Glial swelling following human cerebral contusion: an ...JournalofNeurology,Neurosurgery,andPsychiatry 1991;54:427-434 Glial swellingfollowinghumancerebral contusion: anultrastructural

Journal ofNeurology, Neurosurgery, and Psychiatry 1991;54:427-434

Glial swelling following human cerebralcontusion: an ultrastructural study

R Bullock, W L Maxwell, D I Graham, GM Teasdale, J H Adams

AbstractThe ultrastructural features of cerebralcontusion seen three hours to 11 daysafter head injury were studied in 18patients undergoing surgery. Massiveastrocytic swelling ("cytotoxic" oedema)was seen three hours to three days afterinjury, maximal in perivascular footprocesses, and compressing some of theunderlying capillaries. The tight junc-tions were not disrupted. Neuronaldamage was most marked three to 11days after ijury. The patho-physiological mechanisms leading tooedema formation and neuronal degen-eration are discussed.

Cerebral cortical contusions are generallyregarded as the pathological hallmark of headinjury and are clinically important for severalreasons.`5 Extensive swelling of the corticalribbon and underlying white matter mayoccur acutely after cerebral contusions andmay cause mass effect, brain shift and raisedintracranial pressure, and these events maycause secondary ischaemic brain damage ordeath.6 When widespread, contusions causeextensive destruction of the cortical ribbonparticularly at the tips of the temporal andfrontal lobes and the orbital surface ofthe frontal lobes, and this may lead topost traumatic epilepsy, and to the neuro-psychological and behavioural sequelae,

Institute oftypical of severe head injury.7

Neurological Sciences, Although previous studies have focusedGlasgow upon the pathogenesis,45 histopathology,1 dis-R Bullock tribution' 3 and quantification of cerebral con-

D IGraham tusion, many aspects remain unresolved.3GM Teasdale Why do some cerebral contusions swell andJ H Adams generate raised intracranial pressure, whileCorrespondence to: others do not? What proportion of theMr R Bullock, Department neuronal necrosis which is seen adjacent toof Neurosurgery, Institute of

Neurological Sciences, cerebral contusions is due to primary contactSouthemn GReer Hospitw, damage, and how much of it is due to secon-Govan Road, Glasgow G514TF, UK dary factors, such as delayed reduction inReceived 31 January 1990 blood flow to the contused area?and in final revised form In studies in which we have mapped30 July 1990.Accepted 18 September 1990 cerebral blood flow following human cerebral

Table IA Control patients-clinical data

AgeCase (years) Tissue biopsied Indication for surgery

1 65 Cortex and white matter right Deep temporal glioma, seizurestemporal tip

2 56 Left frontal cortex Partial left frontal lobectomy forglioma

3 36 Posterior right temporal cortex and Right temporal lobectomy for gliomawhite matter

4 22 Right temporal cortex Right temporal lobectomy for severeepilepsy

contusion, zones of profoundly reducedregional blood flow (rCBF) have been foundadjacent to contusions, and this may be afactor in the ischaemic neuronal necrosiswhich is a prominent finding in pathologicalstudies. Alterations to the microvascular bedadjacent to cerebral contusion may thereforebe an important determinant of such changes.8The purpose of this study was to determine

ultrastructurally the changes in neuronal, glialand microvascular morphology which occurwith time following human cerebral con-tusion, and to determine their role in thepathogenic processes which are known todamage the brain after severe head injury.

Patients and methodsEighteen patients with severe cerebral con-tusion were studied. Clinical details are shownin table lB. In each patient, contusions wereshown on CT scan, and in seven patients,mass effect, acute clinical deterioration andassociated haematomas necessitated urgentoperation, at which the contusion was resec-ted, and biopsy taken from the margins of theresected tissue. In a further four patients, ICPmonitoring was initially performed to ascer-tain the need to remove the contusion, andICP rose necessitating surgery. In the remain-ing seven patients, delayed neurologicaldeterioration occurred up to 11 days afterinjury, precipitating the operation. Specimensof resected cerebral contusion from thesepatients were compared with specimensprocessed in an identical manner, from fourcontrol patients, with macroscopically andradiologically normal brain tissue at the site ofbiopsy, who had either temporal lobectomyfor epilepsy or resection of deep temporalglioma. Clinical details of these patients areshown in table 1A. Rongeur specimensaround 2 cubic mm in size were taken fromcontused grey and white matter andimmediately immersed in buffered glutaralde-hyde (350 m osm). Specimens were thenstored at 4°C for two to seven days untilprocessing for electronmicroscopy. The sitefrom which the biopsy was taken was recordedby the surgeon and related as closely as possi-ble to the CT appearance on the pre-operativescan. The study was approved by the EthicsCommittee of the Institute of NeurologicalSciences, Glasgow.

