662

THE NATURAL HISTORY OFHYDROCEPHALUS

K. M. LAURENCE, M.A.(Cantab.), M.B., CH.B.(Liverpool)Senior Lecturer in Paediatric Pathology, Welsh National School of Medicine, Llandough Hospital, Penarth

IntroductionAlthough hydrocephalus has always been a

fascination and a challenge to the medical pro-fession,, until the last decade or so the diagnosisand investigation of the condition has been ratheran academic exercise. Surgical methods of treat-ment were not often successful as is suggestedby the numerous different forms of treatmentintroduced and usually abandoned again.With the gradual elimination of nutritional ill-

nesses and of infective children's diseases due tobetter prophylactic measures and the use of anti-biotics, congenital malformations, the so-calledcongenital hydrocephalus amongst them, haveassumed an increasing importance. Intensiveresearch during the last decade and the develop-ment of new plastics have resulted in the intro-duction of a number of new surgical procedureswhich now seem to offer a reasonable chance ofarresting the progress of the disease. Amongstthese is the Spitz-Holter valve operation which invarious hands appears to arrest the condition in70-90% of the cases.", 14, 15 Early diagnosis andaccurate assessment of the cases is therefore nowessential if the best results are to be obtained.A knowledge of the natural history of hydro-cephalus is of very great importance both indeciding when to recommend a case for surgicaltreatment and later in assessing its results.

Hydrocephalus is, however, only a physicalstate, which has a number of anatomical causes.Each of these tends to run a separate and distinctclinical course, and have its own prognosis.

PathologyContrary to the general' belief, hydrocephalus

is more often acquired than truly congenital inorigin. This fallacy has probably arisen becausethe condition so frequently develops within thefirst few months of life, sometimes without anydefinit ½t4ecedent history.

Hydrocephalus, the accumulation of excessiveamounts of cerebrospinal fluid within the cranialcavity, may result from overproduction, obstruc-

TABLE ICAUSES OF HYDROCEPHALUS

Over-production of Choroid plexus papilloma and,C.S.F. villus hypertrophy.

Obstruction to the Within ventricular system:C.S.F. pathway Foramina of Munro.

Aqueduct.Exit foramina.

Subarachnoid space:Arnold-Chiari malforma-

tion.Basal cistern block.

Outside C.N.S.:Skull abnormalities.

Defective absorption of ? Venous sinus thrombus.C.S.F.

tion of the cerebrospinal pathway, or failure ofabsorption of cerebrospinal fluid (Table i). Over-production undoubtedly occurs in the compara-tively rare cases of choroid plexus papilloma,7though most of these are complicated by a co-existent basal cistern block. Failure to absorbfluid as a cause of hydrocephalus has not beenwell authenticated and it is usually possible tofind an obstructive lesion as the cause of the con-dition in cases of dural venous sinus thrombosisin which this under-absorption is thought to occur.

Obstruction is by far the most important causeboth numerically and pathologically, and is usuallyof malformative or inflafmmatory origin or oftena combination of the two.

Spina bifida cystica, which occurs in one or twocases per i,ooo live births, is frequently associatedwith hydrocephalus and this accounts for nearlyall the cases where the condition is due to mal-formation. Here the hydrocephalus may occa-sionally be due to congenital aqueduct malforma-tions, i.e. forking or stenosis. More frequentlythe Arnold-Chiari malformation of the hind brain,which is found in nearly all cases of myelocele,accounts for the hydrocephalus by obstruction ofthe basal cisterns due to overcrowding of theposterior fossa structures (Fig. i), or a block ofthe exit foramina which are displaced down the

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from

November I960 LAURENCE: The Natural History of Hydrocephalus 663..

............

.....

............ .....

.. .......

..

.....

.........

........

..... ......

.......

...

..

FIG. i.-Severe hydrocephalus affecting the wholeventricular system, with hydromelia. An Arnold-Chiari malformation is present with a cerebellartongue and a largely intraspinal medulla, as well asa lumbo-sacral myelocoele, distorted by in-flammatory changes.

cms

......................

