Journal ofNeurology, Neurosurgery, and Psychiatry 1989;52:923-930

Proceedings of the 113th meeting of the Society ofBritish Neurological Surgeons and the joint meetingwith the British Society of Neuroradiologists, Glasgow22-24 September 1988COMA SCALING AND SCORING: USES AND been recruited since 1981 who had not hadLIMITATIONS any seizures pre-operatively and who hadTeasdale GM (Glasgow) been followed up for a mean time of 31

months, with a minimum follow up of 24The author reviewed the development and months in survivors. No therapy provedapplication of the Glasgow Coma Scale effective in preventing or changing thewhich was first presented to the Society in natural history of post-operative seizures1974 and was aimed at achieving a simple and 11% of those in the active treatmentstandardised approach suitable for a wide groups experienced acute drug reactions.range of circumstances. The scoring systemderived from the scale-Glasgow Coma ReferencesScore-was found to be of value mainly forsummarising data of groups of patients 1 Foy PM, Copeland GP, Shaw MDM. Actaprovided appropriate statistical circumspec- Neurochir 1981;55:253-64.tion was used. Its use to determine risk of 2 Foy PM, Copeland GP, Shaw MDM. Actation wntracraaslu. I semtomaindetemineurik o Neurochir 1981 ;57:15-22.an itracranial haematoma in head injure 3 Shaw MDM, Foy PM, Chadwick DW. Jadults and children and prognosis after Neurol Neurosurg Psychiatry 1983;46:368.head injury was illustrated with data from 4 Shaw MDM, Foy PM, Chadwick DW. Actathe collaborative study between Glasgow, Neurochir 1983;69:253-8.Edinburgh, Liverpool and Southampton.Finally, the Glasgow Coma Score had beensuccessfully applied to the assessment of POST CRANIOTOMY HAEMATOMA RECURRENCEpatients with subarachnoid haemorrhage AMFER HEAD INJURY-AN AUDITand was the main component of the World Bullock R, Rehman S, Hannemann CO,Federation of Neurosurgical Societies scale. Murray L, Teasdale GM (Glasgow)

Craniotomy re-opening is required in 8-DO PROPHYLACTIC ANTICONVULSANT DRUGS 19% of patients who undergo removal ofALTER THE PATTERN OF SEIZURES FOLLOWING post traumatic mass lesions, and a secondCRANIOTOMY? haematoma may mar recovery by causingFoy PM, Chadwick DW, Johnson AL, ischaemic cerebral damage. Despite this,Shaw MDM (Liverpool) factors responsible for post craniotomy

haematomas (PCH) are poorly understood.Supratentorial craniotomy for non- 59 patients who underwent removal of atraumatic conditions is associated with a PCH, over an 8 year period in Glasgow17% overall incidence of seizures.' Ninety- were compared to 762 patients who did nottwo percent of those having seizures develop a PCH, after removal of aexperienced the first fit within 2 years of traumatic mass. (table).surgery.' Six high risk groups have been Results: The incidence of PCH was 7%. Inidentified'2 and the relative ineffectiveness 75% of patients with PCH the initialof anticonvulsant treatment in these high haematoma was intradural, but the secondrisk groups during the first six postoperative haematoma was extradural in 69%,months demonstrated.3' However, the suggesting that the potential space under-original study used historical controls and lying a large craniotomy bone flap may be athe present randomised study was expanded factor leading to the development of theto include a control group in addition to PCH. In 39 patients, ICP monitoring wasthose treated with either phenytoin or car- used after the first craniotomy, and in 22bamazepine for 6 or 24 months. Two hun- (56%) an ICP rise prompted diagnosis ofdred and seventy six high risk patients had the PCH. In 14 patients, ICP failed to rise,

923

despite clinical deterioration. PCHs werediagnosed significantly earlier when ICPmonitoring was successful (21, SD 20 hoursdelay) than when the diagnosis was clinical.(51, SD 47 hours). Sixty-one percent ofPCH patients were severely disabled ordead at follow up, in comparison to 45% ofthose who underwent a single craniotomy.

CLINICAL COMPARISON OF SUBDURALPRESSURE MEASUREMENTS USING A CATHETERTIPPED CAMINO TRANSDUCER WITHVENTRICULAR PRESSURE MEASUREMENTSSinar EJ, Chambers IR, Mendelow AD,Modha P (Newcastle)

Fluid filled ventricular catheters give themost reliable recording of intracranial pres-sure (ICP) but have disadvantages includinginfection, haemorrhage and epilepsy. Sub-dural pressure screws and fluid filled cath-eters have proved inaccurate'2 but a newcatheter tipped transducer (Camino) hasbeen developed for measuring ICP from thesubdural space and the authors had com-pared this new transducer with recordingsfrom a ventricular catheter in 10 headinjured patients. ICP was monitored for 4-7to 77-7 hours. The correlation coefficientbetween the two techniques ranged from0-39 to 0-98, the gradient ranged between0-41 to 1-06 and the intercept ranged from-4 4 to 9 4. Overall the Camino pressurewas recorded as 5 5 mmHg lower than theventricular pressure with a mean correlationcoefficient of 0-95 and a gradient of 1-04.Ninety-nine percent of readings were within16 mmHg of each other. Camino trans-ducers placed within the ventricle correlatedless well than a subdural transducer withventricular fluid pressure recorded througha standard catheter.

References

I Mendelow AD, Rowan JO, Murray L, KerrAE. J Neurosurg 1983;57:45-50.

