testingof patients: guide to - bmj · the computerised infusion test32 is a method forthe accurate...

J7ournal ofNeurology, Neurosurgery, and Psychiatry 1996;60:549-558

Testing of cerebrospinal compensatory reserve inshunted and non-shunted patients: a guide tointerpretation based on an observational study

Marek Czosnyka, Helen Whitehouse, Piotr Smielewski, Sreten Simac, John D Pickard

MRC CambridgeCentre for BrainRepair and AcademicNeurosurgical Unit,University ofCambridge, UKM CzosnykaH WhitehouseP SmielewskiS SimacJ D PickardCorrespondence to:Dr Marek Czosnyka,Academic NeurosurgicalUnit, PO Box 167,Addenbrooke's Hospital,Cambridge CB2 2QQ, UK.Received 20 November 1995Accepted 26 January 1996

AbstractObjective-To design a computerisedinfusion test to compensate for the disad-vantages of Katzman's lumbar infusionmethod: inadequate accuracy of estima-tion of the resistance to cerebrospinalfluid outflow and poor predictive value innormal pressure hydrocephalus.Methods-Accuracy was improved byintracranial pressure signal processingand model analysis for measurement ofcerebrospinal compensatory variables.These include the CSF outflow resistance,brain compliance, pressure-volume index,estimated sagittal sinus pressure, CSFformation rate, and other variables.Infusion may be made into the lumbarspace, ventricles, or, when assessing shuntfunction in vivo, the shunt chamber.Results and conclusions-The comput-erised test has been used for five years in amulticentre study in 350 hydrocephalicpatients of various ages, aetiologies, andstates of cerebrospinal compensation.The principles of using the test to charac-terise different types of CSF circulatorydisorders in patients presenting with ven-tricular dilatation, including brain atro-phy and normal and high pressurehydrocephalus, are presented and illus-trated. Previous studies showed a positivecorrelation between cerebrospinal com-pensatory variables and the results ofshunting, but such a prediction remainsdifficult in idiopathic normal pressurehydrocephalus, particularly in elderlypatients. The technique is helpful in theassessment of shunt malfunction, includ-ing posture-related overdrainage, over-drainage related to the nocturnal B waveactivity, and proximal or distal shuntobstruction.The appendix presents an introduction

to the mathematical modelling of CSFpressure volume-compensation includedin computerised infusion test software.

U Neurol Neurosurg Psychiatry 1996;60:549-558)

Keywords: hydrocephalus; cerebrospinal fluid; infusiontest; modelling; hydrocephalus shunts

Hydrocephalus is more complex than a simpledisorder of the CSF circulation. 1-5 As shuntingis a purely mechanistic treatment, the biome-chanics of a patient's pressure-volume com-pensation should be ideally examined before a

shunt is implanted. Such an examination,although invasive, may help with the decisionas to whether to shunt, and with possible com-plications such as shunt blockage and underand overdrainage.The role of a shunt is to correct the distur-

bances of CSF circulation or outflow bya reduction of the resistance to CSF outflowto about 5-10 mm Hg/ml/min-the typicalhydrodynamic resistance of most shunts. If apatient has normal intracranial pressure and anormal resistance to CSF outflow (6-10 mmHg/ml/min6 7), a shunt theoretically cannothelp because the resistance cannot be reducedfurther without the risk of overdrainage. Thefundamental role of the volume-pressure studyis to measure the resistance to CSF outflowalong with other compensatory variables andconsider whether the disturbed cerebrospinalcompensation can be improved by shunting.Therefore, the examination may help to avoidunnecessary shunting or shunt revision.811

Almost all authors7 8 111-7 agree that in hydro-cephalus the drainage of CSF is disturbed,which may be expressed quantitatively by araised resistance to CSF outflow. The limit ofnormal resistance is around 12 or 13 mmHg/ml/min.7 9 11121416 However, raising of thisresistance is not always correlated with a goodoutcome after shunting18-20-particularly inidiopathic normal pressure hydrocephalus inelderly patients.13 Our experience suggests thatnot only the resistance but also other compen-satory variables are helpful in the diagnosis ofhydrocephalus. 1121-24

As the rate of shunt failure is around20%-30% during the first year after implanta-tion and continues at a rate of 5% per year,25the probability of shunt malfunction over timeis high. A blocked shunt does not always pro-duce the dramatic onset of clinical symp-toms,26 27 that justifies a shunt revision withoutperforming an invasive examination. Notuncommonly the CSF circulatory reservedeteriorates gradually and a question aboutthe shunt function may be very difficult toanswer. In such cases, pressure-volume studiesmay help to assess the severity of malfunctionby comparing the current test to the study per-formed before shunting.26 28 Therefore animportant role of a preshunting test is to estab-lish the baseline for comparison with post-shunting investigations performed to test theshunt's fimction in vivo.The main aim of this observational study is

to present the methods of computer supportedanalysis of CSF pressure used for diagnosis ofhydrocephalus, and show how the results of

549 on F

ebruary 10, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.60.5.549 on 1 May 1996. D

ownloaded from

http://jnnp.bmj.com/

Czosnyka, Whitehouse, Smielewski, Simac, Pickard

0)ICDE0-

IEE0-0

10l A

0601

D

0

lU 13 lb 2U 23Time (min)

Figure 1 Mean pulse amplitude ofCSF pressure (AMP)and mean pressure (ICP) recorded during infusion test (xaxis: time in minutesfrom the beginning of recording). Theinfusion of 1k5 mllmin starts in the fifth minute andfinishes around the 22nd minute.

computerised infusion tests are able to distin-guish different types of disorders presentingwith ventricular dilatation. The computerisedinfusion test is faster, less invasive, and doesnot demand specialised equipment as requiredby other methods described.29-3' The princi-ples of interpretation of the pressure-volumetests performed to assess shunt function invivo are also discussed.

Materials and methodsThree hundred and fifty infusion tests were

performed in three different centres (136 inchildren) at the Child's Health Centre inWarsaw (Poland), Medical Academy Hospitalin Warsaw (Poland), and at Addenbrooke'sHospital in Cambridge (UK). All patients pre-

sented with ventricular dilatation and hadhydrocephalus diagnosed by other clinicalexaminations. In 59 patients the test was per-formed to assess shunt function in vivo.

