experimental hydrocephalus after haemorrhage · h. a. wilkinson", rene b. wilson, p. p. patel,...

Journal of Neurology, Neurosurgery, and Psychiatry, 1974, 37, 224-229

Corticosteroid therapy of experimental hydrocephalusafter intraventricular-subarachnoid haemorrhage

H. A. WILKINSON", RENE B. WILSON, P. P. PATEL, AND M. ESMAILI

From Beth Israel Hospital and Boston City Hospital, Boston, Massachusetts, U.S.A.

SYNOPSIS Symptomatic hydrocephalus after subarachnoid haemorrhage seems to result bothfrom mechanical obstruction of arachnoid villi and basilar cisterns and from an inflammatorycellular reaction in the villi. Subarachnoid haemorrhage was induced in rabbits using whole bloodinjected through an implanted intraventricular needle. Control rabbits receiving intraventricularmethyl prednisolone acetate but no blood, developed ventricular dilatation significantly more oftenthan untreated controls. Eighty-three per cent of rabbits with untreated experimental subarachnoidhaemorrhage developed moderate to severe hydrocephalus. Intramuscular steroid therapy signifi-cantly reduced the incidence of hydrocephalus.

In 1965 Hakim and Adams (also Adams et al.,1965) called attention to the surgically treatablenature of the syndrome of symptomatic hydro-cephalus with normal cerebrospinal fluid (CSF)pressure. This type of acquired hydrocephalus,usually of the communicating type, has beendocumented in association with a variety ofaetiologies (Messert et al., 1966; Wilkinson et al.,1966; Ojemann et al., 1969), but the majority ofcases are encountered after subarachnoid haemo-rrhage (SAH). The resulting hydrocephalususually causes a more or less uniform clinicalsyndrome characterized chiefly by the pro-gressive evolution of akinetic dementia (Messertet al., 1966).An excess of CSF accumulates after sub-

arachnoid haemorrhage chiefly because of anobstruction by erythrocytes of the absorptivepathways, especially the arachnoid villi (Shaboand Maxwell, 1968b; Ellington and Margolis,1969). CSF enters the villi from the subarachnoidspace and passes through the membrane (Shaboand Maxwell, 1968a) or through a labyrinth oftubular channels within the villus (Welch andFriedman, 1960) into the blood stream. Erythro-cytes contained in the CSF block the absorptivesurface or channels of the villi mechanically

1 Reprint requests to Dr. Harold A. Wilkinson, Beth Israel Hospital,330 Brookline Avenue, Boston, Mass. 02215, U.S.A.

224

(Ellington and Margolis, 1969) and by initiatinga leucocytic or phagocytic inflammatory response(Shabo and Maxwell, 1968b; Alksne and Lovings,1972).The present study was designed to explore the

possibility of preventing the evolution of signifi-cant hydrocephalus after SAH by attempting tomodify the induced inflammatory response in thearachnoid villi by adrenocorticosteroid therapy.

'\NIo''X'

FIG. 1. Schematic drawing ofcoronal section throughrabbit's cranium showing implanted intraventricularneedle attached to repuncturable, sealed nipple.

guest. Protected by copyright.

on January 28, 2021 byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.37.2.224 on 1 F

ebruary 1974. Dow

nloaded from

http://jnnp.bmj.com/

Experimental hydrocephalus after intraventricular-subarachnoid haemorrhage

METHODS

One hundred and thirty-nine 5 kg rabbits wereaseptically operated under droperidol and fentanyl(Innovar) or thiopentone anaesthesia. A no. 21gauge needle was introduced into the lateral cerebralventricle through a small drill hole placed 2 mmrostral to the coronal suture and 2 mm lateral to themidline, entering the ventricle at a depth of 5 or 6mm. The needle was fixed to the skull with stainless

steel wires and methylmethacrylate plastic and wasattached through a short length of tubing to arepuncturable, sealed nipple (Fig. 1). Blood injectedinto the ventricle through this needle passed into thesubarachnoid space in all animals, as seen at necropsyand as confirmed by intraventricular injection ofindigo carmine when the animals were killed.