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

Proceedings of the 113th meeting of the Society of British Neurological SurgeonsTable I Age, sex, conscious level, mannitol administration and outcome in patients with and without a post-craniotomy haematoma

Outcomet

Mean Good Badage Mannitoll(years) Sex Patients in Number* given before Mild/(mean coma on with evidence first moderate Severe,SD) M F admission ofalcohol operation GR disability disability Dead/ Veg

Post-craniotomy 47 (17) 45 14 40 (68%) 32 (54%) 31(52%) 17 6 12 24haematoma n = 59 23 (39%) 36 (61%)

Nopostcraniotomy 41 636 126 411 + + 225+ (34%) 279(36%) 284 129 111 239haematoman = 762 (61%) 413 (55%) 350 (45%)

+ No data for 85 cases. + + No data for 102 cases.lp < 0 01- Chi squared testtp < 001 - Chi squared test Comparison between those with and without post craniotomy haematomas.*p < OOS- Chi squared test J

2 Miller JD, Bodo H, Kapp JP. Inaccurate pres- more rapidly if the rate of blood flowsure readings from subarachnoid bolts. through it rises' but heat clearance as aNeurosurgery 1986;19:253-5. method for assessing cerebral blood flow fell

into disrepute because of drift, sensitivity to

INTRASELLAR PRESSURE AND PITUITARY the proximity of large vessels and its inherentTUMOURS: THM ORIGIN, invasiveness. The Flowtronics ThermalCOMPARTMENTALISATION AND CORRELATION Diffusion Monitor' proved in laboratoryWITH ENDOCRINOLOGY AND RADIOLOGY studies on cadaveric brain to be free ofLees PD,* Fahlbusch R,t Zrinzo A,t significant drift and gave values close toLovick AHJ,* Pickard JD* (Southamp- "zero flow". The authors had used theton,* and Erlangen, West Germanyt) technique in patients perioperatively with

aneurysms, tumours and head injuries. WithThe blood supply to the adenohypophysis is hypercapnia, there was a brisk increase inportal venous and not arterial and hence recorded flow within 20 seconds. In aadenohypophysial perfusion may be com- preliminary series, patients who developedpromised by intrasellar pressure (ISP) with ischaemic deficits postoperatively had bloodimpairment of hypothalamic control of flows recorded of less than 20 ml/l00 gm/minanterior pituitary function.' One hundred for periods from 3-30 minutes. The tech-and eighteen patients undergoing trans- nique might play a useful role in the man-phenoidal hypophysectomy of a pituitary agement ofunconscious acutely ill patients intumour were studied either at the Wessex the neurosurgical unit.Neurological Centre (59) or at theNeurosurgical Clinic of the University of ReferencesErlangen-Nurnberg (59). Intra-operative I Gibbs FA. A thermoelectric blood flow recor-ISP was measured before bony decompres- der in the form of a needle. Proc Soc Expriotsion of the pituitary fossa. Pulsatile Med 1933;31:141-6.intrasellar pressure was recorded in 91% Of 2 Carter LP, Erspamer R, Bro W. Cortical bloodpatients with a mean intrasellar pressure flow: thermal diffusion vs isotope clearance.(108 patients) of 20, SEM I (range 3-51 Stroke 1981;12:513-8.mmHg) and mean pulse pressure of 3 mmHg SEM I (range 1-12 mmHg). ISPsometimes varied at different measuring OMENTAL TRANSPOSITION IN CHRONIC SPINALsites within the same patient ("compart- CORD INJURYmentalisation"). G Neil-Dwyer (Southampton), H Goldsmith

(Boston), L Barsoun (London)Referenee

The omentum has been used over the yearsI Lees PD, Pickard JD. J Neurosurg 1987; for a variety of purposes. In the past it was

67:192-6. shown that placing the omentum on thebrain and spinal cord led to an extensive

CEREBRAL BLOOD FLOW BY THERMAL development of vascular connections at theDIFFUSION-REVISITED omental/CNS interface. Recent publicationsChoksey MJ, Crockard HA (London) have shown that placement of the omentum

on the injured spinal cord of cats canAny perfused tissue conducts heat away produce an improvement in neuroelectrical

activity as well as an improvement of spinalcord blood flow. The omentum containsboth an angiogenic factor, which is responsi-ble for the neovascularization after applica-tion of the omentum to the spinal cord, and avariety of neuro-transmitters and gastricpolypeptides, which may play a part inneuro-transmission. Four years ago fivetetraplegic and five paraplegic patientsentered a feasibility study to see if omentaltransposition to their spinal cord mightresult in clinical benefit. All the patients hadwell defined cord lesions. The length of timefrom their injury to operation was 2-15years. All patients had little if any motor andsensory function below the level of the cordlesion. The operation required surgical leng-thening of the pedicled omentum followed byits placement in a subcutaneous tunnel. Anextensive laminectomy was then performedfollowed by wide opening of the dura. In allcircumstances the cord showed changes con-sistent with previous trauma. The omentumwas laid directly on the underlying spinalcord. Two of the tetraplegic patients and oneof the paraplegic patients have shownmarked improvement over the years relatedmore to amelioration of lower motor neurondysfunction than to improvement in cordfunction itself.

A STUDY OF UNFUSED CLOWARD'SPROCEDURESJellinek DA, Marks JC,* Illingworth RD,*Rice-Edwards JM,* Shawdon Ht (Depart-ments of Neurosurgery* and Radiology,tCharing Cross Hospital, London)

A retrospective study of 35 patients who hadundergone anterior cervical decompressionfor brachalgia or cervical cord compressionwith the use of the Cloward drill, but withoutbone grafting, was undertaken to determine

924

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

Proceedings of the 113th meeting of the Society of British Neurological Surgeonswhether this had any deleterious effects onoperative results. Patients in the study weresent a questionnaire and asked to assess thesuccess of the operation in relieving theirpresenting symptomatology and whetherthey had any recurrence of their preoperativecomplaints. Cervical radiology was perfor-med at the time of the survey to assess bonyfusion and the presence of any subluxation atthe operation site in 14 of 16 patients hadradiological evidence of bony fusion. Onlyone patient had (minor) subluxation. Bytheir own assessment, 42% (15) consideredtheir operative result as "excellent", 21% (7)as "good", 15% (5) as "satisfactory", 15%(5) as "poor" and 7% (3) as "dreadful".Eighty six percent (30) reported some symp-tomatic recurrence, but of these only 32%described this as "moderate" or "severe".These results compare favourably with thosereported'2 for fused Cloward's proceduresuggesting satisfactory results may beobtained without autologous bone graftingthereby reducing the discomfort of theprocedure for the patient.

