brain-stem lesions after head injury

J. clin. Path., 23, Suppl. (Roy. Coll. Path.), 4, 154-165

Brain-stem lesions after head injury

B. E. TOMLINSONNewcastle upon Tyne

A full description of brain-stem injuries shouldinclude their mechanisms and types of damage,distinguishing those which are primary and occurat the time of the injury, from those which aresecondary and result from changes elsewhere inthe brain or in other parts of the body; and thevarious lesions should be related to the clinicalcondition of the patients. Unfortunately, know-ledge is insufficient to allow any ofthese objectivesto be satisfactorily fulfilled. Morphologicalstudies on human material are necessarily limitedto fatal cases, and largely to examples of severeinjury. Knowledge of brain-stem changes afterrelatively minor head injuries is fragmentary. Itdepends on the occasional case recovering orrecovered from a non-fatal head injury and dyingof some other cause coming into the hands of apathologist sufficiently interested to study thebrain-stem in detail. This paper consists primarilyof an account of the morphological changes inthe brain-stem in injured patients; mechanismsare not discussed and reference to other factorsconcerned is made only when this is essential.Many accounts of brain-stem injury have been

restricted to descriptions of the haemorrhageswhich occur in rapidly fatal cases, and the lesions,often haemorrhagic, which result from increaseof supratentorial mass. Important as these are,it cannot be too strongly emphasized that lesionsof different kinds and origins occur in the brain-stem after injury. The full spectrum of changecan be appreciated only when cases survivingdifferent periods of time are extensively examinedusing a variety of staining procedures. Severelesions may be present though the brain-stemappears normal or nearly so to the naked eye.

Primary haemorrhages are first described andthen the secondary lesions from tentorial hernia-tion which tend to differ in distribution and areoften partly or largely ischaemic. Other less well

recognized changes and their distribution arethen described. Often they are the result ofprimary injury but are recognizable by presenthistological techniques only in those who survivehours or days. Finally, the changes which occurin the brain-stem of patients surviving in comafor long periods are briefly stated and related tothe other findings.

Primary Brain-stem Haemorrhages

Haemorrhages are usually the only evidence ofinjury to the brain-stem in those dying immedi-ately or within a few hours. They may be visibleto the naked eye or only microscopically. Failureto realize this has led to underestimates of thefrequency of brain-stem haemorrhage after injury,and indeed to the frequency of brain-stem lesionsof all kinds. Thus, Jellinger (1967) reported thatonly 43-5% of 415 fatal head injuries showedevidence of brain-stem lesions, but extensivehistology was done only in the 38 cases dyingafter prolonged coma. Tandon and Kristiansen(1966) report a group of 37 fatal cases with brain-stem symptomatology in whom no structuraldamage was seen at necropsy, but histologicalstudies were not apparently done. In cases withbrain-stem symptomatology, Zoltan (1966) foundno lesions at that level and was prepared to seekthe origin of the symptoms in cortical and sub-cortical damage. By contrast, Matsuoka, Sakaki,and Okamoto (1967) in a careful histologicalstudy found haemorrhagic lesions in the brain-stem in 22 out of 25 cases, and Mayer (1967)found them in all his 25 cases dying within anhour of injury. Mayer plotted primary brain-stemhaemorrhages only from such a group.

Primary haemorrhages are usually small and


tend to have a special distribution. First, they areoften present in the subependymal tissues aroundthe third and fourth ventricles and in the peri-aqueductal grey matter; extension of bleedinginto the ventricles is frequent. Second, they oftenaffect the lateral structures of the brain-stem, andthough any part may be damaged, the tissuesnear the lateral mesencephalic sulcus, thesuperior cerebellar peduncles, the colliculi, andthe basis pedunculi are probably most frequentlyinvolved. In addition, the midline ventral struc-tures of the rostral brain-stem and hypothalamusmay be affected. Microscopy is frequently neededto detect many of the haemorrhages. Figure 1shows diagrammatically the distribution ofhaemorrhages in a case dying within minutes ofinjury and in which only the larger rostral lesionswere visible to the naked eye.

Microscopically the bleeding may be peri-arterial, perivenous, or pericapillary or it maydisrupt the tissues without any apparent relationto a vessel in a single section. To assume that thetype of vessel giving rise to the bleeding can bededuced from the site of the extravasated bloodas Matsuoka et al (1967) have done would seemunwise, but Mayer (1967) stated that the laterallysituated haemorrhages are predominantlyarterial and those near the ventricles are mostlyvenous. In multiple sections he identified tornarteries in 17 and torn veins in 21 of his 25 cases.In a few cases capillaries were apparently tornfrom arterioles.

In the writer's experience, haemorrhages intothe rostral brain-stem are usually more numerousand severe than those in the medulla in rapidlyfatal injuries. Indeed, the lower stem may escapecompletely and the most caudal haemorrhagesmay be situated at the junction of the mid-brainand diencephalon. However, laceration or severecontusion of the medulla or pons may be foundin those with severe bone damage in the vicinityof the foramen magnum, such as fracture dis-locations of the atlanto-occipital joint or fractureswhich run into the foramen. Primary haemor-rhages, though widely distributed, are usuallysmall in those who die quickly which suggeststhat they suffer an injury incompatible with lifefor more than a brief interval. As a corollary,and perhaps more important, this finding suggeststhat patients who are alive even a few hours aftera head injury have not suffered widespreadprimary brain-stem haemorrhage, even of micro-scopic type. Further evidence for this is that evenminute haemorrhages are only occasionally foundin the medulla in those who survive more than afew hours, though small haemorrhages with dis-tribution characteristic of primary bleeding,may be found rostral to the mid-pons even incases surviving several days. The presence in arapidly fatal case of several microscopic haemor-rhages in multiple sections supports the notionthat a brain-stem injury had occurred whichhad been incompatible with life for more than a6

155

2s

f (1/ c-AlS= C

5).1----(..ll %a,!..