Processingfor electronmicroscopyAfter immersion fixation, specimens were cutinto 200 p thick sections using a vibrotome,and ultrasonicated to dislodge non-adherentblood components from the vessel lumina.

427

on June 14, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.54.5.427 on 1 May 1991. D

ownloaded from

http://jnnp.bmj.com/

Bullock, Maxwell, Graham, Teasdale, Adams

Table lB Patients with cerebral contusion-clinical data

Case

2

Age(years)

67

33

Interval betweeninjury and biopsy

3 hours

6 hours

3 68 13 hours

4 18 16 hours

5 52 17 hours

6 18 23 hours

7 29 27 hours

8 17 28 hours

9 59 31 hours

10 55 33 hours

11 68 37hours

12 42 42 hours

13 46 44 hours

14 59 49 hours

15 40 72 hours

16 64 77 hours

17 44 5 days

18 72 11 days

CT diagnosis and clinical details

Fall. Right temporal contusion, small acute subduralhaematoma, hemispheric swelling, high ICP (died)Fall. Right frontal lobectomy for massive brain swellingwhen acute subdural haematoma removed. (Alcoholic withclotting defect-died)RTA. Right temporal contusion with deterioration ofconsciousness. Good recoveryRTA. Large traumatic, right frontal intracerebralhaematoma and overlying contusion, deterioration ofconsciousness. Good recoveryFound unconscious and drunk. Left temporal contusion("Burst lobe") and acute subdural haematoma. DisabledRTA. Bifrontal contusions. Left frontal intracerebralhaematoma, deteriorating consciousness. DisabledAssault. Focal right parietal contusion and depressedfracture. Good recoveryRTA. Left hemispheric swelling and temporalcontusions-ICP monitoring -+ high ICP. DisabledRTA. Acute right extradural, subdural haematomas andright frontal contusion. Deteriorating consciousness.DisabledFell while drunk. Delayed deterioration. Coma. Bifrontalcontusions. Thin left acute subdural haematoma. GoodrecoveryRTA. Right temporal contusion, small acute subdural.ICP monitoring -+ high ICP. Good recoveryFound unconscious. Delayed deterioration. Bitemporalcontusions. This left acute subdural. Good recoveryFall. Delayed confusion. Right temporal contusion. ICPmonitoring -. high ICP. Good recoveryAlcoholic epileptic. ? Fall. Delayed confusion. Bifrontalcontusions, mass effect. DisabledFall. Right parietal acute subdural. Delayed ICP risenew low density. Right temporal -* re-opening craniotomyand right temporal lobectomy. DisabledFall. Multiple contusions. Left hemisphere, smallextradural haematoma. Right parietal. Delayeddeterioration. Partial left temporal lobectomy. DisabledHydrocephalic with shunt. Fell out of bed -. bifrontalcontusions. Delayed coma. ICP high. Right frontalcontusion evacuated. DisabledFall. Persistently confused. Right frontal subacutesubdural, small temporal contusion. (Burst lobe atsurgery). Disabled

Specimens were then fixed in 2% osmiumtetroxide in phosphate buffer, and the sliceswere critical-point dried, mounted on scan-ning electronmicroscopy stubs with silverpaint and sputter-coated before viewing in aJEOL 300 TS scanning electronmicroscope.Blood vessels were selected for particularstudy and micrographs were taken at a mag-nification of 7000 and 1500 times to allowdemonstration of endothelial surface structureand measure vessel size. After scanning elec-tronmicroscopy specimens were carefullyremoved from the SEM stubs and washedthree times in 100% acetone to remove thesilver paint. The material was then embeddedin araldite resin and cut into semi-thin sec-tions stained with toluidine blue and areas of