There were almost no serious complicationsafter the test (one case of epidural abscess).

COMPUTERISED INFUSION TEST

The computerised infusion test32 is a methodfor the accurate analysis of the traditional con-

stant rate infusion.'02833 This method allows an

infusion to be made into any accessible fluidcompartments. Lumbar infusion, even if it hasunderstandable limitations, is less invasive and,therefore, more often performed. The secondmost frequent approach is an intraventricularinfusion into a subcutaneously positionedreservoir, connected to an intraventircularcatheter. In such cases two hypodermic nee-

dles (gauge 23 or less, as advised by manufac-turer of a specific type of reservoir) are used:one for the pressure measurement and the sec-

ond for the infusion.

THEORETICAL FRAMEWORKThe model of cerebrospinal volume compensa-tion investigated during this test was describedby Marmarou and coworkers3435 and slightlymodified later.3'38 Apart from resting CSFpressure and the resistance to CSF outflow,the pressure-volume index (PVI), cere-

brospinal compliance, CSF formation rate,and the pulse wave amplitude of CSF pressure

I' 10 B

E

< 0

40'I

E

-Ei DS ra-

4 7 10 13 16 19 22 2

Time (min)

Figure 2 (A) Example of the lumbar infusion test madethrough one lumbar needle (G19). The end of the needlewas partially obstructed, therefore the pressure gradient D= 25 mm Hg across the needle was recorded during theinfusion (rate 1.5 ml/min). Infusion was discontinuedaround the fifth, seventh, and 12th minute to control thereal level ofCSF pressure. (B) Example of an infusioninto the Ommaya reservoir. Rapid pressure increase D wasaround 1 5 mm Hg indicating that the ventricular catheterwas patent.

are measured (see appendix for definitions).During the infusion, the computer calculatesmean pressure and pulse amplitude and pre-

sents both variables with time along the x axis(fig 1). The resistance to CSF outflow is calcu-lated as the difference between the value of theplateau pressure during infusion and the rest-ing pressure, divided by the infusion rate.However, in many cases strong vasogenicwaves or an excessive increase of the pressureabove the safe limit of 40 mm Hg do not allowthe precise measurement of the final pressureplateau. Computerised analysis, on the otherhand, produces results even in difficult cases

when the infusion is terminated prematurely.The algorithm utilises a complex time seriesanalysis for volume-pressure curve retrieval,38the least mean square model fitting,32 and an

examination of the relation between the pulseamplitude and the mean CSF pressure (fordetails see the appendix).24 3739

ONE NEEDLE TESTA lumbar infusion performed using two sepa-rate needles inserted one or two intervertebralspaces apart is the most precise way of pres-sure-volume testing in communicating hydro-cephalus. Exceptionally, in difficult cases thelumbar infusion may be performed using thesame needle for the infusion and pressure mea-surement, although the gauge should not behigher than 21. Normally, the resistance of theneedle is low (around 1-2 mm Hg/ml/min),but this may increase considerably after theinsertion into the lumbar CSF space if the end

0I

0)

IEcL

E

C-)

40401

550 on F


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Testing of cerebrospinal compensatory reserve in shunted and non-shunted patients: a guide to interpretation based on an observational study

Figure 3 Example ofsimultaneous recording ofCSFpressure (ICP) andMCA bloodflow velocity(FV) using a TCDultrasonograph. Strong Bwaves seen in ICP areobviously in phase withfluctuations of bloodflowduring periods A and C.During period B, thefrequency offluctuations ofbloodflow increased,causing a decrease in theamplitude ofB waves inICP. During period D, theB waves disappeared.

25

IEEa.-

060

en

E

U.,U-

0

Time (1 hour period)

of the needle is touching neural roots or anyother tissue. An infusion rate of 10 or1.5 ml/min may produce a sudden increase inthe measured pressure with a subsequentlyslow rise (fig 2A). The initial rapid pressureincrease represents the pressure gradient acrossthe needle. The slow increase (> 5 min) repre-sents the real rise of the pressure inside theCSF compartment which may be measuredusing short breaks during infusion (< 30 s).For analysis, the pressure recorded during theinfusion should be decreased by the pressuregradient across the needle.

INFUSION INTO A SUBCUTANEOUS RESERVOIR ORSHUNT CHAMBERIn a pressure-volume study performed usingtwo needles inserted into a reservoir or shuntchamber the analysis may be complicated byan additional pressure gradient across the ven-tricular catheter. The resistance of a patentintraventricular catheter is lower than 1 mmHg/ml/min. An infusion rate of 1 0-1 5 ml/minproduces almost negligible initial rapidincrease of the pressure (less than 1-0-1 5 mmHg; fig 2B). If the initial rise is higher, the datashould be analysed as in the one needle infu-sion study-that is, the initial rapid increase inthe pressure should be subtracted from theslow asymptotic increase. Again, two or threeshort intermediate breaks in infusion (everyfive minutes) are recommended to control thereal increase in the pressure within the ventri-cles. If the fast initial increase in the pressure ishigher than 3 mm Hg, it suggests a partialobstruction of the ventricular catheter.

LONG TERM MONITORING OF CSF PRESSUREThe presence of "B waves" in the recordedCSF pressure40 41 is reported to be a good prog-nostic factor in shunting.'842 43 There is still no

agreement about the nature of these waves,which have a character of periodic oscillationsof duration from 20 seconds to one or twominutes. This kind of oscillation can berecorded in cerebral blood flow velocity in nor-mal volunteers and seems to be correlated withan REM phase of sleep.44 Similar waves can beseen in arterial blood pressure and heart rate inpatients with severe head injury with low cere-bral perfusion pressure.4546 Whatever thesource of these waves, they are associated withoscillations of the cerebral blood volume.When the cerebrospinal compensatory reserveis low they produce synchronised oscillationsin CSF pressure. In the classic flow chart forthe investigation of hydrocephalus in adults,presented by Gjerris et al,42 shunting is advisedwhen the B waves are present for more than50% of monitoring time or the baseline pres-sure is raised above 15 mm Hg. However, it isnot clear at which amplitude B waves shouldbe interpreted as pathological: waves of ampli-tude from 1 mm Hg to 25 mm Hg can be seenin the same patient (fig 3) and detected usingcomputer data processing.47-49