References

1 Jeffreys RV. J Neurol Neurosurg Psychiatry1986;49:353-61.

2 Phillips DG. J Neurol Neurosurg Psychiatry1973;36:879-84.

CAUSES OF FAILURE IN LUMBARMICRODISCECTOMYFindlay GFG (Liverpool)

Lumbar microdiscectomy has become anestablished part of spinal surgical practiceand facilitates early mobilisation, but overallresults in the long term are comparable withthe more conventional technique. In a seriesof 324 microdiscectomies over 5 years (agerange 15-78 years; length of history 1-60months; follow up 2-60 months; 140 L 4/5and 184 lumbosacral discs), 88-3% made agood or excellent recovery and 4% a poorrecovery and there were 25 failures (persis-tence or worsening of the original symptomsin 12 patients and recurrence of symptomsafter initial improvement in 13 patients).Seventeen patients were re-explored with themost common finding being root canal sten-osis in 11 patients (other causes-recurrentdisc, a new disc at a new level, extraduralfibrosis, unknown). Following re-explora-tion, 14 patients improved. Recurrent discherniation is rare following microdiscectomyand the commonest cause of failure is thedevelopment or non-recognition of bonynerve root entrapment.

HYPERTERMA FOR MALIGNANT BRAIN serum albumin at 48-72 hours intervals forTUMOURS 14 days. Normal plasma volumes wereCetas TC, Hynynen K, Iacono RP, Shetter measured in outpatients six months later orA, Cassady JR, Roassman K, Shimm D, predicted from an early measurement ofStea B, Guthkelch AN (Arizona USA) total body water. Nine patients (36%) were

found to be hypovolaemic (> 10% fall inDetails were presented of a Phase 1 trial of predicted normal plasma volume) within 96hyperthermia with radiation therapy for hours of the haemorrhage (mean plasmamalignant hemisphere tumours (glioblas- volume -18 + 2%) compared with 16tomas), aiming for intratumoural tem- patients (64%) who were normovolaemic orperatures of >42-5°C. Two hyperthermia hypervolaemic (mean plasma volumemodalities had been used: ferromagnetic +9 ± 2%, p < 001). The hypovolaemia,induction (FMI; four patients) and scanned, which was not clinically evident, persisted forfocussed ultrasound (FUS; eight patients). an average of 6 days. Compression orIn the FMI protocol, self-regulating obliteration of the basal cisterns wasferromagnetic seeds were inserted into cath- observed on the admission computedeters stereotactically implanted into the tomographic (CT) scans of all hypovolaemictumour at a spacing of 1-1-3 cm. These were patients compared with 12-5% of normo-heated within a high-frequency magnetic volaemic patients (Chi-square 14-52, p <field for 60 minutes, then replaced by 0-01). The conditional probabilities of aradioactive 192 Ir seeds for brachytherapy. patient being hypovolaemic if the CT scanThe catheters were then withdrawn. Two indicated raised intracranial pressure (ICP)patients became drowsy during treatment were high: moderate and severe hydro-but responded to steroids. No other sig- cephalus, p = 0-80; compression ofthe basalnificant complications were observed. In the cisterns, p = 0-82; compression of the basalFUS protocol, preliminary craniectomy cisterns associated with intracerebralprovided a window for the ultrasound beam. haematoma or midline shift, p = 1. TheFour FUS treatments were given at weekly presence of significant hypovolaemia shouldintervals on an escalating time scale, curren- be considered when managing SAH patientstly 30, 45, 60 and 60 minutes, together with with clinical evidence of raised ICP.external beam radiation, generally on a dailybasis. Temperatures of >42 5°C were Referencesattained at 46% of up to 49 measuredintratumoural points over all treatments, I Maroon JC, Nelson PB. Neurosurgeryand at 65% of points in the best treatment. 1979;4:223-6.Six patients completed the protocol without Solomon RA, Kalmon D, Post MD, McMurtrySixpaientscompltedth protcol wthout JD, et al. Neurosurgery 1984;15:354-1.significant complications, one patient with-drew after two treatments and one wasjudged too ill to continue, also after twotreatments. AN EVALUATION OF THE APPLICABILITY OF

BRAIN SLICE METHODOLOGY TO THEINVESTIGATION OF THE PATHOPHYSIOLOGY OFFOCAL CEREBRAL ISCHAEMIA

Clough PH, Lye RH (Manchester)HYPOVOLAEMIA FOLLOWING SUBARACHNOIDHAEMORRHAGE: IS RAISED INTRACRANIAL The authors investigated the applicability ofPRESSURE REPONSIBLE? brain slice methodology to the study ofNelson RJ, Rubin C, Roberts J, Walker V, cellular events occurring in the vicinity of anAckery DM, Pickard JD (Southampton) area of focal cerebral ischaemia in the rat

forebrain. Eighteen Sprague-Dawley ratsHypovolaemia, which may occur insidiously were randomly assigned to three groups offollowing a subarachnoid haemorrhage' six. In the lesioned group, strokes were

(SAH), is associated with an increased risk of produced by occlusion of the middle cerebralcerebral ischaemia in patients with symp- artery. After recovery from the anaesthetic,tomatic cerebral vasospasm.2 The aetiology the rats were killed and coronal brain slicesof hypovolaemia remains uncertain prevent- were prepared to include an area of infarc-ing the identification and appropriate man- tion and then mounted in a slice chamber.agement of high-risk patients. We have The presence of an ischaemic focus in thestudied 25 consecutive unselected SAH slice was confirmed by histochemical tech-patients to determine those factors niques. After equilibration for one hour atassociated with the early development of 35°C in artificial cerebrospinal fluid, elec-hypovolaemia. Plasma volume (PV) was trical activity in the slices was monitoredmeasured using radio-iodinated human using microelectrodes. The findings were

925

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

926

compared with brain slices prepared fromthe control (non-anaesthetised) and shamoperated groups. Spontaneous electrical dis-charges were recorded in all preparations. Incomparison with those from the othergroups, discharges were ofgreater amplitudeand duration in the contralateral hemis-pheres of brain slices prepared from lesionedanimals. The potential of this technique toexplore the possible role of excitotoxicamino-acids in cerebral ischaemia was dis-cussed.