Table 2 The time intervalsfor various EMfeatures after cerebral contusion

Time interval at which EMoccurred after injury EMfeature

3 hours-3 days Astrocyte swelling3 hours-3 days Astrocyte podocyte swelling3 hours-3 days Astrocytc podocyte disruption3 hours-7 days Vessel lumen compression3 hours-5 days "Swiss cheese vacuolation" of neuropil on TM3, 16 hours Red cell diapedesis6 hours-3 days Endothelial fenestration on SEM16 hours-3 days Intravascular clots adherent to walls16 hours-3 days Gross tissue disruption + haemorrhage16 hours-5 days White matter "loosening"16 hours-l days Peri-vascular red cells3-11 days Pyknosis of neurons5-7 days Sheet reaction-glycogen granules

-pseudopodial processes from astrocytes5, 11 days Leucocytic phageocytosis

11 days Collagen deposition

interest selected for transmission electron-microscopy. Thin sections were cut for thetransmission EM and routinely stained withlead citrate and uranyl acetate and examinedin a JEOL 100 S transmission electronmicros-cope.

ResultsBiopsies were obtained at time points rangingfrom three hours to 11 days after injury. In allthe specimens the quality of fixation as judgedby preservation of intracellular organelles wasacceptable.

Control patientsNormal glial, neuronal and vascular ultra-structure was present in all control patients,and the neuropil and white matter density wasnormal (fig 1). Scanning electronmicroscopydemonstrated pit-vesicle activity on theendothelial surface of the lumen of all vesselsin the control patients (table 2).

Transmission EM findings in patients withcerebral contusionA) GROSS DISRUPTION OF TISSUEIn five specimens taken from three patients thearchitecture of the cortical neuropil was dis-rupted by red blood cells and plasma. When-ever this was seen, the adjacent neuropil wasgrossly abnormal, demonstrating markedvacuolation and lucency of astrocyte cyto-plasm (fig 2).

B) GLIAL CELL SWELLINGThe most consistent finding in both grey andwhite matter in patients with cerebral con-tusions was severe swelling affecting par-ticularly astrocytes. This was associated with"loosening" and apparent pseudo-vacuolationof the neuropil to produce a "Swiss cheese"appearance, particularly striking on SEM.This appearance developed rapidly (figs 3, 4a)within three hours of injury but was stillpresent in specimens taken three days after

V,. .

' ...:

Figure I Thin section of cortical neuropilfrom controlpatient KW. Preservation of intracytoplasmic organellesis satisfactory and vacuolation of neuropil is absent.(Magnification: x 2975).

428


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Glial swelling following human cerebral contusion: an ultrastructural study

Figure 2 Thin sectionthrough cortical neuropil16 hours after injury. Notegross swelling of astrocyticcytoplasm (lower right)and vacuolation ofneuropil caused by swellingastrocytic process. Notemitochondria apparentlylyingfree within thecytoplasm of swollenastrocytes (arrow).(Magnification: x 5250).

Figure 3 Scanning EMof a section throughcortical neuropil to showvacuolation surrounding ablood vessel (arrow)("Swiss cheese"appearance).(Magnification: x 1000).

injury. In the worst affected astrocytes, cyto-plasmic disruption was present with completerupture of the cell membrane (fig 5). Organe-lles such as mitochondria were, however, stillrecognisable. In the specimens taken five daysafter injury, astrocytic swelling was lessmarked although still apparent (fig 6a, b) andthe astrocytes now contained deposits ofglycogen, typical of the "sheet reaction" des-cribed by Torvick (fig 6).9

Figure 5 Thin section through white matter capillaryshowing gross astrocytic foot process swelling (a), withdisruption of the foot process limiting membranes andnarrowing and compression of vessel lumen (L).(Magnification: x 3750).