Fig. 1 Diagrammatic representation of thedistribution ofprimary haemorrhages in the brain-stem from a girl of 15 years who died within 15minutes of a road accident. A.fracture in theoccipital bone reached the left side of the foramenmagnum. All the haemorrhages in the fouir cauidalsections were microscopic.

brief period. It is not suggested that such haemor-rhages are responsible for death; they merelyindicate damage of a type and distribution whichis incompatible with life.

Secondary Brain-stem Lesions following TentorialHerniation

The secondary destructive lesions in the brain-stem are not confined to head injuries but mayfollow any increase of supratentorial mass. Theirimportance in head injuries is that they mayproduce a fatal issue in patients without brain-stem damage at the time of the injury. Forexample, patients who are not unconscious afterinjury or with only a short period of coma fol-lowed by a lucid interval, may develop subduralor extradural haemorrhage; then tentorialherniation and death from brain-stem lesionsmay occur, both potentially preventable by earlysurgery.

Great variation in brain-stem movement anddistortion and consequent pathological changeoccurs with supratentorial space-occupyinglesions. In cases of head injury, these are usuallyextra- or intracerebral haemorrhage, brain swell-ing, or a combination of these. In uncomplicatedsubdural or extradural bleeding the size of the

B. E. Tomlinson

haematoma is obviously of great moment,though the volume of skull unoccupied by brainis also important. Thus, the cerebral atrophy ofold age often allows a considerable amount ofextracerebral haemorrhage to occur withoutsevere effects on the brain-stem (Aronson andOkasaki, 1963). The speed of development of thesupratentorial mass is also of significance, anobservation experimentally confirmed by Wein-stein, Langfitt, Bruno, Zaren, and Jackson (1968).Herniation may occur within a very short timeof injury and then it may be very difficult orimpossible to decide which of the resulting brain-stem haemorrhages are primary or secondary.The situation of the supratentorial mass and itsunilateral or bilateral distribution largely deter-mine whether or not significant lateral displace-ment of the stem will occur, but Sunderland(1958) and Corsellis (1958) have stressed theimportance in variations in size and shape of thetentorial opening on the results of herniation.Sunderland (1958) detailed and beautifully illus-trated the varying anatomical relations attentorial level and discussed how these couldaffect the results of raised intracranial mass. Allthese considerations determine whether move-ment of the rostral brain-stem is mainly lateralor downwards or a combination ofthese with side-to-side compression. The importance of hernia-tion of the medial temporal lobe has probablybeen overstressed; its occurrence depends interalia on space being available at the tentorialopening.Downward displacement produces marked

elongation, distortion, and stretching of the mid-brain and often of the thalamus and hypo-thalamus. Howell (1961) has stressed that fixationof the medulla and upper cervical spinal cordgreatly limits their possible downward displace-ment so that the upper brain-stem shortens andangulates backwards and literally buckles undersufficient pressure. Sunderland (1958) alsostressed posterior displacement and rotation ofthe mid-brain and Weinstein et al (1968) showedthat its dorsal surface was often displaced morecaudally than the ventral surface.

Brain-stem herniation is therefore a com-plicated phenomenon. It is hardly surprising thatthe lesions found in fatal cases vary considerably,that different features are emphasized by differentauthors, or that disagreement still exists aboutthe origins of lesions.Some of the above features will not be appre-

ciated at a routine necropsy when the brain isremoved from above after cutting the tentoriumat its attachments. Even then elongation anddownward movement of the cerebral pedunclesis often visible since the peduncles are moreprominent than usual on the ventral surface,protruding abnormally from the hemispheres.Some lateral shift may also be seen. Markeddownward displacement of the diencephalon withsome lateral movement or twisting is often appre-

156

ciated, particularly on examining the mammillarybodies. Haemorrhages or necrosis in the hypo-thalamus or into the herniated medial temporallobes or the occipito-temporal lobe are often,visible in the uncut brain. The pons is oftenswollen. Transverse sections often confirm thedownward displacement and elongation of themid-brain and caudal diencephalon. Lateral shiftis revealed by the eccentric position of theaqueduct which is almost invariably greatlynarrowed and slit-like or totally closed. The side-to-side compression of the mid-brain and theanteror-posterior elongation with greater down-ward displacement of the dorsal in relation to theventral structures is often easily appreciated.Distortion and downward movement are detect-able usually only in the mid-brain and dience-phalon though antero-posterior elongation maybe visible in the upper pons.The major effects of these movements and com-

pressions are often visible macroscopically assmall or gross haemorrhages, haemorrhagesassociated with infarction, or as infarction alone.Sometimes microscopy is needed to detect earlyinfarcts which may be the only evidence of dis-tortion of the stem. Haemorrhages are the mostcommon obvious abnormality. The principalhaemorrhages are usually midline in the upperpons and throughout the pontine and mid-braintegmentum (Figs. 2 and 3) though they often

