Effects of hypertonic saline on intracranial pressure variables in

subarachnoid haemorrhage patients

Gunnar Kristoffer Bentsen

Ph.D. thesis

Division of Anaesthesiology and Intensive Care Medicine Rikshospitalet University Hospital

University of Oslo Oslo, Norway

2009

© Gunnar Kristoffer Bentsen, 2009 Series of dissertations submitted to the Faculty of Medicine, University of Oslo No. 714 ISBN 978-82-8072-765-7 All rights reserved. No part of this publication may be reproduced or transmitted, in any form or by any means, without permission. Cover: Inger Sandved Anfinsen. Printed in Norway: AiT e-dit AS, Oslo, 2009. Produced in co-operation with Unipub AS. The thesis is produced by Unipub AS merely in connection with the thesis defence. Kindly direct all inquiries regarding the thesis to the copyright holder or the unit which grants the doctorate. Unipub AS is owned by The University Foundation for Student Life (SiO)

3

”Following the intravenous injection of 30 per cent sodium chloride, the normal convexity of the brain in the trephine opening disappears soon after the injection is begun, so that the brain is seen to lie flat. As the intravenous injection of salt is continued, the brain falls away from the skull until the surface presented becomes concave. The maximum shrinkage has been observed usually in from fifteen to thirty minutes after the completion of the injection, when the brain lies flaccid, 3 to 4 mm below the inner table of the skull.”

LH Weed and PS McKibben, 1919.

4

5

CONTENTS

Page Acknowledgements 6 Abbreviations 8 Publications included in the thesis 9 Introduction 10 Subarachnoid haemorrhage 10 Definitions, incidence, and outcome 10 Mechanisms of injury to the brain 11 Aspects of treatment 12 Osmotherapy 14 Context definition 14 Mechanisms 14 History 16 Mannitol 17 Hypertonic saline 18 Animal studies 20 Human studies 22 Intracranial pressure monitoring 24 Mean ICP 24 ICP waveform analysis 26 Aims of the thesis 28 Methods 29 Study design 29 Data collection 30 Statistical methods 31 Synopsis of results 33 Paper I 33 Paper II 34 Paper III 35 Discussion 37 The relevance of ICP and CPP 38 Different ICP variables 39 Effects of HSS on mean ICP and CPP in SAH patients 40 Temporal pattern of change 41 Study solution 43 Dose and practical use 44 Adverse effects 46 Effects of HSS on mean ICP wave amplitude in SAH patients 47 Future clinical research questions 49 Main conclusions 51 Implications for clinical practice 53 References 55 Papers I - III

6

ACKNOWLEDGEMENTS

The present work has been carried out at the Division of Anaestesiology and Intensive Care

Medicine, Rikshospitalet University Hospital, which has also financed the work, and the

University of Oslo, Faculty of Medicine, during the period 2002 – 2008. During the whole

period I have been surrounded by generosity both at work and at home, and I want to express

my sincere gratitude to all who have helped and supported me these years.

First of all, Professor Audun Stubhaug, my principal supervisor, always supportive and

enthusiastic. Even though heavily tied up in work at all times, finding enough time to couch

me through all stages of this doctoral thesis in a exemplary way, lending me access to his vast

knowledge in research, statistics, and medicine. My sincerest thanks also to my two other

supervisors; Professor Harald Breivik, indeed a man of letters, contributing from his

overwhelming experience, and with his sharpened skills in manuscript preparation. Øyvind

Skraastad, now Head of our Division, former leader of our paediatric anaesthesia team,

supporting me all the way and helping me combine research with clinical work. The same

goes to Einar Hysing, former Head of our Division, now Hospital Medical Director, for

encouragement and preparing the ground for my combination of research and clinical work.

The principal building blocks of a thesis are of course the papers behind it. The great

involvement from all my co-authors has been a rewarding experience for me. I thank

Professor Tryggve Lundar and Per Kristian Eide from the Department of Neurosurgery for

their contribution. Per Kristian Eide has also developed the method for monitoring mean ICP

wave amplitude investigated in Paper III, and has introduced me to the fascinating world of

intracranial pressure waves. I also want to thank William Sorteberg for valuable input and

support.

7

During the time I have worked with this thesis I have been fortunate to enjoy the

privilege of working full- or part-time at our paediatric anaesthesia team. I want to thank

Marius Asplin, Even Fagermoen, Bjørn Aage Feet, Torleiv Haugen, Sidsel Hetland, Bjarne

Morisbak, Roger Roscher, Gunn Kari Sangolt (late), Anne-Beate Solås, Sjur Sponheim, and

Kari Wagner for transfer of clinical skills and knowledge, for encouragement and bright days.

Our collegium as a whole is a group I am proud to be a part of, so many devoted and

skilled physicians working together in a relaxed atmosphere topped with solid portions of

humour and self-irony. Many have helped me recruit patients at all hours, I thank you all.

Sincere thanks also to the entire nursing staff at our intensive care units, which

faithfully and with a smile has included the extra load of work clinical studies carry, attitudes

truly worthy a university clinic.

For technical assistance, I am very grateful to Jan Olav Høgetveit who designed the

LabView application making harvesting of data to Paper I and II possible. Likewise had data

harvest to Paper III been impossible without Trond Stadheim.

Last but foremost, I thank my family, my most important source of help, joy, and

support. My fantastic wife, Elin, with whom I can share all my challenges and wins. My

equally fantastic kids, Hanna and Lars, unconditionally loving and a constant source of

happiness. Our dog, Mila, for not having eaten any important material.

May 2008. Gunnar Bentsen

8

ABBREVIATIONS

ABP Arterial blood pressure

BBB Blood-brain-barrier

CBF

CBV

Cerebral blood flow

Cerebral blood volume

CDO2 Cerebral oxygen delivery

CI Cardiac index

CNS Central nervous system

CPM

CPP

Central pontine myelinolysis

Cerebral perfusion pressure (CPP = MAP – ICP)

CSF

CVP

DIND

ELWI

Cerebrospinal fluid

Central venous pressure

Delayed ischemic neurological deficit

Extravascular lung water index

EVD External ventricular drain

FFT Fast Fourier transformation

GCS Glasgow Coma Score

GOS Glasgow Outcome Score

HR Heart rate

HS Hypertonic saline

HSS 7.2% (72 mg/mL) saline in 6% (60 mg/mL) hydroxyethyl starch

200/0.5 (sodium content: 1.23 mmol/mL)

ICP Intracranial pressure

ICU Intensive care unit

ITBI Intrathorasic blood volume index

IVH Intraventricular haemorrhage

MAP Mean arterial pressure

NS Normal saline

SAH Subarachnoid haemorrhage

SpO2

TBI

Peripheral oxygen saturation

Traumatic brain injury

9

PUBLICATIONS INCLUDED IN THE THESIS

The thesis is based on the following articles, which are referred to in the text by their roman

numerals:

I. Bentsen G, Breivik H, Lundar T, Stubhaug A. Predictable reduction of intracranial

hypertension with hypertonic saline hydroxyethyl starch: a prospective clinical trial in

critically ill patients with subarachnoid haemorrhage. Acta Anaesthesiol Scand 2004;

48: 1089-1095.