C) SWELLING OF PERIVASCULAR PODOCYTICPROCESSESSwelling of astrocytic perivascular podocyticprocesses were the most obvious manifesta-tion of glial swelling. This was sufficientlysevere to suggest compression of the adjacentvessel lumen in some cases (fig 4a, b). Disrup-tion of foot processes was seen in somespecimens taken at periods from three hoursto three days after injury (fig 5). In specimenswhich were taken at later periods after injury,astrocytic swelling appeared to be less exten-sive and at five days astrocytic processes con-tained glycogen deposits and pseudopodialfolds typical of the "sheet reaction" ofTorvick (fig 6a, b).9 Comparison between thechanges seen in cortex (fig 4a, b) and those inwhite matter (figs 4, 5, 6) should, however, bemade with caution.

D) VASCULAR CHANGES (TRANSMISSION ELEC-TRONMICROSCOPY)In the majority of vessels seen within con-

VA .41 *a. .Figure 4 (A) Thin section through a cortical capillary, showingflattening of vessel luwnen (L) and massive swelling ofperivascular astrocytic foot processes (a). The lumen is narrowed to such an extent that passage of red cells would beimpossible. (Magnification: x 10 500); (B) ScanningEM preparation of a larger blood vessel in cortical neuropilshowing collapse of the vessel lumen andfolding of the tunica intima with adherent clot (arrow).(Magnification: x 460).

4z29

lp*P.

I

,i..., ----. ,A -,..4i I"

n -t-S

AW

.,A ..

I1*

a


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



Figure 6 (A) Thinsection of white mattertaken five days aftercontusion-the swelling ofastrocytic foot processes(a) is less marked withglycogen accumulation.The sheet reaction isapparent.(Magnification: x 8200);(B) Thin section ofcortical neuropil in adifferent patient, operatedupon five days after headinjury. Swollen astrocyteprocesses and accumulationofglycogen granulestherein are seen. ( The so-called "sheet reaction").(Magnification: x 8750).

tused tissue, endothelial cell morphology wasgrossly normal and in those vessels whereadjacent astrocytic swelling was not noticed,vessel lumina were not disrupted or occluded.No disruption of inter-endothelial tight junc-tions was noted, and in the majority of vesselsstudied, the morphology of the perivascularbasement membrane was also normal.An increased number of cytoplasmic folds

and processes on the luminal aspects of endo-thelium were seen within vessels and thintransmission sections showed a very irregularluminal profile to the endothelium at two days(fig 7). Inter-endothelial junctions were notdisrupted despite gross disruption of theadjacent neuropil. This irregular endothelialprofile was also seen at five days when theastrocytic sheet reaction had become wellestablished (fig 8).9Counts of endothelial pinocytotic vesicles in

the vessels from contusion specimens taken atdifferent times after injury, showed a declinewith time and the highest numbers of vesicleswere seen in the control and acute patients(table 3).

Early white matter changes on transmission EMTwelve specimens of white matter from con-tused tissue were studied. Preservation ofmyelin was remarkably good. In some patients,swelling of axoplasm was noted, but this wasnever severe. Cytoplasmic swelling of glial cellswithin white matter was present but was lessmarked than that seen in the grey matter.Accumulation of interstitial fluid within whitematter was marked (fig 5).

[%*wo ; % s- * , ;

;

F^fs i&4>2K

Figure 7 High magnification thin section ofpart of avessel wall two days after injury to show a thickenedirregular profile of the endothelium. Microvilli arepresent. Note that the interendothelial tightjunctions areintact (arrow). Astrocytic foot processes on the abluminalsurface of the basal lamina are grossly swollen.(Magnification: x 20 000).

