damageafter febrile convulsions - bmj

BRAIN DAMAGE AFTER FEBRILE CONVULSIONSBY

MALCOLM FOWLERFrom the Adelaide Children's Hospital, South Australia

(RECEIVED FOR PUBLICATION NOVEMBER 5, 1956)

Despite the suggestions of several writers on theimportance of convulsive seizures in infancy as anaetiological factor in the development of cerebralpalsies, little attention is given to this association inthe standard texts in general medicine and pathology,or even in the more special treatises on diseases ofthe nervous system. It is not altogether surprising,therefore, that convulsions occurring in the course offebrile illnesses are often not treated with thegravity and urgency they deserve. Because con-vulsions are so frequent an accompaniment of rela-tively innocuous infections and in most cases leave thepatient apparently normal, the paediatrician mayregard both a convulsion of purely febrile originand the resulting signs of nervous disease as commonmanifestations of some other intracranial process.The following five cases are presented in the hope

that the potentially lethal nature of convulsions willbe more widely acknowledged.

Materials and MethodsThe pathological material was obtained from necropsies

performed at the Adelaide Children's Hospital during theyears 1954 to 1956 inclusive. Before immersion of theintact brain pieces were removed for fixation in formol-ammonium bromide, and in two cases samples were keptfor animal inoculation. After prolonged fixation com-plete coronal slices of the cerebral hemispheres andsmaller pieces of brain-stem and cerebellum were em-bedded in celloidin and paraffin, and stained by Weil'smethod for myelin sheaths, with Nissl's stain, and byhaematoxylin, counterstaining with both eosin and vanGieson's mixture. Frozen sections were used for themetallic impregnation of neurofibrils and astrocytes usingBielschowsky's and Cajal's methods, respectively.

Case ReportsCase 1. The patient, previously a healthy boy of

2 years, developed measles two days before admission.He had not been very sick but had been coughing a lot.On the day of admission he was found in his cot at7.45 a.m. in a convulsive attack, blue and frothing at themouth. This continued until 9.30 a.m. when the con-vulsion was controlled by ether anaesthesia in the

Casualty Department of the Adelaide Children's Hos-pital where his colour was described by the AdmittingOfficer as 'almost black'.On examination in the ward his temperature was

103° F., he was slightly cyanosed, with a cold skin onwhich there was a fading measles rash. The pupils wereequal and reacted sluggishly to light. The fundi ocu-lorum were normal. The fauces were moderatelyreddened, the right tympanic membrane was dull andslightly bulging. Air entry into the lungs was good butthere were many rhonchi. The abdomen was normal.Reflexes were absent in the limbs. Slight neck stiffnesswas present for an hour after the convulsions had ceased.

SPECIAL INvEsTIGAIoNs. The cerebrospinal fluid wasunder a pressure of 100 mm. of water, contained 28 mg.protein per 100 ml. and no cells. The white blood countwas 24,000 per c.mm., of which 81% were polymorpho-nuclear leucocytes.

COURSE. In the evening of admission he went pale,with a blood pressure of 70 mm. systolic and 40 mm.diastolic. This collapse was treated by intravenousserum infusion. The next morning he had a furtherconvulsive attack which lasted for 10 minutes. He wascomatose from the time of his first convulsion, respondingby groaning to painful stimuli, and died three days afteradmission.POST-MORTEM FINDiNGS. The necropsy (No. 41/56)

was performed 22 hours after death. The body was well-nourished and on the skin of the back and face were theremains of a macular rash. In the lungs were numeroushaemorrhagic areas of consolidation and these on micro-scopic examination were confirmed as foci of confluenthaemorrhagic bronchopneumonia. Staphylococcus aureuswas cultured from the lungs. There was severe centri-lobular necrosis in the liver, involving about one half thethickness of the lobule. The liver cords were disorgan-ized, and the cells were intensely eosinophilic andshrunken, with pyknotic nuclei. Many cells were dis-rupted into granular fragments. There was no inflam-matory reaction in the portal areas and there was novascular congestion.

In the spleen there were large reactive centres andreduced numbers of lymphocytes in the Malpighianbodies.

Central Nervous System. No abnormality apart frommuch congestion was noted in the brain on naked-eyeexamination. Microscopically, however, there was awidespread necrosis of neurones. The affected cells

67 2

copyright. on O

ctober 23, 2021 by guest. Protected by

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.32.162.67 on 1 April 1957. D

ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOOD

FIG. 1.

F...3|#. .: .. k....

FIG. 3.