EXPERIMENTAL SUBARACHNOIDHAEMORRHAGE IN A SMALL ANIMAL MODEL:CHANGES IN CEREBRAL BLOOD FLOW,CEREBRAL PERFUSION PRESSURE,INTRACRANIAL PRESSURE ANDNEUROTRANSMITTER STATUSJackowski A, Crockard HA, Burnstock G(London)

No animal preparation replicates all thefeatures of human subarachnoid haemorr-hage but certain features may be reproduced.The acute effects of injection of 0 3 ml ofautologous arterial blood into the cisternamagna of the rat on cerebral blood flow(hydrogen clearance using chronically-implanted cortical electrodes), intracranialpressure (recorded from the cisterna magna),cerebral perfusion pressure, ultrastructure ofthe cerebral arteries and perivascularneurotransmitter levels were studied.Cerebral blood flow was immediatelyreduced by 50% and remained low for thenext three days. This was accompanied by anacute rise in intracranial pressure to 48mm Hg (normally 4 mm Hg) falling to 16mm Hg at 15 minutes until 3 hours andremaining slightly elevated until the thirdday. The intracranial pulse pressure wasacutely increased on which C-waves weresuperimposed. The fall in cerebral perfusionpressure was insufficient to account for thereduction in cerebral blood flow. Scanningand transmission of electron microscopyshowed only minor structural changes invessel walls, with focal areas of subintimalmedial necrosis. Macrophage transforma-tion of pial arachnoid cells may have beenresponsible for the rapid removal of subara-chnoid clot. Cerebral perivascular 5-Hydroxytryptamine content increasedaccompanied by a fall in Neuropeptide Yconcentration, both levels returning towardsnormal by three days.

Proceedings of the 113th meeting of the Society ofBritish Neurological SurgeonsCARCINOMATOUS MENINGITIS: DIAGNOSIS BYRADIOIMMUNOASSAYMoseley RP, Oge K, Shafqat S, MoseleyCM, Sullivan N, Badley A, Burchell J,Taylor-Papadimitriou J, Coakham HB(Bristol)

Neoplastic meningitis occurring as a mani-festation of metastatic carcinoma (carcino-matous meningitis), has been recognisedincreasingly following improved methods ofsystemic therapy. The diagnosis is confirmedby the demonstration of malignant cells inthe CSF and standard cytological techniqueshave been significantly enhanced by mono-clonal antibody immunocytology.' How-ever, cytological confirmation of neoplasticmeningitis is still occasionally achieved withgreat difficulty and often reflects very ad-vanced meningeal disease. Biochemical assayof a tumour marker might overcome thisproblem but no such assay exists that com-bines adequate sensitivity and specificity forthe reliable diagnosis of neoplastic menin-gitis. Serum and CSF levels of an epithelialassociated glycoprotein (polymorphic epith-elial mucin) were assayed in a group of 50patients, 10 of whom had carcinomatousmeningitis-polymorphic epithelial mucinwas found in CSF from all 10 patients withcarcinomatous meningitis but was not detec-table in CSF samples from the 40 controlpatients (normals, demyelination, inflam-matory meningitis, systemic carcinomatosiswithout neurological symptoms, and non-carcinomatous neoplastic meningitis). Thehigh molecular weight of this glycoproteinappeared to prevent its transgression of theblood brain barrier reducing the possibilityof false positive results arising from leakagefrom serum in the CSF.

Reference

I Coakham HB, Brownell B, Harper El, GarsonJA, Allan PM, Lane EB, Kemshead JT.Lancet 1984;1:1095-8.

RECURRENT SUBARACHNOID HAEMORRHAGEIN PATIENTS WITH WRAPPED ANEURYSMSTodd NV, Tocher J, Jones PA, Miller JD(Glasgow and Edinburgh)

Wrapping an aneurysm is now reserved for alimited number of surgical situations includ-ing the unusual aneurysm that cannot beclipped without clipping adjacent vessels, theportion of an aneurysm that may remainoutside a clip, and infundibula that are toosmall to be clipped. The rate of rebleedingfollowing such wrapping remains ill-defined.Between 1965 and 1974, 370 patients had an

intracranial operation for aneurysm in Edin-burgh ofwhom 60 had wrapping as the soletreatment of a single anterior circulationaneurysm (middle cerebral artery aneurysm55%, carotid aneurysms 8%). Follow updata for the subsequent decade were availa-ble for 54 of these patients. Recurrentsubarachnoid haemorrhage occurred in 11patients with wrapped aneurysms, fivewithin six months ofsurgery and six betweensix months and ten years; hence the knownearly rebleed rate from wrapped aneurysmswas 8-6% (95% CI 1.4-16%) and the knownlate rebleed rate was 14% (95 CI 3-24%) thatis 1 5% per annum. When compared with therebleed rates expected in untreatedpreviously ruptured aneurysms, wrappingsignificantly reduces early rebleeding but isinferior to clipping. The sample size was toosmall to demonstrate any significant effect onlate rebleeding.

INTRAOPERATIVE ULTRASOUND GUIDEDRESECTION OF BRAIN TUMOURS ANDCOMPARISON OF MARGINS AND VOLUMES TOCT FINDINGSLe Roux P, Silbergeld D, Mack L, Berger M(Seattle, USA)

The volume of an intrinsic cerebral tumouror a cerebral metastasis was estimatedboth by computed tomography andintraoperatively using a portable RTS sectorscanner (ATL MIC8; 3 5, 7 5 and IOMH).Peritumoural oedema was found to be non-echogenic. In nine patients with previoussurgery, with or without radiotherapy forintrinsic tumours, intra-operative ultra-sound overestimated the volume of tumourfound at re-operation (correlation co-efficient 1-41, SD 0-4). There was a bettercorrelation in 16 patients at the time of initialsurgery (correlation co-efficient 102, SD0 25).