Delayed changesA PYKNOSIS AND SHRINKAGEPyknosis and shrinkage of both neurons andglia was noted when biopsies were taken atthree or more days after injury (fig 9). Pyknoticneurons were frequently surrounded bymicroglia which appeared to be carrying outphagocytosis of damaged neuropil and cells. Inspecimens from the two patients who werestudied five and 11 days after injury, extensiveleucocyte and microglial phagocytosis was seenand deposition ofcollagen fibres was prominent

Table 3 Microvillus counts (SEM) and pinocytotic vesicle activity (TEM) after contusion

Endothelial pinocytotic Luminal microvillus counts per 25 p2Time interval vesicle count mean (SD)after per p2injury mean (SD) Capillaries (1Op)

Controls - 6 105 (3-51) 9-66 (3 77) 12 916 (3 81)Contusion 3-6 hours 6 125 (2-99) 2-92 (1 59) 22 19 (5-6)

13-23 hours 0-75 (0 45) 5 0 (2 5)24-48 hours 4-72 (3-01) 5-343 (2 44) 4 914 (1-35)2-5 days 1 5 (0 91) 7 166 (2 23) 8 45 (3 45)

11 days 7 91 (2 33) 12 31 (3 63)

430 on June 14, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.5.427 on 1 M

ay 1991. Dow

nloaded from


Glial swellingfollowing human cerebral contusion: an ultrastructural study

Figure 8 Highmagnification thin sectiontaken five days aftercerebral contusion. Notenumerous endothelialmicrovilli andaccumulation ofglycogengranules in astrocyticperivascular foot processes(a). (Magnification:x 26 000).

Figure 9 Thin section ofa pyknotic cortical cell,probably a neuroncontaining lipofuscingranules, taken 72 hoursafter head injury. Notemarked shrinkage andperi-cellular vacuolation.(Magnification:x 11 000).

'~~1

.. .i

Figure 10 Thin sectionfrom cortical neuropil inpatient RS 11 days afterhead injury. Noteleucocytic phagocytosis(M) collagen fibredeposition (outline arrow)andfibroblast (F).(Magnification: x 9250).,

(fig 10), indicating the early stages of scar tissueformation. In the patient studied 11 days afterinjury, perivascular foot process swelling wasabsent and the numbers of perivascularmicroglia were increased, suggesting that thestage ofmaximal cell swelling had been passed.

Findings on scanning electronmicroscopy inpatients with cerebral contusionA NEUROPILThe vacuolated "Swiss cheese" appearance of

the neuropil in tissue adjacent to cerebralcontusions was particularly noticeable on scan-n.g EM in comparison with the controlpatients (fig 3).

B VASCULAR AND PERIVASCULAR CHANGESVessel morphology was remarkably normal inall specimens studied. Tears in vessel wallswere not seen. Blood clots occluding the lumenofvessels were seen only occasionally, althoughthis may relate to the vibration and ultrasonica-tion process which was used to visualise thelumina. In many vessels, however, adherentred cells and leucocytes were found suggestingan abnormality of endothelial adhesion charac-teristics. Red cells were seen moving throughfenestrations in the endothelial membrane toleave the vessel lumen, by a process ofdiapedesis (fig 1 la, b).

Perivascular accumulations ofred blood cellswere a frequent finding and in severalspecimens, these caused streak-like disruptionof tissue architecture.

Pit-vesicle activityPit-vesicle activity was present in both controland contusion specimens, but was less exten-sive in the patients with contusions. In someareas, the endothelial luminal membrane wasdisrupted by coalescent vesicles to formirregular holes in the vessel walls, and in twospecimens, red cells were seen passing throughthese (fig 1 la, b). Pit-vesicle activity in theendothelium was always accompanied byparallel development of microvillus formation,but the proportion of pit-vesicle andmicrovillus activity was variable fromspecimen to specimen, and no clear relation-ship was seen, either to time after injury, or toother microscopic changes. When the numberof luminal microvilli per 25 gim2 were countedusing SEM, and correlated with the number ofvacuoles within the endothelial cytoplasm seenon transmission electronmicroscopy, (table 3)no relationship was seen in either the contusionor control material.The mean number of endothelial micropin-

ocytotic vesicles in control patients was6-105 jim2 and 6-125 gm2 in contusion patientsat three hours after injury, 4-72 jim2 at 24-48hours after injury, and 1-5 pm2 at five days afterinjury. Thus there was a slight (but not statis-tically significant) decline over time in thenumber of endothelial pinocytotic vesicles indamaged tissue.Marked changes in luminal microvillus

activity were seen in the contusion material, atvarious times after injury (table 3). Incapillaries, microvilli numbers dropped to theirlowest levels when astrocytic and foot processswelling was most marked, at 13-24 hours andslowly rose to near control values by 11 days.By contrast, in larger vessels microvillusactivity rapidly rose to almost twice controllevels in the first six hours after injury, thendropped to nearly a third ofcontrol levels at 24-48 hours, and were approaching control levelsby 11 days.