All animals were allowed a 7 to 10 day post-operative recovery period. Four animals which dis-played persistent neurological deficits during this

FIG. 2. Photographs-of repre-sentative coronal sectionsthrough frontal horns andbodies of lateral ventriclesshowing no ventricular enlarge-ment (A) and (B), mild ven-tricular enlargement (C) and(D), and moderate/severeventricular enlargement (E)and (F).

225



j.com/

J Neurol N


ebruary 1974. Dow

nloaded from


H. A. Wilkinson, Rene B. Wilson, P. P. Patel, and M. Esmaili

period were excluded from the study. All animalswere killed at the end of an additional two weekperiod and their carcasses were frozen. The entirefrozen head was sectioned with a fine-toothed sawand ventricular size was observed and photographedfor later measurement. The degree of ventricularenlargement was determined from the average of aseries of measurements of lateral ventricular widthtaken from sections through frontal horns andthrough the body of the ventricles; measurements ofthe width of the occipital horns were not includedsince this portion of the ventricular system frequentlyand variably displayed an enlargement dispropor-tionately greater than that seen in the remainder ofthe ventricle. An average width of less than 05 mmwas taken to represent 'normal' ventricles. A widthof 05 to 1 2 mm was considered 'mild' enlargement,a width of 1-3 to 2-0 mm was considered 'moderate'enlargement, and a width over 2-0 mm was termed'severe' enlargement (Fig. 2). Eight animals withimproperly placed needles, infection, or haematomaswere excluded from the study.

TABLE 1SEVERITY OF HYDROCEPHALUS IN 'OPERATIVE' (OR UN-TREATED) AND 'THERAPY' CONTROL RABBITS-ALL WITH

NO INJECTIONS OF BLOOD

Severity of Therapyhydrocephalus

None Intraventricular Intramuscular

(no.) (°/%) (no.) (Y.) (no.) (Y.)

None 9 82 5 29 7 64Mild 2 18 6 35 3 27Moderate/Severe 0 0 6 35 1 9

Totals 11 17 11

Statistical significanceof difference(therapy v. no therapy) P < 0-001 P= 0 7

Twelve rabbits died of anaesthetic complicationsbefore, during, or immediately after surgery. Theirbrains were examined as normal anatomical con-

trols. Fifty-one animals were used in a preliminarystudy to determine the optimum dose of intra-ventricular blood necessary to produce hydro-cephalus, while still being compatible with a reason-

able two week survival. Twelve of these rabbits plusan additional 72 rabbits constituted the main studygroup.Within the main study group 39 animals served as

'.operative' and 'therapy' controls, receiving no

intraventricular blood. Eleven of these animals wereoperated on but received no therapy and either nointraventricular injections or injections of 0 45 ml.Ringer lactate only. Seventeen animals receivedintraventricular therapy of four alternate day injec-tions of methylprednisolone acetate (Depomedrol,Upjohn Co.) 10 mg (0-25 ml.) plus 045 ml. Ringerlactate intraventricularly. Eleven animals receivedintramuscular therapy of four alternate day injec-tions of methylprednisolone 10 mg, and intra-ventricular injections of 0-45 ml. Ringer lactate.The remaining 40 rabbits in the main study group

all received four alternate day injections intra-ventricularly of 0-25 ml. lightly heparinized wholeblood obtained from the rabbits' own ear veins,plus 0 25 ml. Ringer lactate. Twelve of these animalsreceived no therapy, 15 received intraventricularmethylprednisolone therapy, and 13 received intra-muscular methylprednisolone therapy on the sameschedule as detailed above.

