of dorsiflexion of the foot having beenobserved to persist, the swelling of the peroneal andpopliteal divisions appeared to be that usuallyencountered. On section a proliferation ofendoneurium, perineurium and nerve fibres throughthe opening was seen to have occurred (Fig. lOAand B). The spindle-shaped swelling on the peronealnerve was partly a true neuroma. Myelinated fibreswith irregular neurofibrillar regeneration bulgedthrough the opening (Fig. 1OA). The myelin in theremainder of the bundle was swollen and irregularon all fibres. Demyelination, leaving naked axiscylinders, was present in scattered fibres, throughoutthe unherniated portion (Fig. 10c) with a high pro-portion of regeneration. Though endoneural

88



j.com/

J Neurol N


ownloaded from



a...

.,

or

~~~~~~~~t | 4st '= = = = =~~~~-

-A-i

r-

8B~~~~~~~~~~~

a _o__~~~~~~~~~~~~~~~~~~~~~~~.

_ _ =~~~~~~~.-'U--1-

- I

-- r --C- e_ * v~~~~~~~~~~S_~~

- -

8o-

8 - -- - i_~~~a.w9 raJ

FIG. 8A.-Sciatic nerve at level of severe percussion three months earlier, causing severe paralysis for 30 days withcomplete recovery in 60 days. Spielmeyer's method to show very thin myelin sheath and small branches in uppermostfibre.FIG. 8B.-Popliteal nerve three months after mild percussion causing partial weakness for three days, to show thin

myelination at level of injury (no change distally). The ruled line is equivalent to 0-1 mm.FIG. 8c.-As 8B. Gros-Bielschowsky method for axis cylinders to show irregularity of staining. The ruled line is

equivalent to 0 01 mm.FIG. 8D.-Popliteal nerve, percussed three days earlier and ligatured 2 cm. proximal to level of section. Iron

hematoxylin. To show loss of myelin at level of percussion.FIG. 8E.-As 8D. Gros-Bielschowsky method at level of percussion.

H

89

r ---'-

I.-."'odpop"Sommki4poe io -omp-

mmmubk-..-mommu



j.com/

J Neurol N


ownloaded from


90 D. DENNY-BROWN AND CHARLES BRENNER

---wA -W-1. it -40.1P'*! v - -"W ok .O

k'r

-4 A-lQw4vwcb-.

J

p~~~~~~~A

'r '- - t,w.F- - V .~~~ - _4

- U -

011,11W., v'o 4mftmb-e .,,%..4F, ., .0 I

Ww"; -4 aJA

--

FIG. 9A.-Sciatic nerve, Sudan III, 24 days after section between ligatures 4 cm. proximal to region shown.FIG. 9B.-As 9A, but 1 cm. farther distally, where the nerve had been percussed three days previously.

FIG. 9c.-As 9A, phosphotungstic acid hematoxylin.FIG. 9D.-As 9B, phosphotungstic acid hematoxylin.



j.com/

J Neurol N


ownloaded from


THE EFFECT OF PERCUSSION OF NERVEcedema was much less than in a pseudo-neuroma thenaked axis cylinders were greatly swollen. Histio-cyte response was present throughout the nerve atthis level, but these cells were closely packedbetween the nerve fibres. There had therefore beenspontaneous decompression of the perineurium.In other examples encountered it was observed thatdegeneration of all the fibres herniated through theopening and many of those near it had occurred bythe fifth day. In one example, kept for six weeks,the proliferation of the hernia resulted in the forma-tion of a large spindle-shaped mass with profusionof branching fibrils. There was still no evidence ofrecovery. Experiments of longer duration are inprogress.