II. Bentsen G, Breivik H, Lundar T, Stubhaug A. Hypertonic saline (7.2%) in 6%

hydroxyethyl starch reduces intracranial pressure and improves hemodynamics in a

placebo-controlled study involving stable patients with subarachnoid hemorrhage. Crit

Care Med 2006; 34: 2912-7.

III. Bentsen G, Stubhaug A, Eide PK. Differential effects of osmotherapy on static and

pulsatile intracranial pressure (ICP). Crit Care Med; accepted for publication.

10

INTRODUCTION

Subarachnoid haemorrhage

Definitions, incidence, and outcome

Subarachnoid haemorrhage (SAH) is characterized by the presence of blood in the

subarachnoid space; the space covering the central nervous system that normally only

contains cerebrospinal fluid. The bleeding can be either traumatic or spontaneous. This thesis

only deals with the spontaneous, non-traumatic bleedings, which in about 80 to 85% of the

cases are due to the rupture of a cerebrovascular aneurysm (van Gijn and Rinkel, 2001).

The overall incidence of spontaneous SAH is in meta-analyses reported to be

somewhere around 10:100,000 annually, but with variations from six to eight in USA and

most of Europe to around 20:100,000 in Finland and Japan (Linn et al., 1996; Teunissen et al.,

1996; Menghini et al., 1998; van Gijn and Rinkel, 2001). Women have a 1.6 times higher risk

than men (Linn et al., 1996), and black people twice the risk of whites (Broderick et al., 1992)

of experiencing a SAH. Smoking, hypertension, and alcohol abuse increase the risk for SAH

(Teunissen et al., 1996).

Case fatality rates are high in SAH compared with other forms of stroke, varying

between 32% and 67%, with a weighted average of 51%, but the fatality rate has decreased

over the last three to four decades by 0.5 to 1.0% annually (Hop et al., 1997; Johnston et al.,

1998; Koffijberg et al., 2008). Even though SAH accounts for only 3% of all strokes (Sudlow

and Warlow, 1997), 4.4% of cerebrovascular deaths are due to SAH (Johnston et al., 1998).

SAH also accounts for 27% of total years of potential life lost before age 65 due to

cerebrovascular disease (Johnston et al., 1998). This is due to the high fatality rate and the fact

that the population of SAH patients is younger than the populations with cerebral infarction

and intracerebral haemorrhages.

11

About 50% of those who die, die within the first two days after the stroke (Stegmayr et

al., 2004). This represents a challenge to the acute management of these patients. Of vital

importance is prompt diagnosis, rapid transfer to a neurosurgical unit, early securing of the

aneurysm, and optimal intensive care (Bakke and Lindegaard, 2007).

Of the survivors, approximately one third need lifelong care (Hop et al., 1997), and

almost a half of the survivors suffer life-long cognitive impairment (Hackett and Anderson,

2000; Mayer et al., 2002). Hopefully, improved intensive care with avoidance of secondary

insults can help reduce also these figures.

Mechanisms of injury to the brain.

Brain injury after SAH is a biphasic event.

First there is the initial impact of the bleeding itself causing cerebral ischemia. A sharp

increase in intracranial pressure (ICP) leads to a dramatic reduction in cerebral perfusion

pressure (CPP) (Talacchi, 1993). But even after recovery of CPP, there is evidence of

prolonged hypoperfusion with flow measured in different studies at 20 – 80% of baseline

levels (Jackowski et al., 1990; Piepgras et al., 1995; Prunell et al., 2004). There is also

evidence of loss of autoregulation, disruption of the blood-brain-barrier (BBB), cellular

swelling leading to cerebral oedema, and initiation of apoptotic processes, but the

understanding of the underlying mechanisms and the ultimate relevance of these early

changes is still incomplete (Cahill et al., 2006; Mocco et al., 2007; Schubert and Thome,

2008).

Secondly, given that the patient has survived the impact of the bleeding, there is a

more prolonged phase. Many patients experience bleeding into the ventricular system

affecting drainage of cerebrospinal fluid (CSF). This may lead to the development of

hydrocephalus and subsequent rise in ICP leading to ischemia within hours (Bakke and

12

Lindegaard, 2007). Cellular swelling and disruption of the BBB induces cerebral oedema,

which also increases ICP. This process typically culminates after two to four days (Thiex and

Tsirka, 2007), but may cause distortion and displacement of brain tissue resulting in

compression of vital structures (“herniation” syndromes). Claassen et al. have reported the

presence of cerebral oedema to be associated with poor outcome. Global oedema was found at

admission in 8% of patients with an additional 12% developing delayed oedema (Claassen et

al., 2002). Reaching peak incidence at about a week, delayed vasospasm may also cause

seriously impaired cerebral blood flow (CBF) and ischemia. Radiographic evidence of

vasospasm can be found in more than 50% of patients, but only half of these experience

symptoms of delayed ischemic neurological deficits (DIND) (Keyrouz and Diringer, 2007).

Somewhere along the time line, there is also a periprocedural phase, where the source

of bleeding is secured either by surgical clipping or endovascular coiling. These procedures

also carry some risk. Complications may arise due to thrombosis, bleeding, induced

vasospasm, or excessive traumatisation of brain tissue (Bakke and Lindegaard, 2007).

Aspects of treatment

In short, the goals of treatment are; meeting the neuronal metabolic needs at all times,

avoiding secondary deterioration, and avoiding further bleeds.

The avoidance of re-bleeding is achieved by surgical clipping or endovascular coiling

of the aneurysm. The trend is to perform this intervention at an early stage. The risk of re-

bleeding is closely related to time and the occurrence of re-bleeding is strongly associated

with poor outcome (Ohkuma et al., 2001; Bakke and Lindegaard, 2007). The risk of re-

bleeding is highest during the first 24 hours after the initial bleeding and immediate

administration of tranexamic acid upon diagnosis of SAH may reduce the risk of early re-

bleeding (Hillman et al., 2002).