RESULTS

In the preliminary study of optimum dosage ofintraventricular blood, 23 rabbits received dailyintraventricular injections of blood in amountsof 0i25 to 0 50 ml. This caused severe hydro-cephalus in 18 (78%)0 but 17 (740%) failed to sur-vive the two week test period. Seven rabbitsreceived alternate day injections of 0 40 to 0 50ml. blood; all of these developed moderate tosevere hydrocephalus, but none survived the fulltest period. Twelve rabbits received alternate dayinjections of 0-25 ml. blood2. Ten of thesedeveloped moderate to severe hydrocephalus,and 86% survived the full test period. A dosageof 0 25 ml. blood every other day for four injec-tions was adopted as a standard for the mainstudy.The results of the main study are presented in

Tables 1 and 2. All ofthe control rabbits receivingno blood and no therapy displayed normal (82%)or mildly enlarged ventricles (180%) according tothe measurement criteria used in the study-despite the presence of the intraventricularneedle. None of these rabbits developed moder-ate or severe hydrocephalus. Only one controlrabbit receiving intramuscular methylpredni-solone therapy developed moderate ventricularenlargement, but moderate or severe enlarge-ment developed in six (350%) of 17 rabbitsreceiving intraventricular methylprednisolone2 These rabbits served as untreated SAH controls for the main study.

226



j.com/

J Neurol N


ebruary 1974. Dow

nloaded from



TABLE 2SEVERITY OF HYDROCEPHALUS AFTER REPEATED INTRA-VENTRICULAR INJECTIONS OF BLOOD-WITH AND WITHOUT

THERAPY

Severity of Therapyhydrocephalas

None Intraventricular Intramuscular

(no.) (54) (no.) (Y4) (no.) (Y4)

None 1 8 3 20 5 38Mild 1 8 4 27 4 31Moderate/Severe 10 83 8 53 4 31

Totals 12 15 13

Statistical significanceof difference(therapy v. no therapy) P= 03 P = 003

TABLE 3INCIDENCE OF PREMATURE MORTALITY IN RABBITS RELATED

TO SEVERITY OF HYDROCEPHALUS

Severity of hydrocephalus

None Mild Moderate/Severe

Incidence ofpremature mortality (54) 11 40 38

therapy, a statistically significant difference fromthe untreated controls. Of the animals receivingintraventricular injections of blood but notherapy, 10 out of 12 (83%) developed moderateto severe hydrocephalus. With intraventricularmethylprednisolone therapy this was reduced to53%, though the degree of reduction did notachieve statistical significance. With intramuscu-lar methylprednisolone therapy this was reducedstill further to 31%, a reduction which is statis-tically significant. Although mortality was notthe major end-point in this study, a definitely in-creased mortality was seen in those rabbits whichdid develop hydrocephalus (Table 3).

DISCUSSION

Symptomatic hydrocephalus with akinetic de-mentia is an unfortunate, but not uncommon,complication after subarachnoid haemorrhageof differing aetiologies. The hydrocephalus may

evolve rapidly or insidiously and poses a severeadditional burden in this patient populationwhich is already at high risk, frequently necessita-ting operative, mechanical drainage of CSFthrough implanted shunts. Hydrocephalus hasbeen reported to occur in man in approximately3500 of SAH due to ruptured aneurysms, withapproximately 20% suffering neurological defi-cits (Galera et al., 1970; Krayenbuhl et al., 1972).In these rabbit experiments the incidence ofhydrocephalus was 830% in the untreated group.This suggests that with repeated SAH the inci-dence of hydrocephalus increases.The exact pathophysiological mechanism in-

volved in this chronic and dementing form ofhydrocephalus (often with 'low' CSF pressure)is unclear. Hakim and Adams (1965) formulateda 'hydraulic press' hypothesis emphasizing theincreased force exerted by a constant pressureover a larger area-the walls ofa dilated ventricu-lar system. Geschwind (1968) rejected this hypo-thesis as being oversimplified, and stressed a needfor consideration of the structural properties ofthe ventricular walls since, pathologically, dilata-tion of the ventricles is accompanied by a loss orcompression of periventricular white matter(Yakovlev, 1947). Others have speculated on theincreased absorptive surface area of dilatedventricles, allowing a return to normal CSFpressure as long as the ventricular surface arearemains expanded. Presumably the accumulationof an excess of CSF initiates the pathologicalprocess, with acute distention of the cerebralventricles. Once so disturbed, it seems then to bepossible for a lower-or normal-CSF pressureto maintain this abnormal state.