FiG. 2.

were shrunken, their cytoplasm was brightly eosinophilic,all Nissl material having disappeared, and the nucleuswas small, irregular in shape and pyknotic (Fig. 1).Often the nucleolus was enlarged and eosinophilic. Theblood capillaries in these areas were moderately orgreatly dilated, but there was no haemorrhage or soften-ing and only occasionally were a few small round cellsseen in the perivascular space. In addition to neuronaldamage, the astrocytes in the affected grey matter werealso abnormal in that they failed to impregnate with goldalthough their nuclei did not look abnormal using

FIG. I.-Case 1: frontal cortex showing shrunken neurones withpyknotic nuclei, and pericellular vacuolation.

Haematoxylin and eosin x 700.

FIG. 2.-Case 1: cerebellum showing shrunken Purkinje cells withpyknotic nuclei.


FIG. 3.-Case 1: cerebellum showing a normal and a necroticPurkinje cell.


haematoxylin stains. Astrocytes in the white matterunderlying the cortex stained normally but those on thejunction were incompletely impregnated and oftenappeared to have fragmented processes. In most areasthere was much perivascular and pericellular vacuolation.

All parts of the cerebral cortex were involved butparticularly the temporal and frontal gyri. The occi-pital cortex, the hippocampus and dentate fascia weredamaged on both sides, but far more severely on the leftside. The thalamus and amygdaloid nucleus on bothsides showed an equally severe degree of neuronal

68

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

BRAIN DAMAGE FOLLOWING FEBRILE CONVULSIONSnecrosis, but the lesions in the caudate nucleus were moreextensive on the left. The putamen and globus pallidusshowed only pericellular oedema on the right but wereseverely damaged on the left. Small focal lesions werepresent in the left claustrum. In the cerebellum therewas selective damage to the Purkinje cells, other cells ofthe cortex and the central nuclei being unaffected. Onthe whole there was not much damage in the cerebellum,but two areas involving several adjacent folia werepresent, one in the vermis, the other in one lateral lobe.In these foci the majority of the Purkinje cells wereshrunken, with brightly eosinophilic cytoplasm and con-tracted densely staining nuclei (Fig. 2). Many hadcompletely disappeared and others were merely faintoutlines. Often a normal Purkinje cell would be foundamongst degenerate neighbours (Fig. 3). In other partsof the cerebellar folia the Purkinje cells showed patchychromatolysis.

Case 2. This child, a boy aged 1 year, was admittedwith a two days' history of anorexia, drowsiness andrunning nose. The night before admission he had had aconvulsion unrelieved by intramuscular phenobarbitonebut stopped by ether anaesthesia. The attack lastedabout half an hour. He had begun convulsing againin the morning at 6.15 o'clock and was still convulsingwhen first seen in the Casualty Department at 7 a.m.The child was very cyanosed. After treatment withphenobarbitone and ether the convulsions stopped atabout 7.15 a.m.On examination in the ward he was unconscious and

pale, with cyanosed lips. His temperature was 1040 F.There was some residual intermittent twitching of theright arm and leg. The fundi oculorum were normaland the ear drums slightly reddened. There was no neckstiffness, and the thorax and abdomen were normal.The knee and ankle jerks were absent and the plantarreflexes were flexor.

SPECIAL INVESTIGATIONS. A sample of cerebrospinalfluid, taken immediately after admission, was not underincreased pressure, and contained 14 white blood cells,and 1,060 red blood cells per c.mm., 60 mg. protein, and35 mg. sugar per 100 ml. Further lumbar punctures onthe fifth and eleventh days of his illness yielded fluid againcontaminated with blood. The blood contained 21,000white cells per c.mm. of which 80% were polymorpho-nuclear leucocytes. Radiological examination of thechest showed no sign of disease. Blood culture wasnegative and no abnormal organisms were found in thefaeces. The urine was normal.

COURSE. Two days after admission the features ofa right hemiplegia, previously indefinite, emerged clearly,the facial and peripheral musculature on this side beingparalysed, with exaggerated tendon reflexes and extensorplantar response. His temperature was consistently ator above 1030 F. A few days before death the comalightened enough for him to be able to respond by groan-ing to painful stimuli. He died on the twelfth day ofillness in a state of dehydration, deeper coma and sub-normal temperature.PosT-MORTEM FiNDINGs. The necropsy (No. 94/55)

was performed 18 hours after death. There was a

purulent right otitis media, and confluent broncho-pneumonia. Yellow streaks were found in the pectoralisminor muscles. There was necrosis of all papillae in theleft kidney, but of only one papilla in the right. All theselesions were confirmed microscopically, the musclechanges being those of necrosis and calcification, pro-bably secondary to rupture during the convulsion. Thepulmonary and renal changes were both very recent.

Central Nervous System. The parietal lobes, particu-larly on the left, were hyperaemic. On section the corticalgrey matter of the left cerebral hemisphere was pepperedwith fine red dots. The dural sinuses were normal.