13

Securing optimal oxygen and nutritional supply has many aspects. First, one important

measure is to reduce the oxygen need. This can be achieved by avoiding hyperthermia and

sedating the patient, in the final instance by inducing a pentobarbital coma. Sedation of the

patient may warrant securing the airway and the institution of mechanical ventilation.

Mechanical ventilation is often also needed because the patient is unconscious to begin with

or because pneumonia impairs gas exchange to the degree that the patient is hypoxemic. By

mechanical ventilation the oxygen supply and carbon dioxide removal can most often be well

controlled. For the oxygen taken up in the lungs to be delivered to the brain we must ensure

that the haemoglobin content is sufficient, and that the brain is perfused. This means that the

CPP has to be adequate, and the blood pressure and cardiac output can be controlled by

adequate volume supply and the use of vasoactive drugs when needed.

As is obvious from the previous section, raised ICP is often a threat to adequate

perfusion and cell function. Intracranial hypertension can be attenuated by sedation, head

elevation, and temperature control. For short time relief of increased ICP, reduction of arterial

CO2 content by hyperventilation can be very effective, but aggressive hyperventilation can

induce grave vasoconstriction leading to ischemia in itself. Especially in patients suffering

from acute hydrocephalus caused by intraventricular haemorrhage (IVH), external drainage of

cerebrospinal fluid (CSF), either by a ventricular catheter or by the lumbar route is very

helpful in rapidly decreasing the ICP. In extreme cases of cerebral swelling, the procedure of

decompressive hemicraniectomy can be life saving (Schirmer et al., 2007; Hutchinson et al.,

2007). This brings me to the subject of osmotherapy, which is the last of the commonly

applied strategies used to attenuate raised ICP.

14

Osmotherapy

Context definition

The use of osmotically active solutions to reduce the volume of the intracranial content

(Paczynski, 1997).

Mechanisms

The water content of the brain approximates 80% (Bhardwaj, 2007). When we add that a

small 1.6% reduction in water content yields a 90 mL reduction in brain volume (Cascino et

al., 1983), we have the foundation for understanding osmotherapy.

The osmotic pressure of a solution is proportional to the number of particles in a given

volume (Rapoport, 1976), but osmotic effectiveness of solutions depends as much on the

properties of the membrane(s) separating the different compartments of solutions as their

respective molarities. This can be expressed as the membrane’s osmotic reflection coefficient

(�) for a given solute. Given a reflection coefficient of zero, the membrane is completely

permeable for that solute. On the other hand, if the membrane is impermeable, the value is

one (Zornow, 1996).

The properties of the BBB differ greatly from the properties of capillaries elsewhere in

the body. Unlike “ordinary capillaries”, the BBB is virtually impermeable to small solute

molecules like sodium and chloride. It is these compounds that regulate the water balance

across the BBB. The importance of the colloid osmotic pressure, known from the Starling

equation, is less in the uninjured brain; the flow of water is tightly controlled by the

crystalloid osmotic pressure.

Another important property of the BBB, is the relative impermeability of molecules of

water (Raichle et al., 1976). On 9 October 1991 the first water channel, Aquaporin-1, was

discovered (Preston et al., 1992). Subsequently other aquaporins have been identified, and

15

Aquaporine-4 located in the brain may control most of the water transport across the BBB,

and represent a rate limiting factor for this process. However, the importance of Aquaporine-4

in human cerebral oedema is still undefined, but Aquaporine-4 has a potential for future

therapeutic interventions (Agre et al., 2002; Chen et al., 2007; Bloch and Manley, 2007).

The intravenous infusion of a hypertonic solution establishes an osmotic gradient

between the intravascular space and the extracellular volume of the brain. This gradient

provides a potent force to move water from the brain’s extra- and intracellular space into the

capillaries. In this way, the volume of the brain is reduced and thereby the ICP (Zornow,

1996).

There are other theories as to how the osmotherapeutics reduce ICP. There are several

variants of “haemodynamic” theories, all emphasizing the importance of dynamic changes in

cerebral blood volume (CBV) (Paczynski, 1997). Muizelaar et al. have reported, from a study

in cats, a rapid reduction in the diameter of arterioles and venules on the surface of the brain

immediately after an intravenous mannitol bolus of 1 g/kg given during 1 minute (Muizelaar

et al., 1983). Subsequently they reported how this effect on ICP was greater in patients with

intact autoregulation, than in patients without (Muizelaar et al., 1984). I must, however,

express methodological concerns regarding this last study as there seem to be repeated

measurements in some patients and inclusion of patients with ICP above the stated inclusion

criteria. The reduction in CBV was believed to be caused mainly by a vasoconstriction

induced by reduction in blood viscosity improving blood flow (Muizelaar et al., 1983). This

cerebral vasoconstriction has, however, been shown not to occur in other studies (Ravussin et

al., 1985; Auer and Haselsberger, 1987). Auer et al. gave mannitol during 15 minutes, which

is more in line with clinical practice than the one minute used by Muizelaar et al., and

Ravussin et al. found reduced ICP despite increased CBV.

16

It seems reasonable to speculate that instant, haemodynamic changes may account for

some of the early (less than 15 minutes) effect seen on ICP after osmotherapy, whilst the

osmotic effect is slower in onset, gradually dehydrating the brain towards a peek effect at 15

to 30 minutes after administration of a hypertonic solution (Ziai et al., 2007).

History

The first description of osmotherapy applied to the central nervous system (CNS), is attributed

to Weed and McKibben in 1919 (Weed and McKibben, 1919a; Weed and McKibben, 1919b).

As research fellows in the Army Neurosurgical Laboratory at the Johns Hopkins Medical

School, they were attempting to measure the transport of sodium salts from the blood into the

CSF in cats. Upon intravenous injection of small volumes of 30% hypertonic saline (HS), they

measured a marked and sustained decrease in ICP (Weed and McKibben, 1919b). This they

found was due to shrinkage of the brain. They went on to describe how they through a

craniotomy defect could observe the brain shrink away after HS injection with maximum

effect 15 to 30 minutes after completion of the injection (Weed and McKibben, 1919a).

In 1927, intravenous delivery of osmotic agents for clinical practice was first

formalized by Fremont-Smith and Forbes (Fremont-Smith and Forbes, 1927). They first used

concentrated urea. 11 years later they were followed by Hughes et al. who demonstrated that

concentrated solutions of human plasma proteins could reduce raised ICP (Hughes et al.,

1938). The use of plasma proteins was however limited because of fear of allergic reactions

and high cost.

