Cerebrospinal pulse pressure and relationships · J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.8.687 on 1 August 1979. Downloaded from Cerebrospinal fluidpulse

Journal ofNeurology, Neurosurgery, and Psychiatry, 1979, 42, 687-70 0

Cerebrospinal fluid pulse pressure and intracranialvolume-pressure relationshipsC. J. J. AVEZAAT, J. H. M. VAN EIJNDHOVEN,AND D. J. WYPER

From the Departments of Neurosurgery and Electro-neurology, Academic Hospital Rotterdam-Dijkzigt, Erasmus University, Rotterdam, the Netherlands, and the Department of Clinical Physics,Institute of Neurological Sciences, Glasgow

S U MM AR Y In six anaesthetised and ventilated dogs the CSF pulse pressure was compared withthe volume-pressure response (VPR) during continuous inflation of an extradural balloon. Bothpulse pressure and VPR increased linearly with the ventricular fluid pressure (VFP) up to a

mean VFP of 60 mmHg. At this pressure a breakpoint occurred above which the CSF pulsepressure showed a steeper linear increase, while the VPR remained constant. It is suggestedthat the breakpoint is related to failure of autoregulation, and that in non-autoregulating patientsthe CSF pulse pressure is a better parameter of the clinical state than the VPR.

The rapid fluctuations in the cerebrospinal fluid(CSF) pressure synchronous with the heart beat(CSF pulse) have for a long time excited theinterest of numerous workers. This interest con-cerns the origin, site of transfer, configuration, andmagnitude of the pulse as well as its role in thepathogenesis and diagnosis of intracranial disease.The subject of the present study is the amplitude ofthe CSF pulse wave, hereafter termed the CSFpulse pressure. Table 1 contains a list of definitionsof abbreviations and symbols used.

It is an old observation that the CSF pulsepressure increases concomitantly with the intra-cranial pressure (ICP) (Leyden, 1866; Mosso,1881). This phenomenon can be explained by theexponential nature of the relationship betweenvolume and pressure within the craniospinal com-partment (Marmarou et al., 1975). With risingICP, rapid intracranial volume changes such asthose produced by the transient increase in cerebralblood volume (CBV) during each cardiac cyclewill result in increasingly larger pressure changes(Fig. 1). From this concept the question ariseswhether the CSF pulse pressure can be used as adiagnostic tool to assess the intracranial volume-pressure relationships. All hitherto describedvolume-pressure parameters require volume ad-

Address for reprint requests: Dr C. J. J. Avezaat, Department ofNeurosurgery, Academisch Ziekenhuis Rotterdam-Dijkzigt, ErasmusUniversity, Dr Molewaterplein 40, Rotterdam 3002, the Netherlands.Accepted 7 February 1979

Table 1 Definitions of abbreviations and symbols

Abbreviationl Definitionsymbol

ICP Intracranial pressureVFP Ventricular fluid pressureCMP Cisterna magna pressureSAP Systemic arterial pressureA SAP Systemic arterial pulse pressureCPP Cerebral perfusion pressurePeq Equilibrium pressure, physiological steady state ICPA P CSF pulse pressure or volume-pressure responseV Change in total volume of the craniospinal

compartmentVeq Equilibrium volume of the craniospinal compartment

at which ICP= PeqA V Injected volume during a volume-pressure test or the

transient increase in CBV per cardiac cycleCBV Cerebral blood volumeVPT Volume-pressure testVPR Volume-pressure responseE Elastance of the craniospinal compartmentE, Elastance coefficientPVI Pressure-volume indexCBF Cerebral blood flowCVR Cerebral vascular resistanceHR Heart rate

ditions to the CSF compartment (Shulman andMarmarou, 1971; Miller and Garibi, 1972; Lofgrenet al., 1973). The most widely used parameterclinically is the volume-pressure response (VPR)which has been defined as the immediate change inICP resulting from a uniform change in ventri-cular CSF volume (volume-pressure test; Millerand Garibi, 1972). The VPR is a measure of theintracranial elastance-that is, the instantaneous

687

Protected by copyright.

on July 5, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.42.8.687 on 1 A

ugust 1979. Dow

nloaded from

http://jnnp.bmj.com/

C. J. J. Avezaat, J. H. M. van Eijndhoven, and D. J. Wyper

Fig. I The CSF pulse pressureis a pressure response (AP) tothe transient increase inintracranial blood volumeduring a cardiac cycle (AV).The exponential shape of theintracranial volume-pressurecurve explains why the CSFpulse pressure increases withrising intracranial pressure(ICP). Abbreviations definedin Table 1.

Iv IvVolume (ml)

craniospinal comportment

slope (dP/dV) of the exponential volume-pressurecurve. The VPR has been claimed to show a bettercorrelation with the clinical state of the patientthan the magnitude of ICP alone (Miller et al.,1973; Miller and Pickard, 1974). However, someobjections against the volume-pressure test (VPT)can be raised. The test increases infection hazardsand cannot be used in epidural ICP measurements.In clinical practice the VPR shows rather a widevariability necessitating the use of multiple tests.Pressure waves may be provoked, while thealternative, aspiration of CSF, may lead toblockage of the ventricular catheter.These disadvantages led us to devise a new

method of evaluating intracranial volume-pressurerelationships by using the CSF pulse pressure inrelation to the ICP as an index (Avezaat et al.,1976), and others have been working along thesame lines (Castel and Cohadon, 1976; Norneset al., 1977; Szewczykowski et al., 1977). Thismethod was based on the analogy between theCSF pulse pressure and the VPR, both beingpressure responses to rapid changes in intra-cranial volume. In a group of patients the CSFpulse pressure was found to be related linearlyto the ICP. However, since the CSF pulse pressureunderlying volume change varied considerablybetween patients, no correlation between CSFpulse pressure and VPR could be demonstrated.