Krl,

Fig. 2 Diagrammatic representation ofsecondaryhaemorrhage and infarction in the brain-stem froma man of58 years who developed bilateral subduralhaematomata without any history of trauma. Thesolid black areas represent haemorrhage, and thelines ischaemic necrosis.


spread almost to the lateral limits of the mid-brain. They may appear as numerous, small focalhaemorrhages, but are often massive and con-fluent, spreading from the central rostral ponsthroughout the tegmentum to produce a butter-fly-shaped haemorrhage. Sometimes they pre-dominate on the lateral aspect of one cerebralpeduncle on the side opposite to the supra-tentorial mass when the latter is predominantlyunilateral; these result from pressure against thefree edge of the tentorium. Extensive haemor-rhage may extend into or rostral to the superiorcolliculi, involving the medial parts of each rednucleus and substantia nigra, and even disruptthe posterior thalamus and structures in the floorof the posterior third ventricle. When haemor-rhage is gross, evidence of infarction may beminimal in surrounding tissues, but when haemor-rhages are less massive, although extending overthe same territory, they are almost invariablyaccompanied by surrounding macro- or micro-scopic infarction. Occasionally, infarction with-out haemorrhage occurs. The haemorrhages aremostly pericapillary or interstitial when they aresmall or not extensive, though perivenoushaemorrhage around dilated veins may also beseen. In those surviving a few days, many axonretraction balls are often present in the vicinityof the haemorrhages or infarcts; in the occasionalcase with long survival, softenings with or with-out iron-laden macrophages will be found in thesame areas (vide infra).The frequency and severity of central lesions

in the pons and tegmentum following herniationand the relative sparing of lateral mid-brainstructures can be readily confirmed in most casesdue to trauma or non-traumatic causes. Massivehaemorrhage is probably a terminal event andmay disrupt the entire rostral stem. In less severecases and in those showing predominant infarc-

157

tion, central and paracentral structures areusually heavily involved (Wolman, 1953). Tom-linson (1964) illustrates the extensive mid-linedestruction which may be found in those surviv-ing in coma for long periods after herniation,and Strich (1969) shows a similar case of longsurvival in 'coma vigil' with smaller, widespread,but still predominantly central and paracentrallesions. Laterally placed pontine or mid-brainnecrosis or haemorrhage due to herniation israre in the absence of central lesions, thoughperipheral haemorrhages certainly occur in thecontralateral cerebral peduncle due to lateraldisplacement by herniation, and other laterallesions also may be seen.Disagreement still exists about the origin of

the mid-brain lesions in tentorial herniation. Anyexplanation must account for the frequentinfarction, sometimes with little or no haemor-rhage, and for the fact that some patients in deepcoma for a short time from tentorial hernia-tion recover completely on removal of the supra-tentorial clot. This evidence denies the place ofhaemorrhages from venous obstruction as theprimary cause in many cases, though the presenceof dilated veins and perivenous haemorrhage hasproduced advocates for the venous origin of thebleeding (Scheinker, 1945; Poppen, Kendrick,and Hicks, 1952; Stroobandt, Brucher, and vande Voorde, 1967). Infarction without haemor-rhage indicates an arterial origin. There is muchto support this view (Blackwood, 1963), includingthe evidence of torn small arteries obtained byinjection studies (Johnson and Yates, 1956) andof considerable arterial distortion (Hassler, 1967).The coma in those recovering completely musthave been due primarily to distortion of the stemand its functional disturbance without the irre-versible ischaemia or haemorrhage which mayoccur subsequently.

Fig. 3 Typical central pontine and mid-brain haemorrhages found after tentorial herniation.from a man of28 years who died within one and a halfhours ofa severe head injury. Also present were multiple corticalcontusions and massive hemisphere swelling. The rostral mid-brain photograph suggests moderate antero-posterior elongation.

6*

B. E. Tomlinson

( .

158

Fig. 4 Diagrammatic representation of brain-stem-.< lesions in a woman of 73 years who died in coma

seven days after a head injur*v. There was no evidenceof herniation: the brain was small (1,080 g) with aconsiderable gap between brain and skull; there wasno sub- or extradural bleeding and only smallhaemorrhages in the cerebrum, particularly in the

--\) < '-/corpus callosum. The solid black areas represent/,' haemorrhage, the lines foci of ischaemic necrosis,

and the stippled areas the sites ofaxon retractionballs.

7?X,0

Fig. 5a Numerous axon retraction balls in the Fig. 5b Middle cerebellar peduncle showing lessupper pons in a woman of 73 years who died in coma numerous but more densely stained axon retractionseven days after a head injury. Haematoxylin and balls in a silver impregnation. From a man of 41eosin, x190. years dying seven days after head injury. Other

evidence of tract degeneration also seen. Glees andMarsland, x 190.

--A; --7.'I

l,' 1 1/

Braint-stem lesions after head injury

Delayed Changes from Primary Injury

DAMAGE TO AXONS

Though some primary haemorrhages may befound in the rostral stem in cases surviving fordays or weeks, other primary changes becomeapparent, particularly in those who had been incoma from the time of the injury and withoutevidence of tentorial herniation. Damage tonerve fibres is important. This was first describedby Strich (1956 and 1961) particularly in thehemispheres. Axon damage can be detected inhaematoxylin and eosin sections in most cases;it tends to occur in certain sites, those whereprimary haemorrhages are often found in rapidlyfatal cases, and to involve certain fibre tracts, par-ticularly in the mid and rostral pons and the mid-brain (Fig. 4). In the early days after injury, fibredamage is manifest by the appearance of axonretraction balls (Fig. Sa and b), usually inter-preted as axoplasm escaping from the ends oftorn axons. In addition, many axons are swollenor grossly distorted. Retraction balls may be fewor many in the descending corticospinal andcorticopontine fibres of the upper pons, andoccasionally are numerous in pontocerebellarfibres. Not infrequently these tracts show large

c C)

I1.:..