Wider attention was brought to the use of urea by Javid et al in the late 1950s (Javid

and Settlage, 1956; Javid, 1961). With its low molecular weight of 60 Daltons, slow

elimination from blood, and relatively slow BBB penetration (� = 0.48) (Qureshi and Suarez,

17

2000), it was a potent osmotic agent. It had however a significant potential for rebound

cerebral oedema and side effects like phlebitis and haemolysis.

During early 1960s two other compounds replaced urea; mannitol and glycerol.

Glycerol, introduced in 1964 is a trivalent alchohol (1,2,3-propanetriol), partially metabolized

to CO2 and water (Cantore et al., 1964). With a reflection coefficient (�) of 0.59 (Qureshi and

Suarez, 2000), it has a significant potential for causing rebound cerebral oedema. Side effects

include haemolysis, haemoglobinuria, renal failure, and hyperosmolar coma (Bhardwaj,

2007). Possibly due to tradition, glycerol has been used quite frequently throughout Asia and

some centres in continental Europe, whereas mannitol has been the drug of choice in the UK

and the Americas (Paczynski, 1997). Mannitol has also been the preferred choice in the

Nordic countries. Only since late 80s has hypertonic saline, the compound that started it all,

gained renewed interest.

Mannitol

As mentioned previously, mannitol has been the osmotic agent of choice in clinical practice in

many countries for nearly five decades (Brain Trauma Foundation et al., 2007a). Mannitol has

a molecular weight of 182 Daltons, and is an alcohol derivate of the sugar mannose. As

opposed to urea, it has a short plasma half-life of 2 to 4 hours. The reflection coefficient (�) is

0.9 (Qureshi and Suarez, 2000), which is much better than those of urea and glycerol. It does,

however, leave the door open for rebound elevation of ICP upon repeated administrations.

This has been much debated during the last two decades, and has reduced its use at least for

repeated administrations (Garcia-Sola et al., 1991; Kaufmann and Cardoso, 1992; Kofke,

1993; Polderman et al., 2003). In a study by Rudehill et al. the level of mannitol in the CSF

increased after a single intravenous administration, and did not start to fall during the first

eight hours after administration (Rudehill et al., 1993). Other side effects are acute renal

18

failure, particularly if serum osmolarity is > 320 mOsm/L combined with hypovolaemia

(Worthley et al., 1988; Bullock, 1995). Mannitol is also of course foremost a diuretic, and will

upon repeated administration lead to hypovolaemia if volume is not adequately substituted.

Hypertonic saline

After Velasco et al. in 1980 showed how bled dogs could be haemodynamically resuscitated

with 4 mL/kg of 7.0% saline (Velasco et al., 1980), HS was evaluated for rapid resuscitation

of patients with haemorrhagic shock during the 1980s and 90s. A small study by De Felippe Jr

et al. (de Felippe Jr et al., 1980) was followed by a series of large randomised studies

conducted in the pre-hospital setting (Holcroft et al., 1987; Vassar et al., 1991; Mattox et al.,

1991; Vassar et al., 1993a; Vassar et al., 1993b). In two of the studies by Vassar et al. (Vassar

et al., 1991; Vassar et al., 1993a), improved survival was reported after HS as compared to

isotonic resuscitation in patients suffering from traumatic brain injury (TBI); survival to

discharge being 32 versus 16% in the first study and 34 versus 12% in the second. This

boosted the wave of studies looking at HS and cerebral effects from the 1990s.

The reflection coefficient (�) across the BBB for NaCl is 1.0 (Rapoport, 1976;

Zornow, 1996; Qureshi and Suarez, 2000). In that respect, HS is an ideal osmotic agent. The

main effect on elevated ICP is due to the “dehydrating” effect on the brain (Qureshi and

Suarez, 2000; Ziai et al., 2007). The “dehydrating” effect of hypertonic saline (HS) has been

shown both in animal studies (Todd et al., 1985; Zornow et al., 1989; Bacher et al., 1998;

Toung et al., 2007; Chen et al., 2007) and in a MRI-study in humans (Saltarini et al., 2002).

Theoretically most of the effect would be found in uninjured parts of the brain. This is

supported by most studies (Zornow et al., 1989; Battistella and Wisner, 1991; Shackford et

al., 1992). Normal brain tissue is “dehydrated” to accommodate the increase in intracranial

19

volume that results from oedema or haemorrhage. The effect in parts of the brain with injured

BBB is more unpredictable (Chen et al., 2007).

Cerebrovascular endothelial cells and red blood cells are also dehydrated, improving

CBF and oxygen delivery due to increased vessel diameter and reduced size of the blood cells

(Mazzoni et al., 1990; Shackford et al., 1992; Doyle et al., 2001). And like mannitol, HS

lowers blood viscosity which may induce vasoconstriction with reduced CBV (Burke et al.,

1981; Muizelaar et al., 1983).

Another advantage of HS is that it does not have the potentially detrimental diuretic

effect like mannitol, which may lead to hypovolaemia and impaired cerebral perfusion (Arai

et al., 1986; Worthley et al., 1988).

There are a number of additional effects attributed to HS that deserve a brief

mentioning even though most of the documentation is based on animal studies and has to date

uncertain clinical implication. HS may have positive neurochemical effects, as HS solutions

restore normal membrane resting potential by normalizing intracellular concentrations of

sodium and chloride (Nakayama et al., 1985). HS may also have positive immunomodulatory

effects on the inflammatory process generated by brain trauma, reducing leukocyte adherence

and migration to injured parts of the brain (Hartl et al., 1997b). Several animal studies indicate

a protective effect against bacterial infections (Coimbra et al., 1997; Shields et al., 2003; Chen

et al., 2006), but very limited conclusions can be drawn as to the effect in humans (Kolsen-

Petersen, 2004).

The saline concentration in the solutions that have been investigated ranges from 1.4 to

29.2%, and both solutions with and without the addition of a colloid have been used (Tables

1-2, page 21)(Table 3, page 23). The addition of a colloid, typically 6% dextrane or

hydroxyethyl starch, has in haemodynamic resuscitation studies prolonged the haemodynamic

effect compared with HS alone (Smith et al., 1985; Kramer et al., 1986; Velasco et al., 1989).

20

For the attenuating effect on raised ICP, the addition of a colloid has not altered the effect

compared with HS alone (Vassar et al., 1993a). Other studies have shown that intravenous

infusion of 6% dextrane or hydroxyethyl starch solutions without HS do not reduce ICP

(Gunnar et al., 1986; Ducey et al., 1989). This fits well with what can be expected from

theoretical considerations based on the different properties of the BBB and capillaries

elsewhere in the body.