431

CA

.. I11

.Alk

.14..o


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



Figure 11 (A) ScanningEMpreparation of theluminal surface of acortical arteriole taken 17hours after head injury.Note the large fenestr:ae inthe endothelial surfacethrough which erythrocytesare seen to pass (arrow).(Magnification: x 1780),(B) High magnificationview of the endothelialsurface 17 hours after headinjury. Note theendothelial microvilli anderythrocyte passingthrough the fenestration.(Magnification: x 9700).

ii. . . ' ....................... ': S,'l B l l F '' ........................................................................... !.: :.t Srsh t!M ; | l l 4 vSX w tleKi;Mes t ..s.X.'B','''.il5.E.s.iy l >lE .. v ;-. w* * r v . S: uS ' ....... Y ] X l | S X .. . ..... . .>m A

...... < , . t: - :oR} ,i . g ,S - eSr ; .; ..

.. .8 ......... SX mt' .......................... ze e / * . _.'. 't S e S, ,g,.,>,> SS ' -; S '' A A. .. , ,, .. W .. ,,, .. ,,

Glial swellingfollowing human cerebral contusion: an ultrastructural study

studied within the first few days, these changesshould be interpreted with caution. Serialbiopsies were not taken from the same patientand the severity of the contusion was variablefrom patient to patient.For these reasons the mechanisms by which

perivascular astrocytic foot process swellingmay resolve after a contusion cannot beinferred from this study. Although specimenstaken at five days or more after the injury showevidence of astrocytic process sprouting, andapparent reconstitution of the foot processlimiting membranes, (fig 6a) this may representthe response to a less severe lesion. To improveour understanding of these fundamental repairmechanisms, serial biopsy studies will beneeded, in an animal model, with a controlledseverity of injury.

Early perifocal oedema has been demon-strated in microgravimetric studies afterhuman contusion.8 19 The prominence ofastrocytic swelling early after cerebral con-tusion, and the integrity of the inter-endoth-elial tight junctions seen in this study and inanimal models,1011 2021 suggests that this is acytotoxic process. The relative paucity ofprotein-rich extracellular fluid (except wheregross tissue disruption was present), (fig 5)within the neuropil in this study is alsoevidence against a vasogenic mechanism.This agrees with the results of studies of

blood-brain barrier permeability which wehave performed at various time intervals afterfocal cerebral contusion in 20 patients. SPECTscanning after intravenous 9'Tc pertech-netate, showed that blood-brain barrier chan-ges were seldom detectable before three daysafter injury, and were maximal at about 10days.8The micro-anatomical basis for the delayed

development of increased vascular per-meability after contusion is difficult to deter-mine from this study, or from the animalstudies of other authors. Povlishock, using afluid percussion model of head injury, demon-strated that Horseradish peroxidase (HRP)passed rapidly through intact endothelial cellsafter intravascular injection, and accumulatedbeneath the basement membrane, and withinendothelial vesicles.2" 22 HRP was found chieflywithin endothelial cytoplasmic vesicles and thenumber of vesicles increased rapidly within 15to 30 minutes after trauma.2' In contrast tothese experimental observations, we havedemonstrated that after a contusion micro-pinocytotic vesicles decline in number withtime (table 3). Therefore, it is unlikely thatmicropinocytosis is an important mechanism indelayed post-contusional oedema.The role of pit-vesicle activity and

microvillus formation after head injury is notunderstood. Its presence within hours of injuryafter "pure" diffuse axonal injury suggests thatit may be a non-specific response to the invaria-ble hypertensive surge which occurs withinseconds after severe head injury.2 23-25We have observed coalescence of vesicles to

form ragged holes in the endothelial luminalmembrane, through which intact erythrocytesare seen to pass (fig 11). This partly accords