A. N

159

masses of axon balls in a single section, suggest-ing that many axons were torn at approximatelythe same level. In sections a few millimetres aboveand below, there may be little or no evidence ofaxon destruction, strongly supporting thelocalization of injury. Similar evidence of severeaxon damage may be found in the superiorcerebellar peduncles, with involvement of theneighbouring lateral lemniscus and central teg-mental tract, in the medial lemniscus, and in thedecussation of the brachium conjunctivum. Thetissues adjacent to the lateral mesencephalicsulcus are often most severely involved. Herenumerous tracts may be damaged, including thelateral spinothalamic, spinotectal, and the lateraltectopontine tracts, the lateral lemniscus, and thelateral limits of the central tegmental tract.Damage may not be limited to torn axons: itranges from isolated retraction balls to almosttotal destruction of all elements, with or withoutsome haemorrhage. Retraction balls may befound more rostrally in the basis pedunculi, inthe peduncle of the inferior colliculus, in thecolliculus itself, and in the caudal diencephalon.Figure 4 shows the distribution of damaged axonsin a patient who died from pneumonia seven daysafter a head injury. Comparison of the distribu-

,) K.K'

1.

4.

iF I

(

.K I

Fig. 6 Diagrammittiatic representation of brain-stemlesions from a man of 25 years who died four monthsafter a head injury which produced immediate comafrom which recovery was insignificant. The stippledareas represent degenerating fibre tracts, the thinlines are areas of ischaemic softening, and the heavyblack lines are areas of haemosiderin deposition.

Fig. 7 Diagrammatic representation of thedistribution of axon retraction balls in the brain-stemof a woman of 83 years. She was unconscious forfour days after a head injury, recovered slowly, andwas able to answer simple questions by the fourteenthday when she diedfrom a pulmonary embolism.

7-pi--

c- --.,

1

. c)`1.--1

'I.,

1, - __)(\

'

B. E. Tomlinson

tion of the laterally placed lesions in Fig. 4 withthose of the laterally placed haemorrhages in therostral brain stem in Fig. 1 shows a markedsimilarity. This suggests that similar sites in theupper pons and mid-brain are particularly liableto injury in cases of different severity. Further,the tracts which show retraction balls are virtuallythe same as those showing Wallerian degenera-tion in many cases dying after prolonged periodsof coma (Fig. 6). Weight is added to this argu-ment by the distribution of retraction balls shownin Figure 7. The sections were from a patient

Fig. 8a Foci of ischaemic necrosis in the rightsuperior cerebellar peduncle from a woman of 73years surviving in coma seven days after head injury.Neighbouring cerebellar folia also show someischaemic necrosis. Loyez, x 3.

unconscious for four days after a head injurymaking a slow but progressive recovery andanswering simple questions by the 14th day, whendeath occurred from pulmonary embolism.Many axon balls were present in single sectionsat the sites indicated though not in the quantityof concentrations found in cases remaining indeep coma.Ofcourse, not all cases dying in coma of several

days' duration show evidence of axon damage inall these sites; brain-stem herniation in some casesproduces massive ponto-mid-brain lesions andthen it may be impossible to distinguish evidenceof primary injury. However, some evidence ofaxonal damage in the rostral pons, mid-brain, ordiencephalon will be found in most cases dyingin coma from the time of the accident and with-out clinical or pathological evidence of hernia-tion. Often it is widespread and severe andinvolves many of the sites mentioned above.

ISCHAEMIC NECROSISDestruction of tissue with the appearance ofischaemic necrosis may also be found in the mid-brain in cases dying after the second or third dayand in cases which survive many months andeven make some degree of recovery. Typicalchanges of necrosis with macrophage production,gliosis, and cystic change may be seen. Probablythe most frequent site for such lesions is again thetissues bordering the lateral mesencephalic sulcus,though the superior cerebellar peduncles, andparts of the colliculi are not infrequently involved(Fig. 8a, b, and c). Lesions may also occur in morecentral upper pontine and mesencephalic struc-tures and in the diencephalon. These are usually

Fig. 8b Large softening in the right brachiumconjunctivum and smaller lesions on the left in a Fig. 8c Typically placed ischaemic necrosis in thesimilar position and in the decussation ofbrachium tegmentum adjacent to the left lateral mesencephalicconjunctivum from a woman of 53 years unconscious sulcusfrom a man of 41 years unconscious for sevenfor two weeks after head injury and making a days after head injury. A similar smaller area ofconsiderable recovery. She diedfrom pneumonia nine ischaemic necrosis is present in the centre of theweeks after injury. Loyez, x 3. right caudal substantia nigra. Loyez, x 3.