FIG. 2.-A hemisected skull of a case of basal cisternblock. The lateral ventricle, seen through thefenestrated septum pellucidum, and the thirdventricle are severely dilated. A terminal stenosishad occurred in the aqueduct (not shown in thissection). The cisterna magna is very grosslydilated, compressing the cerebellum.

spinal canal. Ascending infection from the rawspina bifida sac, which so very frequently occurs,may lead to inflammatory obstructive lesionssuperimposed upoii what may already be an em-barrassed cerebrospinal fluid circulation. Aque-duct malformations may rarely be found withoutassociation with spina bifida cystica.

Inflammatory lesions, apart from those super-imposed upon malformations, form the otherimportant, though smaller, group. Probably thecommonest single cause of post-inflammatoryhydrocephalus is extravasation of blood into thecerebrospinal fluid pathway, resulting from directbirth trauma or perinatal anoxia, so frequentlyseen in premature infants or, occasionally, fromspontaneous subarachnoid haernorrhage or headinjury in older children. Intracranial infectionappears to be a rather less important cause.Frequently, however, there is no antecedenthistory of cerebral ' birth injury '- or infection,though in somne.- ofthese- cases haemosiderindeposits in the meninges, found at post-mortemexamination, suggests a ' traumatic ' origin. In

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from

664 POSTGRADUATE MEDICAL JOURNAL November I960

ACE-OF ONSET AND AET!OLOGY182 CASES

40 v~JaknkIoivinkurninhfction

( 3 trwmemulfmorution

0-20

1aJL

O -I -2 -3 -6 9 -12 12+ unknownAGE Of ONSET in MONTHS

FIG. 3.-The age of onset of hydrocephalus in i82cases.

nearly all these cases an occlusive fibrosis in thesubarachnoid space leads to a basal cistern block(Fig. 2). Some cases are complicated by theblock of the exit foramina of the fourth ventriclewhere the fibrosis has involved these orifices.The inflammatory agent. often damages the vul-nerable ependymal lining, and when this occurs inthe narrow aqueduct, the subependymal gliosiswhich results may occlude the lumen, while inother instances blood clot or pus will cause theaqueduct block.Tumours anywhere within the cranium may

.also lead to obstruction of the cerebrospinal fluidpathway.

Fuller accounts of the pathology have beengiven elsewhere. 9, 12

Natural HistorySince hydrocephalus is a condition with a

number of causes the disease is naturally notuniform in character. When due entirely to acongenital malformation, such as aqueduct stenosisor severe Arnold-Chiari malformation in asso-ciation with spina bifida, head enlargement isusually present at birth or becomes apparent very-soon afterwards, whereas if it results from ' birthhaemorrhage' it is often not noticed for four tosix weeks (Fig. 3). When the condition, asso-ciated with spina bifida cystica, is not present atbirth but develops later due to a supexi.nposedinfection, it rarely does so later than the age ofsix months, as by then the spina bifida sac has

usually fully epithelialized and infection becomesexcluded.Many cases progress rapidly with unremitting

head enlargement. Terminally, spasticity, vomit-ing, pressure sores and occasionally blindnessdevelop. Death often occurs before the patient isa year old. The commonest immediate causes ofdeath are cardio-respiratory failure, broncho-pneumonia and intracranial infection, the latterespecially in cases of spina bifida cystica.

Other cases may progress rapidly for a whileand then stabilize, though not before the headhas become too large to be supported properly orbefore additional brain damage has occurred.,Either these cases are placed in an institution ortthe long-suffering families will nurse them with"their severe spasticity, mental defect, oculoblindness and often complete helplessness. Suchtragedies are remembered and they praccount for the general belief that used to bheld by the medical profession, that few patientwith this condition survive, and that those whodo are doomed to permanent institutional care.For example, Ford,2 in his textbook on ' Diseaeof the Nervous System in Infancy, Childhoodand Adolescence,' agrees that in some childrenthe disease seems to become arrested, but addsthat in such cases the child goes through life nmreor less-mentally defective. Holt and MacIntosh,4in their textbook on ' Paediatrics,' state that con-genital hydrocephalus may become arrested atany stage but go on to say that such an outcomeis rare. Clinicians who commonly deal with thisproblem are, however, well aware that naturalarrest is quite frequent. Ingraham and Matson5briefly refer to it, but make no mention of itsincidence, and Jolly6 discusses its 'importance.On the other hand, Bucyl put the spontaneousarrest rate at up to i6o and Macnab"1 reportsI) cases of spontaneous remission amongst hisi6o patients. A much higher arrest rate was firstindicated by Forrest, Laurence and Macnab,3who found that 30% of their 70 patients seen in1954 and 1955 had arrested spontaneously.

