superficial haemosiderosis of central nervous system · suiperficial haemosiderosis ofthe central...

J. Neurol. Neutrosurg. Psychiat., 1964, 27, 332

Superficial haemosiderosis of the central nervous

systemB. E. TOMLINSON AND JOHN N. WALTON

From the Departments ofNeuropathology and Neurology, Regional Neurological Centre, the General Hospital,Newcastle upon Tyne

In 1940 Noetzel reported two cases of slowlyprogressive neurological illness in which, at necropsy,the meninges overlying the brain and spinal cord andthe peripheral areas of the brain, cerebellum, andspinal cord were found to be rust-brown in colour.Iron pigment was found in the meninges, glial cells,within mesodermal cells and around blood vessels inthe cortex, and in the lining of the cerebral ventricles,and the pigmented cerebellar convolutions weredisorganized. The author concluded that this condi-tion was due to the diffusion of blood pigment intothe marginal zone of the central nervous system andsuggested that the condition was probably the resultof chronic subarachnoid bleeding. Subsequently,however, additional cases were reported by Leweyand Govons (1942), Cammermeyer (1947), Neumann(1948, 1956), Garcin and Lapresle (1957), Rosenthal(1958), and McGee, van Patter, Morotta, andOlszewski (1962). Neumann (1948, 1956) concludedthat the pathological changes in the central nervoussystem which she observed in her two cases were theresult of haemochromatosis of the central nervoussystem. Others, however, have pointed out that thepigment deposits consist of haemosiderin and suggestthat in most, if not all, cases this disorder resultsfrom repeated or long-continued bleeding into thesubarachnoid space. In all, 11 cases of this type havenow been reported in the world literature. They haveshown a remarkably consistent clinical picturecharacterized particularly by progressive dementia,nerve deafness, and cerebellar ataxia, and the patho-logical changes in the reported cases have also beenstrikingly uniform in character. Table I summarizesdetails of the cases reported to date and we givebelow detailed clinical and pathological observationsin the first case of this type to be reported from GreatBritain.

CASE REPORT

CLINICAL DETAILS An untrained children's nurse, aged16 years, was admitted to the Newcastle General Hospital

for investigation on 13 May 1959. For two years she hadsuffered from paroxysmal headaches with vomiting whichhad become much more severe in the last six months.During the headache the patient would often experienceblurring of vision and transient right-sided hemianopia.Independently of these headaches she had begun toexperience brief episodes lasting only a few seconds inwhich she lost the power of speech and would go 'blank'for a few seconds. Tongue-biting and incontinence hadnever occurred. Recently too she had seemed to becomeslow and disinterested and intermittently drowsy and herappetite and weight had diminished considerably.On examination on admission the patient was conscious

and coherent but seemed dull, lethargic, and disin-terested. She complained intermittently of right-sidedheadache. The optic fundi were normal, but examinationof the visual fields revealed a congruous right upperquadrantic field defect. There were no other clear-cutneurological signs on examination. Survey views of theskull showed no significant bonv abnormality; the pinealgland was faintly calcified but could not be accuratelyidentified in the antero-posterior view. A lumbar puncturewas not performed. During her stay in hospital thepatient was observed to have recurrent attacks of minorepilepsy. An E.E;G. was abnormal; it showed generalizedparoxysms of slow activity of both theta and delta fre-quency suggestive of epilepsy of central origin. There wasno clear evidence of a focal lesion in either cerebralhemisphere. Bilateral carotid arteriography was per-formed on 19 May. No angioma or other vascular mal-formation and no pathological circulation was demon-strated in either hemisphere, but Dr. G. L. Gryspeerdtcommented that the appearance suggested a degree ofdilatation of both lateral ventricles.The patient was treated with anticonvulsants and

ergotamine preparations; the headaches improved andshe became more alert; she was discharged home on 27May 1959. She was seen again on 11 July and haddeteriorated. Her speech had become slurred and her gaitunsteady, she was continuing to lose weight, and her'blackouts' were more frequent. The patient was re-admitted to hospital on 14 July. She was found to beconfused, disorientated, and lethargic and formal intel-lectual testing indicated a degree of early dementia. Therewas some neck stiffness and the visual field defect wasunchanged. Psychometric tests showed that she appearedto be functioning at the 'high-grade feeble minded' level