Animal studies

Todd et al. were among the first to re-examine the cerebral effects of hypertonic saline

solutions (Todd et al., 1985). They demonstrated in rabbits a decrease in brain water content,

a decrease in ICP, and an increase in cerebral blood flow (CBF) following infusion of

hypertonic Ringers’s solution (osmolality of 480 mOsm/kg). A number of animal studies

followed. Studies looking at the effect of HS on ICP, CBF, or cerebral oxygen delivery

(CDO2) are listed in Tables 1 and 2 (page 21).

Although there are indications of possible problems with rebound increase in ICP in

the material from Prough et al. from 1999 (Prough et al., 1999), the overall picture is that of

reliable reduction in ICP and improvement in CBF. The later findings by German groups of a

possible neuroprotective effect of HS is also encouraging (Heimann et al., 2003; Zausinger et

al., 2004; Bermueller et al., 2006). These findings are supported by other studies not looking

at effect on ICP (Hamaguchi and Ogata, 1995; Hamaguchi et al., 2002; Thomale et al., 2004).

21

Table 1. Animal studies investigating cerebral effects of HS without cerebral trauma.

↓ ICP, ↑ CBF, ↑ CDO21,5% SL, RLSSPigSchmoker et al., 1991

↓ ICP, ↓ ICE6%S,6%HE,NS,WBSSPigDucey et al., 1989

↓ ICP after HS, ↓ CBF and CDO2both groups

7.5% S, RLSSDogPrough et al., 1986

� ICP after HS, ↑ICP after NS, D40

3%S, NS, 6%DSSDogGunnar et al., 1986

↓ ICP, � CBF and CDO27.5% S, RLSSDogPrough et al., 1985

↓ ICP, ↑CBF, cerebral dehydration1.6% S, NSNoneRabbitTodd et al., 1985

Results after HSStudy solutionsInjuryAnimal

↓ ICP, ↑ CBF, ↑ CDO21,5% SL, RLSSPigSchmoker et al., 1991

↓ ICP, ↓ ICE6%S,6%HE,NS,WBSSPigDucey et al., 1989

↓ ICP after HS, ↓ CBF and CDO2both groups

7.5% S, RLSSDogPrough et al., 1986

� ICP after HS, ↑ICP after NS, D40

3%S, NS, 6%DSSDogGunnar et al., 1986

↓ ICP, � CBF and CDO27.5% S, RLSSDogPrough et al., 1985

↓ ICP, ↑CBF, cerebral dehydration1.6% S, NSNoneRabbitTodd et al., 1985

Results after HSStudy solutionsInjuryAnimal

↓, decreased; ↑, increased; �, unchanged; SS, sirculatory shock induced by bleeding; HS, hypertonic saline; S, saline; NS, 0.9% saline, RL, lactated Ringers; SL, sodium lactate; D, Dextrane 40; HE, hetastarch; WB, whole blood; ICP, intracranial pressure; CBF, cerebral blood flow, CDO2, cerebral oxygen delivery; ICE, intracranial elastance

Table 2. Animal studies investigating cerebral effects of HS with cerebral trauma.

↓ICP, HSD also ↓morphological damage

HSD, 7,2%S/HES, M, NS

SAHRatBermueller et al., 2006

↓ICP, HSD also ↑ neurological recovery

HSD, 7.5%S, NSSAHRatZausinger et al., 2004

�CBF, ↓infarct size7.5%S/10%HES, 10%HES, NS

VORatHeimann et al., 2003

↓ICP, rebound ↑ ICP7.2%S, 7,2%S/20%HES, 20%HES, 0.8%S

SS, SBDogPrough et al., 1999

↓ICP, BWC equal7.5%S, RLH, CryoSheepAnderson et al., 1997

↓ICP, ↑ ScO27.5%S, RLSS, CryoPigletTaylor et al., 1996

↓ CDO2 for both3%S, 10%HEH, PFICatDeWitt et al., 1996

↓ICP, ↑ CBF, ↑ CDO2, ↓contralat BWC

1.5%SL, RLCryoPigShackford et al., 1992

↓ ICP, ↑ CBF7.2% S, 0.8% SSS, SBDogPrough et al., 1991

↓ ICP, ↑ CBFHSD, RLSS, CryoPigWalsh et al., 1991

↓ ICP, ↓ contralat BWC7.5%S, RLHT, CryoSheepBattistella and Wisner, 1991

Less ↑ ICP, ↓ contralat BWC1.4%RLHD, CryoRabbitZornow et al., 1989

↓ ICP, ↓BWC3%S, NS, 10%DSS, EBDogGunnar et al., 1988

Results after HSStudy solutionsInjuryAnimal

↓ICP, HSD also ↓morphological damage

HSD, 7,2%S/HES, M, NS

SAHRatBermueller et al., 2006

↓ICP, HSD also ↑ neurological recovery

HSD, 7.5%S, NSSAHRatZausinger et al., 2004

�CBF, ↓infarct size7.5%S/10%HES, 10%HES, NS

VORatHeimann et al., 2003

↓ICP, rebound ↑ ICP7.2%S, 7,2%S/20%HES, 20%HES, 0.8%S

SS, SBDogPrough et al., 1999

↓ICP, BWC equal7.5%S, RLH, CryoSheepAnderson et al., 1997

↓ICP, ↑ ScO27.5%S, RLSS, CryoPigletTaylor et al., 1996

↓ CDO2 for both3%S, 10%HEH, PFICatDeWitt et al., 1996

↓ICP, ↑ CBF, ↑ CDO2, ↓contralat BWC

1.5%SL, RLCryoPigShackford et al., 1992

↓ ICP, ↑ CBF7.2% S, 0.8% SSS, SBDogPrough et al., 1991

↓ ICP, ↑ CBFHSD, RLSS, CryoPigWalsh et al., 1991

↓ ICP, ↓ contralat BWC7.5%S, RLHT, CryoSheepBattistella and Wisner, 1991

Less ↑ ICP, ↓ contralat BWC1.4%RLHD, CryoRabbitZornow et al., 1989

↓ ICP, ↓BWC3%S, NS, 10%DSS, EBDogGunnar et al., 1988

Results after HSStudy solutionsInjuryAnimal

↓, decreased; ↑, increased; �, unchanged; SS, circulatory shock induced by bleeding; H, moderate haemorrhage; HD, hemodilusion; HT, hypotension; EB, epidural balloon; SB subdural balloon; Cryo, cryogenic lesion; PFI, fluid-percussion injury; VO, occlusion of cortical veins; SAH, subarachnoid haemorrhage; HS, hypertonic saline; S, saline; NS, 0.9% saline, RL, lactated Ringer; SL, sodium lactate; D, Dextrane 40; HSD, 7.5%S in 6%D;HE, hetastarch; HES, hydroxyethyl starch; M, 20% mannitol; ICP, intracranial pressure; CBF, cerebral blood flow, CDO2, cerebral oxygen delivery; BWC, brain water content ; ScO2, cerebrovascular oxygen saturation

22

Human studies

Several human studies have been performed looking at the effect of HS on ICP (Table

3, page 23). Most of these were published prior to the publication of the studies included in

this thesis, but also a handful during the time during completion of this thesis. I have included

all of these in Table 3 (page 23) for completeness.