In a more recent investigation largely publishedas a preliminary communication,8 the authorexamined this problem further by following up alarge unselected series of hydrocephalic patientsseen over a period of 20 years by Mr. WyllieMcKissock at the Hospital for Sick Children,Great Ormond Street, and the Atkinson-MorleyHospital, Wimbledon. This surgeon felt uncon-vinced of the merits of any form of surgery thenavailable and therefore very rarely operated upon-these cases. The results were wholly unexpected,as of the i82 patients 8i were found to be alivewith the hydrocephalus arrested spontaneously.A. further three were not traced, but the-available-

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from

November I960 LAURENCE: The Natural History of Hydrocephalus 665

TABLE IITHE NATURAL HISTORY OF HYDROCEPHALUS

(i 82 patients)No. of Cases

Alive:Hydrocephalus arrested .. .. I

Untraced (presumed alive) .. *. 3 46.2%Still progressive.. .. .. .. 99

Dead .. .. .. .. .. 89 48.8%

TABLE IIIINTELLIGENCE ASSESSMENTS

(8i patients)No. of Cases

Normal intelligence (I.Q. 85) 33 (4I%)9Below normal intelligence, but 59 (73%)

educable (I.Q. 50-84) .. 26 (32%)JIneducable (I.Q. 50) .. .. 22 (27%)

INTELLIGENCE TESTS81 CASES AVERAGE I.Q. 69.7

--mIucs -504- E.S.Nt--85-Monnqg Imtellme-4-22 Zpta 26$. 33

20.

IS, IQ of normal poplaotion.16

1~.4

12

10 i 30 40 O 60 70 6090 0010 110 120 130 40INTELLIGENCE QUOTENI

FIG. 4.-Distribution of intelligence-assessment resultsin 8i cases (average I.Q. 69.7). The superimposedline represents the distribution of the same numberof test results in a normal population.

evidence suggested that they were still alive withthe disease also arrested; 89 patients had died andin nine, all then under the age of five years, thecondition was still progressing (Table 2). Theintellectual performance of the survivors and theirphysical condition were further investigated asupon these factors depends their future place inthe community. This yielded further surprisingresults.

It was possible to obtain intelligence assess-ments on all the 8i patients and it was found that33 had an I.Q. of 85 or above. A further 26patients were below normal intelligence buteducable and the remaining 22 were ineducable(Table 3). The average intelligence of the wholegroup was just below 70 with a distribution whichwas approximately normal (Fig. 4), but a standard-deviation of about twice t-he normal. The Schonellreading testl3 was used on 2I of the 59 educablechildren who had an average chronological age of

HEAD SIZE. PHYSICAL STATEAND 10.

x120 X

as5 x0

0 0 U0 I00X X X NOfa

01X

0O a a XEAN.

60

g 75 -X----a..........-a- -a--X---------- -v----------

30o INEDICALEIso* 0- * *

1.

0 2 3 4 5 6 7 8 9 IoINCHES

PHYSICAL STATEL X*NrNrml. o.5lightly Hen4icoPed. U'S-nrsly HANicNpp.. **InoapucitatN.

FIG. 5.-The distribution of intelligence assessmentresults when plotted against the head measurement,measured in inches, above the normal for the ageand sex.

I I years 5 months (the remaining 23 children hadnot yet reached school age, and I5 patients werenot tested). The average reading age of the 21children was 9 years io months. This suggeststhat these children were not only educable butwere being educated.

Unfortunately, some of these patients sufferfrom physical disability in addition. The severesthydrocephalus is, however, not always associatedwith the severest physical handicap. Of the sur-vivors, 27 (i.e. one-third) suffered no physicaldisability at all. A further I5 cases did sufferdisability such as slight spasticity, a squint, per-haps slight imbalance, which did not preventthem from getting about unaided; most of themwere not forced to attend special schools onaccount of their physical handicap alone. 22 caseswere severely handicapped and always had toattend special schools or even in some instanceshad to receive home schooling. Only I7 caseswere completely bedridden and were generallyconfined to institutions. Altogether 25 childrenhad normal intelligence, i.e. I.Q. 85 or abovewith negligible or no physical handicap.