In order to study more precisely and under con-trolled conditions the concept of the CSF pulsepressure as a parameter of intracranial volume-pressure relationships, a series of experiments wasdesigned in dogs. During rising ICP the relation-ships of the CSF pulse pressure and the VPR tothe ICP were studied and compared with eachother. In another set of experiments the influenceon these relationships of variables affecting intra-cranial haemodynamics, such as blood pressureand arterial Pco2, was investigated. These resultswill be discussed in a subsequent paper.

Methods

ANALYTICAL METHOD

The intracranial volume-pressure relationship re-flects the elastic properties of the craniospinalsystem, if the pressure inside the system is com-pared with the total volume of the system or,since the total volume is not accessible to directmeasurement, with additions to this volume. Mostauthors have approximated the volume-pressurecurve by an exponential function (Ryder et al.,1953; L6fgren et al., 1973; Marmarou, 1973). Thisfunction was, therefore, chosen as the most simplemodel to proceed from (Fig. 1):

P =PeqeEiV (1)

ICP

PeqVeq

688



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


Cerebrospinal fluid pulse pressure and intracranial volume-pressure relationships

where P=ICP, P.,qa constant equal to the ICPat the equilibrium point (the normal, physiologicalsteady state ICP), E1=elastance coefficient, andV= change in total volume of the craniospinalcontents with regard to the equilibrium volume(Veq). The elastance (E) of the craniospinal systemdefined as the slope of the volume-pressure curveis the derivative of Equation 1:

E =dp/dV =ElPeqeEiV =ElP (2)So El determines the elastance at a given pressureand is, therefore, termed the elastance coefficient.E1 is the inverse of the pressure-volume index(PVI) defined by Shulman and Marmarou (1971).When the intracranial volume increases by an

amount AV, the intracranial pressure increases bya pressure response AP:

P+AP =Peqe[Ei(v+Av)] (3)Using equation 1 this equation can be rewrittenas:

AP=P [eEAV-1] (4)Equation 4 is applicable to the relationships ofboth the VPR and the CSF pulse pressure to theICP if AV is defined as the injected volume duringa volume-pressure test and as the transient in-crease in intracranial blood volume per cardiaccycle respectively. It is the equation of a straightline if E1 and AV are constant. The elastancecoefficient is constant if the exponential volume-pressure model is valid. Since in the case of thevolume-pressure test, AV is constant by definition,a linear correlation between VPR and ICP oughtto be accepted as proof of a mono-exponentialvolume-pressure relationship. The elastancecoefficient can be calculated using equation 4:

El 1=I AP 1 (5)

where AP/P is the slope of the VPR versus ICPrelationship.With regard to the CSF pulse pressure, however,

AV is unknown. Here, the volume change equalsthe transient increase in CBV per cardiac cycleand is caused by the difference in pulsatile shapeof the intracranial arterial inflow and venous out-flow of blood. This concept is illustrated in Fig. 2.If CSF pulse pressure and ICP are relatedlinearly it can be inferred that, in case of a con-stant E1, the change in CBV (AV) is also constant.It can be calculated using the following equation,derived from equation 4:

AVE=Iln [AP 1 (6)

where AP/P is the slope of the CSF pulse pressure-ICP relationship.

orterialinflow

+ t m AVmax

V,,venousoutflow

cardiac cycle cardiac cycle cardiac cycle

(a) (b) (c)

Fig. 2 Diagram illustrating the concept of thetransient increase in intracranial blood volume percardiac cycle (AV). The cerebral arterial inflow andvenous outflow curves are scheematically drawn; thechange in cerebral blood volume is represented bythe dashed curve. The magnitude of AV is determinedby the interaction between inflow and outflow. Ifthe flow profiles were identical, the resulting volumechange would be nil. (a) If inflow and outflow werecompletely separated, AV would be maximal.(b) Inflow and outflow, however, have differentpulsatile flow patterns, thus producing an intermediatevolume change (c).

In the experimental study the CSF pulse pressurewill be compared with the VPR, so that con-clusions can be drawn on AV and its behaviourwith rising ICP.

EXPERIMENTAL METHODPreparation and measurementsThe experiments were performed on six adultmongrel dogs of both sexes, weighing from 10 to19 kg. Anaesthesia was induced with intra-venously administered thiopentone (30 mg/kg),and maintained with a mixture of nitrous oxideand oxygen (2: 1) supplemented by chloralose(50 mg/kg) intravenously. The animals wereventilated using a modified Starling pump, thestroke volume of which was adjusted to maintainnormocapnia (PaCO2 37-43 mmHg). Body tem-perature was maintained between 360C and 380Cby heating lamps. Throughout the experimentnormal saline was slowly infused intravenously.The animals were placed in a stereotaxic frame

in the sphinx position. The scalp and temporalmuscles were reflected. A balloon for cerebralcompression, fashioned from glove rubber, wasplaced epidurally through a burrhole in the rightfrontotemporal region and connected to a variablespeed infusion pump (Sage Instruments, model351). The ventricular fluid pressure (VFP) wasmeasured bilaterally from needles (inside diameter

689

AV



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



1.5 mm) introduced with the continuous infusionmethod into the lateral ventricles through smallburrholes in the parietal region. All bone defectswere sealed with acrylic dental cement. Thecisterna magna pressure (CMP) was measuredthrough a needle inserted percutaneously into thecisterna magna. Systemic arterial pressure (SAP)was monitored from an aortic catheter introducedvia the femoral artery. Central venous pressure(CVP) was measured in the inferior vena cavathrough catheterisation of the femoral vein. Theelectrocardiogram (ECG) was recorded via needleelectrodes placed in the limbs of the animal. Allpressures were measured continuously by pressuretransducers (Bell and Howell) placed at the levelof the right atrium. The frequency response ofthe ventricular needle to manometer system wasflat up to 15 Hz with resonance frequency of42 Hz and a damping ratio of 0.25. The varioussignals were displayed on two chart recorders(Devices M4 and M2) and stored on a tape re-corder (Philips Analog 7). Arterial blood gaseswere measured at 30 minute intervals through-out the experiment using a direct reading electrodesystem (Radiometer BMS 3), and values werecorrected for temperature differences between theanimal and recording system (Severinghaus, 1966).