The main finding in these reports is that of a reliable reduction in ICP after HS

administration. In studies comparing HS to mannitol, HS is either equal or superior to

mannitol. The magnitude of the maximum effect is described, but information about the time

course of the effect is most often lacking. When is maximum effect reached? How long does

the effect last, and is there rebound in patients with damaged BBB?

Only two retrospective studies from Baltimore report negative effects. Increased in-

hospital mortality was found in one study (Qureshi et al., 1999) and loss of effect on ICP after

3 – 4 days in another (Qureshi et al., 1998). Both these studies applied continuous infusion of

3% sodium chloride / sodium acetate during several days. The indication for HS was cerebral

oedema with or without intracranial hypertension. Mean ICP before initiation of HS was < 20

mmHg in these studies and infusion was continued regardless of ICP. There was also a greater

incidence of penetrating injury in the HS group in the study from 1999. My interpretation of

these studies is that without signs of intracranial hypertension or decreased intracranial

compliance, there is no clear indication for continuous infusion of HS.

Many of the studies are conducted with mixed patient populations, although there is a

majority of TBI patients. All together there are only ten SAH patients included in the

prospective studies. The effect of HS on ICP may very well differ between patient groups.

Patients with localized BBB disruption will most probably respond much better to

osmotherapy than those with generalized BBB disruption.

23

Table 3. Studies exploring the effect of hypertonic saline on ICP in humans.

Reversal of TTH in 75% of cases,�ICP

30-60 mL 23.4% S(bolus)

68(76)

Retrospective, cohort

8stroke, 29ICH, 16 SAH, 8 BT, 7other;TTH

Koenig et al.,2008

�ICP, �CPP, �PI300 mL 3% S, (daily bolus, 20 min. if ↑ ICP)

18(38)

Prospective, observational

TBI; ICP > 20mmHg

Huang et al., 2006

↓ ICP, equal reduction, but longer duration

23,4% S vs M (bolus, S after M)

13RetrospectiveTBI; ↑ICP resistant to conv. therapy

Ware et al., 2005

↓ICP, mean 57% vs 48% from baseline

HSH vs 15% M (1.4 vs 1.8mL/kg)

32Prospective, randomized

10TBI, 9SAH, 7CI, 4ICH, 2other; ICP > 20mmHg

Harutjunyan et al., 2005

↓ ICP, greater reduction and longer duration

100ml HSD vs. 200ml 20% M(bolus, 5 min.)

9Prospective, randomized,cross-over

TBI; ICP > 20mmHg

Battison et al., 2005

Less time with ↑ ICP, less clinical failure

7.5% S vs. 20% M 2 ml/kg

20Prospective, randomized

TBI; ↑ICP resistant to conv. Therapy

Vialet et al., 2003

↓ ICP, � CVP. M: ↓CVP

3% S vs. M(boluses, 3 days)

30Prospective, randomized

BT; surgeryDe Vivo et al., 2001

↓ ICP, no adverse effects, 3 dead due to ↑ ICP

3% S (cont. infusion)

68RetrospectiveTBI, childrenPeterson et al., 2000

↓ ICP, no adverse effects, peak serum sodium 157-187 mEq/L

3% S (cont. infusion 4 -18 days)

10Prospective, observational

TBI, children; ↑ICP resistant to conv. therapy

Khanna et al., 2000

↓ ICP, mean ~30% from baseline, ↑CPP, ↑ CO

1.5 mL/kg, 7,2%S (bolus, 15 min.)

14Prospective, observational

TBI; GCS < 13, ICP >15mmHg

Munar et al., 2000

↑ in-hospital mortality3% S vs. NS(infusjon)

36 46c

Retrospective case-control

TBIQureshi et al., 1999

↓ICP, mean 42% from baseline, mean duration 3 hours

2 mL/kg 7.5% S (bolus)

10(48)

Prospective, observational

6TBI, 4SAH; ↑ICP resistant to conv. therapy

Horn et al., 1999

Improved haemodynamics, no adverse effects

3% S (cont. infusion)

29RetrospectiveSAH with vasospasm and hyponatremia

Suarez et al., 1999

↓ ICP, ↓ventilator time, ↓RDS, ↑survivale

1,7% S vs. RL(cont. infusion)

32Prospective, randomized

TBI, childrenSimma et al., 1998

↓ICP, mean 62% from baseline, mean duration several hours.

30 mL 23,4%S (bolus)

8(20)

Retrospective5SAH, 3 other; ↑ICP resistant to conv. therapy

Suarez et al., 1998

↓ICP, mean 43% from baseline, mean duration 101min.

100mL 10% S(bolus, 5 min.)

6(42)

Prospective, observational

5TBI, 1SAH; ↑ICP, ↓CPP, no effect of M, THAM, Sorbitol

Schatzmann et al., 1998

mean ΔICP: HSH: -11 mmHg, M: -6 mmHg

100mL HSH, vs. 200mL, 20% M (bolus, 15 min.)

9(30)

Prospective, randomized,cross-over

Stroke; ↑ICP >25 mmHg or pupil dilatation

Schwarz et al., 1998

↓ ICP, ↓ oedema, rebound after 3-4 days

3%S(cont. infusion)

27(30)

Retrospective8TBI, 4BT, 1SAH, 8ICH, 6CI

Qureshi et al., 1998

Equally safe, HS group more severely ill

1.6%S vs. RL34Prospective, randomized

TBI – resuscitationShackford et al., 1998

↓ ICP, ↑ CPP, 5 survivedHSH (bolus)

6 (32)

Prospective, observational

TBI; ↑ICP resistant to conv. therapy

Hartl et al., 1997a

Immediate ↓ ICP 7.5% S (bolus)

1Case reportTBI; ↑ICP resistant to conv. therapy

Einhaus et al., 1996

↓ ICP 3% vs. 0.9%S (10mL/kg bolus)

18Prospective, rando-mized, cross-over

TBI, childrenFisher et al., 1992

Immediate ↓ ICP29.2% S (bolus)