Microscopically the basic lesion was degeneration ofneurones. In the most severely affected areas the nervecells had disappeared but in other places many were stillpresent and showed degenerative changes-shrinkage,loss of Nissl substance, acidophilia of the cytoplasm,pyknosis and enlargement of the nucleolus which wasweakly acidophilic (Fig. 4). Coincident with loss of thenerve cell body was a great scarcity of axonal processes,and many showed beading and fragmentation. In thecentral white matter of the cerebral hemispheres myelinsheaths were finely beaded. In the affected grey mattermicroglial proliferation was obvious, sometimes asamoeboid forms, but more conspicuous was the pleo-morphic rod-cell stage. Even in areas of cortex wheredegenerate neurones were still visible normal astrocytescould not be demonstrated by Cajal's gold impregnationmethod, except in the subpial layer where they wereincreased. Astrocytes were present in large numbers inthe immediately subcortical white matter, and here theywere impregnated clearly (Fig. 5). Abnormal formscould be seen in the junction between cortical grey andwhite matters and in the deeper parts of layer I in thecerebral cortex (Fig. 6). The capillary vessels in theaffected grey matter were extremely numerous and theirendothelial and adventitial nuclei were very prominent(Fig. 7). There was a very occasional slight infiltrationof lymphocytes around larger vessels. The meningeswere thickened and contained an excess of macrophages.The grey matter of the left cerebral hemisphere was

very severely affected. The lesions in the cerebralcortex were diffuse and of approximately equal severityin all the frontal and temporal gyri, the anterior andposterior central gyri, the gyrus cinguli, the insula, theuncus and the fusiform gyrus. The hippocampalgyrus had lost layer III of its cortex for a short segment.There was almost complete loss of neurones in the hippo-campus, particularly where its cortex enters the con-cavity of the dentate fascia. The granule layer of thedentate fascia showed much less outfall, exhibitingscattered small foci and segments of partial or completeloss. In the thalamus, putamen, caudate nucleus andamygdaloid nucleus only a few recognizable cells remain-ed. The globus pallidus showed lesions of lesserseverity although in one section all cells had been lostfrom its inner division. A small focal lesion was seen inthe anterior perforated substance, and there were moder-ate numbers of such foci in the claustrum.On the right side lesions were very much less pro-

nounced. The cortex was normal, apart from damage

69

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

70 ARCHIVES OF DISEASE IN CHILDHOOD

FIG. 4.-Case 2: frontal cortex showing shrunkenneurone with pyknotic nucleus, enlarged pale nucleolus

and faint cytoplasm.Haematoxylin and eosin x 790.

FIG. 5.-Case 2: numerous astrocytes in the white matterimmediately beneath the frontal cortex.

Cajal's gold chloride x 400.

FIG. 6.-Case 2: abnormal astrocytes on the border of grey and whitematters of the frontal cortex.Cajal's gold chloride x 400.

FIG. 7.-Case 2: temporal cortex showing capillaryproliferation.


copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

BRAIN DAMAGE FOLLOWING FEBRILE CONVULSIONS

FIG. 8.-Case 2: insular cortex showing loss of neurones in layer III.Haematoxylin and eosin x 30.

to the superior part of the superior temporal gyrus,

inferior portion of the insular cortex and to the hippo-campal gyrus, and in these portions layer III had beenselectively damaged. There was moderate destruction ofcells in the anterior nucleus of the thalamus, a small

focal lesion in both the hippocampus and the centralpart of the dentate fascia. The putamen, globuspallidus, caudate nucleus, claustrum and the rest of thethalamus were normal.

Structures of the midbrain, medulla oblongata andcerebellum were normal.

In the cerebral cortex, layers III, V and VI were mostseverely damaged, and, as stated above, the neurones oflayer III were sometimes involved alone (Fig. 8). Layer IIoften was visible when all deeper layers were completelydisorganized.

Case 3. This girl, aged 21 months, had been irritable,feverish and anorexic for a week before admission.On the day before admission she had vomited once andretched several times. At 4.45 p.m. on the day ofadmission her mother had found her in a convulsiveseizure involving all limbs; her skin was blue, cold andsweaty. The fit lasted an hour and was then followed byunconsciousness in which the child's respirations wereslow and grunting.On examination in the ward she was in unrousable

coma and breathing irregularly. Her temperature was

102.20 F. and her pulse 160. The eyes were deviated to

the left and there was a fine nystagmus to the left. Thepupils were moderately dilated, equal and reacted tolight. The fundi oculorum were normal. The thoraxand abdomen were normal. The legs were hypertonic,without tendon reflexes but with extensor plantarresponses. The arms were flaccid, with active reflexesequal on both sides. There was a fine erythematous rashon the trunk and limbs. Four hours after admission shebegan to convulse and this fit lasted two hours.