332

Protected by copyright.

on August 17, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.27.4.332 on 1 August 1964. D

ownloaded from

http://jnnp.bmj.com/

Suiperficial haemosiderosis of the central nervous system 333

TABLE I12 REPORTED CASES OF SUPERFICIAL HAEMOSIDEROSIS OF THE CENTRAL NERVOUS SYSTEM (INCLUDING PRESENT CASE)

9r and Age and Clinical FeaturesSex ofPatient

Pathological FeaturesOther than Haemosiderosis

Duration Cerebellar Nerve Dementia Sphincter Spastic ityof Illness Ataxia Deafness Dis- Subarachnoid

turbance BleedingAetiology

zel 1940 M 47 Not Notknown known

+ + Notknown

zel 1940 F 59 Not Not Not Not Notknown known known known known

- Xanthochromia Carcinomatosis ofmeninges.Ca of stomach

Xanthochromia('Haemorrhagicmeningitis')

Not found atnecropsy

Positive C.S.F.Wassermann re-action, fixed pupils,absent reflexes

Absent knee jerks,sluggish pupils

M 36 7 years + + + + + Xanthochromia,bleeding suspected

nann 1948 M 47 12 years + + -rv + +

nann 1956 F 55 17 years + + + -V T -_ None found

nermeyer

in and-sle 1957

*Not described +

Not found atnecropsy

'Cherry-sized massin fourth ventricle'

? Severe closedhead injury

Aneurysm ofvertebral artery

F 47 10 years T -V- ±-L Suspected, not Not discovered Diabetes mellitus,proved loss of position sense

in legs; cirrhosis ofliver

rithal M 67 Notknown

+ + Notknown

± Not found Pyelonephritis,chronic bronchitis,emphysema

nthal M 27 Not Because of widespread dissemination of the causalknown neoplasm, signs and symptoms due to haemosiderosis

could not be properly assessed

M 52 10months

M 59 10 years

_ -V - +(? due totumour)

++ - Notknown

++

_ -V+

+±+

2 years + + + + + + +Terminal Terminal

Oligodendroglioma,subarachnoid andintraventricularspread

Ependymoma offilum terminale

Intracranialangioma

Ependymoma oflateral ventricle

Absent ankle jerks.? due to tumour

(I.Q. 67). Her recent memory was severely impaired,there was marked fluctuation in her performance, andsome of the test results suggested recent organic deteriora-tion. Neck stiffness increased and the patient was now

developing a progressive bilateral deafness of perceptivetype. There was moderate symmetrical cerebellar ataxiaof the arms and legs, but the tendon reflexes were activeand equal, the plantar responses flexor. A lumbarpuncture on 12 August gave fluid under pressure of120 mm. cerebrospinal fluid; the fluid was uniformly anddeeply bloodstained. The packed cell volume in the fluidwas 80% and the protein content of the deeply xantho-chromic supernatant fluid was 5.0 g. per 100 ml. AnE.E.G. was more abnormal than the previous recording;it was now somewhat asymmetrical, showing more slowactivity on the left side and particularly in the lefttempoial region, but no clear-cut focal abnormality wasdemonstrated. Ventriculography was performed by Mr. J.Hankinson on 14 August 1959. The intracranial tensionwas incdeased; the lateral ventricles wele encountered at

a depth of 5 cm. The cerebrospinal fluid was deeply blood-stained with a mixture of fresh and recent blood. Airpassed freely from the left to the right ventricle and a

total of 40 ml. of cerebrospinal fluid was replaced withair. Radiographs demonstrated multiple filling defects inthe lateral ventricles which were interpreted as being dueto patchy coagulum and it was decided that no surgicaltreatment was possible. Shortly afterwards the patientdeveloped retention of urine with overflow and bilateralextensor plantar responses. She gradually lapsed intocoma and died on 3 September 1959.