Results and interpretationsDIFFERENTIATION BETWEEN BRAIN ATROPHYAND NORMAL PRESSURE HYDROCEPHALUSPatients with predominantly brain atrophy (48patients were diagnosed) had a normal CSFcirculation, by contrast with hydrocephalicpatients. Typically, the opening pressure, theresistance to CSF outflow, and the pulseamplitude were low (ICP < 12 mm Hg, RCSF< 12 mm Hg/ml/min, amplitude < 2 mm Hg).The pressure volume index was high (PVI > 20ml), reflecting low elasticity of the atrophicbrain. There were no vasogenic waves seen inthe pressure recording (fig 4A). In this situa-

551

on February 10, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.60.5.549 on 1 M

ay 1996. Dow

nloaded from



Figure 4 Typical infusiontests performed in patientspresented with: (A) brainatrophy; (B) normalpressure hydrocephalus;(C) non-communicatinghydrocephalus (lumbarinfusion); (D) acutecommunicatinghydrocephalus (aftersubarachnoidhaemorrhage).

I

...

..,.,,,,, , *. ...

'..''"....

4-

tion the CSF spaces are being inflated anddeflated passively, like a balloon.Normal pressure hydrocephalus (101

patients) was characterised by a normal open-ing pressure (ICP < 15 mm Hg) and high PVI(> 20 ml). Again, as in brain atrophy, the CSFsystem was very compliant." 14 The increasedresistance to CSF outflow (> 12 mmHg/ml/min) and B waves seen during infusion,particularly prominent when ICP increases,allow normal pressure hydrocephalus to be dis-tinguished from atrophy (fig 4B).

NON-COMMUNICATING HYDROCEPHALUSLumbar infusion is not recommended in non-communicating hydrocephalus both becauseof the subsequent risk of herniation andbecause of the misleading results obtained.However, this type of hydrocephalus may notalways be detected by initial CT. In those fewpatients in whom lumbar infusion was per-formed (12), the resistance to CSF outflowwas normal, because the lumbar infusion wasnot able to detect the proximal narrowing inCSF circulatory pathways. The resting pres-sure, pulse amplitude, and paradoxically, PVIwere all high (ICP > 12 mm Hg, pulse ampli-tude > 4 mm Hg, PVI > 18 ml; fig 4C).The same type of hydrocephalus investi-

gated using ventricular infusion (via a reser-voir, seven patients), showed high restingpressure and high resistance to CSF outflowhigh (ICP > 15 mm Hg, RCSF > 13 mmHg/ml/min). The PVI was low (< 13 ml) andthe pulse amplitude was high (> 4 mm Hg)indicating a poor compensatory reserve. Acute

communicating hydrocephalus (after sub-arachnoid haemorrhage), when problems withCSF circulation arise from insufficient reab-sorption or circulation of CSF over the brainconvexity (39 patients) gave the same patternof variables, regardless of whether the lumbaror ventricular approach was used (fig 4D).

IS CLASSIFICATION ALWAYS POSSIBLE?In 84 patients unqualified differentiationbetween atrophy, normal pressure hydro-cephalus, and non-communicating and acutehydrocephalus was not possible. Some 56patients fell into the border range betweenhydrocephalus and atrophy. In 32 patients fur-ther overnight ICP monitoring with continu-ous assessment of the magnitude andfrequency of B waves helped in making thedecision about shunting. In 24 patientsovernight ICP dynamics were limited withvery little vasomotor pressure activity.

CHANGE IN THE PROFILE OF CEREBROSPINALCOMPENSATION AFTER SHUNTINGThe computerised infusion test was used toassess shunt function in vivo in 59 shuntedpatients. With a shunt functioning properly(26 patients), the resistance to CSF outflownever exceeds 10 mm Hg/ml/min, an excep-tion being the flow regulating Orbis-Sigmavalve.50 The resting pressure remained at orbelow the shunt's opening pressure. Testsrepeated after shunting should always be con-sidered in comparison with the test performedbefore surgery. Abnormal cerebrospinal com-pensatory variables such as high resting pres-

552

i



j.com/

J Neurol N


ay 1996. Dow

nloaded from



15

0

15

Before shLinting

E0

I

4 8 12 15 19 22 26

)' X

A ,W

A

1 2 3_4 5

IKA I 1I

Time (duration 45 min)After shunting

.....

0^...-....,,:..:-,.i'Z'''' ~50

4 24 8 12 15 19 22 26

Time (min)

Figure 5 Infusion test made in a patient (aftersubarachnoid haemorrhage) before and after shunting(Pudenz medium pressure valve). The change in openingpressure was minimal (8 mm Hg before, 7 mm Hg after),but resistance to CSF outflow decreasedfrom 20 mmHglmllmin to 6 mm Hglmllmin. PVI did not change.

sure, increased resistance to CSF outflow, ordecreased PVI should return to normal aftersuccessful shunting. In valves having a lowhydrodynamic resistance and well definedopening pressure a sharp plateau of the pres-sure trend is seen about 1-5 mm Hg above thelevel of shunt's opening pressure (fig 5).1328

DETERMINATION OF SHUNT PATENCYThere are three main methods to assess CSFdynamics in shunted patients. Firstly, the sim-plest and the least invasive way is a measure-ment of intraventricular pressure using a

previously implanted subcutaneous CSFreservoir. Lumbar puncture in communicatinghydrocephalus (27 patients) or the measure-ment of the pressure inside the chamber of theshunt (12 patients)-when the shunt's con-

struction allows such a procedure-were alsopossible.

Pressure measurement and infusion into theshunt chamber is only possible with shuntshaving a CSF sampling reservoir proximal tothe valve. Therefore the method is not usefulfor the testing of all burr hole valves.With the pressure measurement inside the

shunt chamber, the presence of a CSF pres-sure pulse wave and a pressure increase inresponse to coughing should confirm pressuretransmission between the needle and the CSFspace, proving patency of the ventricularcatheter. However, the CSF flow through theshunt may itself attenuate these pressure fluc-tuations. In cases of absent pulse waveform or aweak response to coughing it is useful to blockthe shunt distally by depression of an

occluder, or distal section of outlet catheter toconfirm that during proximal obstructionwaves do not appear.