SPECIAL INVESTIGATIONS. The cerebrospinal fluid wasunder a pressure of 260 mm. of water, and contained4 white cells per c.mm. and 22 mg. protein per 100 ml.The urine contained many granular casts. The whitecells in the blood numbered 20,000 per c.mm., of which62% were polymorphonuclear leucocytes. Neither lead,antimony, arsenic nor mercury was present in the gastriccontents. The urine and blood showed no evidence oflead in abnormal amounts. Radiological examinationof the skull and chest showed no abnormality. Bloodculture was not done.

COURSE. She remained deeply comatose and her limbsbecame more spastic, the legs in extension, and shegradually developed opisthotonos. After two days at1020 F. her temperature fell to below 1000 F. Twofurther examinations of the cerebrospinal fluid showednormal pressure and no increase in protein; 7 cells perc.mm. were present on one occasion, none on the other.The urine a week after admission was normal. Noorganisms of a pathogenic nature were isolated from thefaeces. She died four weeks after admission in a state ofdecerebrate rigidity.POST-MORTEM FINDINGS. The necropsy (No. 13/56)

was performed five hours after death. Apart from anearly but widespread bronchopneumonia, confirmedmicroscopically, there were no significant changes exceptin the brain.

Central Nervous System. There was a slight increasein the space between the cerebral hemispheres and theskull. The frontal lobes were very pale and their con-volutions slightly angular. The parietal and temporallobes were very spongy to touch, and the grey matter ofthe temporal lobes (even after fixation) tended to fallaway spontaneously under the most gentle handling,and was very granular in appearance, with the consistencyof stiff porridge. The occipital lobes were firm andvascular by comparison. The central white matter con-contained conspicuous capillaries but otherwise wasnormal. The cerebellum, brain-stem and dural sinuseswere normal.

Microscopical examination revealed loss of neuronesof variable degree throughout the cerebral cortex andbasal nuclei.

In the most severely damaged areas of cortex allneurones had disappeared, resulting in a complete lossof stratification, and the grey matter tended to be frag-mented, even in the celloidin sections (Fig. 9). In manyareas the thickness of the cortex was already reduced andearly tenting of the surface was recognizable. Therewas much astrocytic proliferation in the white matterimmediately beneath the affected cortex, but the astro-cytes in the grey matter, apart from those of the subpial

71

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOOD

FIG. 9.-Case 3: temporal cortex showing fragmentation.Weil stain x 25.

; 40>, ,j ?3..jj ,?Ftw? gl?- ?issw ? s<v it, .?(:!l '? fI?:+tp .:i.....5

FiG. 10.-Case 3: caudate nucleus showing loss of neurones andmicroglial proliferation.


layer, were fragmented and impregnated poorly. Axiscylinders were grossly reduced in number, a merescattering of faintly impregnated fragments remaining.Many areas had suffered minimal loss, perhaps of onlyo,te cortical layer, and in this case astrocytic proliferationwithin the grey matter was marked. Microglia, mostlyin the rod-cell stage, were often present in large numbers,except in areas of most severe damage (Fig. 10). Betweenthese extremes all grades of neuronal loss were encoun-tered. A few small strips of cortex were completelynormal but the vastly greater part showed some lesions.Capillaries were slightly increased in damaged areas butthere was no perivascular or meningeal infiltration byinflammatory cells.Of the cortex, that in both temporal lobes suffered

most severely, all neurones having been lost from all thegyri in the anterior portions of these lobes. Smallsegments of complete loss were scattered throughoutother areas of cortex. All frontal gyri on both sideswere severely affected, and lesions were also found in theanterior and posterior central gyri, the cingulate gyrus,insular cortex, the hippocampal gyri and the orbitalgyri. A small area, about I cm. long, of moderatelysevere cell loss was present in the cortex of the hippo-campus and some lesser damage in the concavity of thedentate fascia. Layers II and IV were the most oftenpreserved, and sometimes for long stretches all otherlayers had disappeared while these remained. Layers Vand VI were on the whole the most severely and frequentlyinvolved.The lateral and anterior parts of the thalamus were

considerably damaged, particularly the anterior nucleus,and on one side the amygdaloid nucleus had lost nearlyall its cells. The head of the caudate nucleus had sufferedmuch loss on one side whereas only its mid-portion onthe other contained lesions and these were very slight.The claustrum on one side showed a few focal areas ofloss. The putamen, globus pallidus, substantia nigra,red nucleus and subthalamic nucleus on both sides werenormal, nor were any changes found in the cerebellumor medulla. The choroid plexus, meninges and whitematter were normal.