PATHOLOGICAL FINDINGS At necropsy the body was thatof a fully developed but very thin girl, 5 ft. 2 in. tall andweighing 44 kg. (5 st. 4 Jb.). There was no skin pigmenta-tion. In the internal organs, apart from broncho-pneumonia with abscess formation, there were no signi-ficant abnormalities outside the central nervous system.No pigment deposition or fibrosis was present in any ofthe general organs. The scalp and skull and dura mater

Other Features

y and)ns 1942

Alcoholism

ec, vanr, Morotta,)lszew ski

inson and F 16n 1964



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


B. E. Tomlinson and John N. Walton

were normal. In the subarachnoid space of the spinalcord and brain, particularly in the basal cisterns and overthe pons and medulla and anterior third of the cerebellum,there was deep brown pigmentation with numerous ad-hesions and with brown staining of the outer layers of thecerebral cortex and spinal cord (Fig. 1). There was nofresh blood in the subarachnoid space. The basal vesselswere embedded in adhesions and pigment but were allthin-walled and apparently normal. The origins of all thecranial nerves were pigmented, the pigmentation stoppingabruptly at their exit from the subarachnoid space. Somenerves were only slightly pigmented, the fifth and sixthbeing only yellow at the origins, but the seventh andeighth were nigger brown, the latter for 7-8 mm. andthe pigmentation extended for 2-3 mm. into the brainsubstance. The olfactory bulbs and the optic nerveswithin the skull were deep brown in colour. In the latterpigmentation extended to the posterior surface of thebulb but gradually diminished from the optic foramenperipherally. The origins of all the spinal nerves weredeeply pigmented. Arachnoid fibrosis, which was markedat the base of the brain, had apparently obliterated theforamina of Lushka but not the foramen of Magendie.

There was considerable dilatation of both lateralventricles and intense brown pigmentation of a 1 mm.layer of subependymal tissue in all the ventricles. Multipledeep brown, ependymal granulations were presentthroughout all the ventricles and the left posterior hornwas filled by a solid, vascular, granular tumour (Fig. 2).Two small deposits of tumour were visible in the anteriorend of the left lateral ventricle, and a further deposit,5 mm. across, lay in the hypothalamus at the level of thetuber cinereum. Pigmentation was approximately i mm.deep over most parts of the cortex, 1 mm. deep beneaththe ependyma, 2-3 mm. deep over the periphery of thepons, cerebral peduncles, the quadrigeminal bodies, andthe medulla and the whole of the cord (Figs. 3 and 8). Theentire thickness of the hypothalamus was pigmented.Pigmentation of the cervical cord left only a central un-pigmented zone 4 mm. across and the thoracic andlumbar cord showed an even smaller central zone ofunpigmented tissue. The anterior third of the cerebellumwas pigmented to a depth of 1 cm., with visible shrinkageand disorganization of folia, particularly those of thevermis and of the structures in the roof of the upper endof the fourth ventricle. The central lobule and flocculuswere not identifiable, being reduced to a deeply pig-mented mass of honeycombed glial tissue.The tumour in the lateral ventricle was a papillary type

of ependymoma with a vascular stroma which includedmany thin-walled blood vessels, some of which werethrombosed. Within the tumour itself there was verylittle pigment; it was present in the interstitial tissue, butnot within the tumour cells. The granular appearance ofthe ventricles was due to subependymal gliosis, withmuch pigment in macrophages and encrusting astrocytes.Such granulations were present throughout all theventricular system, in the aqueduct and in the floor androof of the fourth ventricle (Figs. 2 and 3). The ependymalcells contained either no pigment or only a few minuteparticles, the pigment being in the subependymal tissues.The choroid plexuses were deeply pigmented.

The pigment was almost entirely iron-containing,giving a positive Perls's reaction. In unstained sections alarge quantity of pigment occurred as yellowish-brownor brown granular particles up to 4 or 5 j* in diameter,all apparently within macrophages. Fine pigment granuleswere present within macrophages in the pia arachnoidbut some lay free and a little was present within fibro-blasts. Very fine particles of iron-positive pigment,approximately 1 ,u in size, were present within the cyto-plasm of numerous hypertrophied astrocytes and in somemacrophages. In astrocytes these particles were mostnumerous at the base of the astrocytic processes, andsimilar particles were seen in the adventitia and outermedia of blood vessels in the pia arachnoid and in theperiphery of round, hyaline bodies which formed aprominent feature of the histological preparations (Figs.4 and 5). Some non-granular iron-positive material waspresent within these hyaline bodies and macrophages. Iniron-stained preparations almost no unstained pigmentapart from a little lipochrome within an occasionalneurone could be identified. The pigment was removedfrom sections by treatment with 5% oxalic acid. It is con-cluded, therefore, that almost all of the pigment depositedin the central nervous system in this case was haemo-siderin. Attempts to identify other pigments were almostcompletely unsuccessful. A very occasional macrophagein the spinal cord and cerebellar cortex contained sudano-philic material and a very occasional neurone a smallquantity of lipofuscin. Haematoidin (bilirubin), calcium,and melanin were not found. The pigment will, therefore,from now on be called haemosiderin.The hyaline bodies, previously mentioned, were present