Relationships were looked for between intelli-gence and other features. It was found thatthere was no relationship whatsoever betweenintelligence quotient and head size (Fig. 5).Further, in 37 cases where reliable informationon the thickness of the cerebral mantle was avail-able, it was found that this could not be relatedto intelligence (Fig. 6). For example, one childwith a cortical thickness of o.5 cm., estimatedfrom a ventriculogram, had an I.Q. of 85 andanother case had one of ioo (Fig. 7). Indeed,some of the most severely mentally handicappedpatients had a thick but presumably damaged

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from

666 POSTGRADUATE MEDICAL JOURNAL November I960

CORTICAL THICKNESS, I.0.,AND PHYSICALSTATE.

25 37 CASES

x LNOR100 0 U

NOR

U 0

x-- - --D---x - --x----------

75 0x

U I

50 S-X---e------X--°--f---°---------X-X------ ----

0s° INEDUCABLE

2 4X2 3 4 4+

CMSPHYSICLSTATE: X-Nrmd o-Slightl Handicpwd. O-S# iwo"w. .lbuioFUd.

FIG. 6.-The distribution of intelligence assessmentresults when plotted against the thickness of thecerebral'mantle in the 37 patients in whom reliableencephalographs were available.

G. 7.-A ventriculogram of a case of severe hydro-cephalus with a ' cortex' of less than i cm. thick.The I.Q. was 88 and there was no physical dis-ability.

pallium, damage probably produced by the sameagent that led to the hydrocephalus. This suggeststhat the oft-quoted minimum of z cm. of ' cortex 'required for normal intelligence should be revised.A low correlation existed between the I.Q. and

age, the scores tending to rise slightly in theolder children. The reasons for this may betwofold. Firstly, younger children tend to be moredeprived as they will have spent a greater propor-tion of their life in institutions. Secondly, withage these patients may overcome some of theirphysical handicap and will therefi7re be able tohave more intelligence available for measurementby the tests which were employed.

PROGNOSISACCORDING TO SITE OF BLOCK

36 8 78 34 26 N io Uses100

8 no|| |thcw or

2p p E ~non rX

4 i- Ef mb t

tU

20

15 5

Aqueduc 8.ui cisterns Uskrmu'Exit WNo, mwommN3ft

FIG. 8.-Prognosis regarding handicap and survivalaccording to site of block.

There was a close correlation between physicaldisability and intelligence quotient, those whowere severely handicapped tending to score verymuch lower. This was partly caused by thefailure of the tests available to measure theremaining mental ability in the presence of severephysical handicap, as most of the tests require atleast some physical skill. Severely handicappedchildren also have difficulty in communicatingwith others of the same age either because theyare not able to play freely with others, isolated athome or in an institution, or because they maylack one or more of the five senses. Older childrenmay require so much hospital treatment for cor-rection of deformities, that schooling suffers andso they do not receive the same educationalopportunities as their less physically handicappedcontemporaries.The prognosis of the whole group seems to be

that should a hydrocephalic child survive, toarrest spontaneously, for which he has a 46%chance, he then has a 75% chance of beingeducable and a 57% chance of being taught(apart from his physical handicap) in a school fornormal children.

Hydrocephalus has, however, a number ofdifferent causes and each group has a differentprognosis. Unfortunately, the individual aetio-logical groups in this investigation were too smallto draw definite conclusions, but an attempt wasmade to compare them. It appeared that caseswith a basal cistern block or those in which thesite of the block is unknown, being due to eitheran unknown aetiological agent or infection, seemedto have the best outlook (Figs. 8 and 9)3 On the

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from

November I960 LAURENCE: The Natural History of Hydrocephalus 667

ACCORDING TO AETIOLOGY

100 rL SD 42' 4 34 W4 of COs

100

not traced orprogressive

4 61n r o

60

Maldaf lion Infction UnknTrmou Tumour

FIG. 9.-Prognosis regarding handicap and survivalaccording to aetiology.

other hand, cases where the hydrocephalus wasdue to malformation in the aqueduct or in thehind brain (Arnold-Chiari malformation) seemedto carry the worst prognosis.

Unfortunately, some cases after spontaneousarrest has taken place may, months or even yearslater, become progressive once more. In otherinstances an acute episode of intracranial hyper-tension may kill the patient within a very shorttime. Such events may be precipitated by minorhead injury or some infection. In other instancesno precipitating cause can be discovered. Thesepatients then, although the progress of the hydro-cephalus may have arrested, are never quite freefrom danger.