'aI

E

E

M:I

E

E

a-

Time (duration 25 min)

Figure 6 Two examples of tests performed to check for theshunt patency using both recordingfrom and infusion intothe shunt's chamber. (A) Peritoneal catheter or valveblocked: 1-baseline recording with good pulse amplitude(AMP); 2-infusion with distal occluder depressed(resistance to CSF outflow = 14 mm Hglmllmin); 3-after the proximal occluder was pressed the pulse amplitudedisappeared; 4-infusion with proximal occluder closed:the pressure increased immediately to 90 mm Hg; 5-slowdecrease in pressure with proximal occluder closed, showingthat the shunt was not blocked totally but rather partiallyobstructed. (B) Ventricular catheter blocked: 1-baselinerecording with the pulse amplitude absent; 2-distaloccluder closed; 3-infusion with the distal occluder closed:the pressure increased quickly to 80 mm Hg and resolvedimmediately after the release of the distal occluder; 4-baseline monitoring with proximal occluder closedfollowedby the infusion 5-showing patent valve and distalcatheter (resistance 4 mm Hglmllmin).

When the shunt or chamber have proximaland distal occluders or may be occluded bypercutaneous pressure on the inlet or outlettubes an infusion into shunt chamber is per-formed first with the distal occluder closed.The resistance to CSF outflow and PVIshould not be dramatically different from thevalues measured before shunting. If RCSF ishigher (> 20%) and PVI much lower (< 50%)than before shunting, a partial obstruction ofthe proximal drain, or obstruction of the CSFspace (as in slit ventricle syndrome), are likely(five patients). The infusion is repeated withthe proximal occluder closed. Pressure shouldnot rise higher than 4-5 mm Hg above theshunt's opening pressure (again, the flow regu-lating Orbis-Sigma valve is an exception50). Ifit does, an obstruction of the distal catheter orvalve blockage is likely (four patients; fig 6).

OVERDRAINAGE RELATED TO POSTUREThis may be assessed using a tilting armchair(four patients). When the baseline pressuremeasured in the horizontal body position islow (usually negative), overdrainage is possi-ble. A change of posture to sitting, with thetransducer kept at the same level relative to

-aI

C):

IE

-

CL

EE

ClI

a_

EE0-

I

N

553

oIni;bUI



j.com/

J Neurol N


ay 1996. Dow

nloaded from



ICP during REM sleep is usually manifestedby a period of low (negative) pressure and alow amplitude of the ICP pulse wave (fig 8).Early morning headaches should not beassumed to be due to "high pressure". Theymay be a consequence of low pressure causedby nocturnal overdrainage.

20

0,

I

EE0-

u

I

EE

I

EEa-

u-

-20 16 . ............

Flat Flat

SittingB

5

A

0

2020

Flat 10 min Flat 12 min

Sitting 15 minFigure 7 Tests for posture related overdrainage. (A) no overdrainage (patient with an

antisyphon device put in-line with the shunt); pressure decreased in sitting position to - 5mm Hg, AMP did not react to change in the body position, after 15 minutes sitting ICPreturned to the baseline value. (B) Overdrainage manifested by a decrease in ICP insitting position below - 10 mm Hg, with simultaneous decrease in AMP. After the returnto the horizontal position, ICP andAMP were significantly lower than at baseline.

the patient's ear, usually produces a furtherdecrease in the pressure. If the pressuredecreases to a value lower than - 10 mm Hg,with a gradual decrease in pulse amplitude,overdrainage is possible. If, after 10-15 min-utes of sitting, the pressure and pulse ampli-tude in a horizontal position are lower than atthe baseline, overdrainage is confirmed (fig 7).

OVERDRAINAGE RELATED TO NOCTURNAL

VASOMOTOR WAVES (TWO PATIENTS)Most contemporary valves usually have a low

5152whydrodynamic resistance, a property whichmay result in the overdrainage related to theperiodic oscillations of the cerebrovascularvolume. The expanding cerebrovascular bedacts like a membrane of a water pump with a

distal low resistance valve. Overdrainage seen

after a period of moderately raised oscillating

DiscussionDoes lumbar infusion supported by computer

data analysis give as precise measurement ofthe CSF outflow resistance as other methods?Borgesen et al53 compared the computerisedinfusion test with their lumboventricular per-

fusion,54 a method giving an accurate estimateof the resistance at multiple intraventricularpressure levels. Both tests were compared in a

heterogeneous group of patients presentingwith ventricular dilatation. Strong correlation(r = 0-98; P< 0-001) between the results ofthe tests was found. They concluded thatcomputerised lumbar infusion is as accurate inthe measurement of the conductance of CSFoutflow as the lumboventricular perfusion butis much less invasive.The purpose of the test in clinical practice is

to measure the cerebrospinal compensatoryvariables and use them to predict the results ofshunting. The test is helpful in such a predic-tion,'42' but no single variable is predictive on

its own.'8 Recent studies showed that in cer-

tain groups of patients, such as those withadult hydrocephalus syndrome, there is no

good predictor of outcome.33 55 Also, idio-pathic normal pressure hydrocephalus is a dif-ficult area for outcome prediction,'9 but withsome encouraging reports.42

Often, an improvement after shunting is dif-ficult to assess and may be very short lastingand misleading in making comparisonbetween patients. Three components con-

tribute to postshunt improvement: accuratediagnosis, surgery, and shunt performance invivo.

For diagnosis, a hierarchy of examinationsand tests should form an integrated picture, inwhich the pressure-volume study is only one ofmultiple components. Clinical examination,brain imaging (CT and MRI to assess whitematter lesions'855 in complicated cases), andneuropsychological tests are essential to makea proper decision about shunting. Pressure-volume testing should be preceded byovernight ICP monitoring if suitable access

(Ommaya or Rickham reservoir) is available.However, even after positive diagnosis, a shuntmay not resolve the clinical symptoms,because of postsurgical complications, shuntobstruction, or a mismatch between thepatient's own CSF circulatory reserve and theshunt's pressure-flow performance. It has beenassessed that 30% of shunts develop failureduring the first year with a subsequent rate ofmalfunction of 5% per year,52 giving a totalrate of shunt revisions of 80% within 12 years.As shunt related complications are likely and a

pressure-volume study is a reliable test ofshunt function in vivo, a baseline measure-

ment performed before shunting is very help-

A5j

I

EE0-cL

554

11



j.com/

J Neurol N


ay 1996. Dow

nloaded from



B1h 0

Figure 8 An overdrainage related to the nocturnal vasomotor waves; patient with the valve equipped with a siphoncontrol device (Delta, performance level 2). Two episodes ofa strong, around 30 minutes long, B wave activity wererecorded (the raised ICP, pulse amplitude AMP, and the computer detected power ofB waves). After each episode thebaseline pressure fell to negative values (0) for at least one hour.

ful. Also, some indications for choosing a par-ticular type of the shunt may be made usingpressure-volume tests.