EXPERIMENTAL PROTOCOL AND DATA COLLECTION

After all pressures had stabilised the extraduralballoon was distended by continuous inflation withnormal saline at a rate of 1.0 ml per 15 minutes.As a method of acquiring volume-pressure datathe volume-pressure test (VPT) as described byMiller et al. (1973) was used. At intervals of fiveminutes a cluster of four tests was done by rapidinjections of 0.05 ml of normal saline into the leftventricle using a three-way stopcock interposedin the tubing connecting the ventricular needle tothe pressure transducer. The VPR was read fromthe injected ventricle and calculated from theimmediate change in the mean VFP in correspond-ing parts of the VFP respiratory cycle. All meanpressures including SAP were calculated from thediastolic pressure plus one-third of the pulsepressure. In order to reduce variability from errorsinherent in the manoeuvre, the mean VPR of eachcluster of four tests was taken and related to themean preinjection VFP (mean baseline VFP).After completion of the four tests the VFP alwaysreturned to the steady state level before the nexttime interval was reached.During the volume-pressure tests the speed of

the chart recorder was increased, so that theamplitude of the CSF pulse could be distinguished

clearly. Eight pulse pressures were calculated bytaking the average value of the pulse pressuresover the last respiratory cycle before and the firstcycle after each VPT. The means of the fourpulse pressures before and after each test wererelated to the mean baseline and peak VFPrespectively. As a result of this procedure eachtime interval yielded one VPR and two pulsepressures.When changes in SAP occurred, as they some-

times did in the advanced stages of cerebral com-pression, no special measures were taken to keepthe SAP at a constant level. However, sincechanges in SAP are known to affect both theintracranial volume-pressure relationship (Lof-gren, 1973; Leech and Miller, 1974b) and theCSF pulse pressure, all values of both VPR andpulse pressure obtained at levels of SAP deviatingmore than 15% from the mean SAP werediscarded.The experiment was terminated when the VFP

had reached the level of the blood pressure, exceptfor three animals in which it was continued farbeyond this level. The brains were transected atthe foramen magnum (brain weights between 72and 92 g), fixed in formalin and sectioned coronallyat 5 mm intervals for pathology. Proper placementof the ventricular needles was verified and majorbrain lesions from surgical preparation wereexcluded.

Results

RELATIONSHIP BETWEEN VOLUME-PRESSURERESPONSE AND VENTRICULAR FLUID PRESSUREIn the plots of VPR versus VFP (Fig. 3) threemore or less clearly definable zones could beobserved. At first the VPR increased linearly withthe VFP. Next the VPR levelled off or even de-creased, and finally at the advanced stage ofcerebral compression the response started to riseagain, but more rapidly than before. The re-gression equations of the first zone are given inTable 2. These linear relationships are in accord-ance with the mono-exponential volume-pressuremodel, implying a constant elastance coefficient(E1). E1 was calculated from the slopes of theregression lines using equation 5 and the resultsare given in Table 2. The data with regard to thebreakpoint between first and second zone aregiven in Table 3. It occurred at a mean VFP of60 mmHg, but with a considerable variation be-tween the animals. There was a positive linearcorrelation between VFP and SAP at the break-point (r=0.82, P<0.05), and consequently the

690



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



Volume PressureResponse(mmHg)

dog 1

40 80 120

dog 2

dog 5

0xZ-11,~~~

160 40 80 120

dog 3

7~~~~~~~~

*00/ .-:.*O.

dog 6

40 80 120 160

Ventricular Fluid Pressure (mmHg)Fig. 3 Plots of the volume-pressure response (VPR) against the ventricular fluid pressure

(VFP) during continuous inflation of an extradural balloon in six animals. Three differentzones can be recognised. In the first zone the VPR increases linearly with the VFP. Theregression lines are drawn and the equations are given in Table 2. In the second zone theVPR remains constant and/or decreases, and finally the VPR increases again but more rapidlythan in the first zone.

Table 2 Relationship between volume-pressure response (VPR) and ventricular fluid pressure (VFP):regression equations and elastance coefficients (E,) below and mean VPR (±SD) and elastance above thefirst breakpoint

Animal Below first breakpoint Above first breakpoint

Regression n r E, VFP range Mean VPR n Elastance VFP range

equations ( l/m) (mmHg) (mmHg) (mmHg/ml) (mmHg)

1 y=O.lOx- 0.1 18 0.95 1.9 17.1-66.3 5.541.0 8 110 66.3-111.82 y=0.22x-1.5 11 0.98 4.0 13.4-52.0 7.1+0.6 2 142 65.4- 72.53 y=0.08x+0.7 6 0.97 1.5 14.3-41.6 3.7±0.6 9 74 49.4- 87.24 y=0.10x+0.0 10 0.99 2.0 14.8-57.7 5.7±0.6 5 114 57.7-126.55 y=0.08x+0.8 7 0.89 1.5 24.0-60.1 4.7+0.2 5 94 60.1- 85.06 y=0.07x+0.6 13 0.99 1.3 8.8-81.0 3.2+0.9 5 64 85.3-145.5

Mean y=0.11x+0.1 2.0 15.4-59.8 5.0±1.4 100 63.9-104.8

variation was much narrower if the cerebral per-

fusion pressures (CPP) were considered. No rela-tion could be demonstrated between thebreakpoint and transtentorial herniation (Table 4).Early herniation was defined as the appearance ofa pressure gradient between VFP and CMP. When

CMP did not rise any further and no VPR couldbe detected in the CMP tracing, tentorial obstruc-tion was considered complete.The second zone in the plot of VPR versus VFP

poses a problem with regard to the analysis of thedata. Less data points were available, as the

691

16-

12

8-

4-

12 -

8-

4-

.-

,I,, I,r I

] ~~~dog 4

] .~~~~~~~~~~~~...