DiscussionPercussion of peripheral nerve evokes a tissue

response characterized by localized vascular stasisand cedema, with the appearance of large numbersof histiocytes and with proliferation ofendoneurium.The myelin for a millimetre or more on either sideof the level of percussion undergoes dissolution,and the histiocytes in the endoneural cedema becomefilled with fat droplets. This process is complete inthree to five days. It differs completely from the chiefhistological feature of Wallerian degeneration, wherephagocytes enter the sheath of Schwann and slowlyingest large globules or segments of myelin in thecourse of 2-25 weeks. Doinikow (1911), however,observed that in Wallerian degeneration some solu-tion of myelin also took place, with the appearanceof fat in surrounding endoneural cells and histio-cytes. The rapid disintegration of all myelin by aprocess of solution following percussion is, however,sufficiently distinct to merit special attention. Itclosely resembles the rapid dissolution of myelin ina segment of nerve that has been frozen, as describedby Bielschowsky and Valentin (1922). We haverecently (Denny-Brown and Brenner, 1944b),described an identical dissolution of myelin con-sequent on compression of nerve, these beingrelated to the effect of ischimia under andimmediately distal to the compressing agent. Inthat case there was also cedema and inflammatoryresponse. In mild degrees the ischaemic lesion wasassociated with preservation of axis cylinders, and ablock of conduction of motor fibres.

In the experiments here described a moderatedegree of percussion led to localized disappearanceof myelin without anatomical interruption of axiscylinders. Disturbance in function was then chieflyif not entirely related to motor conduction, and,though complete for three to four days, made rapidrecovery. This milder type of lesion resembles theischmemic lesion induced by compression, thoughit is of briefer duration. The possibility of acommon mechanism in both compression and per-cussion lesions therefore arises.Though percussion of a degree sufficient to cause

impairment of conduction often led to some effusionof blood into the adventitia of the larger veins in theepineurium, these hemorrhages are insignificant inrelation to the remainder of the lesion and the

changes described were independent of them. Inother experiments, sub-perineural hemorrhage,induced by gentle kneading, has in any case beenfound to be innocuous. No evidence of thrombosisof vessels was found early or late after percussion.All vessels in the affected area were greatly con-gested. If the perineurium was ruptured by theoriginal injury the swelling or pseudo-neuroma wasdecompressed by the rupture, with herniation ofendoneurium and nerve fibres through the tear.The damage to the myelin and the cellular response,however, still occurred throughout the nerve,though not to the same extent. The naked axiscylinders were then greatly swollen. Demyelina-tion therefore does not depend directly on theinternal pressure developed by cedema confinedwithin intact perineurium, though it was lessextensive when perineural rupture had occurred,nor does spontaneous decompression prevent theappearance of the fibrillary regeneration of somefibres throughout the nerve bundle. Lastly, whereascompression of nerve affected all sizes of nervefibres alike (Denny-Brown and Brenner, 1944b), thedemyelination from percussion was in all but themost severe degree selective for the largest mye-linated fibres.

Escape of some of the largest fibres from demye-lination, as in Fig. 2B, was related to slight asym-metry of the lesion, for they then lay to one or otherside of the nerve bundle. In general we weresatisfied that in the percussion lesion the measurableextent of demyelination in any given fibre bore arelation to the size of fibre, and was thus longest inthe largest fibres. In ischemic lesions the demye-linated gaps were by contrast surprisingly uniformin length (Denny-Brown and Brenner, 1944a and b).The ischemic lesion also lacked the Schwann cellproliferation which was a late feature in the per-cussion lesion, and the demyelinated axon stainedless well with silver. The essential nature of thepercussion response, however, appeared to be thesame in all kinds of medullated nerve fibres andvaried only in extent. Non-medullated nervefibres often presented a beaded appearance in pass-ing through a pseudo-neuroma, but this in somespecimens was the result only of a series of ovalswellings on each such fibre, and was sometimesseen in fine medullated fibres. It is possibly only anexpression of cedema.The finding of preservation of sensory conduction,

even to light touch in one experiment, where nomyelin was found to pass continuously across thepseudo-neuroma, indicates that the dissociation ofconductile function was not merely the presence orabsence of demyelination. The rapid recovery offunction, in spite of persistence of the histologicalchange, also warns against close identification ofloss of conduction with loss of myelin. We havealready indicated a similar discrepancy in ischemiclesions. It was notable that the transient motorparalysis, when partial, was more complete for distalmuscles (spreading), a feature which is probablyrelated to the finding in partial lesions in human

91



j.com/

J Neurol N


ownloaded from



Jr~~ ~ ~ ~ ~ ~~~n~ ~ ~ ~_- _-

' f

_ . *~~~~~~~~~~.0_W _0 -*V

WN_' wv_,,_-.At

~

4.