Overdrainage related to posture56 57 may leadto serious complications,58-60 particularly inpatients with the combination of gross ventricu-lar dilatation and a high pressure-volume index(> 26 ml). Therefore, siphon controllingdevices should be considered in such patients.Shunts having a high hydrodynamic resistance(flow regulating, mitre valves, lumboperitonealshunts) should be avoided in patients witha low cerebrospinal compliance and largeamplitude of ICP B waves recorded duringthe test or overnight. The opening pressure ofprogrammable valves should be matched to therecorded resting CSF pressure with acareful stepwise reduction in time to avoid sub-dural collections, and lead to normalisa-tion of ventricles and resolution of clinicalsymptoms. Underdrainage is likely in rarepatients with increased formation rate ofCSF or with very high resistance to CSF out-flow (> 20 mm Hg/ml/min) in whom a flowregulating device, which is able to drain CSFwith a maximal rate 0 3 ml/min, is implanted.

AppendixMATHEMATICAL MODELLING OF CSF CIRCULATIONIn normal conditions, without long term fluctuations ofthe cerebral blood volume, production of CSF and

extemal infusion into CSF spaces is balanced by itsstorage and reabsorption:

Production ofCSF + external infusion =storing ofCSF + reabsorption ofCSF (1)

Production of CSF is almost constant.61-63Reabsorption is proportional to the gradient betweenCSF pressure (p) and pressure in sagittal sinuses (p,,).

Reabsorption = P- PR

(2)

Coefficient R is named the resistance to CSF reab-sorption or outflow (units: mm Hg/ml/min). Normalvalue of this resistance is 6-10 mm Hg/ml/min.612 Aresistance above 12-13 mm Hg/ml/min can beinterpreted as raised.710111416 Storing of CSF is pro-portional to the cerebrospinal compliance C (units:mm Hg/ml/min).

Storing = C. dp

dt(3)

The compliance of the cerebrospinal space isinversely proportional to the gradient of CSF pressure

Po (4).3435

1

CE. (p- Po)

(4)

Some authors suggest that relation (4) is valid onlyabove the certain pressure level named the "opitmalpressure"3637 64; however, this is still a point of somedispute. Coefficient E is the cerebral elasticity (units:ml-l). Elevated elasticity (> 0-18 ml-l) signifiesa poor pressure-volume compensatory reserve." 1221The reference pressure po is a parameter of uncertain

0)

IEE0-- OC

50)

IEEm

0)

m

10

IEE0-

Z:

B

555

W--*Y-N



j.com/

J Neurol N


ay 1996. Dow

nloaded from



Time (20 min period)

8

7

4

3

2

i++ * . . * . w. 4 #* |w;~.~.

I dm .*.1...................2 3. 4. 5. 6 7.;y (ml) .. ... ..

..

l............ ................... ................... ..................

1 2 3 ~~~~45 6 7

A,y (ml)

D

I

E

a-CL

2

0

I.1.. In

I a Ion. . .'.-

.Is .7

Pa

Popt :'d,'

,...;,,iI-'

II,>./Z

20

P (mm Hg)

Figure 9 Different techniques usedfor the analysis of the pressure-volume tests. (A) Example ofCSF pressure recorded during an infusion test and a

modelling curve described by (6). The least mean square fit of the analytical curve to the real pressure enables an accurate measurement of thecompensatory parameters. (B) Example of the pressure response to the bolus injection of saline. Pressure is increasingfrom Pb to Pp after a very rapidaddition of volume (1). (C) Real volume-pressure relation detected during an infusion test. Volume load (1) is calculated as the balance of the volumeinfused, produced, and reabsorbed during infusion. Pressure rise (y axis) is expressed as the ratio of actual pressure and baseline pressure. (D) Relationbetween pulse amplitude (AMP) and mean pressure (P) recorded during the infusion test. This is one case in which the lower breakpoint, marked as P,,,was detected. In such a case all equations (4-8) are valid only for P > P,,p, which is taken into account in the computer calculations (such a breakpointcan be detected in 1-3% of infusion tests).

significance. Some authors suggest that it is a pressure inthe venous compartment and may be equal to ps.37Others assume that this variable can be neglected.3435The relation (4) reflects the most important law of

the cerebrospinal dynamic compensation: the compli-ance of the brain decreases when the CSF pressureincreases.

Combination of (1) with (2) and (4) gives a finalequation (5):

1 dp + P-P+ = I(t) (5)

E. (p-po) dt R

where I(t) is the rate of external volume addition and Pbis a baseline pressure.

Equation (5) can be solved for various types of exter-nal volume additions I(t). The most common in clini-cal practice is:

(a) a constant infusion of CSF (I(t) = 0 fort < 0 and I(t) = Imf for t> 0): see fig 9A:

|Inf + R |[Pb Po]P(t)= po (6)

PoPo~~1 .[e[ POPoIt

-E |+ Iinf |-t

R LThe analytical curve (6) can be matched to the actualrecording of the pressure during the test, which results inaccurate estimation of unknown parameters: R, E, and

PO.