in both neuronal and non-neuronal areas and in someparts of the outer cortex were very numerous (Figs. 4 and5). They stained with differing degrees of eosinophilia andvaried in size from 15 to 50 IA; they were oval or round,and usually well defined, appearing to have a limitingmembrane. In many, fine particles of haemosiderin werepresent, particularly in the periphery, and all gave a posi-tive reaction for haemosiderin, though the intensity variedfrom a faint to a very strong reaction. The larger bodiescontained no nucleus, but an occasional small body hada minute flattened peripheral nucleus. They contained nodemonstrable lipid, showed no metachromasia withtoluidine blue but were slightly metachromatic withmethyl violet, only faintly positive with periodic-acidSchiff reagent, and were negative with carbol fuchsin orin attempts to stain by Schmorl's method. They showedonly a faint affinity for silver. It appears likely that themajority are degenerating macrophages since occasionallysmall groups of them were found in perivascular spaces.Although very numerous in this condition and in someof the previous cases reported, identical bodies aieoccasionally found in old haemorrhagic infarcts andpresumably result from bleedingand release of breakdownproducts of red cells in the central nervous system.

In all the pigmented areas there was macrophagereaction, marked astrocytic proliferation, and in someplaces severe gliosis. In some areas there was actualnecrosis and wherever pigmentation was heavy, damageto neurones, myelin sheaths, and neurofibrils was present.In many areas marked endothelial swelling in small

334



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Superficial haemosiderosis of the central nervous system

2i._

FIG. 1. Pigmentation and arachnoid thickening is particu-larly prominent over the pons, medulla, and anterior thirdofthe cerebellar hemispheres. The bifurcation of the basilarartery is hidden by pigment which has been removedfromthe interpeduncular cistern.

FIG. 2. Horizontal brain section showing the ventriculardilatation, pigmented granular ependymitis, and the tumourfilling the posterior horn of the left lateral ventricle.

FIG. 3. Peripheral haemosiderin deposition at the level ofthe rostral limit of the fourth ventricle. Note the sub-ependymal pigmented granulations in the roof and floorof the fourth ventricle. Perls x 4.

I I(... 'I

335



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



F7# ,*

Se "*I. fr

4d *fo

A1*Fs

0$

.*

if I

FIG. 4 FIG. 5

FIG. 4. Apex of a frontal lobe convolution. The cellular, heavily pigmented pia arachnoid is adherent in one placeto the grossly disorganized superficial cortex. Haematoxylin and eosin x 58.FIG. 5. Higher-power view of the cortex of Fig. 4 showing a network of capillaries, heavily pigmented macrophages,and several hyaline bodies. A hyaline bodv and a neurone lie adjacent to an arteriole showing cellular subintimalhyperplasia. Haematoxylin and eosin x 220.

arterioles was present, suggesting that some of thenecrosis might be secondary to vascular occlusion.

In all areas of the cerebral cortex, an occasionalmacrophage and iron-encrusted astrocytes were presentin the molecular layer. Pigmentation was heaviest overthe apices of the convolutions, reaching the third layer ofthe cortex, and masses of pigmented macrophages werepresent around blood vessels, and in clumps and singlyin the cortex, and throughout these areas there wasmarked astrocytic proliferation. In such areas there wassevere loss of neurones (Figs. 4 and 5). Occasionalshrunken or very pale-staining neurones could be foundin the affected areas but in no identifiable neurone couldiron pigment be demonstrated. Similar appearances werepresent in all exposed neuronal areas of the brain, suchas the quadrigeminal bodies, and the pulvinar of thethalamus and the hypothalamus.