CLOSED HEAD INJURIES: WHERE DOES DELAYOCCUR IN THE PROCESS OF TRANSFER TONEUROSURGICAL CARE?Maurice-Williams RS, Marsh H, Hatfield R(London)In the United Kingdom, most patients withhead injuries are admitted to a district gen-eral hospital and only selected patients sub-sequently transferred to a neurosurgical unitwith the possibility of delay in treatment ofahaematoma or inappropriate management.In a prospective analysis of the emergency

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

Proceedings of the 113th meeting ofthe Society ofBritish Neurological Surgeonstransfer of 117 consecutive patients withsuspected traumatic intracranialhaematoma, the authors defined the timesinvolved in the process of secondary referralfrom 14 district general hospitals lying bet-ween 3 and 9 miles by road from theneurosurgical unit (15-90 mins). Avoidabledelay occurred at the district general hospitaleither from failure to institute appropriatetreatment for non-cranial injuries or fromfailure to realise that transfer was necessary.Once the decision to transfer a patient hadbeen made, the process of transfer consumedrelatively little time, regardless of the dis-tance from the district general hospital to theneurosurgical unit. The geographical disper-sal of neurosurgical services would notimprove the outlook of patients with headinjury and the authors proposed that theoptimum outcome would be achieved byconcentrating head injury admissions toselected district general hospitals each equip-ped with a CT scanner linked to a neuro-surgical unit and a standby ambulance forthe transfer of head injured patients.Immediate scanning following resuscitationat the district general hospitals of all headinjured patients with a high or intermediaterisk of traumatic intracranial haematoma(that is skull fracture, with or withoutdepression ofconscious level) and immediatetransmission ofthe pictures for inspection bystaff at the neurosurgical unit would permitthe early identification of the patient with a

significant haematoma. Theatre facilitieswould then be mobilised at the neurosurgicalunit. The time required for transfer was so

short that such a system would allow surgeryat the neurosurgical unit to be carried out as

quickly as at the district general hospitalitself.

AVOIDABLE SECONDARY BRAIN DAMAGE TOHEAD INJURED PATIENTS TRANSFERRED TONEUROSURGICAL UNITSGentleman D, Jennett B (Glasgow)

Patients in coma following head injury are atrisk of secondary brain damage fromhypoxia (arterial PG2 < 70 mmHg) andhypotension (systolic blood pressures < 100mmHg) produced by avoidable adverse fac-tors before and during transfer to theNeurosurgical Unit. The frequency of suchfactors has been compared in 200 patientsarriving at the Glasgow Neurosurgical Unitin coma after head injury in 1986/87 with a

study of 150 patients six years previously.'Two-thirds of both series reached the NSUwithin six hours of injury. There had beenmodest but non-significant improvements in

the care of comatose patients during trans-fer; more had a medical escort, more were

intubated, fewer arrived hypoxic, and fewerhad inadequate extracranial injuries. Non-etheless, one-third of patients in the recentseries had no mechanical airway protection,15% arrived hypoxic, 10% had inadequatelymanaged extracranial injuries and 7%arrived in shock. Over a third suffered an

adverse event before reaching the Neurosur-gical Unit-epilepsy, aspiration or res-

piratory arrest. Referring doctors who dis-patch patients in coma to neurosurgical unitsshould be more aware of the potentialhazards of ambulance transfer and policiesand routines need to be agreed to reducetheir occurrence.

Reference

I Gentleman D, Jennett B. Lancet 198 :ii:853-5.

CLINICAL ADVANTAGES OF LOW COST CT

GUIDED STEREOTAXYMackenzie Al, Bell BA, Marsh HT, Uttley D(London)

In October 1986 a standard pre-CT Leksellstereotaxic frame was modified at low costfor use in a GE 9800 CT scanner for burrhole brain biopsies. The procedure requiredan average of 84 minutes of theatre time ofwhich 15 minutes was spent in the CTscanner. One hundred and ten biopsiesincluding seven of the posterior fossa hadbeen performed in 100 patients (mean age 56,range 15-78; 59 males, 41 females) by 12surgeons of varying experience. His-tologically abnormal tissue was obtained in83% of biopsies with benign lesions found in7% and management changed as a result in14%. Persistent deterioration occurred infour patients and temporary deteriorationwas noted in seven patients. Two patientsdied both with high grade gliomas. Secondbiopsies were performed in ten patients andoverall a positive diagnosis was obtained in90 of the 100 patients. Comparison of theseresults with published studies ofother biopsymethods'2 provoked considerable discussionover accuracy and rate of positive biopsies.

References

I Heilbrun PM, Roberts TS, Apuzzo MUJ, WellsTH, Sabshin JK. J Neurosurg 1983;59:217-22.

2 Goldstein S, Gumerlock MK, Neuwelt EA. JNeurosurg 1987;67:341-8.

927STEREOTACTIC RADIOSURGERY IN VEIN OFGALEN MALFORMATIONSDias PS, Forster DMC, Battersby RDE,Bergvall U, Powell T (Sheffield)

Vein of Galen malformations (VGM) con-stitute a group of midline, high flowarteriovenous malformations with secon-dary aneursymal dilatation of the vein ofGalen. It is relatively rare and usuallypresents in early life, though cases have beenreported in later life. The reported results arepoor, with an overall mortality of 55-6%(91-4% in neonates) with an overall surgicalmortality of 30 3% and a significant mor-bidity in 515%. Johnston' concluded thatimproved surgical techniques alone wereunlikely to alter the outcome substantiallyand staged surgery and embolisation weresuggested as adjuncts to treatment. An infantwith a vein of Galen malformation presentedat the age of 9 months with macrocephalyand a cranial bruit. He was treated withstereotactic radiosurgery at 10 months ofage. A clinical response was apparent withinmonths of the treatment and partialangiographic resolution of the malformationwas demonstrated one year after treatment.Theoretically such stereotactic radiosurgeryhas the advantage that gradual occlusion ofonly the fistulous component may be bettertolerated and be less likely to interfere withthe function of the normal circulation. Inparticular, it may reduce the tendency topostocclusive thrombosis of the venousoutflow tract following single stage interven-tion.