(b) A bolus injection of CSF (volume A V): see fig 8:

E [aV+ PR Pt]pl3 ) (P=Cb Po) e

P (7)E pb- PoE. PoPot

I EAV R1

The bolus injection can be used for calculation of theso-called pressure-volume index (PVI), having an

intepretation of the volume added externally to pro-duce the tenfold increase in the pressure3435:

def AV 1PVI = PVI

Pp-Po 0 434. Elog'o Ph PO

(8)

PVI is theoretically an inverse of the brain elasticity E.The pressure-volume compensatory reserve is insuffi-cient when PVI < 13 ml. The value ofPVI above 26 mlsignifies an "overcompliant" brain.The formula (7) for time t = 0 describes the shape of

the relationship between the effective volume increaseA V and the CSF pressure, called the pressure-volumecurve (see fig 9C):

P = Pb- po). eEA V+ p (9)

Finally, the equation (7) can be helpful in theoreticalevaluation of the relationship between the pulse wave

amplitude of CSF pressure and the mean CSF pres-sure. If we presume that the rise of the blood volumeafter a heart contraction is equivalent to a rapid bolus

A

I

556

1-( LI/

I

I

-i



j.com/

J Neurol N


ay 1996. Dow

nloaded from



addition of CSF fluid at the baseline pressure Pb, thepulse amplitude (AMP) can be expressed as:

AMP=pP-Pb=(Pb-pO) .(e _ 1) (10)In almost all cases, when the CSF pressure is being

increased by extemal volume addition the pulse ampli-tude is rising. The gradient of the regression linebetween AMP and p is proportional to the elasticity.The intercept, theoretically, marks the reference pres-sure po. The mathematical modelling of CSF dynamicshas several applications. In all pressure-volume testingtechniques model (5) is identified using variousalgorithms and various volume-adding techniques:

6566 rtpressure-controlled servoinfusion, constant rateinfusion,933 bolus injection,203536 multiple rate perfu-sion,54 computer controlled drainage,67 etc.The presented model has, however, limited scope: it

cannot interpret dynamic interactions between the ris-ing CSF pressure, expanding ventricles, and cerebralblood volume. More complex,68 or even multicompart-mental (Blood-CSF), dynamic models are helpful insimulation of such phenomena.69-7'

Many thanks to Mr Peter Spittaler for valuable comments andcritical revision and to the following colleagues from hospitalsin Poland, Denmark, and the United Kingdom who con-tributed to refining of the computerised infusion test and itsclinical verification: Dr L Batorski, Dr SE Borgesen, Professor FGjerris, Dr E Guazzo, Dr A Hong, Dr W Maksymowicz, Dr PKirkpatrick, Dr W Koszewski, Dr P Laniewski, Dr NTisavipat, Dr J Tomaszewski, Dr A Walencik.MC arid PS are on leave from Warsaw University of

Technology, Poland.

1 Schmidt JF, Andersen AR, Paulson OB, Gjerris F.Angiotensin converting enzyme inhibition, CBF autoreg-ulation, and ICP in patients with normal pressure hydro-cephalus. Acta Neurochir (Wien) 1990;106:9-12.

2 Vorstrup S, Christensen J, Gjerris, et al. Cerebral bloodflow in patients with normal-pressure hydrocephalusbefore and after shunting. J Neurosurg 1987;66:379-87.

3 Jagust WJ, Friedland RP, Budinger TF. Positron emissiontomography with ["'Efluorodeoxyglucose differentiatesnormal pressure hydrocephalus from Alzheimer typedementia. J Neurol Neurosurg Psychiatry 1985;48: 109-16.

4 McIntyre AW, Emsley RA. Shoplifting associated with nor-mal pressure hydrocephalus: report of a case. Jf GeriatrPsychiatry Neurol 1990;3:229-30.

5 Graff J, Radford NR, Rezai K, et al. Regional cerebralblood flow in normal pressure hydrocephalus. Jf NeurolNeurosurg Psychiatry 1987;50: 158-996.

6 Albeck MJ, Borgesen SE, Gjerris F, Schmidt JF, SorensenPS. Intracranial pressure and cerebrospinal fluid outflowconductance in healthy subjects. Jf Neurosurg 1991;74:597-600.

7 Borgesen SE, Gjerris F. The predictive value of conduc-tance to outflow of CSF in normal pressure hydro-cephalus. Brain 1982;105:65-86.

8 Morgan M, Johnston IH, Spittaler PJ. A ventricular infu-sion technique for the evaluation of treated and untreatedhydrocephalus. Neurosurgery 1991;26:832-7.

9 Lundar T, Nomes H. Determination of ventricular fluidoutflow resistance in patients with ventriculomegaly. JNeurol Neurosurg Psychiatry 1990;53:896-8.

10 Costabile G, Probst C. Intrathecal infusion test anddecrease in shunt revisions and infections. Neurochirurgia(Stuttg) 1988;31:134-5.

11 Czosnyka M, Batorski L, Roszkowski M, et al. Cerebro-spinal compensation in hydrocephalic children. ChildsNerv Syst 1993;9:17-22.

12 Ekstedt J. CSF hydrodynamic studies in man. Method ofconstant pressure CSF infusion. Jf Neurol NeurosurgPsychiatry 1977;40: 105-19

13 Malm J, Kristensen B, Karlsson T, Fagerlund M, ElfversonJ, Ekstedt J. The predictive value of CSF dynamic tests inpatients with the idiopathic hydrocephalus syndrome.Arch Neurol 1995;52:783-9.

14 Maksymowicz W, Czosnyka M, Zabolotny W, KoszewskiW. The communicating hydrocephalus versus brain atro-phy-is a reliable differentiation possible? In: Nagai H,Kamiya K, Ishii S, eds. Intracranial pressure IX. Berlin:Springer-Verlag, 1994:568-9.

15 Shapiro K, Fried A. The theoretical requirements of shuntdesign as determined by biomechanical testing in pedi-atric hydrocephalus. Childs Nerv Syst 1988;4:348-53.

16 Price DJ. Attempts to predict the probablity of clinicalimprovement following shunting of patients with pre-sumed normal pressure hydrocephalus. In: HoffJT, BetzAL, eds. Intracranialpressure VII. Berlin: Springer-Verlag,1989:390-93.

17 Gjerris F, Borgesen SE, Sorensen PS, et al. Resistance tocerebrospinal fluid outflow and intracranial pressure inpatients with hydrocephalus after subarachnoid haemor-rhage. Acta Neurochir (Wien) 1987;88:79-86.

18 Pickard JD, Newton H, Greene A, et al. A prospectivestudy of idiopathic normal pressure hydrocephalus-

guidlines for outpatient investigation. In: Proceedings ofthe Society of British Neurological Surgeons with theNew England Neurosurgical Society. J Neurol NeurosurgPsychiatry 1992;55:513-21.