In the cerebellum more intense destruction hadoccurred, the midline roof structures in the rostral endof the fourth ventricule being reduced to pigmented glialscars, as were many of the collapsed folia (Figs. 6 and 7).

The incompletely destroyed folia had a characteristicappearance; pigmentation, not necessarily heavy, waslimited to the apex of the folium where neuronal destruc-tion was out of proportion to the degree of pigmentation,and resulted in marked thinning and collapse of themolecular layer with complete loss of the Purkinje cells.Usually the cap of granular cells was also destroyed, buta thick layer of proliferated Bergmann astrocytesremained, their processes encrusted with haemosiderin.The periphery of the spinal cord was heavily pigmented,

haemosiderin being within macrophages and astrocytes,but pigment was particularly heavy in the anterior horns(Fig. 8), though the neurones appeared to be entirelynormal. The pigmented periphery of the cord showedpartial demyelinization with numerous beaded andswollen myelin sheaths and loss and fragmentation ofneurofibrils.

Large clumps of macrophages and some astrocyteswere present at the points of exit of cranial nerves and atthe origins of ventral and dorsal nerve roots in the cord.The exit zones of the third and fourth nerves were only

A.:0.I..:..A

0: 4e

W- . f..

i:t.

'A:.

.a.F

336

*

.,....

.1

p

.t

4

*.f 4'

'iW



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



*; FIG. 7

FIG. 6. Destruction of the cerebellar hemisphere. Normalj ,̂ ' <folia with heavily stained zones of granule cells lie deep to

the eroded surface, where the outline ofdestroyedfolia andthe characteristic loss ofgranhle cells at the apex ofpartly

.-APwiaffected folia are visible. Haematoxylin and eosin x 5.

FIG. 7. Varying destruction of cerebellar folia. In theF...t .h .. f 14o>4wt *< -r,<.<centralfolium, the solid mass of granule cells which form

the apex of a normal folium (at the top) has disappeared;the outline of the folium is altered from the normal club-shaped, to aflattened, mushroom-shaped top; the remainingcap of cells is hypertrophied Bergmann astrocytes. Thethin, acellular molecular layer is just visible. The folium atthe bottom has been completely destroyed. Haematoxvlinand eosin x 50.

II1( i. IY FIG. 8. Perls's reaction on a section of the thoracic cordshowing the heavy peripheral pigmentation, the smallcentral unpigmented zone, and the heavily pigmentedanterior horns. x 10.

337



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



slightly involved, but the second and eighth showedextremely heavy deposits of iron with considerable myelinloss, and iron deposition and astrocytic proliferationextended into the superficial areas of the dorsal andventral cochlear nuiclei. The optic nerves showed littleloss of myelin or nerve fibres.

DISCUSSION

The symptoms and the physical signs in this caseclearly resulted from the combined effects of thetumour and the haemosiderosis. Repeated smallhaemorrhages were presumably responsible for therecurrent severe headaches. Subsequently hydro-cephalus, progressive dementia, and chronic sub-arachnoid bleeding brought to mind the possibilityof an intraventricular tumour such as a choroidplexus papilloma, but the findings on ventriculo-graphy were inconclusive. The haemosiderosis of thecentral nervous system was almost certainly respon-sible for the cerebellar ataxia, the nerve deafness, andfor much of the mental deterioration. This lattercould be accounted for by the widespread thoughrelatively superficial cortical destruction, but theassociated hydrocephalus, possibly due to obstruc-tion of the fourth ventricular foramina, could havebeen a contributory factor. The involvement of theeighth nerve at its origin and of the peripheral partsof the cochlear nuclei presumably caused the nervedeafness and the severe cerebellar destructionaccounted for the cerebellar symptoms and signs.