Reference

I Johston IH, Whittle IR, Besser M, MorganMK. Neurosurgery 1987;20:747-58.

THE EFFECT OF FLUDROCORTISONE ACETATEON PLASMA VOLUME & SODIUM BALANCE INPATIENTS WITH SUBARACHNOIDHAEMORRHAGELindsay KW, Hasan D, Vermeulen M, Wij-dicks E, Murray G, Brouwers P, Hatfield R,van Gijn J (London, Rotterdam, Utrecht &Glasgow)

The authors demonstrated in a previousmulticentre trial' that the antifibrinolyticagent tranexamic acid reduced rebleedingafter subarachnoid haemorrhage but in-creased the incidence of cerebral infarction.In one centre (Rotterdam), hyponatraemicpatients, thought to have inappropriateADH secretion and treated with fluid restric-tion, developed a high incidence of cerebralischaemia.2 A subsequent study suggested

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

928that hyponatraemia was not dilutional, butresulted from sodium loss and wasassociated with a fall in plasma volume thatmight contribute to cerebral ischaemia.3Increase in the fluid input failed to preventthe fall in plasma volume. In a randomisedstudy the effect of fludrocortisone acetate onsodium balance in plasma volume was inves-tigated in 91 patients with subarachnoidhaemorrhage, 46 of whom receivedfludrocortisone. Fluid and sodium balancewere measured daily and the plasma volumedetermined on days 1, 6 and 12 after admis-sion. In the first six days, a negative sodiumbalance correlated with a negative fluidbalance and a fall in plasma volume.Fludrocortisone reduced the incidence ofpatients with a negative sodium balance(p = 0-04) but did not significantly reducethe fall in plasma volume nor the incidence ofcerebral ischaemia.

References

1 Vermeulen M, Lindsay KW, Murray G. N EngiJ Med 1984;311:432-7.

2 Wijdicks E, Vermeulen M, Hijdra, et al. AnnNeurol 1985;17:137-40.

3 Wijdicks E, Vermeulen M, ten Haaf JA, et al.Ann Neurol 1985;18:211-6.

Proceedings of the 113th meeting of the Society ofBritish Neurological SurgeonsCHANGES IN PATIENTS AFTER FOCAL HEADINJURYStatham P, Bullock R, Patterson J, TeasdaleGM, Teasdale E, Wyper D (Glasgow)

Haematomas or contusions complicate atleast one third of severe head injuries and area major cause of preventable mortality andmorbidity. Many such patients undergosecondary deterioration, but the mechan-isms responsible are poorly understood.Although ischaemic neuronal damage isassociated with haematomas and contusionsin human post mortem and animal studies,non-tomographic 133 Xe cerebral blood flow(CBF) measurements have failed to showsufficiently low CBF to cause ischaemicneuronal damage in survivors. Serial MRI(015 resistive) and single photon emissioncomputed tomography (SPECT) scanningwere used to map regional CBF (HMPAO)and blood brain barrier lesions (99Tm per-technetate) both early (one to ten days) andlater (seven to 90 days) after head injury inorder to analyse pathophysiological con-sequences of focal lesions. Examples of theresults of the various imaging modalitieswere presented (8 subdural and 4 extraduralhaematomas; 11 contusions) together with acritical discussion of the problems ofquantification.

EMERGENCY SURGERY FOR HAEMATOMA

FORMING ANEURSYMAL HAEMORRHAGES

Page R, Richardson PL (Manchester)

Management ofan intracerebral haematomafollowing aneursymal rupture remains con-

troversial and the authors reported theirexperience in seven patients with depressedlevel of consciousness and proven haema-toma on CT scan (4 males, 3 females; age 26-51 years; six middle cerebral arteryaneurysms and one anterior communicatinganeurysm). The haematoma was removed asan emergency and the aneurysm clipped atthe same operation following angiography.One patient died. One patient (anterior com-municating aneurysm) was still in hospitaland required constant nursing care but fivepatients had returned to work or were

independent at home (epilepsy-2; hemian-opia-1; mild dysphasia-2). The authorssupported the current trend towards emer-gency surgery for selected patients whowould otherwise have a protracted period ofhigh intracranial pressure with its attendanthigher morbidity and/or mortality.

SERIAL TOMOGRAPHIC MAPPING OF CEREBRALBLOOD FLOW AND BLOOD BRAIN BARRIER

DORSAL ROOT ENTRY ZONE LESION IN THETREATMENT OF PAIN IN PATIENTS WITHDAMAGE TO THE SPINAL CORD OR CAUDAEQUINATeddy P, Vafadis J, Fairholm D (Oxford)

Lesions made in the dorsal root entry zone ofthe spinal cord may prove beneficial inrelieving central pain following avulsion ofthe roots of the brachial plexus but theresults achieved using this operation in thetreatment of central or deafferentation painfollowing spinal cord injury or damage to theroots of the cauda equina are less clear cut.The results of a combined series of 30patients was reviewed, of whom 21 sufferedblunt or penetrating injuries to the spinalcord or cauda equina and nine had pain innumb areas following surgery for lumbardisc or from presumed chemical arach-noiditis. The results from the first group weregenerally good, particularly in those withcomplete spinal cord injuries but the resultsin the second group were universally verypoor in terms of pain relief and wereassociated with a high incidence of unaccep-table side effects.