Clotted blood, or red blood cells and cellulardebris, in the basal cisterns or at the arachnoidvilli obstructs these channels of CSF resorptionand results in excessive accumulation of CSFafter SAH, leading to the development of hydro-cephalus. It is frequently observed clinically thatblood in the CSF elicits a polymorphonuclearinflammatory response in the CSF, meningism,and fever. A focal inflammatory response toSAH has also been documented in the arachnoidvilli (Shabo and Maxwell, 1969; Alksne andLovings, 1972). It has been suggested that thisinflammatory response to the presence of bloodin CSF also may be an important part of themechanism of obstruction in the acute phase of

227



j.com/

J Neurol N


ebruary 1974. Dow

nloaded from


H. A. Wilkinson, Rene B. Wilson, P. P. Patel, and M. Esmaili

post-SAH hydrocephalus. Fibrous adhesionscausing permanent obstruction develop late inthe subarachnoid space, with significant scar

production requiring more than four weeks todevelop (Schemm et al., 1968). The initiatingmechanism which causes the acute ventriculardilatation after SAH seems to depend on a com-

bination of mechanical obstruction by red bloodcells and debris, plus an induced inflammatoryresponse.The adrenocorticosteroids are frequently em-

ployed in clinical medicine for their anti-inflammatory effect. One long-acting steroidpreparation, methylprednisolone acetate (De-pomedrol, Upjohn Co.) has been used intra-thecally to provide a high CSF concentration innon-infectious inflammatory diseases of thearachnoid membrane-adhesive arachnoiditis,nerve root fibrosis, etc. (Smith and Ross, 1959;Sehgal et al., 1962; Mack, 1964; Kulick, 1965).Since the arachnoid villi are actually extensionsof the arachnoid membrane, the treatment insimilar fashion of an inflammatory disease of thevilli caused by blood in the CSF would seem tobe a logical adjunctive measure in an attempt toprevent symptomatic hydrocephalus after SAH.

In this study, adrenocorticosteroids were

shown to have a definite effect in reducing theincidence of hydrocephalus after subarachnoidhaemorrhage. Direct intraventricular injectionof methylprednisolone was less effective thanintramuscular methylprednisolone, and, in fact,caused a significant incidence of hydrocephalusin rabbits receiving no injections of blood. Thelack of significant complications in reports ofhuman intrathecal methylprednisolone therapysuggests that this might represent a speciessensitivity, though it may represent simply a

mechanical interference with CSF resorption bythis relatively insoluble depository adrenocortico-steroid. Kulick (1965) did report occasional post-injection CSF pleocytosis in man, and Sehgal etal. (1962) advised against intracisternal injectionbecause of severe nausea and vomiting. Byinjecting the methyl prednisolone intramuscu-larly this hydrocephalus-inducing effect was

apparently avoided and a significant (P=0O03)reduction in ventricular dilatation was effected.The effectiveness of adrenocorticosteroids in

lessening the incidence of significant hydro-cephalus after subarachnoid haemorrhage in

this study relates presumably to their known anti-inflammatory effect. Their possible effect onlong-term evolution of fibrous scarring of sub-arachnoid pathways or arachnoid villi remainsconjectural (and will form the basis for con-tinued studies). Clinically significant post-SAHhydrocephalus in humans seems to evolve onlyafter an initial episode of acute ventriculardilatation of limited duration, presumably due toboth the mechanical and the inflammatoryeffects of subarachnoid blood. If this acute phasecan be eliminated or modified by adrenocortico-steroids, it seems reasonable to hope that areduction in long-term, clinically significantpost-SAH hydrocephalus with akinetic dementiamay also be realized.