In the case ofNeumann (1948) marked destructionof anterior horn cells had apparently occurred as aresult of the haemosiderin deposition. In that case,however, the illness was protracted, lasting more than12 years. We observed marked pigmentation of theanterior horns at necropsy in our case, but signs oflower motor neurone weakness were not detectedduring life and there was no neuronal degeneration,though the latter change might have developed hadthe illness lasted longer. In the case we have describedthe partial demyelination and fibre damage in theperiphery of the cord could well have given rise tosensory disturbances, e.g., loss of vibration and posi-tion sense in the legs, but the patient became sodemented that accurate sensory testing was notpossible. However, signs of pyramidal tract dys-function only appeared as a terminal event.The clinical findings in the 12 cases reported to

date, including our case, may therefore be sum-marized by saying that eight showed bilateral nervedeafness, seven bilateral cerebellar ataxia, eigl4 hadevidence of bilarteal pyramidal tract disease, fourretention of urine with overflow, four absent reflexesor impaired vibration sense in the legs, and at leastseven showed progressive dementia in the terminalstages. Several of these features may have been

present in the remaining cases but were not com-mented upon specifically in the case reports or elsethey were obscured by other neurological findings.Turning to the pathological findings, all cases showedthe characteristic brown staining of the meninges andof peripheral areas of the central nervous system,usually with a striking degree of cerebellar degenera-tion and atrophy. In seven there was definiteevidence of subarachnoid bleeding and in three it wassuspected; only in two was there no evidence ofmeningeal haemorrhage. In one case the haemor-rhage was due to an intracranial aneurysm, in twoto the presence of an ependymoma (of the filumterminale and ofthe left lateral ventricle respectively),in one to a cerebral vascular malformation, in oneto carcinomatosis of the meninges, and in one to anoligodendroglioma. In the remaining cases no definitesource of haemorrhage was identified (Table I).

There is no evidence in this case, as there has beennone in the others in the literature, of generalizedhaemochromatosis and there is no reason to believethat this is a form of that disease localized to thecentral nervous system. Cammermeyer (1947),McDougal and Adams (1950), and Adams (1956)pointed out that in haemochromatosis deposition ofiron in the central nervous system is extremelylimited. The choroid plexuses are intensely pig-mented, but the deposition of iron pigment is con-fined otherwise to a few small foci in relation to theventricular system and to neuro-secretory areas,e.g., area postrema, pineal, and infundibulum. Pig-mentation of the type described in the 12 cases wehave discussed has never been found in associationwith systemic haemochromatosis.

In cases of superficial haemosiderosis, by contrast,the iron deposition has been found in those areasbathed by circulating cerebrospinal fluid. Thus, thecerebral hemispheres, the cerebellum, and the outerlayers of the cord and the brain-stem are involved.The subependymal layers around the ventricles havebeen found to be pigmented in a number of cases;in each of these a focus of intraventricular haemor-rhage was present. In our case the haemorrhage wasclearly arising from the ependymoma in the lateralventricle. The rarity of this condition in no sensedetracts from the view that it is the result of frequentor continuous haemorrhages occurring into the sub-arachnoid space or cerebral ventricles. Although sub-arachnoid bleeding is common, the cases in which itoccurs repeatedly or over a long period of time mustbe rare. It should also be noted that the changes inthe pia arachnoid and in the superficial cortex closeto a ruptured aneurysm are similar, though muchless severe. Thus there is fibroblastic proliferation inthe pia arachnoid with deposition of haemosiderinin macrophages as well as in some fibroblasts and in

338



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



the adventitia of blood vessels. Often there is con-siderable astrocytic proliferation in the outer layersof the cortex; the astrocytes are often aligned in rowsand in some places their processes are encrusted withiron pigment. The curious large hyaline andhaemosiderin-positive bodies found in our case andin a number of the reported instances are also to befound occasionally in relation to areas of old cerebralhaemorrhage or haemorrhagic infarction. Further-more, Iwanowski and Olszewski (1960) have studiedthe effects of multiple injections of heparinized blood,of haemolysed red cells, and of the iron preparationImferon into the cisterna magna of dogs. They foundthat similar lesions in the cerebellar folia, as well aspia arachnoid pigmentation, incrustation of subpialastrocytes with iron-containing pigment, and, in someinstances, foci of cortical necrosis occurred whenmultiple injections of blood were given. The evidence,therefore, that this condition is associated withhaemorrhage into the subarachnoid space orventricles seems to be conclusive. It cannot, however,be stated categorically that it is the deposition ofhaemosiderin itself which results in degeneration.This deposition may merely be the sign of haemor-rhage rather than the explanation for necrosis. In-deed one piece of evidence strongly suggests that itis not haemosiderin itself which causes the necrosis;although neurones in many areas are apparentlydestroyed, and in the cerebellum necrosis of alltissues except the hypertrophied astrocytes andmacrophages is to be found, little, if any, haemo-siderin is found within neurones. In our case haemo-siderin deposition was at its heaviest in the anteriorhorns of the spinal cord, yet the neurones themselveswere intact and contained no haemosiderin. It thusseems probable that some other breakdown productof the blood and plasma liberated into the sub-arachnoid space is responsible for tissue damagerather than the haemosiderin itself and that the latteris merely an obvious and striking indication thatrepeated previous bleeding has occurred.The relatively constant occurrence of the three