Case 4. Until the present illness at the age of 13months, this patient had developed normally and wasalert, intelligent and affectionate. One week before herillness she had not been herself, being restless and irritableafter having a haemangioma on the temple treated withcarbon dioxide snow. On the day before the convulsionshe was hot and drowsy. The next day she had a tem-perature of 1030 F. in the morning but ate a good lunchand had a normal motion. Apart from being hot andhaving a foul breath she was normal at 3 o'clock in theafternoon. At 4 o'clock she gave a cry and went into aconvulsion which lasted two and a half hours, duringwhich her colour was bluish-grey. The doctor took hertemperature just after 4 p.m. and found it to be 1050 F.She remained unconscious, with intermittent clonicmovements all night but her colour was normal. Thenext morning she passed foul fluid motions full of bloodand mucus. She was admitted unconscious to a private

72

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

BRAIN DAMAGE FOLLOWING FEBRILE CONVULSIONS

ward. The sample of cerebrospinal fluid taken justafter admission was normal. Six weeks later when shewas transferred to a public ward, spasticity in flexion ofboth arms and slight spasticity of the right leg werepresent, and she lay in bed showing no awareness of hersurroundings.Ten days after admission her temperature rose to

1050 F. and the left tympanic membrane bulged a little.This settled down with penicillin treatment but there wasintermittent pyrexia over the next month, occasionalcrepitations being heard in the lungs. Six weeks afteradmission she had a typical attack of measles and threedays after the rash had faded her temperature rose again,as a result of a right otitis media, both drums at this stagebeing opened under anaesthesia. From the pus obtainedpneumococci werp cultured. Her temperature felltemporarily, only to rise with the appearance of wide-spread crepitations in the lungs. Her general conditiondeteriorated and death occurred 18 weeks after the onsetof her illness. She had remained unconscious through-out.

SPECIAL INVESTIGATIONS. Apart from examinationsof the cerebrospinal fluid at the beginning of her illnesstwo further samples, both normal, were taken at threeweeks and 11 weeks after the onset.The Mantoux reaction was negative. Blood examina-

tion showed little significant abnormality, except for araised white blood count (21,000 per c.mm.) on oneoccasion. Sera collected four weeks and 10 weeks fromthe onset of the illness gave negative complement-

FIG. 11.-Case 4: atrophy of convolutions of frontal and anteriorparietal lobes.

FIG. 12.-Case 4: frontal cortex showing loss of architecture andsurface irregularity.

Haematoxylin and van Gieson x 25.

fixation reactions against mumps, lymphocytic chorio-meningitis and Murray Valley encephalitis viruses.POST-MORTEM FINDINGS. The necropsy (No. 32/54)

was performed two hours after death. Apart from generalwasting of the body there was a bilateral purulent otitismedia and mastoiditis, and terminal bronchopneumonia.

Central Nervous System. Both frontal lobes and thepost-central gyri were shrunken with a correspondingincrease in cerebrospinal fluid over them. The gyri werenarrow and angular and were separated by wide sulci(Fig. 11). The temporal and occipital lobes and cere-bellum looked normal. Coronal slicing of the brain inthe affected areas revealed irregularity and reduction inthickness of the cortical grey matter and mild reduction inbulk of the central white matter. The basal ganglia werenormal.

Microscopic examination showed a mild cellularthickening and increase of collagen fibres in the arachnoidmembrane, but no inflammatory cells. Neurones haddisappeared in greater or less degree throughout thecortex of the gyrus cinguli, the frontal gyri, anteriorcentral and posterior central gyri. The mid-portion ofthe caudate nucleus, the lentiform nucleus, the hippo-campus, amygdaloid nucleus, mammillary bodies,lateral part of the thalamus, brain-stem and cerebellum,and the temporal gyri were normal. The head of thecaudate nucleus, midbrain, occipital cortex and otherparts of the parietal cortex and thalamus were notexamined microscopically. Large parts of the cortexwere extremely thin, with almost complete loss ofneurones and with surface tenting (Fig. 12). In areas oflesser severity the layering of neurones was severely dis-turbed, layers III and V apparently suffering most.Even when other layers had completely disappearedlayer II was often still recognizable. No Betz cells wereseen. Coincident with the disappearance of neuroneswas an extreme reduction in nerve fibres, and an increasein astrocytes, particularly in layers V and VI and in theadjacent white matter. In the white matter there was adiffuse pallor together with vacuolation of remaining

73

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOODmyelin sheaths and there were numerous grape-likebodies. Capillaries had increased slightly in number,but were not accompanied by any infiltration of inflam-matory cells in their adventitial zones.