160


microscopic in size but they may be visible to thenaked eye. Some haemorrhage may be associatedwith the infarcts in cases dying during the first10 days and the surrounding area may be thickwith retraction balls, but often no haemorrhagesare present. In patients dying after some weeks ormonths, it may be impossible to demonstratehaemosiderin in the resulting softening. Throm-bosed arterioles and venules and fibrinoid necrosisof other small vessels were found in three caseswith haemorrhagic infarcts. Serial sections fromcases with infarcts have not been examined, butthrombosis of blood vessels damaged by stretch-ing or distortion at the time of injury may beresponsible for some of the lesions.

Infarcts, then, though small, are a relativelyfrequent sequel of severe injury to the mid-brain;many occur at the same sites as primary haemor-rhages in cases dying quickly and at similar sitesto the concentrations of axon balls in cases dyingafter several days. The relative constancy of the

Fig. 9 Focus of microglial proliferation in therostral pons from a man of53 years dying ofbronchopneumonia 16 days after an apparently trivialhead injury which produced deep unconsciousnessfor only five minutes though he remained aggressiveand confused until shortly before death. Two or threesimilar foci, some associated with small haemorrhages,were present in each section examined between themid-pons and inferior colliculus. Haematoxylin andeosin, x 480.

161

sites of small infarcts can bejudged by comparingthe illustrations in this paper with those of Strich(1961 and 1969), Denny-Brown (1962), Zulch(1966), and Jellinger and Seitelberger (1969). Ofcourse, they may be found in brains with centrallesions resulting from herniation, but laterallesions alone are rarely found in cases of hernia-tion without the more typical central lesions.

FOCI OF MICROGLIAL PROLIFERATIONThe infarcts show the usual macrophage response,and these cells are loaded with sudanophiliclipoid after seven to 10 days. By contrast,cellular response to axon retraction balls is oftenminimal or apparently absent even when themajority of fibres in a tract are damaged. Never-theless, foci of microglial proliferation arepresent in the brains of many cases (Fig. 9).These are often situated around or close tovessels, sometimes in association with smallhaemorrhages or degenerating axons, but some-times without other evidence of injury. They varyin size from 100 microns to a millimetre in diam-eter. They occur both in fatal cases and inpatients recovering from relatively minor injuriesbut dying of other causes (Oppenheimer, 1968).They also occur in the brain-stem, particularly inthe tegmentum and brachium conjunctivum andthe lateral and ventral surfaces of the pons(Oppenheimer, 1968). Oppenheimer made aspecial study of microglial foci in the injuredbrain and illustrates their development from thetime of their first appearance at about 15 hoursafter injury. Although they are distributed likeprimary lesions they appear to be more widelyscattered than the other primary effects.

LESIONS IN PATIENTS DYING AFTERMONTHS IN COMABrain-stem lesions are also found in most subjectswho die after having been in coma for weeks ormonths, and in some who partially recover, butcome to necropsy after a long interval. Most ofthese are related to the lesions already described.Thus areas of old ischaemic softening in the

rostral pons and tegmentum, usually pre-dominantly midline, are found in those whosurvive tentorial herniation (Wolman, 1953)(Fig. 2). Occasionally they are very extensiveand the pons is visibly shrunken from loss of itscentral tissues (Tomlinson, 1964). Frequently,however, many of the changes found are theresult of primary damage to the brain-stem, with,in addition, degeneration of descending fibretracts from destruction of tissue above thetentorium and retrograde atrophy of neuronemasses following interruption of their axons.

Degenerating corticopontine and corticospinalfibres in the peduncles, pons, and medulla areparticularly prominent in cases of coma of longduration. With experience, these changes are

B. E. Tomlinson

readily recognized in paraffin sections stainedwith haematoxylin and eosin, appearing asmarkedly vacuolated tracts containing manyempty, irregular spaces with rather ragged delicatemargins: under high power these are recognizableas macrophages. The tract abnormality is readilyrecognized by pallor in preparations stained formyelin, in frozen sections stained for neutral fat,and in Marchi preparations (Fig. 10a, b, and c).Pyramidal tract degeneration is often visible tothe naked eye in the medulla in cases survivingseveral months where one pyramid may begreyish brown as a result of demyelination. Bothpyramids are usually degenerate, but, as Strich(1956 and 1961) has emphasized, equal bilateral

lOb

lOaFig. lOa Degeneration of descending corticalfibres in the pons from a man aged 24 years whosurvived 15 weeks from the time of injury. He wasdecerebrate when first seen, deeply unconscious andonly showed slight lightening of his conscious stateup to the time of death. Haematoxylin and eosin,x 80.

Fig. lOb Degeneration of descending corticalfibres (same case as Fig. IOa). Marchi, x 48.

Fig. 10c Massive degeneration in the brachiumconjunctivum, the dark structures being fat-filledmacrophages. From the same case as Figure 6.Fat Red 7B, x 120. lOc