loc~~~~~~~~~~~~~~~~~~~~~~~-Or' - --w - _

G t. 0 , jA - -p- &..

FIG. lOA.-Peroneal nerve at midsciatic level. Percussion with herniation through perineurium 12 days earlier.Spielmeyer's method for myelin. To show the herniation and decompression.

FIG. IOB.-As IOA. Gros-Bielschowsky method. The ruled line is equivalent to 0 5 mm.FIG. 1Oc.-Fibres in the centre of the nerve in IOA stained by Gros-Bielschowsky method.

92

I~Ilbilw C.

a -WM- ---- l-,-

4 AWAMNMllvm.- ow,

0



j.com/

J Neurol N


ownloaded from



nerves that the proximal muscles may recoverrapidly, the distal only after long delay or not at all(Claude and Dumas, 1917, Platt, 1921). Though itis tempting to relate such disturbance to selectiveeffect of fibre size the evidence is lacking at present.

Seddon (1943) has recently introduced the termneurapraxia to describe disturbances in whichparalysis occurs from a nerve injury which does notresult in peripheral degeneration. The transientdisturbances of function following percussion ofnerve fall into this category. On account of thedifferences between the transient disturbance pro-duced by percussion, and that of ischemia or com-pression, we prefer to use the term neurapraxiafor the clinical phenomeonon in cases of nerveinjury when the actual mechanism is unknown. Wehave also avoided the use of the term contusion, forthis implies the presence of extravasated blood, whichis not an essential part of the percussion lesion. Wehave in fact been able to demonstrate that intra-fascicular bleeding does not disturb function, ifproduced by kneading or other gentle physicalmanceuvre. The red blood corpuscles are hemo-lysed and rapidly absorbed.The initial mechanism of stasis and histiocyte

response requires explanation. The almost com-plete absence of additional aedema and cellularchange following percussion of degenerated nerve,and the striking mobilization of the histiocytes thenalready )resent, led to the conclusion that per-cussion fed to rupture of the sheath of Schwann.Longitudinal continuity of the Schwann tube wasretained, so that such rupture must therefore havebeen a longitudinal splitting. The argument is alsoapplicable to the immediate disturbance of intactnerve by percussion. Splits in the myelin andSchwann sheath were observed on the third day.The breakdown of myelin is stated to release power-ful tissue irritants such as cholesterol and thus inducestasis and histiocyte response in the central nervoussystem (Marsden and Hurst, 1932). This irritativephenomenon in peripheral nerve can also be relatedto breakdown of myelin by rupture of the Schwannsheath. Some of the bizarre bulbs, sprays, and side-branches of the axis cylinder which lie in theendoneurium at the level of a severe percussion mayalso represent herniations of the axon through theSchwann sheath at the time of injury. Like acuteinflammatory reactions in other tissues the stasisand histiocyte response fail to appear in the absenceof an intact blood supply. We have presentedevidence that the fluid is not derived from anycirculation of normal endoneural fluid, and oncepresent does not migrate proximally or distally.We have found the " barbed " axis cylinders and

fibrillary cords described by Tinel (1917), andrecognize the former as cedematous fibres stillretaining myelin, the latter as regenerating fibrils.We have in addition described the truly demye-linated but otherwise intact axis cylinder, the survivalof which is associated with brief block of motornerve impulses. The lesion is never as pure as thecompression lesion. A small proportion of

degenerated fibres, and a larger proportion of intactmyelination, is always present. An initially completeparalysis was associated with some degeneration,an incomplete paralysis with some unaffected fibres,but the number in each case was very small, andallows transient paralysis to be attributed to demye-lination. It will be noted that the loss of myelin isapparent for two to three weeks after a four-dayparalysis, and that the renewed myelin film was notobserved earlier than the thirteenth day. We havenoted this disparity between the slow process ofremyelination and return of function in relation toischimic lesions also (Denny-Brown and Brenner,1944a and b).Apart from the cedema and tissue response