2Case-seriesTBI; ↑ICP resistant to conv. therapy

Worthley et al., 1988

Results after HSStudy solutionsNStudy designPatient population

Reversal of TTH in 75% of cases,�ICP

30-60 mL 23.4% S(bolus)

68(76)

Retrospective, cohort

8stroke, 29ICH, 16 SAH, 8 BT, 7other;TTH

Koenig et al.,2008

�ICP, �CPP, �PI300 mL 3% S, (daily bolus, 20 min. if ↑ ICP)

18(38)

Prospective, observational

TBI; ICP > 20mmHg

Huang et al., 2006

↓ ICP, equal reduction, but longer duration

23,4% S vs M (bolus, S after M)

13RetrospectiveTBI; ↑ICP resistant to conv. therapy

Ware et al., 2005

↓ICP, mean 57% vs 48% from baseline

HSH vs 15% M (1.4 vs 1.8mL/kg)

32Prospective, randomized

10TBI, 9SAH, 7CI, 4ICH, 2other; ICP > 20mmHg

Harutjunyan et al., 2005

↓ ICP, greater reduction and longer duration

100ml HSD vs. 200ml 20% M(bolus, 5 min.)

9Prospective, randomized,cross-over

TBI; ICP > 20mmHg

Battison et al., 2005

Less time with ↑ ICP, less clinical failure

7.5% S vs. 20% M 2 ml/kg

20Prospective, randomized

TBI; ↑ICP resistant to conv. Therapy

Vialet et al., 2003

↓ ICP, � CVP. M: ↓CVP

3% S vs. M(boluses, 3 days)

30Prospective, randomized

BT; surgeryDe Vivo et al., 2001

↓ ICP, no adverse effects, 3 dead due to ↑ ICP

3% S (cont. infusion)

68RetrospectiveTBI, childrenPeterson et al., 2000

↓ ICP, no adverse effects, peak serum sodium 157-187 mEq/L

3% S (cont. infusion 4 -18 days)

10Prospective, observational

TBI, children; ↑ICP resistant to conv. therapy

Khanna et al., 2000

↓ ICP, mean ~30% from baseline, ↑CPP, ↑ CO

1.5 mL/kg, 7,2%S (bolus, 15 min.)

14Prospective, observational

TBI; GCS < 13, ICP >15mmHg

Munar et al., 2000

↑ in-hospital mortality3% S vs. NS(infusjon)

36 46c

Retrospective case-control

TBIQureshi et al., 1999

↓ICP, mean 42% from baseline, mean duration 3 hours

2 mL/kg 7.5% S (bolus)

10(48)

Prospective, observational

6TBI, 4SAH; ↑ICP resistant to conv. therapy

Horn et al., 1999

Improved haemodynamics, no adverse effects

3% S (cont. infusion)

29RetrospectiveSAH with vasospasm and hyponatremia

Suarez et al., 1999

↓ ICP, ↓ventilator time, ↓RDS, ↑survivale

1,7% S vs. RL(cont. infusion)

32Prospective, randomized

TBI, childrenSimma et al., 1998

↓ICP, mean 62% from baseline, mean duration several hours.

30 mL 23,4%S (bolus)

8(20)

Retrospective5SAH, 3 other; ↑ICP resistant to conv. therapy

Suarez et al., 1998

↓ICP, mean 43% from baseline, mean duration 101min.

100mL 10% S(bolus, 5 min.)

6(42)

Prospective, observational

5TBI, 1SAH; ↑ICP, ↓CPP, no effect of M, THAM, Sorbitol

Schatzmann et al., 1998

mean ΔICP: HSH: -11 mmHg, M: -6 mmHg

100mL HSH, vs. 200mL, 20% M (bolus, 15 min.)

9(30)

Prospective, randomized,cross-over

Stroke; ↑ICP >25 mmHg or pupil dilatation

Schwarz et al., 1998

↓ ICP, ↓ oedema, rebound after 3-4 days

3%S(cont. infusion)

27(30)

Retrospective8TBI, 4BT, 1SAH, 8ICH, 6CI

Qureshi et al., 1998

Equally safe, HS group more severely ill

1.6%S vs. RL34Prospective, randomized

TBI – resuscitationShackford et al., 1998

↓ ICP, ↑ CPP, 5 survivedHSH (bolus)

6 (32)

Prospective, observational

TBI; ↑ICP resistant to conv. therapy

Hartl et al., 1997a

Immediate ↓ ICP 7.5% S (bolus)

1Case reportTBI; ↑ICP resistant to conv. therapy

Einhaus et al., 1996

↓ ICP 3% vs. 0.9%S (10mL/kg bolus)

18Prospective, rando-mized, cross-over

TBI, childrenFisher et al., 1992

Immediate ↓ ICP29.2% S (bolus)

2Case-seriesTBI; ↑ICP resistant to conv. therapy

Worthley et al., 1988

Results after HSStudy solutionsNStudy designPatient population

↓, decreased; ↑, increased; �, unchanged; N, number of patients (interventions); ICP, intracranial pressure; CPP, cerebral perfusion pressure; PI, cerebrovascular pulsatility index; BO, brain oedema; TBI, traumatic brain injury; SAH, subarachnoid haemorrhage; PO, postoperative; BT, brain tumour; ICH, non-traumatic intracerebral haemorrhage; CI, cerebral infarction; TTH, transtentorial herniation; GCS, Glasgow coma score; CO, cardiac output; HS, hypertonic saline; S, saline; NS, 0.9% saline; HES, hydroxyethyl starch; RL, lactated Ringer; HSH, 7.2-7.5% HS in 6% HES; HSD, 7.5%S in 6%Dextran; M, mannitol; RDS, respiratory distress syndrome

24

Intracranial pressure monitoring

Mean ICP

Current state-of-art technology computes mean ICP during short time windows (e.g. 5 – 15 s

duration). ICP can be measured by a number of different techniques, unfortunately all

invasive. This means that measuring ICP must be weighed against the risks of infection and

haemorrhage (Ghajar, 1995). The measurement of ICP via an intraventricular catheter, is

internationally considered “the gold standard” (Guillaume and Janny, 1951; Lundberg, 1960;

Czosnyka and Pickard, 2004; Steiner and Andrews, 2006). The advantages of this technique

are the possibility of CSF drainage and the possibility to perform re-zeroing of the system, i.e.

re-calibration towards atmospheric pressure. This technique requires, however, meticulous

attention to adjustments of the zero level as the patient’s position is changed, and the

monitoring is of course lost if the drain is clotted. The alternative technique is the use of a

miniature, solid, intraparenchymal sensor. These allow continuous pressure monitoring at the

same time as CSF can be drained through a ventricular catheter. The main disadvantage of

most of the available systems is that they can not be re-calibrated once inserted, and that the

zero pressure level may drift (Morgalla et al., 2001; Piper et al., 2001). Other techniques like

subarachnoid or epidural probes, lumbar CSF pressure, tympanic membrane displacement, or

transcranial Doppler, have not shown sufficient reliability for clinical use (Steiner and

Andrews, 2006).