DORSAL COLUMN STIMULATION FOR PAIN: AREVIEW OF 60 CASESSimpson BA (London and Cardiff)

Experience with implantation of Dorsalcolumn stimulators for pain relief in 62patients in the London Hospital betweenSeptember 1978 and April 1987 was presen-ted-all electrodes were inserted via a lamin-ectomy and secured epidurally. The outcomein two cases was not available. In 60 patients(34 male and 26 female; age range 21-74years) symptoms had been present for fromthree months to fifty years and follow-upranged from 2 months to 9 years (12 deadand 32 remaining for follow-up). Cervicalelectrodes were employed in 25 and thoracicin 35. The average number of operations perpatient was 2-8 and 29 required no re-opera-tion. Four result categories were employed:made worse (4), no effect (12), modest benefit(14), and significant benefit (28). Twopatients were made worse after initial benefit.Benefit was therefore achieved in 70%overall in a wide range of conditions butpost-herpetic and post-thalamotomy painwere unlikely to be helped.

DOES FOCAL BRAIN OEDEMA CAUSE CEREBRALDYSFUNCTION?Whittle IR, Miller JD (Edinburgh)

The relationship between brain dysfunctionand focal brain oedema is complex. It isinfluenced by factors such as neural tissuedamage associated with the oedemogeniclesion, altered ICP dynamics, and ischaemia.An experimental model using focal brainoedema generated by direct intracerebralinfusion was utilised to minimise these com-pounding variables and assess the effects offocal brain oedema, generated by infusatescontaining various substances, on brainphysiology. The infusion method of brainoedema' was used in adult mongrel cats. Theanimal was anaesthetised, intubated andventilated. A 23 G needle was stereotacticallyplaced in the right forebrain white matter.Infusates were delivered at 0-2 ml/h for threehours. Infusates studied were saline, 20%protein (Fetal calf serum), arachidonic acid2-20 mg/ml, bradykinin 5-150 mg/ml, andcyst fluid from 4 human gliomas. Parametersstudied were rCBF, cerebrovascular reac-tivity to changes in arterial carbon dioxidetension, ICP, pressure-volume index, braincompliance, somatosensory (SEP) andmotor (MEP) evoked potentials, BBB dis-ruption and brain water content. Forty catswere studied. All infusates increased whitematter water content by 8-12% when com-

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

Proceedings of the 113th meeting ofthe Society ofBritish Neurological Surgeons

pared with the contralateral hemisphere. ICPincreased from baseline means of 5 mmHg to15-25 mmHg depending on infusate, whilstcompliance fell and PVI remained constant.However, despite often major BBB disrup-tion the rCBF remained stable, CO2 re-

activity intact, and SEPs and MEPs repro-

ducible. This study has shown that even inthe presence of severe focal brain oedemaand a moderate rise in ICP, many parametersindicative of brain function remain stable.The addition of putative neuronal, glial or

vasomodulating substances to the oedemafluid, although causing BBB disruption, didnot alter this finding.

Reference

I Marmarou A. Adv Neurol 1980;28:345-58.

THE USE OF TRANS-VENTRICULAR ENDOSCOPY

IN THE MANAGEMENT OF ACQUIRED NON-

COMMUNICATING HYDROCEPHALUS

Rawlinson JN, Coakham HB, Torrens M,Griffith HB (Bristol)

Acquired non-communicating hydrocepha-lus is most commonly due to a tumour withinor adjacent to the ventricular system. Trans-ventricular endoscopy permits establishingthe tissue diagnosis at the same time asproviding emergency drainage. Experiencewith 22 patients was presented. Hydro-cephalus was relieved in all patients. Thehistological diagnosis was achieved in 17patients and visually in three (two colloidcysts and one thrombosed arteriovenousmalformation distorting the opening of theaqueduct). Two colloid cysts were too toughto biopsy or aspirate and were removed bycraniotomy under the same anaesthetic. Sub-sequent arteriography in the patient with a

thrombosed AVM did not reveal the lesionand endoscopy was possibly the only way thediagnosis might have been made. Biopsyfailed in two patients. Two patients diedfollowing haemorrhage from a tumour site.There were no cases of infection.

THE TREATMENT OF INTRA-CRANIAL

ANEURYSMS BY INTERVENTIONAL AND

NEUROVASCULAR TECHNIQUESHigashida RT (San Francisco, USA)

The author presented his experience with 186patients (104 females, 82 males; age range15-72) with surgically inaccessibleaneurysms using local anaesthesia and con-

tinuous neurological monitoring. Tech-

niques included proximal ligation, trappingand embolisation of the aneurysm itself. Thevessel could not be preserved in 119 patients.Immediate complications including transientischaemic attacks (16), stroke (12), haemorr-hage (8), and death (8).

CT INFUSION SCANNING FOR THE DETECTIONOF ANEURYSMS CAUSING INTRACEREBRALHAEMORRHAGELe Roux PD, Newell DW, Silbergeld DL,Stimac GK, Winn HR (Seattle, USA)

Computer tomographic scanning (CT) is theprocedure of choice in the diagnosis ofintracerebral haemorrhage (ICH), but doesnot always demonstrate a responsibleaneurysm. In the deteriorating patient thereis insufficient time for angiography: thus weprospectively studied the use of a continuousinfusion of contrast material during CTscanning to detect aneurysms in patientswith ICH. CT infusion scans were done on a

GE 9800 scanner. Patients were given a

constant infusion of 80-100 ml of sodiumdiatriazote (Hypaque 60). Scans wereobtained in the dynamic mode from the floorof the sella to a point above the anteriorcommunicating artery (15-18 1 5 mm slices).Images were photographed at intermediatewindows (level 80W; 400 Hounsfield units).Eight patients received infusion scans. Scanswere normal in three patients and aneurysmsdetected in five patients. These findings wereconfirmed at surgery and post-operativeangiography. All aneurysms were success-

fully clipped at the time of ICH evacuation.Surgical planning was facilitated by theinformation obtained. CT infusion scanningmay thus be useful in rapidly detectinganeurysms in cases of ICH.