I I I



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


C. J. J. Avezaat, J. H. M. van Eiindhoven, and D. J. Wyper

Table 3 Ventricular fluid pressure (VFP), systemic arterial pressure (SAP), cerebral perfusion pressure(CPP), and balloon volume at first and second breakpoints

Animal First breakpoint Second breakpoint

VFP SAP CPP Balloon VFP SAP CPP Balloon(mmHg) (mmHg) (mmHg) volume (ml) (mmHg) (mmHg) (mmHg) volume (ml)

1 66.3 162.3 96.0 5.7 111.8 136.7 24.9 8.32 52.0 128.3 76.3 3.3 - - - -

3 41.6 126.6 85.0 1.7 104.6 116.0 11.4 6.04 57.7 156.0 98.3 3.0 130.3 153.7 23.4 5.05 60.1 134.3 74.2 2.0 112.5 131.3 18.8 5.36 81.0 164.0 83.0 4.0 163.9 165.7 1.8 6.7

Mean 59.8 145.3 85.5 3.3 124.6 140.7 16.1 6.3

Table 4 Ventricular fluid pressure (VFP) and balloonvolume at early and complete tentorial obstruction

Animal Early tentorial Complete tentorialobstruction obstruction

VFP Balloon VFP Balloon(mmHg) volume (ml) (mmHg) volume (ml)

1 52.3 5.3 72.7 6.72 16.3 0.7 24.7 1.73 24.5 1.0 54.3 2.74 35.0 2.3 114.0 4.05 48.5 1.3 79.0 2.76 48.3 2.7 113.0 4.7

Mean 37.5 2.2 76.3 3.8

animals passed more rapidly through this phase ofthe experiment and the blood pressure had some-times fallen below acceptable levels. In addition,the VPR in the injected ventricle was often smallin contrast with a much higher response in theother ventricle, the latter consisting of a highpeak with a very rapid return to the preinjectionlevel (Fig. 4). In those cases there was neverreason to doubt the patency of the ventricularneedle. The high peak response is apparentlymissed in the injected ventricle because of thetime factor involved in turning the three-waystopcock in the recording position. The rapid

return of the pressure to the baseline suggests avery rapid compensation of the volume added. Inview of these findings we believe that the fall inVPR, at least in part, is an artefact produced byfailure of the volume-pressure test to detect rapidspatial compensatory mechanisms. Figure 5 showsthe VPR in both ventricles in one of the experi-mental animals plotted against the VFP, demon-strating that the phenomenon described is respon-sible for the decrease in VPR, but does notexplain the levelling off. Unfortunately, theseobservations could not be made in all animalswith falling VPRs as the recording needle in thenon-injected ventricle was sometimes blocked atthis stage. This VFP range was, therefore, con-ceived as a zone with a constant VPR defined bythe mean VPR (Table 2). The implication is thatthe volume-pressure relationship is no longerexponential but linear in nature (constantelastance), and consequently the mathematicalmodel cannot be applied. The elastance was cal-culated by using E=AP/AV=VPR/0.05 mmHg/ml (Table 2).The data with regard to the second breakpoint

in the VPR-VFP relationship at which the VPRbegan to increase again are given in Table 3. Thisoccurred at a mean VFP of 125 mmHg. In each

(mmHg) VPT 1 VPT 2 VPT 3 VPT 4

110FVFPleft

100- %AkL A0qqAL0 qA.jriv "W

80-

1 10 VFPrih100

80L,oh.1PI ~~~~~~~~~~10s

Fig. 4 Comparison of thevolume-pressure response(VPR) in the injected (uppertracing) and the non-injectedventricle (lower tracing) duringfour volume-pressure tests(VPT) performed at highventricular fluid pressures(VFP) in a single animal. Thehigh pressure spike observed inthe non-injected ventricle ismissed in the other ventriclebecause of the time factorinvolved in turning thethree-way stopcock into therecording position.

692



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



CP

0.

0

00 0

0

0.

Fig. 5 Plot of the volume-pressure response (VPR) againstthe ventricular fluid pressure(VFP) in both lateral ventriclesof a single animal (dog 6),*=the injected, O=non-injected ventricle. The fall ofthe VPR in the injectedventricle above a VFP of80 mmHg is almost absent inthe other ventricle.

20120 160 200

VFP (mm Hg)

animal the breakpoint was related to the level ofthe diastolic blood pressure. In three animalssufficient data were available to calculate theregression lines for the pressure range above thisbreak point (Table 5). The slopes were alwayssteeper than in the first pressure zone.

RELATIONSHIP BETWEEN CSF PULSE PRESSURE AND

VENTRICULAR FLUID PRESSUREThe relationship between CSF pulse pressure andVFP is demonstrated in Fig. 6. The pulse pressureincreased linearly with rising intracranial pressurebut more rapidly in the high pressure range al-though this was not equally distinct in eachanimal. When the VFP approached the level ofthe blood pressure the amplitude of the CSF pulselevelled off, and finally decreased above this level.

Table 5 Regression equations of the relationshipbetween volume-pressure response and ventricularfluid pressure (VFP) above the level of the diastolicblood pressure

Animal Regression n r VFP rangeequations (mmHg)

1 y=0.26x-23.6 4 0.98 111.8-155.33 y=0.27x-15.4 5 0.93 104.6-159.26 y=0.31x-44.7 4 1.00 163.9-207.8

Since the VPR versus VFP plots had shown threepressure zones with different volume-pressurerelationships, we decided to analyse the CSFpulse pressure data for each zone separately. Theregression lines are given in Table 6. In all butone animal there was a significant difference inslope on either side of the breakpoint, the slopealways becoming steeper (F distribution, P<0.01in four animals and P<0.05 in one). Variables withpossible effect on the CSF pulse pressure, such as

systemic arterial pressure, arterial pulse pressure,and heart rate, remained stable throughout theexperiment (Table 7).The linearity of the relationship between CSF

pulse pressure and VFP has a different meaningon either side of the breakpoint if the volume-pressure relationship is taken into consideration.As an example, Fig. 7 shows a comparison of thetwo relationships in one of the animals. Belowthe breakpoint, where the concept of a mono-

exponential volume-pressure curve holds true(constant E1), it means according to equation 4,that the change in CBV per cardiac cycle (AV) isalso constant. Using equation 6, AV was calcu-lated for the various animals (Table 6): AV variedconsiderably, which explains the lack of corres-pondence between the slopes of the CSF pulsepressure versus VFP regression lines and the