162


involvement of descending tracts is uncommon(Fig. 11). Sometimes the greater part of thedescending fibre tract degeneration can be tracedfrom above the tentorium and the internal capsulethereby confirming that hemisphere lesions(Strich, 1956 and 1961) have been largely respon-sible. In other cases, extensive lower pontine andmedullary tract degeneration is not associatedwith an equivalent rostral degree of degeneration,suggesting, as the studies of cases dying after afew days indicated, that much disruption offibres often occurs at the level of the brain-stem.Sudanophilic material develops slowly in thedegenerating tracts and strongly positive stainingwith Sudan IV, Scharlach R, or Fat Red 7B takesmonths to develop (Fig. 10c). Much material isin the form of birefringent crystals, well seen inunstained frozen sections; and Marchi prepara-tions may be strongly positive after severalweeks. Cellular response, apart from macrophages,is relatively slight, consisting of occasionalprominent astrocytes in and around the de-generating tracts, and occasional glial clusters,though these may be found in areas not involvedby tract degeneration. The tracts most affectedare those in which retraction balls are seen inlarge numbers in cases dying earlier. Thuscortico-spinal and cortico-pontine tracts, ponto-cerebellar fibres, medial lemniscus, brachiumconjunctivum, lateral and medial tegmentaltracts are commonly involved. Superior andmiddle cerebellar peduncles may show scatteredor large clusters of degenerating fibres. In addi-tion, small cystic foci of softening up to 5 mmacross may be present, particularly in the superiorcerebellar peduncles, the lateral margins of theinferior collicular brachium, and the tissues

Fig. 1 1 Gross demyelination of one pyramid andpartial demyelination of the other from a man of33 years dying 14 months after a head injury. Hewas unconscious for several weeks and made only avery partial recovery, being aphasic, hemiplegic, andtotally apathetic until his death from pneumonia.Numerous other brain-stem tracts showed severedegeneration but there was marked degeneration ofhemispheral white matter so that the pyramidaldegeneration may have resulted from hemispheredamage. Loyez, x 12.

163

adjacent to the lateral mesencephalic sulcus.Occasionally other minute softenings, 100 ,u to500 g diameter, with or without iron-pigmentedmacrophages, may be present elsewhere, andindeed almost anywhere in the upper pons, mid-brain tegmentum, the tectum, peduncles, sub-stantia nigra, or posterior hypothalamus. Peri-ventricular and peri-aqueductal tissues are prob-ably where these small remnants of the originaldisrupted foci are most often found.

In addition, a more diffuse though usuallylight gliosis, and scattered macrophages may bepresent, particularly in the mid-brain. In manycases, marked shrinkage of the whole brain-stemoccurs, easily recognizable to the nakedeye, even inthe uncut specimen; the pons may present slightbilateral concavities on its ventral surface mainlyfrom loss of fibre bulk. Different cases presentwith particular changes, such as traumatic lesionsof various cranial nerves with marked chroma-tolysis of the corresponding neurones. An impres-sion of widespread neuronal loss in the pons andmid-brain is often gained from sections in casesof prolonged unconsciousness, but an objectiveassessment would require an exhaustive andcarefully controlled neuroanatomical study,which so far has not been attempted.

Brain-stem Lesions after Head Injury in PatientsDying of Extracerebral Causes

Those who become conscious after many weeksand survive for months or years with varyingdegrees of physical and mental impairment oftenshow changes similar to those found in casesdying after prolonged continued coma. Thelesions shown diagrammatically in Fig. 6, froma man in coma from injury to death, are verysimilar in distribution and type to those foundin a case in which recovery from coma tookalmost three months. The latter patient eventuallyreturned home for several months, remainedaphasic and hemiplegic (possibly from hemi-sphere lesions), apathetic, and almost completelyhelpless, though apparently aware of his sur-roundings. Figure 8b is from a woman whobegan to recover consciousness two weeks afterinjury; after nine weeks, when she developed arapidly fatal pneumonia, she was feeding herself,-walking a little with the aid of a stick, and talkingsensibly for short periods. In addition to thecharacteristically placed mid-brain softeningsillustrated, severe third nerve degeneration waspresent.The frequency of these changes and their

severity in patients recovering from mild headinjury with short periods of unconsciousness isnot known. Oppenheimer (1968) recorded myelindestruction, axon retraction balls, and microglialclusters in the brain-stem of patients who hadsuffered relatively trivial concussion. In one

B. E. Tomlinson

personally studied case many widely scatteredretraction balls were present in cortico-spinaland cortico-pontine fibres; that patient also hada serious chest injury and died five days after theaccident. In another patient who had been un-conscious for five minutes and died 16 days laterfrom bronchopneumonia, a few retraction ballswere present ventral to one inferior colliculus,occasional small perivenous haemorrhages werefound in the ventral and lateral tissues of theupper pons, and microglial foci (Fig. 9) wereobserved in the rostral pons and caudal mid-brain. These are very similar to the lesions illus-trated by Oppenheimer (1968) who suggestedthat they possibly resulted from displaced vesselsdamaging neural tissue.

Evidence is therefore accumulating that brain-stem lesions occur not only in the great majority,if not all cases of severe head injury with pro-longed unconsciousness, but also in at least somecases of minor head injury associated with tem-porary unconsciousness. In both, the brain-stemdamage is similar in distribution and type thoughthe severity varies.

Conclusions and Summary

Severe injuries produce tearing of various bloodvessels and bleeding in certain sites of the brain-stem. When bleeding is widespread, death isextremely rapid and there is no time for severehaemorrhagic disruption of the stem to occur.In such cases it must be assumed that damage toneural tissue is also severe and widespread; thelesions would be demonstrable with longersurvival, or if histological techniques capable ofshowing immediate neural damage were avail-able. Neural damage produced at the time ofinjury becomes manifest when injury to the stemis not so widespread or severe and the patientsurvives for many hours or several days. Haemor-rhages may be present but they are usually con-fined to the rostral brain-stem; they have a dis-tribution similar to those found in rapidly fatalcases, unless herniation of the brain-stem hasalso occurred. Infarcts develop in many caseswithout evidence of herniation and these arelargely sited in the rostral stem where primaryhaemorrhages are conspicuous. Therefore theseinfarcts probably result from damage to smallblood vessels of a severity less than that whichproduces tearing and haemorrhage. Thrombosisor necrosis of small vessels near the infarcts issometimes demonstrable. Examination of serialsections may show that such injuries to smallblood vessels are frequent and the cause of theinfarcts. Damaged axons are demonstrable inmany cases surviving for several days. They occurin similar sites to the primary haemorrhages andinfarcts and also in certain more centrally placedfibre tracts. The sites of the retraction balls seen