ConclusionsAlthough some of the mental and physical

handicap of hydrocephalic patients is due to'cortical ' thinning and distortion of nerve tracks,much of the deficit is due to associated mal-formations and to parenchymal brain damagecaused by the lesion that originally led to thehydrocephalus, such as infection and perinatalanoxia. The great importance of the inter-relationship of the physical and mental abilitycannot be stressed too much, for a great deal ofthe apparent mental handicap in these patientsis caused by associated severe physical disability,some of which may be correctable by carefultraining and education and by surgery.The series upon which the investigation into the

natural history of hydrocephalus was based wasnot entirely representative of hydrocephalus inthe general population as many of the mostseverelv and the mildly affected cases neverreached this particular large children's hospital.Certain general conclusions can nevertheless be

drawn from it. A much more hopeful prognosisis suggested, as an appreciable proportion ofthese children, whose outlook used to be regardedas completely hopeless without surgical treatment,are now growing up into useful citizens capableof earning their own livelihood.

Care must be exercised in recommendingpatients for surgery, especially those in whomthe condition is at the point of becoming arrestedand who may do well without it. It must bestressed, however, that in cases of progressivehydrocephalus much additional brain damage maybe caused by waiting for spontaneous arrest tooccur, especially when this may never take place.Even in a very slowly progressive disease, earlyand effective surgical- intervention may possiblyleave the patients with less brain damage thanwhen the disease is allowed to run its naturalcourse, even though the best surgery is not with-out its serious risks in this condition. Furtherwork investigating this particular point, using aproperly controlled series, is very urgently needed.

SummaryThe pathology and natural history of hydro-

cephalus is discussed. A very much more hopefulprognosis is suggested by a 46% survival and a73% educability rate of the survivors.

AcknowledgmentsI wish to thank Mr. T. J. Cooke, Mr. Derek

Martin and Mr. L. Williams for illustrations, andthe editors of the Lancet and the Cerebral PalsyBulletin for permission to use Fig. 4, and Figs. 5and 6 respectively. Thanks are also due to theresearch committee of the Hospital for SickChildren which supported most of the work uponwhich this paper was based.

REFERENCESI. BUCY, P. C. (Ig5o), in 'Brennemann's Practice of Pediatrics,'

chap. 3. Prior: Hagerstown.2. FORD, F. R. (I952), 'Diseases of the Nervous System in

Infancy, Childhood and Adolescence,' p. 255. Charles CThomas: Springfield, Ill.

3. FORREST, D. M., LAURENCE, K. M., and MACNAB,G. H. (I957), Lancet, i, 1274.

4. HOLT, L. E., and McINTOSH, R. (I953), 'Pediatrics,'p. 1053. Edward Appleton: New York.

5. INGRAHAM, F. D., and MATSON, D. D. (954), 'Neuro-surgery of Infancy and Childhood,' p. I17. Charles CThomas: Springfield, Ill.

6. JOLLY, H. (Igss), Proc. roy. Soc. Med., 48, 843.7. KAHN, E. A., and LUROS, J. T. (1952), J. Neurosurg., 9, 59.8. LAURENCE, K. M. (I958), Lancet, ii, II 2.9. LAURENCE, K. M. (I959), Ann. roy. Coll. Surg. Ejzgl., 24, 388.

Io. MACNAB, G. H. (I955), Proc. roy. Soc. Med. 48, 846.I I. MACNAB, G. H. (I959), J. Neurol. Neurosurg. Psychiat., 22, 84.I2. RUSSEL, D. S. (1949), 'Observations on the Pathology of

Hydrocephalus,' Special Report, Medical Research Council,Ser. 265. London: H.M.S.O.

13. SCHONELL, F. J., and SCHONELL, F. E. (1952), 'Diagnosticand Attainment Testing,' 2nd edition. Edinburgh: Oliverand Boyd.

14. SPITZ, E. B. (1959), Pediat. Clin. N. Amer., 6, 1215.s5. ZACHARY, R. B., and LAUBER, J. (I960), Personal com-

munication.

copyright. on 1 July 2019 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.421.662 on 1 Novem

ber 1960. Dow

nloaded from