pseudo-neuroma is distinguished from true neuromaby the parallel disposition of nerve fibres, whetherthey be the remaining " barbed " cedematous axons,the fibrillary bundles which betoken regeneration,or the demyelinated, single axis cylinders. Thisparallelism indicates that the sheath of Schwann hasnot ruptured transversely. If the axon is ruptured,regeneration occurs with great rapidity, associatedwith a profusion of secondary sprouts within thesame sheath and proliferation of Schwann cells.Though most of the finer branches adopt a spiralcourse, the intricate spirals of Perroncito do notoccur, and such lesser spirals as are seen are rapidlyresorbed. Some branches run centripetally, andthese already present resorptive knobs by thefourteenth day. The true Perroncito spiral isformed almost entirely of recurrent centripetalbranches, indicating resistance to centrifugal pro-gress. Further, though resorption of the inter-rupted axis cylinder is very rapid in the pseudo-neuroma (in three days following percussion belowa ligature), one of the regenerative branches hadfrequently penetrated far into the peripheral nerveby the fourteenth day, and at the level of the pseudo-neuroma had already attained nearly the proportionsof a mature axon. In contrast, a true neuromapresents essentially a confusion of direction of nervefibres, which, with the persistent Perroncito spirals,indicates a transverse disruption of the sheaths ofSchwann.The cedema associated with pseudo-neuroma is

accompanied by a brisk endoneural reaction.Though increase in number of the pale endoneuralnuclei has been observed in simple oedema of nervewithout damage to the nerve fibres (Denny-Brownand Brenner, 1944b), they are particularly clearlydefined in relation to demyelinated but intact axons.The Schwann nuclei also proliferate, but to a lessremarkable degree and at a later stage. Long afterrestitution of myelin the pale, oval nuclei ofendoneural cells lie in the interspaces between thenerve fibres and mark the original site of the pseudo-neuroma. The endoneural cells are then presum-ably those that surrounded the naked axon in theacute stage of the lesion.When the original lesion had been sufficiently

severe to be followed by a paralysis lasting untilregeneration occurred, a similar increase in number

93



j.com/

J Neurol N


ownloaded from



of endoneural nuclei surrounds the fibrillary bundleof a regenerating axon in a pseudo-neuroma, and isthen less obvious only because of the greater pro-liferation of Schwann cells. A persistent spindle-shaped swelling marked the site of percussion afteras long as three months. The nerve fibre andsheaths were reconstituted (Fig. 8A), but occasionalphagocytes and a large number of endoneural cellsand fibres filled the dilated endoneural spaces. Thephagocytes were fewer than in the degeneratedperipheral segments. The tissue reaction is moreintense than in the transient type of lesion. Wewere not able to identify additional fibroblasts.The traumatic neuroma of " traumatic ulnarneuritis" where trauma is repetitive, is likely to bethe cumulative effect of repetition of either mild orsevere degree of this process. The gross over-growth in the endoneural spaces then strangulatesthe nerve fibres. Such a condition does not appearto arise as a result of a single percussion.These findings throw some additional light on the

phenomena of regeneration of nerve. They indicatethat the phagocytic cells, which are attracted bydegenerating myelin, are true histiocytes, probablyderived from Maximow's " polyblasts " as Nageotte(1932) asserts. The " foam cell" appearance ofthe Wallerian ovoids is but another aspect of thesame phagocytic " Kornzellen " (Doinikow, 191 1)which lie free in the endoneural oedema. Neitheris derived from the Schwann cell. CEdema ofnerve is a characteristic result of trauma as in othertissues, and we find no evidence of migration offluid or phagocytes either proximally or distally.It is difficult to concede the possibility of anendoneural or perineural circulation in the absenceof such migration. The subsequent resorption ofregenerative neurofibrillar outgrowths, excludingthat that has been successful, is also clearly indicated.This phenomenon was suspected by Cajal (1928),and was clearly observed in the present experiments.It would appear desirable to investigate the influenceof such abortive branches upon function of theconductor concerned, for such branching is the onlyhistological evidence which suggests a clue to thephenomenon of hypervsthesia. The behaviour ofendoneural cells in covering the naked axis cylindersindicates a role previously confused with fibro-blastic reaction. The phenomenon appears to haveimportance in the genesis of true neuroma, anddemonstrates an essential difference between theendoneural cell and mesoblastic elements.