VPR(mm Hg)

10 -

8-

6-

4 -

2.

so0so

so* o

8

40 80

693



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



dog 1

CSFpulsepressure

(mmHg)

20-

16 -

12 -

8-

4-

dog 2

/I

dog 3

/

dog 4

/4

40 80 120 160

dog 5

I8/'

40 80120 40 80 120 160

Ventricular Fluid Pressure (mmHg)

Fig. 6 Plots of the CSF pulse pressure against the ventricular fluid pressure (VFP) duringcontinuous inflation of an extradural balloon in six animals. The pulse pressure increaseslinearly with the VFP up to the level of the blood pressure but with a breakpoint. Theregression lines are drawn and the equations are given in Table 6. The slope of the regressionline above the breakpoint was always steeper except for animal 3.

Table 6 Relationship between CSF pulse pressure and ventricular fluid pressure: regression equations,volume change per cardiac cycle (AV) below and increase in AV above the first breakpoint

Animal Below first breakpoint Abovefirst breakpoint

Regression n r V Regression n r Increaseequations (ml) equations in j V

(mllmmHg)

I y=0.09x-0.5 35 0.99 0.047 y=0.45x-27.5 14 0.96 0.00412 y=0.14x-0.9 21 0.99 0.033 y=0.19x- 3.5 5 0.93 0.00133 y=0.13x-0.6 11 0.98 0.082 y=0.13x- 0.3 18 0.97 0.00184 y=0.06x-0.1 19 0.98 0.029 y=O.llx- 4.2 8 0.96 0.00105 y=0.1lx-1.2 13 0.97 0.072 y=0.22x- 8.5 8 0.94 0.00236 y=0.1Ix-1.0 25 0.97 0.081 y=0.17x- 5.3 10 0.92 0.0027

Mean y=0.11x-0.7 0.057 y=0.21x- 8.0 0.0022

20-

16 -

12 -

8-

4-

694



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



Table 7 Systemic arterial pressure (SAP), arterial pulse (ASAP), and heart rate (HR) below and above thefirst breakpoint (M+SD)

Animal Belowfirst breakpoint Above first breakpoint

SAPm JSAPm HR SAPm 4SAPm HR(mmHg) (mmHg) (beats/min) (mmHg) (mmHg) (beats/min)

1 159.5± 8.1 56.3±8.3 114.2±11.4 144.9±11.9 54.146.4 114.0±12.42 126.4± 2.6 39.8±1.3 151.5±10.7 122.0± 2.8 39.5±4.9 134.5± 3.5

3 118.8± 5.8 34.0±1.7 132.0± 3.3 119.6± 4.7 35.1±1.5 134.7± 2.84 151.6± 3.6 46.1±2.2 144.5± 5.3 156.4± 6.2 43.7±2.6 175.0±14.25 135.6± 2.9 53.9±6.6 165.3± 1.6 130.8± 9.3 50.3±2.1 168.3± 2.36 167.8± 1.8 50.8±4.3 168.0± 6.6 158.4± 4.7 46.0±4.2 156.0± 9.5

Mean 143.3±19.4 46.8±8.6 145.9±20.5 138.7±17.0 44.6±7.0 147.1±23.3

VPR 18-(mmHg)

16 -

14 -

12 -

10 -

6 -

4.

2-

CSFpulse

0 pressure(mmHg)

r 32

0

0

0

20 40 60 80 100 120 140 160

VFP (mmHg)

magnitude of the elastance coefficients. Thevariation in AV could not be explained fromdifferences in SAP, arterial pulse pressure andheart rate.Above the breakpoint, where a linear volume-

pressure relationship exists, the increase in CSFpulse pressure is only related to the change inCBV per cardiac cycle according to: AP=EAV,where E (elastance) is constant. Consequently, inthis pressure zone AV is increasing with risingintracranial pressure. The increase in AV permmHg was calculated and the results are givenin Table 6.

In the animals with data above the level of theSAP no specific analysis was possible. A decreasein CSF pulse pressure was generally observed.

-28

-24

-20 Fig. 7 Plot of the volume-pressure response(0) and the CSF pulse pressure (@) against the

-16 ventricular fluid pressure (VFP) in a singleanimal (dog 1).

-12

- 8

K4

Discussion

The present series of experiments was designed tostudy the relation of the CSF pulse pressure toboth ICP and the VPR with the clinically moti-vated objective of investigating whether the ratioof pulse pressure to ICP could serve as a reliableguide in the assessment of the craniospinalvolume-pressure relationship. The results, how-ever, also give rise to a critical review of thevolume-pressure function itself.

VOLUME-PRESSURE RELATIONSHIPSThe validity of the mono-exponential volume-pressure model follows directly from the linearcorrelation of VPR with VFP. This is consistent

695



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


C. J. J. A vezaat, J. H. M. van Eijndhoven, and D. J. WyperVentricular fluid pressureCSF pulse pressure(mm Hg)

40

20

15 30

(a)CSFpulsepressure

(mm Hg) 30

20

10

45Time (min)

Fig. 8 (a) Computer plot of theventricular fluid pressure (VFP,upper tracing) and the CSF pulsepressure (lower tracing) duringVFP monitoring in a patientshowing a classical plateau waveand two "abortive" waves. Themethod of computer analysis hasbeen reported previously (A vezaatet al., 1976). (b) Computer plotof the relationship between theCSF pulse pressure and the VFP ofthe pressure recording shown in(a). A breakpoint occurs at aVFP of about 25 mmHgdemonstrating a steeper linearrelationship during theplateau wave.