REFERENCES

Adams, R. D., Fisher, C. M., Hakim, S., Ojemann, R. G.,and Sweet, W. H. (1965). Symptomatic occult hydro-cephalus with 'normal' cerebrospinal-fluid pressure. Atreatable syndrome. New England Journal of Medicine, 273,117-126.

Alksne, J. F., and Lovings, E. T. (1972). The role of thearachnoid villus in the removal of red blood cells from thesubarachnoid space. An electron microscope study inthe dog. Journal of Neurosurgery, 36, 192-200.

Ellington, E., and Margolis, G. (1969). Block of arachnoidvillus by subarachnoid hemorrhage. Journal of Neuro-surgery, 30, 651-657.

Galera, R., and Greitz, T. (1970). Hydrocephalus in theadult secondary to the rupture of intracranial arterialaneurysms. Journal of Neurosurgery, 32, 634-641.

Geschwind, N. (1968). The mechanism of normal pressurehydrocephalus. Journal of Neurological Sciences, 7, 481-493.

Hakim, S., and Adams, R. D. (1965). The special clinicalproblem of symptomatic hydrocephalus with normalcerebrospinal fluid pressure. Journal of NeurologicalSciences, 2, 307-327.

Krayenbuhl, H. A., Ya$argil, M. G., Flamm, E. S., and Tew,J. M., Jr. (1972). Microsurgical treatment of intracranialsaccular aneurysms. Journal of Neurosurgery, 37, 678-686.

Kulick, S. A. (1965). The clinical use of intrathecal methyl-prednisolone acetate following lumbar puncture. Journalof the Mount Sinai Hospital, 32, 75-78.

Mack, E. W. (1964). Intrathecal steroid administration.Rocky Mountain Medical Journal, 61, no. 8, 33-34.

Messert, B., Henke, T. K., and Langheim, W. (1966).Syndrome of akinetic mutism associated with obstructivehydrocephalus. Neurology (Minneap.), 16, 635-649.

Ojemann, R. G., Fisher, C. M., Adams, R. D., Sweet, W. H.,and New, P. F. J. (1969). Further experience with thesyndrome of 'normal' pressure hydrocephalus. Jouirnal ofNeurosurgery, 31, 279-294.

Schemm, C. W., Bentley, J. P., and Doerfier, M. (1968).Wound healing in the subarachnoid space. Neurology(Minneap.), 18, 862-869.

228



j.com/

J Neurol N


ebruary 1974. Dow

nloaded from



Sehgal, A. D., Gardner, W. J., and Dohn, D. F. (1962).Pantopaque 'arachnoiditis'. Treatment with subarachnoidinjections of corticosteroids. Cleveland Clinic Quarterly,29, 177-188.

Shabo, A. L., and Maxwell, D. S. (1968a). The morphologyof the arachnoid villi: a light and electron microscopicstudy in the monkey. Journal of Neurosurgery, 29, 451-463.

Shabo, A. L., and Maxwell, D. S. (1968b). Electron micro-scopic observations on the fate of particulate matter in thecerebrospinal fluid. Joutrnal of Neutrosuirgery, 29, 464-474.

Smith, J. K., and Ross, L. (1959). Steroid suppression ofmeningeal inflammation caused by Pantopaque. Neutrology(Minneap.), 9, 48-52.

Welch, K., and Friedman, V. (1960). The cerebrospinal fluidvalves. Brain, 83, 454-469.

Wilkinson, H. A., LeMay, M., and Drew, J. H. (1966). Adultaqueductal stenosis. Archives of Neiurology, 15, 643-648.

Yakovlev, P. 1. (1947). Paraplegias of hydrocephalics. (Aclinical note and interpretation.) American Joutrnal of

Mental Deficiency, 51, 561-576.

229



j.com/

J Neurol N


ebruary 1974. Dow

nloaded from


experimental hydrocephalus after haemorrhage · h. a. wilkinson", rene b. wilson, p. p. patel,...

Documents