main symptoms and signs of this condition, namely,nerve deafness, cerebellar ataxia, and dementia,shows that the eighth nerves and perhaps the peri-pheral parts of the cochlear nuclei constantly suffer,as do the cerebellum and the superficial layers of thecerebral cortex. It is, however, clear that other areas,

particularly the anterior horns, roots, and long tractsin the spinal cord, may also be severely affected andmay give rise to some of the predominant symptoms.Thus the impairment of position sense and of vibra-tion sense in the legs reported in the case of Garcinand Lapresle (1957) would presumably be explainedby degeneration ofmyelin and axons in the peripheralparts of the cord, and absent lower limb reflexes

could be accounted for by either root or cord in-volvement. The bladder disturbances seen in ourpatient and reported in a number of those previouslydescribed probably result from involvement of thesacral segments of the cord or sacral roots. The bi-lateral pyramidal tract signs present in the majorityof the cases can be explained either by the involve-ment of the pyramids within the medulla or bydegeneration of the peripheral parts of descendingtracts in the spinal cord. In our case demyelinationwas slight in these areas and signs of pyramidaltract dysfunction were only found terminally.

SUMMARY

The case is reported of a 16-year-old girl who gavea two-year history of recurrent headaches and of arecurrent right-sided visual field defect. Later shedeveloped minor epileptic seizures, nerve deafness,dementia, cerebellar ataxia, urinary retention withoverflow, and finally signs of bilateral pyramidaltract dysfunction. The cerebrospinal fluid was deeplyblood-stained. At necropsy a vascular ependymomaof the left lateral ventricle was discovered and therewas widespread staining of the meninges and ofsuperficial and periventricular areas of the cerebralhemispheres, cerebellum, brain-stem, and spinal cordwith haemosiderin. In the deeply-stained areas therewas usually widespread degeneration of neuronesand of glial cells; superficial demyelination was ob-served not only in the brain-stem and spinal cordbut also in cranial nerves and nerve roots.

This is the twelfth case of haemosiderosis of thecentral nervous system to be reported in the worldliterature. The clinical and pathological findings inthese 12 cases are discussed. The predominant clinicalfeatures of the illness are nerve deafness, dementiaand cerebellar ataxia, with eventually, retention ofurine with overflow and other signs of spinal corddisease. Reasons are given for concluding that thesyndrome results either from repeated small haemor-rhages or from continuous bleeding within thesubarachnoid space over a prolonged period of time.

We are grateful to E. Manns, A. Willis, and Miss R. Allenfor their help.

REFERENCES

Adams, R. D. (1956). J. Neuropath. exp. Neurol., 15, 215.Cammermeyer, J. (1947). Ibid., 6, 111.Garcin, R., and Lapresle, J. (1957). Rev. neurol., 97, 417.Iwanowski, L., and Olszewski, J. (1960). J. Neuropath. exp. Neurol.,

19, 433.Lewey, F. H., and Govons, S. R. (1942). Ibid., 1, 129.McDougal, D. B., and Adams, R. D. (1950). Ibid., 9, 117.McGee; D. A., van Patter, H. J., Morotta, J., and Olszewski, J. (1962).

Neurology (Minneap.), 12, 108.Neumann, M. A. (1948). J. Neuropath. exp. Neurol., 7, 19.- (1956). Ibid., 15, 214.Noetzel, H. (1940). Arch. Psychiat. Nervenkr., 111, 129.Rosenthal, P. (1958). Dtsch. Nervenheilk., 178, 431.

339



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


superficial haemosiderosis of central nervous system · suiperficial haemosiderosis ofthe central...

Documents