NON-INVASIVE INVESTIGATION FOROCULOMOTOR PALSY DUE TO ANEURYSMTeasdale E, Macpherson P, Statham P(Glasgow)

One of the common causes of a painful 3rdnerve palsy is expansion or rupture of an

aneurysm of the internal carotid artery aris-ing at or near the origin of the posteriorcommunicating vessel or more rarely the tipof the basilar artery. The aneurysm responsi-ble for such a 3rd nerve palsy is almostalways greater than 4 mm diameter andshould be within the resolving capabilities ofmodern dynamic CT scanning therebyreducing the need for angiography in the

929

exclusion ofan aneurysm. Sixty patients withsurgically confirmed posterior communicat-ing artery aneurysms were examined toassess the size and location of the aneurysmand the presence or absence of a third nervepalsy. Twenty eight patients were studiedprospectively for the clinical history of 3rdnerve palsy. Niopam 370 (100 ml) was injec-ted as 3 mm overlapping axial or coronal CTslices of the perisellar area were taken andcoronal, sagittal and oblique reformationsstudied. Of the 48 carotids and 24 basilarvessels so examined an aneurysm was iden-tified in four, a small possible aneurysm intwo and an abnormality of the internalcarotid bifurcation in another. Two othercarotids showed marked tortuosity inassociation with basilar ectasia (one patient).Angiography was carried out in 19 of thesepatients and confirmed the aneurysm in threeofthe four cases (one 75 year old lady had noconfirmatory angiography). One small possi-ble aneurysm was identified as a 2 mminfundibulum and the other confirmed as a 3mm x 8 mm aneurysm. The bifurcationabnormality was a 3 mm infundibulum aris-ing in relation to an abnormally highanterior choroidal vessel. The remainder ofthe carotids and basilars were confirmednormal at angiography and there wereneither false negative nor positive examina-tions. Vertebral angiography is unnecessaryifdynamic CT shows the basilar system to benormal. Thin section dynamic perisellar CTwith reformation in many planes wouldappear to be an appropriate non-invasiveway to identify or exclude an aneurysmcausing 3rd nerve palsy thereby saving manypatients the hazard and discomfort ofangiography.

EXPERIENCE WITH CT AND MRI DIRECTEDBIOPSY 1982-1988Thomas DGT, Nouby R (London)

The authors reviewed their experience in 388image directed stereotactic surgicalprocedures performed with the Brown-Roberts-Wells stereotactic apparatus duringa five year period. The series included 292cases ofCT directed biopsy, 17 cases ofMRIdirected biopsy for lesions not shown on CTor an ambiguous or anomalous CTappearance and eight cases of haematomaaspiration. Positive diagnosis was obtainedin 92 8% of cases in which 10% were foundto be non-neoplastic lesions. Ten of the 13cases of MRI directed procedures yieldedpositive diagnosis in which three were foundto be non-neoplastic. Complications in 317cases of biopsy or haematoma drainage

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from

930included haemorrhage, hemiparesis andepilepsy with an overall complication rate inthe first week of 4 4% cent including onedeath. Image directed stereotactic biopsymay be performed at any intra-cranial sitewith a high diagnostic yield and low mor-bidity.

STUDY OF INTRACRANIAL FLUID FLOW BY

ULTRAFAST ECHO-PLANAR IMAGINGFirth JL, Stehling MK, Coxon R, OrdidgeRJ, Howseman AM, Chapman B, MansfieldP (Nottingham)

Echo-planar imaging (EPI) is an ultrafastNMR technique capable of producing snap-shot transectional images in around 1/15thofa second.' This non-invasive technique hasbeen used clinically during various phases ofits development since 1981.2 The considera-ble potential of the technique was demon-strated in a series of head studies includingvarious types of hydrocephalus and thehaemodynamics within an arteriovenousmalformation.

References

I Mansfield P, Morris PG. NMR Imaging in

Proceedings of the 113th meeting of the Society ofBritish Neurological SurgeonsBiomedicine, New York: Academic Press,1982.

2 Rzedzian R, Mansfield P, Doyle M, GuilfoyleDN, Chapman B, Coupland RE, Chrispin A,Small P. Lancet 1983:ii:1281-2.

SUSPECTED POSTERIOR FOSSA LESIONS: DOES

THE FIRST IMAGING MODALITY AFFECT

EVENTUAL PATIENT MANAGEMENV? A

PROSPECTIVE RANDOMISED TRIAL

Hadley DM, Macpherson P, Teasdale E,Grossart K, Lawrence A, Teasdale GM(Glasgow)

Uncontrolled studies of small selectedgroups of patients suggest that an initialMRI is more effective than CT in the assess-

ment of patients suspected of harbouring a

posterior fossa lesion. Utilisation of thesetwo investigations was compared in a prosp-

ective- trial by randomising patients withsigns or symptoms of posterior fossa diseaseto have either CT or MRI as their firstexamination. After imaging, the results werereported to the responsible clinician whocould then either manage the patient on thebasis of the first study or could request thealternative examination. Patients (1020)

were entered into the study over a two yearseries, 511 were randomised to have CT firstand 509 to MRI. One per cent of patientswere precluded from MRI as a result ofclaustrophobia. Patients (887) were man-aged by the imaging technique to which theywere first allocated and a second examina-tion was requested in 126. Of 511 patientswho were allocated to CT first, 97 sub-sequently had MR imaging requested. Bycontrast, of patients whose first examinationwas MRI only, 41 had a further request for aCT. In patients with suspected demyelina-tion, a tumour or a degenerative disease, MRimaging was requested in addition to CT in36%, 22% and 20% of patients respectively.A CT was carried out after MRI in 5%, 9%and 11% of corresponding cases. Cliniciansultimately requested MR imaging in 75% ofpatients diagnosed finally to have eitherdemyelination or a tumour. In patients inwhom a diagnosis was not established, 1 %of those whose initial study was CT also hadMR but only 2% had CT after MR. Hence inthe majority of patients either CT or MRIprovided the information required for initialmanagement but when CT was a firstexamination it was considered insufficientfor the management of a patient three timesmore often than MR.

Protected by copyright.

on January 20, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.52.7.923 on 1 July 1989. D

ownloaded from