19 Delwell EJ, de Jong DA, Avezaat CJJ. The relative prognos-tic value of CSF outflow resistance measurement inshunting for normal pressure hydrocephalus. In: AvezaatCJJ, van Eijndhoven JHM, Maas AIR, Tans JTJ, eds.Intracranial pressure VIII. Berlin: Springer-Verlag, 1993:816-20.

20 Kosteljanetz M, Nehen AM, Kaalund J. Cerebrospinalfluid outflow resistance measurements in the selection ofpatients for shunt surgery in the normal pressure hydro-cephalus syndrome. A controlled trial. Acta Neurochir(Wien) 1990;104:48-53.

21 Maksymowicz W, Czosnyka M, Koszewski W, et al. Therole of cerebrospinal system compensatory parameters inestimation of functioning of implanted shunt system inpatients with communicating hydrocephalus. ActaNeurochir (Wien) 1989:101:112-16.

22 Bradley WG Jr, Whittemore AR, Kortman KE, et al.Marked cerebrospinal fluid void: indicator of successfulshunt in patients with suspected normal pressure hydro-cephalus. Radiology 1991;178:459-66.

23 Foltz EL, Blanks JP, Yonemura K. CSF pulsatility inhydrocephalus: respiratory effect on pulse wave slope asan indicator of intracranial compliance. Neurol Res 1990;12:67-74.

24 Chopp M, Portnoy HD. System analysis of intracranialpressure. Comparison of volume-pressure test and CSF-pulse amplitude analysis. J Neurosurg 1980;53:516-27.

25 Drake JM, Saint-Rose CH, eds. Shunt complications. In:The shunt book. Oxford: Blackwell Science, 1995:23-92.

26 Czosnyka M, Maksymowicz W, Batorski L, Koszewski W.Comparison between classic differential and automaticshunt functioning on the basis of infusion tests. ActaNeurochir (Wein) 1990;106: 1-8.

27 Maksymowicz W, Czosnyka M, Koszewski W, Szymanska0, Zabolotny W. Post shunting improvement in hydro-cephalic patients described by cerebrospinal compen-satory parameters. In: Avezaat CJJ, van EijndhovenJHM, Maas AIR, Tans JTJ, eds. Intracranialpressure VIII.Berlin: Springer-Verlag, 1994:829-32.

28 Sprung C, Collman H, Fuchs EC, et al. Pre- and postoper-ative evaluation of hydrocephalus using the infusion test.Adv Neurosurg 1977;4:463-70.

29 Reilly PL, Savage JP, Doecke L. Isotope transport studiesand shunt pressure measurements as a guide to shuntfunction. BrJ Neurosurg 1989;3:681-90.

30 Seppanen U, Serlo W, Saukkonen AL. Valvography in theassessment of hydrocephalus shunt function in children.Neuroradiology 1987;29:53-7.

31 Hara M, Kadowaki C, Konishi Y. A new method for mea-suring cerebrospinal fluid flow in shunts. J Neurosurg1983;58:557-61.

32 Czosnyka M,Batorski L,Laniewski P, et al. A computer sys-tem for the identification of the cerebrospinal compen-satory model. Acta Neurochir (Wein) 1990;105:112-16.

33 Katzman R, Hussey F. A simple constant infusion mano-metric test for measurement of CSF absorption.Neurology (Minneap) 1970;20:534-44.

34 Marmarou A. A theoretical model and experimental evaluationof the cerebrospinal fluid system [thesis]. Philadelphia, PA:Drexel University, 1973.

35 Marmarou A, Shulman K, Rosende RM. A non-linearanalysis of CSF system and intracranial pressure dynam-ics. JNeurosurg 1978;48:332-44.

36 Szewczykowski J, Sliwka S, Kunicki A, et al. A fast methodof estimating the elastance of intracranial system. JNeurosurg 1977;47: 19-25.

37 Avezaat CJJ, Eijndhoven JHM. Cerebrospinalfluid pulse pres-sure and craniospinal dynamics. A theoretical, clinical andexperimental study [thesis]. The Hague: Jongbloed A,1984.

38 Frieden H, Eksted J. Estimation of CSF outflow resistance:model validation and resistance calculation by themethod of CSF volume accounting. In: Gjerris F,Borgesen SE, Sorensen PS, eds. Outflow of cerebrospinalfluid. Copenhagen: Munksgaard, 1989:198-210.

39 Sliwka S: Static and dynamic cerebrospinal elastance-clin-ical verification. In: Miller JD, Teasdale G, Rowan JO,Galbraith SL, Mendelow AD, eds. Intracranial pressureVI. Berlin: Springer-Verlag, 1986:84-8.

40 Bolander HG. A new method for longterm lumbar pressuremonitoring with a fiber optic catheter. Acta Neurochir(Wien) 1990;105:135-9.

41 Cardoso ER, Piatek D, Del Bigio MR, Stambrook M,Sutherland JB. Quantification of abnormal intracranialpressure waves and isotope cisternography for diagnosisof occult communicating hydrocephalus. Surg Neurol1989;31:20-7.

42 Gjerris F, Borgesen SE. Pathophysiology of CSF circula-tion. In: Neurosurgery. The scientific basis of clinical practice.Crockard A, Hayward A, HoffJT, eds. Oxford: BlackwellScientific, 1992:146-74

43 Pickard JD, Teasdale G, Matheson M, et al. Intra-ventricular pressure waves-the best predictive test forshunting in normal pressure hydrocephalus. In: ShulmanK, Marmarou A, Miller JD, Becker DP, Hochwold GM,Brock M, eds. Intracranial pressure IV. Berlin: Springer-Verlag, 1980:498-500.

44 Klingelhofer J, Sander D. Dynamics in cerebral blood flowvelocities in sleep. Journal of Cardiovascular Technology1995;97: 142-8.

557 on F


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



45 Newell DW, Aaslid R, Stoss R, Reulen HJ. The relation-ship of blood flow velocity fluctuations to intracranialpressure B waves. J Neurosurg 1992:76:415-2 1.