20 40 60

(b) VFP (mm Hg)

696



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



with the findings of other authors (Ryder et al.,1953; Lofgren et al., 1973; Marmarou, 1973;Miller et al., 1973). The volume-pressure para-meter used in this study, E1, determines the slopeof the volume-pressure curve or the elastance ata given ICP and was, therefore, termed the"elastance coefficient." A larger E1 means asteeper slope. Sullivan and his colleagues (1977),using a similar mathematical model, have recentlyreported a progressive elevation of the PVI (theinverse parameter of E1) during cerebral com-pression in cats. This invalidates the mono-exponential model which implies a constant PVI.However, when their elastance data are plottedagainst the ICP, instead of balloon volume, alinear relationship is found confirming the mono-exponential model. This contradiction is causedby the mathematical model used, which impliesthat the linear relationship of elastance (VPR)versus ICP passes through the origin of the axes.As long as the regression lines have their inter-cepts close to the origin, as in this study, themodel will not be seriously disturbed. If not, themathematical model must be extended by a con-stant term as we have previously reported(Avezaat et al., 1976). When this extended modelis applied to the data of Sullivan et al. (1977) thePVI remains constant up to an ICP of about60 mmHg; so the mono-exponential volume-pressure model holds true. The constant term hasthe effect of shifting the volume-pressure curvealong the pressure axis, which was demonstratedby Lofgren et al. (1973) to be related to the centralvenous pressure.

Confusion has been caused in the past when,in discussing volume-pressure relationships, refer-ence was made to curves constructed by plottingthe steady state ICP against the volume of anexpanding intracranial mass (Langfitt et al., 1965;Leech and Miller, 1974a). However clinicallyrelevant they may be, these curves do not reflectthe elastic properties of the system as they involvethe process of spatial compensation. This is inagreement with the results of the present studywhich failed to show a correspondence betweenE1 and the balloon volume accommodated over theICP range for which E1 was valid. For example,animal 2 has a relatively large spatial compen-satory capacity and a steep volume-pressure curve.The large E1 of this animal may be explained bythe early occurrence of tentorial herniation. Otherinvestigators (Lim et al., 1973; Lofgren andZwetnow, 1973; Marmarou et al., 1975) have re-ported an increase of the elastance when thecommunication between the cranial and spinal

compartment was impaired. In this series hernia-tion always started to develop before the break-point, which may account for the generally largeE1 compared to results obtained with rapidinfusion techniques.An important finding in this study was the

validity of the exponential volume-pressuremodel for a restricted ICP range only. At a meanICP of 60 mmHg a breakpoint occurred abovewhich the volume-pressure relationship became alinear function with a slope equal to the elastanceat that point. In order to understand this thequestion should be raised as to why the relation-ship is exponential in the first place. The magni-tude of the change in ICP in response to a rapidvolume addition is determined by two factors con-stituting the so-called elastic properties of thecraniospinal system. The first factor is related tothe physical properties of the craniospinal con-tents and their confining structures. Here, the"give" in the system depends largely on the dis-tensibility of the spinal dural sac (Martins et al.,1972). The characteristics of these structures, likethose of all other soft body tissues, are essentiallynonlinear (King and Lawton, 1950). The secondfactor is related to disturbances in the intracranialhaemodynamics. The volume added compressesthe venous outflow section and, through an in-crease of the outflow resistance, produces a rise inthe upstream intravascular pressure, which inturn is transmitted to the surrounding CSF. Thepressure response is thus determined by the com-pressibility of the venous vascular bed. Accordingto Poiseuille's law the resulting venous pressure isinversely proportional to the square of vascularvolume, which provides further support for thenonlinear nature of the volume-pressure relation-ship. From Poiseuille's law also follows that thesame volume increment will result in a lowerpressure response, if flow is reduced. So progressivereduction of cerebral blood flow (CBF) contributesto the linear shape of the volume-pressure curveabove the breakpoint, and the breakpoint mayconsequently be related to failure of autoregula-tion. Lofgren et al. (1973) described the volume-pressure curve from an ICP of 20 mmHg up to thelevel of the blood pressure as a linear function(high pressure elastance). The curve was estab-lished during rapid infusion in the cisterna magnaof dogs within a time interval too short for anautoregulatory vasomotor response to develop.Their results are, therefore, in agreement withthe concept that a linear volume-pressure relation-ship is related to a decrease in CBF. Furthermore,the ICP at which Johnston et al. (1973), in a

697



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


C. J. J. Avezaat, J. H. M. van Eiindhoven, and D. J. Wyper

similar preparation in the baboon, found a re-

duction of CBF corresponds with the pressure atthe breakpoint of this study: 60 mmHg. Thenarrower range of CPP than of ICP at the break-points is another argument for relating this pointto changes in intracranial haemodynamics.From the difference in VPR between injected

and non-injected ventricles we conclude that thedecrease of the elastance close to the level of theblood pressure, an observation also made by others(Lofgren et al., 1973; Leech and Miller, 1974a),is an artefact. This artefact is caused by a rapidspatial compensatory mechanism, probably dis-placement of venous blood. So the VPR at thislevel of ICP is no longer an accurate measure ofthe elastance. Above the blood pressure level,where CBF has ceased completely, the cranialcompartment fulfils the requirements of theMonro-Kellie doctrine resulting in a very steepvolume-pressure relationship.