164

in cases dying several days after injury coincidewith the common sites ofprimary haemorrhage inthe rostral stem, with the fibre-tract degenerationfound in many cases surviving in coma for weeksor months, and also in some who partiallyrecover. Axon damage of lesser severity but insimilar locations has been demonstrated in somecases with relatively minor head injury. All theabove features point to a common direct resultof the primary injury.Primary damage to structures rostral to the

mid-brain may also lead to fibre-tract degenera-tion, and secondary herniation at the tentorialhiatus may produce further haemorrhagic orischaemic lesions. The latter are predominantlycentral in situation. Jellinger and Seitelberger(1969) attribute all brain-stem necrosis andhaemorrhage in long-surviving cases to tentorialherniation and suggest that the more lateral loca-tion of many lesions 'must result from factorssomewhat different from those operative in acutefatal cases', when the lesions are more centrallyplaced. It is possible that laterally placed lesionsof primary distribution can occur from herniationwithout central lesions, but they must be un-common except for peduncle damage from lateralmovement of the stem. The evidence, contrary tothis view of Jellinger and Seitelberger andgathered from cases dying at different intervalsafter injury and from patients without evidenceof herniation, is strong.

Quite apart from the frequency and variety ofprimary brain-stem lesions presented in this paperand by other authors (Strich, 1956, 1961, and1969; Crompton, Teare and Bowen, 1966a andb; Oppenheimer, 1968), the possibility thatprimary brain-stem damage in man is associatedwith temporary unconsciousness is attractive inthat it forms a possible link with the lesions foundafter experimental concussion in animals.

Early observations recorded chromatolysisand some loss of brain-stem neurones in differentanimals after concussive blows (Windle, Groat,and Fox, 1944; Groat, Windle, and Magoun,1945; Windle, Rambach, Arellano, Groat,and Becker, 1946). After repeated, spaced, con-cussive blows to guinea-pigs, Windle and Groat(1945) showed that up to half the large inter-neurones of the reticular formation in the brain-stem and cells of the lateral vestibular nucleuscould be lost. Similar changes have not beendescribed in man after trivial or even severe headinjury nor, to my knowledge, is the quantitativeinformation of the nerve cell population of thebrain-stem in man sufficient to allow such workto be attempted. Even if it were, the opportunitiesof examining the brain in cases of simple con-cussion are excessively rare. Evidence of myelindegeneration in the guinea-pig brain-stem(Windle, 1948) and damage to nerve fibres withconsequent Wallerian degeneration in cats(Friede, 1961) have, however, also been describedafter experimental concussion and also occasional

Brain-stem lesions after head injury 165

local cellular proliferation (Bryan and Walker,1969). These experimental findings are not dis-similar in character to those seen in some casesof head injury with short periods of uncon-sciousness. Variations in detail, particularly inrelation to the site of the lesions in man andexperimental animals, is to be expected if onlybecause of the neuroanatomical differences.Further work may yet prove that the morpho-logical changes in animals concussed experi-mentally and in man after accidental concussionare essentially similar.

The author wishes to thank Miss D. Kitchenerfor the preparation of the diagrams and thephotography, and Miss M. McNaughton fortyping the manuscript.

References

Aronson, S. M., and Okasaki, H. (1963). A study of some factorsmodifying response of cerebral tissue to subdural haemato-mata. J. Nearosurg., 20, 89-93.

Blackwood, W. (1963). Vascular disease of the central nervoussystem. In Greenfield's Neuropathology, edited by W.Blackwood, W. H. McMenemey, A. Meyer, R. M. Normanand D. S. Russell, 2nd ed., pp. 71-137. Arnold, London.

Bryan, N., and Walker, B. E. (1969). Cellular response in the brainto concussion. Tex. Rep. Biol. Med., 27, 27-33.

Corsellis, J. A. N. (1958). Individual variation in the size of thetentorial opening. J. Neurol. Neurosurg. Psychiat., 21,279-283.

Crompton, R. M., Teare, R. D., and Bowen, D. A. L. (1966a).Prolonged coma after head injury. Lancet, 2, 938-940.

Crompton, R. M., Teare, R. D., and Bowen, D. A. L. (1966b).Prolonged coma after head injury. Lancet, 2, 1137.

Denny-Brown, D. (1962). The mid brain and motor integration.Proc. roy. Soc. Med., 55, 527-538.

Friede, R. L. (1961). Experimental concussion acceleration.Pathology and mechanics. Arch. Neurol. (Chic.), 4, 449-462.

Groat, R. A., Windle, W. F., and Magoun, H. W. (1945). Func-tional and structural changes in the monkey's brain duringand after concussion. J. Neurosurg., 2, 26-35.

Hassler, 0. (1967). Arterial pattern of human brainstem. Neuro-logy (Minneap.), 17, 368-375.

Howell, D. A. (1961). Longitudinal brain stem compression withbuckling. Arch. Neurol. (Chic.), 4, 572-579.