with other studies,22 and may provide amechanism for the peri-vascular accumulationoferythrocytes, which is a frequent histologicalfeature after contusion but which has not beenpreviously explained.'45 Vessel tears maytherefore not be necessary for perivascular"haemorrhage".Although we and other authors have shown

that astrocytic swelling is well established aftercerebral contusion as early as three hoursafter injury, both in this study and in animalmodels, the cause for this swelling remainsuncertain.""20 22When cerebral blood flow falls sufficiently

low to cause failure of neuronal energymetabolism, and membrane depolarisation(around 14-15 ml/100 gm/min) due to globalor focal ischaemia, massive astrocytic swellingoccurs as a buffering mechanism; astrocytesabsorb free K+ ions in an attempt to maintainextracellular fluid homeostasis.26 27 Globalischaemia is, however, unlikely to occur as avery early event after focal human contusionand, moreover, some animal studies haveshown blood pressure and CBF to be transien-tly elevated after head injury.2325Faden et al and Becker et al have demon-

strated massive release of excitatory aminoacids after fluid percussion injury, in tissuesimmediately beneath the impact point, but notdistant from it.2829 Excitatory neuronal mem-brane depolarisation may result from thisglutamate release, to cause subsequent K+ionflooding ofthe extracellular space, and this maycause astrocytic swelling, which itself mayjeopardise focal tissue perfusion adjacent to acerebral contusion, and induce furtherischaemic damage.26 A pathological "viciouscycle" such as that shown in fig 12, may thus bepostulated, as a cause for neuronal death incerebral contusion.

Vasoconstrictor substances such as K+,serotonin, or prostaglandins may also bereleased into the contused tissue, inducing theprofound fall in regional blood flow, which wehave demonstrated on SPECT (8).

Further studies are needed to clarify thecauses of astrocytic swelling, and to explorethe possibilities of therapy for preventing orattenuating this astrocytic response, which maybe detrimental to neuronal survival.

We are grateful to the neurosurgeons at the Institute ofNeurological Sciences, Glasgow who assisted with obtainingcerebral contusion and control specimens, and to Anne Semplefor preparing the manuscript.

1 Adams JH, Gennarelli TA, Graham DI. Brain damage innon-missile head injury: observations in man and sub-human primates. In: Thomas Smith W, Cavanagh JB, edsRecent advances in neuropathology. Edinburgh: ChurchillLivingstone, 1982:165-90.

2 Adams JH, Gennarelli TA, Graham DI, et al. Changes inendothelial morphology of brain vasculature after con-trolled angular acceleration of the subhuman primatehead. Proc Physiological Soc 1987:92.

3 Adams JH, Scott G, Parker L, et al. The Contusion Index: aquantitative approach to cerebral contusions in headinjury. Neuropathol Appl Neurobiol 1980;6:319-24.

4 Lindenberg R, Freytag E. Morphology of cortical con-tusions. Arch Pathol 1957;63:23-42.

5 Lindenberg R, Freytag E. The mechanism of cerebralcontusion. Arch Pathol 1960;69:440-69.

6 Miller JD, Gudeman SK, Kishore PS, et al. Computedtomography in brain oedema. Advances in neurology Vol28. New York: Raven Press; 1980:413-22.

7 Levin HS, Grossman RG, Rose JE, et al. Longtermneuropsychological outcome in closed head injury. J

433


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



Neurosurg 1979;50:412-22.8 Bullock R, Teasdale GM, Wyper D, et al. Tomographic

mapping of CBF, CBV and BBB changes after focal headinjury using SPECT: mechanisms for late deterioration.In: J Hoff, ed. Proc VIIth Int Symposium IntracranialPressure and Brain Injury. Berlin: Springer-Verlag1989:637-9.

9 Torvick R, Skjorten F. Electron microscopic observationson nerve cell regeneration and degeneration, after axonlesions. I. Changes in the nerve cell cytoplasm. ActaNeuropathol 1971;17:248-64.

10 Allen IV, Kirk J, Maynard RL, et al. An ultrastructuralstudy of experimental high velocity penetrating headinjury. Acta Neuropathologica (Berl) 1983;59:277-82.