46 Pickard JD, Czosnyka M. Management of raised intracra-nial presure. J Neurol Neurosurg Psychiatry 1993;56:845-58.

47 Lemaire JJ, Chazal J, Gutknecht JL, Picard P, Irthum B,Boire JY. Effects of acute compliance fluctuation on slowICP waves: frequential aspects. In: Nagai H, Kamiya K,Ishii S, eds. Intracranial pressure IX. Berlin: Springer-Verlag, 1994:498-501

48 Whitehouse H, Czosnyka M, Smielewski P, Pickard JD.Blood flow velocity in the middle cerebral artery duringlumbar CSF infusion tests in patients with ventriculardilatation. In: Nagai H, Kamiya K, Ishii S, eds. Intra-cranial pressure IX. Berlin: Springer-Verlag, 1994:512-13

49 Hara K, Nakatani S. Detection of the B waves in the oscila-tion of intracranial pressure by fast Fourier transform.Med Inf (Lond) 1990;15:125-31.

50 Saint-Rose C, Hooven MD, Hirsch JF. A new approach inthe treatment of hydrocephalus. J Neurosurg 1987;66:213-26.

51 Aschoff A, Kremer P, Benesch C, et al. Overdrainage andshunt technology. Childs Nerv Syst 1995;11: 193-202.

52 Drake JM, Saint-Rose CH, eds. Shunt complications. In:The shunt book. Oxford: Blackwell Science, 1995.

53 Borgesen SE, Albeck MJ, Gjerris F, Czosnyka M.Computerized infusion test compared to conventionallumbar-ventricular perfusion for measurement of resis-tance to CSF outflow. Acta Neurochir (Wein) 1992;119:12-16.

54 Borgesen SE, Gjerris F, Sorensen SC. The resistance tocerebrospinal fluid absorption in humans. A method ofevaluation by lumbo-ventricular perfusion, with particu-lar reference to normal pressure hydrocephalus. ActaNeurol Scand 1978;57:88-96.

55 Graff J, Radford NR, Godersky JC, Jones MP. Variablespredicting surgical outcome in symptomatic hydro-cephalus in the elderly. Neurology 1989;39:160-14.

56 Chapman PH, Cosman ER, Arnold MA. The relationshipbetween ventricular fluid pressure and body position innormal subjects and subjects with shunts: a telemetricstudy. Neurosurgery 1990;26:1819.

57 Pudenz RH, Foltz EL. Hydrocephalus: overdrainage byventricular shunts. A review and recommendations. SurgNeurol 1991;35:200-12.

58 Serlo W, Saukkonen AL, Heikkinen E, von Wendt L. Theincidence and management of the slit ventricle syn-drome. Acta Neurochir (Wien) 1989;99:113-6.

59 Epstein F, Lapras C, WisoffJH. "Slit ventricle syndrome":etiology and treatment. Pediatr Neurosci 1988;14:5-10.

60 Foltz EL, Blanks JP. Symptomatic low intracranial pressurein shunted hydrocephalus. J Neurosurg 1988;68:401-8.

61 Davson H, Welch K, Segal MB. Physiology and patho-physiology of cerebrospinal fluid. Edinburgh: ChurchillLivingstone 1987:189-246.

62 Davson H. Formation and drainage of the CSF in hydro-cephalus. In: Shapiro K, Marmarou A, Portnoy H, eds.Hydrocephalus. New York: Raven Press, 1984:112-60.

63 McComb JG. Recent research into the nature of cere-brospinal fluid formation and absorption. I Neurosurg1983;59:369-83.

64 Szewczykowski J, Dytko P, Kunicki A, et al. Method of esti-mating intracranial decompensation in man. _7 Neurosurg1976;45: 155-63.

65 Sklar FH, Beyer CW, Ramanathan M, et al. Servo-con-trolled lumbar infusions: a clinical tool for determinationof CSF dynamics as a function of pressure. Neurosurgery1978;3: 170-8.

66 Frieden H, Ekstedt J. Instrumentation for cerebrospinalfluid hydrodynamic studies in man. Med Biol Eng Comput1982;20: 167-180.

67 Mielewski P, Czosnyka M, Roszkowski M, Walencik A.Identification of the cerebrospinal compensatory mecha-nisms via computer controlled drainage of CSF. ChildsNerv Syst 1995;11:297-300.

68 Rekate HL, Brodkey JA, Chizeck HJ, el Sakka W, Ko WH.Ventricular volume regulation: a mathematical modeland computer simulation. Pediatr Neurosci 1988;14:77-84.

69 Hoffmann 0. CSF dynamimics: integration of pulsatorycomponents and autoregulation into mathematicalmodel. In: Ishii S, Nagai H, Brock M, eds. Intracranialpressure V. Berlin: Springer-Verlag, 1983:169-73.

70 Ursino M, Di Giammarco P. A mathematical model of therelationship between cerebral blood volume and intra-cranial pressure changes: the generation of plateau waves.Ann Biomed Eng 199 1; 19:15-25.

71 Czosnyka M, Pickard J, Whitehouse H, Piechnik S. Thehypearemic response to a transient reduction in cerebralperfusion pressure-a modelling study. Acta Neurochir(Wien) 1992;115:90-7.

St Apollonia (died about AD 249)

During an Alexandrian riot against the Christians,Apollonia, an elderly deaconess, was repeatedly struck inthe face. Her teeth were torn out and her jaws werebroken but she refused to renounce her Christian faith. Abonfire was built. Apollonia, threatened with being burntalive, was asked once more to return to the Roman gods.Instead, she uttered a short prayer and walked into theflames. Apollonia was canonised about 50 years after herdeath. She became the patron saint of toothache and thepatroness of the dental profession.A stamp issued in 1979 by San Marino for the 13th

stomatological conference (Stanley Gibbons 1114, Scott959) reproduces a 14th century wood carving which ispart of the Wessler Collection at the Royal School ofDentistry in Stockholm. St Apollonia is shown with onehand laid on the swollen cheek of a toothache sufferer.Her insignia, a tooth held by forceps, is shown in herother hand.

L F HAAS

Db a s

.ft_.

a "-tW Lk l> Sib tr F

> os% i

^ .*^

w

w .iD _

^is,_

i @_@ WW ^*i r

; aa .:

Q 'i

be ^..^^.:^|||^..

II O Js~~~~~~~~~~~~~~~~~~~~~~~~~~~~:- . d~~~~~~~~~~~~~~~~~~~~~~~'

.'l

NEUROLOGICAL STAMP

558 on F


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


testingof patients: guide to - bmj · the computerised infusion test32 is a method forthe accurate...

Documents