Rupture of the perineurium is an event which wasmore prone to occur if there had been asymmetryin percussion of the nerve, so that the momentarycompression occurred more on one side than theother. Although the subsequent cedema was absentor greatly lessened, the histological and functionaldisturbance consequent upon such rupture wasdisastrous for nerve fibres near the resulting hernia.Nerve fibres, Schwann cells, endoneurium andperineurium proliferate through the gaping opening.There is formed a true neuroma with a characteristictangle of fibres and cells. The neuroma may be

lateral and single, or multiple, but fused to thefasciculus in such a way as to make it spindle-shaped. Except for some greater adherence tosurrounding tissue it was usually indistinguishablefrom pseudo-neuroma to the naked eye. It wasassociated with great disorder of conduction ofsensation and movement in some fibres for longerthan six weeks, which is longer than the usual dura-tion of paralysis from severe uncomplicated per-cussion at the midsciatic level. This delay inregeneration is related 'to the malalignment ofSchwann tubes due to the perineural hernia, and tothe great degree of lateral fibrillary and endoneuralproliferation thus brought about. Even when thehernia is single and small the remainder of the nervefibres in the nerve show the patchy demyelinationcharacteristic of percussion lesion, with corre-sponding partial or complete block. All combina-tions between the pseudo-neuroma and multipleperineural herniation are therefore to be expected.This complication is likely to be the cause of lateneuroma in cases such as that described by Maculaireand Navarre (1915), in which a projectile passed nearthe median nerve, but did not touch it, and yet alarge neuroma resulted.The descriptions of " spindle neuromata," found

in association with prolonged partial disorder offunction, and particularly with causalgia, suggeststhat multiple perineural rupture in the fasciculiwhich make up the nerve trunk may well be themechanism of such residual disability. The dis-turbance of function associated with simple pseudo-neuroma was in contrast repaired extremely rapidlyand effectively. We therefore suggest that, thoughrepeated mild trauma may lead to permanent over-growth of endoneurium, and consequent cumulativeimpairment of function, a single percussion of nervecan produce disorder in function lasting beyond theperiod of regeneration only through perineuralrupture.

Conclusions1. A single percussion of a nerve trunk resulted

primarily in damage to the myelin sheath, probablywith longitudinal rupture of the Schwann sheathand escape of damaged myelin into the endoneuralspaces. The axis cylinder escaped interruption ifthe blow was moderate. Even in the most severeblows the axis cylinders of the finest medullatedfibres and non-medullated fibres were not inter-rupted, and consequently did not degenerate. Thesheath of Schwann was not divided transversely byintensities of blow sufficient to disrupt both myelinand axis cylinder.

2. A localized swelling or pseudo-neuromadeveloped rapidly at the site of injury and persistedfor some weeks. The swelling was due to cedema,associated with the rapid appearance of largenumbers of histiocytes. This reaction was relatedto the breakdown of myelin, and occurred ifregeneration was prevented by a proximal ligature.It was prevented by interruption of blood supply.

3. In a moderate degree of lesion, whep naked

94copyright.

on June 10, 2021 by guest. Protected by

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.7.3-4.76 on 1 July 1944. Dow

nloaded from


THE EFFECT OF PERCUSSION OF NERVEaxis cylinders of all sizes remained intact, there wasretention of sensation with transient completeblock of impulses in motor fibres. The smallestmedullated fibres escaped or were less severelyaffected. Non-medullated fibres were then beadedbut not interrupted. The dissociation in functionclosely resembled that caused by ischkmia, thoughhyperesthesia to touch was more prominent, andrecovery of motor conduction occurred morerapidly. Both the degree of cedema and extent ofdemyelination were lessened if the oedema wasdecompressed by concomitant rupture of the peri-neurium. Some ischaemia was therefore presumedto accompany the aedema and stasis, but was not anessential feature of the mechanism of the lesion.