Case 5. This girl was first admitted at the age of10 months for treatment of scurvy, having had novitamin additions to her diet of cows' milk and 'farex'.After treatment with ascorbic acid and iron for a fort-night the child was discharged much improved. Thefollowing day, however, she was readmitted at 1 p.m.having been convulsing since 7 a.m. On examinationshe was convulsing and had a temperature of 1050 F.The convulsions involved the left side more than theright. The fundi oculorum and pupils were normal.There was no neck stiffness nor any abnormality notedin the chest, abdomen or limbs. Following treatmentshe came out of the convulsive state but was comatose,responding to painful stimuli by withdrawal, and thelimbs were spastic.

SPECIAL INVESTIGATIONS. The haemoglobin value was8 g. % and the white blood cells numbered 28,000 perc.mm., 92% being polymorphonuclear leucocytes. Theurine contained numerous granular casts. The cerebro-spinal fluid, which was not under increased pressure,contained 21 mg. of protein per 100 ml., and no cells.Culture of the faeces yielded a Salmonella group Corganism. Blood culture was negative. The Wasser-mann reaction was negative. Blood taken five weeksafter the illness began had no complement-fixingactivity against mumps and Murray Valley encephalitisviruses.

COURSE. Her temperature remained elevated until aweek after admission when it returned to normal, andat this time a further culture of the faeces was negativefor Salmonella organisms. The child's state of con-sciousness, however, improved very little. She lay inbed unaware of her surroundings, often breaking intospasms of high-pitched crying. She remained spastic inall limbs, the arms being held in flexion. Her eyesshowed conjugate deviation to the right, and there wasintermittent coarse nystagmus. The pupils reacted tolight and were equal but she showed no awareness ofmoving objects and she did not respond to auditorystimuli. Further examinations of the cerebrospinal fluidyielded normal results. She was finally discharged twoand a half months after the onset of her illness-spasticand comatose, responding only to painful stimuli, andstill prone to outbursts of prolonged crying.

Virus ExaminationApart from complement-fixation tests on sera from

Cases 4 and 5 during life, pieces of frontal cortex, basalganglia, cerebellum and medulla selected from Cases2 and 3 were submitted to the Institute of Medical andVeterinary Science, Adelaide, for virus isolation. Thematerial was inoculated into newborn and weaned mice,on to the rabbit's cornea, into embryonated eggs andHeLa cell tissue culture. No viruses were recovered.

DiscussionScholz (1951) has published his findings in the

brains of children dying after severe febrile con-vulsions, but his monograph is not available to thepresent writer. However, Meyer, Beck andShepherd (1955) refer to his work in their paperdescribing the cerebral lesions in a boy of 9 yearswho suffered from status epilepticus a year beforedeath. They found neuronal damage in variousparts of the cerebral cortex, in the basal ganglia andcerebellar cortex. They accepted Scholz's view thatbrain lesions may be directly due to a convulsivedisorder. Before considering the part played by theconvulsions in the present cases it is first necessaryto exclude the action of other well-recognized causesof cerebral disease.

In this series of cases, all the children were men-tally and physically normal before the onset of theillness, except for Case 5 who has scurvy. Theoldest child was 2 years. The precipitating illnesswas gastro-enteritis in two cases, upper respiratoryinfection in one, measles bronchopneumonia in oneand undiagnosed in one. In all cases the tempera-ture at the onset was above 1020 F. and in two casesreached 1050 F. The length of the convulsions wasnot easy to determine for two patients were alreadyin the attack when found, but the shortest time wasone hour (Case 3) and the longest time at least sixhours. None of the patients recovered conscious-ness and as the residual twitchings faded variousdegrees of spastic paralysis appeared-completequadriplegia in three cases, diplegia and hemiplegiaeach in one case.The usual clinical diagnosis of encephalitis made

in these cases was not confirmed by lumbar puncture,for the cerebrospinal fluid, apart from raised pres-sure in one instance, was normal. In Case 2 onlybloody taps were obtained.One of the children is still alive but pathological

examination of the other four cases revealed damageto various parts of the cortex and basal ganglia.In Case 4 the frontal lobes and post-central gyriwere involved bilaterally, all other areas examinedbeing normal. The temporal and frontal lobeswere severely damaged bilaterally in Cases 1 and 3,unilaterally in Case 2, but in this last case the superiortemporal gyrus was selectively involved on the otherside. Layer III was the most frequent and oftenonly cortical layer showing disease, layers V and VIbeing involved next in frequency. The thalamus,caudate and amygdaloid nuclei were damaged inthree cases, often severely, but in Case 2 unilaterally.The hippocampus contained lesions in three cases,severe in Cases 1 and 2. Lesions in the claustrumtended to be focal and small, unlike the diffuse