CSF PULSE PRESSUREThe linear increase of the CSF pulse pressure withrising ICP is consistent with observations made inpatients (Janny, 1974; Avezaat et al., 1976; Norneset al., 1977; Szewczykowski et al., 1977). In thisstudy we found a breakpoint above which thepulse amplitude increased more rapidly and whichcoincided with a breakpoint in the volume-pressure relationship. From the comparison be-tween pulse pressure and VPR we concluded thatthe change in CBV per cardiac cycle (AV) remainsconstant up to the level of the ICP at the break-point. The important implication is that in thisICP range changes in the slope of the CSF pulsepressure versus ICP relationship truly reflectchanges in the volume-pressure relationship.Above the breakpoint the increase in pulsepressure is, on account of the constant elastance,caused by a progressive increase of AV.Because of the coincidence of the two break-

points it is plausible that the increase of AV, likethe levelling off of the VPR, is caused by haemo-dynamic changes, particularly failure of auto-regulation and reduction of CBF. Symon et al.(1974, 1975), using the pulse transmission frompial artery to Labbe's vein as an index of cere-

brovascular resistance (CVR), found a linearincrease of the pulse transmission up to an ICPof 70 mmHg. From this point onwards auto-regulation began to fail, CBF decreased, and thepulse transmission increased more steeply. We(Avezaat et al., 1976) have reported an increasedratio of CSF pulse pressure to ICP during plateauwaves (Lundberg, 1960). The cause of plateau

waves has been attributed to cerebral vasodilata-tion and increased CBV (Lundberg et al., 1968).A physiological explanation for the increase of

AV can be offered using the concept on the originof the change in CBV during a cardiac cycle(Fig. 2). The magnitude of this volume change isdetermined by the difference in shape between thepulsatile arterial inflow and venous outflow curves.These flow profiles depend on the impedances ofthe arterial and venous sections of the cerebralvasculature. The resistance vessels dilate withrising ICP as part of the autoregulatory process,causing a decrease of the inflow impedance and amore rapid inflow of blood. The outflow im-pedance at the same time increases as a result ofvenous compression, thus delaying the outflow ofblood. As a result, inflow and outflow graduallybecome two more or less separate processes(Fig. 2b). This implies an increase in AV as soonas ICP starts to rise, which, according to theanalytical model, corresponds with a nonlinearrelationship between pulse pressure and ICP.Evidence for this theory can be derived from thework of Guinane (1975) who described this re-lationship by an exponential function. In the ICPrange below the breakpoint the changes in inflowand outflow impedance are probably relativelysmall, so that a linear relationship implying aconstant AV is a good approximation. Whenautoregulation fails and progressive vasoparalysisoccurs (Langfitt et al., 1965) the inflow impedanceis at its lowest. At the same time the resistanceat the outflow tract continues to increase, towhich compression of the dural sinuses might bea contributory factor (Shapiro et al., 1966). Thisexplains the marked increase in AV above thebreakpoint. The paradox of an increase in AV inspite of a reduction in CBF can be solved by look-ing more closely at the cerebral inflow section. Adecrease in CBF does not necessarily mean adecrease in the arterial inflow of blood but simplyan increase in arterial reflux. Nornes et al. (1977),measuring electromagnetic flow in the internalcarotid artery of patients, found a reduction offlow at high levels of ICP but at the same timean increase of pulsatile flow. This increase wasmost marked at a breakpoint where also the ICPpulse increased more rapidly.Another observation was the occurrence of

reverse flow during diastole. This also explainswhy large pulse pressures can still be recordedwhen the ICP has been raised above the level ofthe blood pressure and consequently completecirculatory arrest exists. In this case both inflowand outflow take place at the arterial side.

698



j.com/

J Neurol N


ugust 1979. Dow

nloaded from



CLINICAL IMPLICATIONSThe volume-pressure test is generally recognisedas the most practical method of assessing thevolume-pressure relationship in patients. The dis-advantages of the test have been mentioned. More-over, the test gives only a narrow view of thevolume-pressure curve. The heart, on the otherhand, is generating volume-pressure tests con-tinuously. The relationship between the CSF pulsepressure and the ICP can be processed at the bed-side by simple electronic equipment on the basisof which the volume-pressure function can beassessed continuously. However, since the changein CBV per cardiac cycle varies between patientsas between the animals of this study, a VPT isstill needed at the beginning of an ICP recordingto determine the magnitude of AV. In the courseof ICP monitoring AV may be affected by factorschanging cerebral haemodynamics, such as arterialcarbon dioxide tension, blood pressure, and arterialpulse pressure. The influence of these variables onAV will be dealt with in a subsequent study.An important finding of this study was that a

marked increase in the CSF pulse pressure to ICPratio may indicate loss of autoregulation. In thatcase a relatively low VPR can be expected as aresult of the levelling off phenomenon and therapid spatial compensatory mechanism. Conse-quently, in non-autoregulating patients the CSFpulse pressure is a better parameter of the clinicalstate than the VPR.

This study was supported by a grant from theNetherlands Organisation for the Advancement ofPure Research (ZWO). We would like to expressour appreciation to the staffs of the Institute ofNeurological Sciences and the Wellcome SurgicalResearch Institute, Glasgow for their assistance.

References

Avezaat, C. J. J., van Eijndhoven, J. H. M., de Jong,D. A., and Moolenaar, W. C. J. (I 976). A newmethod of monitoring intracranial volume-pressurerelationships. In Intracranial Pressure III, pp. 308-313. Edited by J. W. F. Beks, D. A. Bosch, and M.Brock. Springer: Berlin.

Castel, J. P., and Cohadon, F. (1976). The pattern ofcerebral pulse: automatic analysis. In IntracranialPressure III, pp. 303-307. Edited by J. F. W. Beks,D. A. Bosch, and M. Brock. Springer: Berlin.

Guinane, J. E. (1975). Cerebrospinal fluid pulsepressure and brain compliance in adult cats.Neurology (Minneapolis), 25, 559-564.

Janny, P. (1974). La surveillance de la pression intra-cranienne en neuro-chirurgie. Aspects graphiques etsemeiologiques de la pression intra-cranienne.Neurochirurgie, 20, 521-534.

Johnston, I. H., Rowan, J. O., Harper, A. M., andJennett, W. B. (1973). Raised intracranial pressureand cerebral blood flow. 2. Supratentorial andinfratentorial mass lesions in primates. Journal ofNeurology, Neurosurgery, and Psychiatry, 36, 161-170.

King, A. L., and Lawton, R. W. (1950). The elasticityof soft body tissues. Scientific Monthly, 71, 258-260.

Langfitt, T. W., Weinstein, J. D., and Kassell, N. F.(1965). Cerebral vasomotor paralysis produced byintracranial hypertension. Neurology (Minneapolis),15, 622-641.