4. The phagocytic reaction was derived fromhistiocytes. The Schwann cells played no part inthe phagocytic process which was fully developedat the end of the first week when the Schwannnuclei began moderate proliferation along ademyelinated axon, or more profuse multiplicationwithin the tube of a regenerating axon. Theendoneural cells surrounded the isolated Schwannsheaths and remained in increased numbers eventhree months later when myelin structure wasnearly completely reconstituted. This increase ofendoneural cells was minimal in ultimate degree aftera single percussion.

5. Transient paralysis lasting three to four dayswithout loss of gross sensation was the effect of apartial lesion. A pseudo-neuroma was then alsopresent. The majority of axis cylinders had lostmyelin in passing through the lesion, but recoveredit below. The extent of loss was greatest in thelargest fibres. The characteristics of the demye-linated axon are described.

6. If the axis cylinder was transversely inter-rupted by the percussion, degeneration occurred inthe distal portion of the nerve fibre. Regenerationwas then extremely rapid (four to eight weeks) andcomplete. Such regeneration of axis cylinders inintact Schwann tubes involved an initial productionof a large number of fibrils by the parent axiscylinder. Early maturation of one of these, pre-

sumably that which penetrated farthest, was fol-lowed by progressive resorption of the remainder.

7. Delay in regeneration resulted only when theperineurium was ruptured by the percussion.Herniation of nerve fibres and endoneurium thenoccurred, with true neuroma formation.

8. Percussion of a nerve trunk in process ofWallerian degeneration resulted in liberation ofintratubal phagocytes (Wallerian ovoids). Thephagocytes then assumed the characters of free fatgranule cells. The mechanism of liberation ap-peared to be traumatic splitting of the Schwann tube.

REFERENCESBielschowsky, M., and Valentin, B. (1922). J. Psychol.

Neurol., Lpz., 29, 133.Cajal, S. Ramon y. (1928). Degeneration and Regenera-

tion of the Nervous System. Trans. by R. May,London.

Claude, H., and Dumas, R. (1917). Bull. Soc. me'd.Hop., Paris, 33, 626.

Dejerine, J., Dejerine, Mme., and Mouzon, M. J. (1915).Pr. mid. No. 31, 245.

Denny-Brown, D., and Brenner, C. (1944a). Arch.Neurol. Psychiat., Chicago, 51, 1.

(1944b). Ibid., 52, 1.Doinikow, B. (1911). Histol. u. histopath. Arbeiten v.

Nissl-Alzheimer, Jena, 4, 445.Guttmann, L., and Medawar, P. B. (1942). J. Neurol.

Psychiat., 5, 130.Holmes, W., and Young, J. Z. (1942). J. Anat., London,

77, 63.Maculaire and Navarre (1915). Paris. mid., 17, 378.Marsden, J. P., and Hurst, E. W. (1932). Brain 55,

181.Medical Research Council Special Report Series No. 54.

The Diagnosis and Treatment of Peripheral NerveInjuries, London, 1920.

Mitchell, S. Weir (1872). Injuries of Nerves and theirConsequences, Philadelphia.

Nageotte, J. (1932). In Section V, of Cytology andCellular Pathology of the Nervous System, Vol. 1,edited by Wilder Penfield, New York.

Platt, H. (1921). The Surgery of the Peripheral NerveInjuries of Warfare, Bristol.

Seddon, H. J. (1943). Brain, 66, 237.Tillaux, P. (1866). Des Affections Chirurgicales des

Nerfs, These de Paris.Tinel, J. (1917). Nerve Wounds. Symptomatology of

Peripheral Nerve Lesions Caused by War Wounds.Trans. by Rothwell, F., and Joll, C. A., London.

Tubby, A. H. (1915). Brit. med. J., 1, 57.Weiss, P. (1943). Anat. Rec., 86, 491.

95copyright.

on June 10, 2021 by guest. Protected by

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.7.3-4.76 on 1 July 1944. Dow

nloaded from

THE EFFECT OF PERCUSSION OF NERVE - BMJ · Tinel (1917) summarizes the findings ofDejerine andhis...

Documents

Transcript of THE EFFECT OF PERCUSSION OF NERVE - BMJ · Tinel (1917) summarizes the findings ofDejerine andhis...