11 Hekmatpanah J, Hekmatpanah CR. Microvascular altera-tions following cerebral contusion in rats. J Neurosurg1985;62:888-97.

12 Jenkins A, Maxwell WL, Graham DI. Experimentalintracerebral haematoma in the rat: sequential lightmicroscopical changes. Neuropathol Appl Neurobiol1989;15:477-86.

13 Little JR, Kerr FWL, Sundt TM. Microcirculatory obs-truction and focal cerebral ischaemia: an electron micros-copic investigation in monkeys. Stroke 1976;7(1):25-30.

14 Adams JH, Doyle D, Ford I, et al. Diffuse axonal injury inhead injury. Definition, diagnosis and grading. Histopath-ology 1989;15:49-59.

15 Tenjin H, Mizukawa N, Yamaki T, et al. The investigationof cerebral haemodynamics in patients with cerebralcontusion: Evaluation of Positron Emission Tomography(PET). J CBF Metabol 1989;9(Suppl 1):S87.

16 Dietrich WD, Busto R, Ginsberg MD. Cerebral endothelialmicrovilli and fenestration following global fore brainischaemia. J Neuropathol Exp Neurol 1984;43:72-83.

17 Mendelow AD, Bullock MRR, Teasdale GM, et al.Intracranial haemorrhage induced at arterial pressure inthe rat. Pt II. Short term changes in local cerebral bloodflow measured by autoradiography. Neurol Res1984;6:189-93.

18 Bullock MRR, Smith R, Favier J, et al. Brain specific gravityand CT density measurements after human head injury. J

Neurosurg 1985;63:64-8.19 Galbraith S, Cardoso E, Patterson J, et al. The water content

of white matter after head injury in man. In: Go KG,Baethmann A. Recent progress in the study and therapy ofbrain oedema. New York: Plenum Press, 1984:323-9.

20 Olsson Y, Rinder L. Studies on vascular permeabilitychanges in experimental brain concussion. ActaNeuropathologica (Berl) 1968;11:183-200.

21 Povlishock JT, Becker DP, Sullivan HG, et al. Vascularpermeability alterations to horse radish peroxidase inexperimental brain injury. Brain Research 1978;153:223-39.

22 Povlishock JT. The morphopathologic responses toexperimental head injuries of varying severity. In: BeckerDP, Povlishock JT, eds. CNS Trauma Status Report.Bethesda, Maryland: NIH, 1986:443-52.

23 Gennarelli TA, Segawa H, Wald U, et al. Physiologicalresponse to angular accelaration of the head. In: Gross-mann RG, Gildenberg PL, eds. Head Injury: basic andclinical aspects. New York: Raven Press 1982:129-39.

24 Maxwell WL, Irvine A, Adams JH, et al. Response of thecerebral microvasculature to brain inmury. J Pathology1988;155:327-35.

25 Nilsson B, Nordstrom CH. Experimental head injury in therat. Pt 3. Cerebral blood flow and oxygen consumptionafter concussive impact acceleration. J Neurosurg1977;47:262-73.

26 Bourke RS, Nelson KM. Further studies in the K +dependent swelling ofprimate cerebral cortex in vivo: theenzymatic basis of the K+ dependent transport ofchloride. J Neurochem 1972;19:663-85.

27 Plum F, Cooper AJL, Craig RP, et al. Glial cells: the silentpartners of the working brain. The Thomas E DuffyMemorial Lecture. J CBF and Metabol 1985;5(Suppl1):S1.

28 Becker DP, Katayama Y, Tamura T, et al. Excitotoxic ionicfluxes and neuronal dysfunction following traumatic braininjury. J CBF Metabol 1989;9(Suppl 1):302.

29 Faden AI, Demediuk P, Panter SS, et al. The role ofexcitatory amino acids andNMDA receptors in traumaticbrain injury. Science 1989;244:798-800.

434


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from

Documents

Glial swelling following human cerebral contusion: an ...JournalofNeurology,Neurosurgery,andPsychiatry 1991;54:427-434 Glial swellingfollowinghumancerebral contusion: anultrastructural