74

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

BRAIN DAMAGE FOLLO WING FEBRILE CONVULSIONS

lesions in the other nuclei. Damage of moderateseverity was found in the lentiform nucleus in twocases. Only in Case 1 did the cerebellum showlesions, which consisted of small scattered groups ofnecrotic Purkinje cells. No lesions were found inthe midbrain, pons or medulla. The choroidplexuses, cerebral arteries and dural sinuses werenormal. There was no evidence of inflammatoryreaction around the blood vessels or in the meninges,nor did the vessels contain thrombi or other signi-ficant abnormality, although in some cases they weresomewhat dilated and in Case 2 greatly increased innumber. No signs of perivascular demyelination orinflammation were found in the white matter andthese were deliberately sought in the patient who hadsuffered from measles. The first four cases illus-trate a progress of the changes from necrosis todisappearance of nerve cells, followed by astrocyticand microglial proliferation. An interesting findingin the three cases of shorter duration was the in-ability to impregnate astrocytes in areas of neuronaldamage, associated with abnormal astrocytic formsin the border zones. In Case 4, however, of 18weeks' duration, astrocytic proliferation was easy todemonstrate.

In Cases 2 and 3 tissue culture and animal inocu-lation failed to provide evidence for the presence ofviruses in brain tissue removed at necropsy, andserum from Cases 4 and 5 gave negative complement-fixation reactions. Case 5 is included, in spite ofthe lack of pathological proof, for the clinicalfindings are very similar to the other cases, and sheis in a state of mental defectiveness and blindnessmost compatible with widespread cortical damage.An alternative diagnosis of haemorrhage due toscurvy is unacceptable for several reasons, the mainones being that there was no evidence of haemor-rhage elsewhere and that the child had respondedwell to treatment. Moreover, haemorrhage greatenough to damage the occipital and frontal cortex ismost unlikely without signs of increased pressure orother abnormalities in the cerebrospinal fluid.

It seems clear, therefore, that these children weresuffering from febrile convulsions, of a prolonged,severe type, and that brain damage occurred duringthe convulsion. The five cases appeared duringa two-year period in a 300-bed hospital so that thecondition cannot be called rare. In contrast, Cary(1956), in a series of 100 cases of infantile febrileconvulsions, found little evidence of serious cerebralimpairment. This failure may well be due to exclu-sion of some grave cases on the grounds that theyrepresent the convulsive prelude to encephalitis orother primarily cerebral disease. It is probablethat some of the cases of mental deterioration after

convulsions reported by Illingworth (1955) areexamples of the same process described here. Inmy Case 4 a little less damage to the frontal lobeswithout involvement of the motor area could haveled to mental changes alone.The pathogenesis of the neuronal necrosis is not

certain. Toxaemia from the infection is unlikely,for changes in the brain of the type and severityunder discussion are not a feature in children dyingof much more severe and prolonged infections.Direct action of microorganisms is a theory notsupported by any of the findings. The two mostacceptable theories are that either the nerve cellsdegenerate from the effects of the general anoxiawhich occurs during a convulsion or that there isa specific mechanism in the brain producing localanoxia. The unilateral character of the degenera-tion in Case 2 may be interpreted in favour of a localmechanism, although it is probable that individualcellular or lobar metabolic activity in the brain mayinfluence the degree of susceptibility to generalanoxia, e.g., the Purkinje cells illustrated in Fig. 3.This regional character of lesions resulting from whatmight reasonably be thought of as a generallyacting disease process has been mentioned bySouthcott and Fowler (1954) in relation to acutehaemorrhagic leucoencephalitis. Against the par-ticipation of local vascular spasm in the causation ofthe changes is the conclusion of Penfield and Jasper(1954) that during an epileptic seizure the cortex,contrary to general opinion, does not blanche andthat its blood flow is actually increased. In Case 1the changes in the liver are very suggestive ofanoxic degeneration, particularly in view of theircentral localization in the lobule, without inflam-matory reaction. The lesions in the liver may.however, have been an effect of toxaemia, althoughI have never seen this phenomenon in other cases ofequally severe and widespread bronchopneumoniadue to Staphylococcus aureus. Anoxia as animportant factor in centrilobular necrosis of therabbit liver, particularly in conjunction with lowblood pressure, has been shown by McMichael(1937) who concluded that due to the low oxygentension in its mixed arterial and venous supply theliver is especially liable to suffer in states of generalanoxia.