Leech, P., and Miller, J. D. (1974a). Intracranialvolume-pressure relationships during experimentalbrain compression in primates. 1. Pressure responsesto changes in ventricular volume. Journal ofNeurology, Neurosurgery, and Psychiatry, 37, 1093-1098.

Leech, P., and Miller, J. D. (1974b). Intracranialvolume-pressure relationships during experimentalbrain compression in primates. 2. Effect of inducedchanges in systemic arterial pressure and cerebralblood flow. Journal of Neurology, Neurosurgery,and Psychiatry, 37, 1099-1104.

Leyden, E. (1866). Beitrage und Untersuchungen zurPhysiologie und Pathologie des Gehirns. Uber Hirn-druck und Hirnbewegungen. Virchows Archiv furPathologische Anatomie und Physiologie, 37, 519-559.

Lim, S. T., Potts, D. G., Deonarine, V., and Deck,M. D. F. (1973). Ventricular compliance in dogswith and without aqueductal obstruction. Journalof Neurosurgery, 39, 463-473.

Lofgren, J. (1973). Effects of variations in arterialpressure and arterial carbon dioxide tension on thecerebrospinal fluid pressure-volume relationships.Acta Neurologica Scandinavica, 49, 586-598.

Lofgren, J., and Zwetnow, N. N. (1973). Cranial andspinal components of the cerebrospinal fluidpressure-volume relationships. A cta NeurologicaScandinavica, 49, 586-598.

Lofgren, J., Essen, C. von, and Zwetnow, N. N.(1973). The pressure-volume curve of the cerebro-spinal fluid space in dogs. A cta NeurologicaScandinavica, 49, 557-574.

Lundberg, N. (1960). Continuous recording and con-trol of ventricular fluid pressure in neurosurgicalpractice. A cta Psychiatrica et NeurologicaScandinavica, 36, Supplement, 149.

Lundberg, N., Cronqvist, S., and Kjallquist, A. (1968).Clinical investigations on interrelations betweenintracranial pressure and intracranial haemo-dynamics. Progress in Brain Research, 30, 69-75.

Marmarou, A. (1973). A theoretical and experi-mental evaluation of the cerebrospinal fluid system.PhD thesis, Drexel University.

Marmarou, A., Shulman, K., and LaMorgese, J.(1975). Compartmental analysis of compliance andoutflow resistance of the cerebrospinal fluid system.Journal of Neurosurgery, 43, 523-534.

Martins, A. N., Wiley, J. K., and Myers, P. W. (1972).

699



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


C. J. J. A vezaat, J. H. M. van Eijndhoven, and D. J. Wyper

Dynamics of the cerebrospinal fluid and the spinaldura mater. Journal of Neurology, Neurosurgery,and Psychiatry, 35, 468-473.

Miller, J. D., and Garibi, J. (1972). Intracranialvolume-pressure relationships during continuousmonitoring of ventricular fluid pressure. InIntracranial Pressure, pp. 270-274. Edited by M.Brock and H. Dietz. Springer: Berlin.

Miller, J. D., Garibi, J., and Pickard, J. D. (1973).Induced changes of cerebrospinal fluid volume.Effects during continuous monitoring of ventricularfluid pressure. Archives of Neurology (Chicago),28, 265-269.

Miller, J. D., and Pickard, J. D. (1974). Intracranialvolume-pressure studies in patients with head in-jury. Injury, 5, 265-268.

Mosso, A. (1881). Uber den Kreislauf des Blutes immenslichen Gehirn. Leipzig.

Nornes, H., Aaslid, R., and Lindegaard, K. F. (1977).Intracranial pulse pressure dynamics in patientswith intracranial hypertension. A cta Neurochirur-gica, 38, 177-186.

Ryder, H. W., Epsey, F. F., Kimbell, F. D., Penka,E. J., Rosenauer, A., Podolsky, B., and Evans,J. P. (1953). The mechanism of the change in cere-brospinal fluid pressure following an induced changein the volume of the fluid space. Journal of Labora-tory and Clinical Medicine, 41, 428-435.

Severinghaus, J. W. (1966). Blood gas calculator.Journal of Applied Physiology, 21, 1108-1116.

Shapiro, H. M., Langfitt, T. W., and Weinstein, J. D.

(1966). Compression of cerebral vessels by intra-cranial hypertension. II Morphological evidence forcollapse of vessels. A cta Neurochirurgica, 15, 223-233.

Shulman, K., and Marmarou, A. (1971). Pressure-volume considerations in infantile hydrocephalus.Developmental Medicine and Child Neurology, 13,Supplement 25, 90-95.

Sullivan, H. G., Miller, J. D., Becker, D. P., Flora,R. E. and Allen, G. (1977). The physiological basisof intracranial pressure change with progressiveepidural brain compression. An experimentalevaluation in cats. Journal of Neurosurgery, 47,532-550.

Symon, L., Pasztor, E., Branston, N. M., and Dorsch,N. W. C. (1974). Effect of supratentorial space-occupying lesions on regional intracranial pressureand local cerebral blood flow: an experimentalstudy in baboons. Journal of Neurology, Neuro-surgery, and Psychiatry, 37, 617-626.

Symon, L., Crockard, H. A., and Juhasz, J. (1975).Some aspects of cerebrovascular resistance in raisedintracranial pressure: an experimental study. InIntracranial Pressure II, pp. 257-262. Edited by N.Lundberg, U. Ponten, and M. Brock. Springer:Berlin.

Szewczykowski, J., Sliwka, S., Kunicki, A., Dytko, P.,and Korsak-Sliwka, J. (1977). A fast method ofestimating the elastance of the intracranial system.A practical application in neurosurgery. Journal ofNeurosurgery, 47, 19-26.

700



j.com/

J Neurol N


ugust 1979. Dow

nloaded from


Documents

Cerebrospinal pulse pressure and relationships · J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.8.687 on 1 August 1979. Downloaded from Cerebrospinal fluidpulse