Jellinger, K. (1967). Haufigkeit und Pathogenese zentralerHirnlassionen nach stumpfer Gewalteinwirkung auf denSchadel. Wien. Z. Nervenheilk., 25, 223-249.

Jellinger, K., and Seitelberger, F. (1969). Protracted post-traumaticencephalopathy: pathology and clinical implications. InThe Late Effects of Head Injury, edited by E. A. Walker,W. F. Caveness, and M. Critchley. Thomas, Springfield,Illinois.

Johnson, R. T., and Yates, P. 0. (1956). Brain stem haemorrhagesin expanding supratentorial conditions. Acta radiol.Stockh.). 46,250-256.

Matsuoka, K., Sakaki, S., and Okamoto, E. (1967). A clinico-pathological study on the causes of hemorrhagic lesionsin the brain stem with head injury. Med. J. Osaka Univ.,17, 257-266.

Mayer, E. T. (1967). Zentrale Hirnschaden nach Einwirkungstumpfer Gewalt auf den Schadel. Hirnstammlasionen.Arch. Psychiat. Nervenkr., 210, 238-262.

Oppenheimer, D. R. (1968). Microscopic lesions in the brainfollowing head injury. J. Neurol. Neurosurg. Psychiat., 31,299-306.

Poppen, J. L., Kendrick, J. F., Jr., and Hicks, S. F. (1952). Brainstem hemorrhages secondary to supratentorial space-takinglesions. J. Nearopath. exp. Nearol., 11, 267-279.

Scheinker, 1. M. (1945). Transtentorial herniation of brain stem;a characteristic clinicopathologic syndrome. Pathogenesis ofhemorrhages in brain stem. Arch. Neuirol. Psychiat. (Chic.),53,289-298.

Stroobandt, G., Brucher, J. M., and van de Voorde, H. (1967).H6morragies c6r6brales multiples d'origine traumatique.Observation anatomoclinique. Acta nearol. belg., 67, 55-72.

Strich, S. J. (1956). Diffuse degeneration of the cerebral whitematter in severe dementia following head injury. J. Nearol.Neurosurg. Psychiat., 19, 163-185.

Strich, S. J. (1961). Shearing of nerve fibres as a cause of braindamage due to head injury. A pathological study of 20cases. Lancet, 2, 443-448.

Strich, S. J. (1969). The pathology of damage due to blunt headinjuries. In The Late Effect of Head Injuries, edited byE. A. Walker, W. F. Caveness, and M. Critchley, pp. 501-524. Thomas, Springfield, Illinois.

Sunderland, S. (1958). The tentorial notch and complicationsproduced by herniations of the brain through that aperture.Brit. J. Surg., 45, 422-438.

Tandon, P. N., and Kristiansen, K. (1966). Clinico-patho-logical observations on brain stem dysfunction in cranio-cerebral injuries. In Proceedings of the 3rd InternationalCongress of Neurological Surgery, Copenhagen, 1965,edited by A. C. de Vet, W. F C Kennedy, and P. J.Vinken, pp. 126-129. (Excerpta med. int. Congr. Ser.,110). Excerpta Medica Foundation, Amsterdam.

Tomlinson, B. E. (1964). Pathology. In Acute Injuries of the Head,edited by G. F. Rowbotham, 4th ed., pp. 93-158. Living-stone, Edinburgh and London.

Weinstein, J. D., Langfitt, T. W., Bruno, L., Zaren, H. Q., andJackson, J. L. F. (1968). Experimental study of patternsof brain distortion and ischemia produced by an intra-cranial mass. J. Nearosarg., 28, 513-521.

Windle, W. F. (1948). Damage to myelin sheaths of the brainafter concussion. (Abstr.) Anat. Rec., 100, 725.

Windle, W. F., and Groat, R. A. (1945). Disappearance of nervecells after concussion. Anat. Rec., 93, 201-209.

Windle, W. F., Groat, R. A., and Fox, C. A. (1944). Experimentalstructural alterations in the brain during and after con-cussion. Surg. Gynec. Obstet., 79, 561-572.

Windle, W. F., Rambach, W. A., Jr., Ramirez de Arellano, M. I. R.de Groat, R. A., and Becker, R. A. (1946). Water contentof the brain after concussion and its non-contributoryrelation to the histopathology ofconcussion. J. Neurosurg.,3, 157-164.

Wolman, L. (1953). Ischaemic lesions in the brain stem associatedwith raised supratentorial pressure. Brain, 76, 364-377.

Zoltan, L. (1966). Clinico-pathological changes in cases oftrau-matic brain stem lesions with long survival. In Proceedingsof the 3rd International Congress of Neurological Surgery,Copenhagen, 1965, edited by A. C. de Vet, W. F. C.Kennedy, and P. J. Vinken, pp. 151-156. (Excerpta med. int.Congr. Ser., 110). Excerpta Medica Foundation, Amster-dam.

Zuilch, K. J. (1966). Anatomical-functional correlations in thebrain: Pathological morphological and clinical considera-tions. In Proceedings of the 3rd International Congress ofNeurological Surgery, Copenhagen, 1965, edited by A. C.de Vet, W. F. C. Kennedy, and P. J. Vinken,pp. 466-480.(Excerpta med. int. Congr. Ser., 110). Excerpta MedicaFoundation, Amsterdam.

brain-stem lesions after head injury

Documents