Several authors have reported the association ofinfantile infections with cerebral disease similar tothe cases described here, but they have interpreted theassociation very differently. Woolf (1955) des-cribed the brain changes in three patients dyingfive, eight and 16 years after the sudden onset of aconvulsive illness associated with infection. Hethought the cerebral changes were due to dural

75

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

76 ARCHIVES OF DISEASE IN CHILDHOODsinus thrombosis, but, owing to the nature of hismaterial, he did not have the opportunity to demon-strate the presence of old thrombi. Hallman,Hjelt and Tahka (1956) report the findings in twoinfants with gastro-enteritis. Both of their patientshad convulsions early in the illness. These authorsattribute the brain lesions to dehydration and theresulting vascular stasis. Kramer's clinical account(1954) of gastro-intestinal disturbance and cerebraldamage includes four cases beginning acutely withconvulsions.A unique case is that of Crome (1952). Encephalo-

pathy in the later stages of gastro-enteritis occurredin the absence of a history of convulsions. Cromeattributed the cerebral changes to anoxia followingdehydration and vascular stasis. The rarity of suchan association is most unusual considering theextreme degrees of dehydration commonly producedby gastro-enteritis in infants. Nevertheless, thepresent cases demonstrate that similar lesions canoccur in the absence of both gastro-enteritis anddehydration. The commonest factor found in allthe reported examples, apart from Crome's case,is the presence of convulsions.

Putting aside the question of pathogenesis, severalfurther aspects of the process are of importance.Meyer, Beck and Shepherd (1955) have interpretedthe degeneration in the thalamus of their case as aretrograde effect, and therefore likely to throw lighton the cortico-thalamic connexions. In Cases 1and 2 the nerve cells of the thalamus and cortex wereat the same stage of degeneration suggesting ratherthat the damage was simultaneous and not con-secutive. Quite apart from this consideration thesewriters emphasize the importance of the lesions asfoci for convulsive seizures in those patients whosurvive, particularly in reference to temporal lobeepilepsy. In their case and in three of the presentcases the temporal lobes suffered severely. It islikely, moreover, that damage following febrile con-vulsions in early infancy may be the basis for manyof the spastic conditions, such as Little's disease,infantile hemiplegia and other variants, with orwithout dementia, and of behaviour disorders.

SummaryFive children, aged up to 2 years, and previously

normal except for one with scurvy, suffered from aprolonged convulsion at the beginning of febrileillnesses comprising gastro-enteritis (two cases),measles bronchopneumonia (one case), upperrespiratory tract infection (one case) and unidentified(one case).The patients remained unconscious after the con-

vulsions, which lasted from one to six hours, butdeveloped various spastic paralyses, and in fourcases death occurred four days, 12 days, four weeks,and 18 weeks from the onset of the convulsion.Pathological examination revealed neuronal necrosisof variable distribution and severity in the cerebralcortex, basal ganglia and cerebellum, with noevidence of the direct action of microorganisms.The changes are considered to be the result ofsystemic anoxia of convulsive origin or, less likely,to be due to a specific mechanism acting locally inthe brain. Alternative explanations put forwardby other observers are discussed, and it is concededthat stagnation anoxia following dehydration ingastro-enteritis may be responsible on rare occasions.The importance of convulsive cerebral degenerationas a precursor of epilepsy and as the cause ofinfantile spastic and behaviour disorders ismentioned.

My thanks are due to the various members of thehonorary medical staff who have kindly allowed me touse their case notes, and to the Board of Management ofthe Adelaide Children's Hospital for permission topublish these investigations. The virus examinationswere kindly undertaken by Dr. Peter Warner and Dr.Alan Duxbury, of the Medical Research Division of theInstitute of Medical and Veterinary Science, Adelaide.

REFERENCES

Cary, W. (1956). Med. J. Aust., 2, 254.Crome, L. (1952). Archives of Disease in Childhood, 27, 468.Hallman, N., Hjelt, L. and Tahka, H. (1956). Ann. paediat. Fenn.,

2, 31.Illingworth, R. S. (1955). Archives of Disease in Childhood, 30, 529.Kramer, W. (1954). Ned. T. Geneesk., 98, 3631. Quoted by The

Year Book of Neurology, Psychiatry and Neurosurgery,1955-56 Series, p. 167.

McMichael, J. (1937). Quart. J. exp. Physiol., 27, 73.Meyer, A., Beck, E. and Shepherd, M. (1955). J. Neurol. Neurosurg.

Psychiat., 18, 24.Penfield, W. and Jasper, H. (1954). Epilepsy and the Functional

Anatomy of the Human Brain. Boston.Scholz, W. (1951). Die Krampfschddigungen des Gehirns. Berlin.

Quoted by Meyer, Beck and Shepherd, above.Southcott, R. V. and Fowler, M. (1954). Med. J. Aust., 2, 65.Woolf, A. L. (1955). J. ment. Sci., 101, 610.

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

damageafter febrile convulsions - bmj

Documents