There are other problems with the measurement of mean ICP in addition to the

technical aspects. We see in clinical practice how mean ICP changes abruptly when the

patient’s position is changed. If an intraparenchymal probe is inserted on the right side, and

the patients head is turned from resting on the left side to the right side, measured mean ICP

might increase as much as 10 mmHg. Looking at trend plots of mean ICP, we regularly see

25

these different ICP levels directly relating to patient position. This can be explained by the

simple fact that mean ICP is influenced by the weight of the brain above the point of

measurement. Eide demonstrated in a small study how mean ICP measurements recorded

simultaneously from two different intraparynchamal sensors differed considerably (Eide,

2006b). A difference of above 5 mmHg was recorded 13% of the time and above 10 mmHg

6% of the time. He describes how one sensor showed mean ICP > 20 mmHg while the other

showed < 15 mmHg. Such a difference might have direct implications on patient treatment, no

matter what the cause of the difference is.

Another problem is deciding on what level of mean ICP should bring about an

intervention to decrease ICP. What is a normal ICP? In young healthy volunteers lying down,

average mean ICP was 11 mmHg (range 7 to 15 mmHg) (Albeck et al., 1991). In the upright

position mean ICP was found to be in the range of -5 to +5 mmHg (Chapman et al., 1990).

Starting measurements from the supine position, the decrease to this level was found to occur

when elevating the head to about 45°, not changing much thereafter. In evaluating different

mean ICP levels to outcome, one should therefore pay close attention to the degree of head

elevation in order to be able to compare results.

Although there is no definite agreement on the threshold for treatment, the guidelines

from the Brain Trauma Foundation recommend treatment of mean ICP > 20 – 25 mmHg in

traumatic brain injury (TBI) patients (Brain Trauma Foundation et al., 2007b). The decision in

an individual patient is of course taken looking at ICP together with other variables like CPP,

transcranial Doppler, clinical status, and radiological findings. It is not within the scope of this

thesis to go into all these considerations, but at this point only to emphasise the need for

further research that can improve the predictive power of our intracranial pressure

measurements.

26

ICP waveform analysis

One path that might prove worth while is looking at the intracranial pressure wave and

intracranial compliance. Intracranial compliance relates to the changes in ICP subsequent to

changes in intracranial volume at different pressure levels. When ICP is low and compliance

is high, a small increase in volume only causes a small increase in pressure. If, however, the

pressure is high and the compliance is low, this same increase in volume yields a far greater

increase in pressure. It has been shown that this intracranial pressure-volume relationship is

better described by a pulsatile than a static ICP (Avezaat et al., 1979; Gonzalez-Darder and

Barcia-Salorio, 1989).

A method for direct monitoring of brain compliance has been implemented in the

Spiegelberg Brain Compliance monitor. This method relies on the evaluation of the pressure

response to known small volume additions by inflating and deflating a balloon inserted within

the cerebrospinal space. Initial trials indicate its usefulness, but implication for outcome

remains to be demonstrated (Piper et al., 1999; Yau et al., 2002).

A lot of work has been done by a group in Cambridge, UK, looking at waveform

analysis of ICP (pulsatile ICP) (Czosnyka and Pickard, 2004). The ICP waveform consists of

several components, but it is especially the amplitude of the component with a frequency

equal to the heart rate that has gained most focus, the intracranial pulse pressure or ICP

amplitude. The Cambridge group has described two ICP derived indexes (Czosnyka and

Pickard, 2004). The first is called RAP (correlation coefficient (R) between AMP amplitude

(A) and mean pressure (P); index of compensatory reserve). When RAP is close to +1, the

amplitude varies directly with ICP, indicating a low intracranial compensatory reserve. The

second index is a pressure-reactivity index called PRx. A positive index is correlated with

poor autoregulation. Abnormal values have been demonstrated to predict poor outcome after

head injury (Steiner et al., 2002).

27

The magnitude of “slow ICP waves” (frequency ½ - 3 per minute) has also been

shown to correlate with outcome, a greater magnitude indicating favourable outcome

(Lundberg, 1960; Balestreri et al., 2004).

My curiosity about waveform analysis and its possible use in clinical practice was

awakened by the work performed at our institution by Eide and co-workers. He has developed

at method of actually continuously measuring the heartbeat to heartbeat intracranial pulse

pressure and displaying this in real time at the bed side (Eide, 2006a). Previously, information

about single ICP waves has been derived from spectral analysis using fast Fourier

transformation (FFT) (Christensen and Borgesen, 1989). FFT does, however, not include an

algorithm for identification of the single ICP waves. The intracranial pulse pressure recorded

using the new method is called “mean ICP wave amplitude”, and it was found to relate

significantly to both the acute clinical state (Glasgow Coma Score, GCS) and the final clinical

outcome (Glasgow Outcome Score, GOS) in SAH patients. This relationship could in that

study not be demonstrated for mean ICP and mean CPP (Eide and Sorteberg, 2006).

Moreover, a case report showed how a long-standing bad clinical state of a SAH patient was

rapidly changed by turning management from being guided by static ICP (mean ICP) to being

pulsatility (or waveform) guided according to the mean ICP wave amplitude (Eide et al.,

2007c). It was also reported how a reduction in intracranial pulsatility could be achieved by

increased drainage of cerebro-spinal fluid (CSF) via an external ventricular drain (EVD) (Eide

et al., 2007c). Given the positive indications that mean ICP wave amplitude could predict

outcome, I found it important to investigate whether a medical intervention with HS could

decrease mean ICP wave amplitude. An additional incentive for contributing research results

regarding this new variable is all the technical and practical difficulties related to mean ICP

that I have mentioned earlier.

28

AIMS OF THE THESIS

1. Describe the effects of bolus infusion of 7.2% (72 mg/mL) saline in 6% (60 mg/mL)

hydroxyethyl starch 200/0.5 solution (HSS) on the intracranial pressure (ICP)

variables, mean ICP and mean ICP wave amplitude, and cerebral perfusion pressure

(CPP) in patients with subarachnoid haemorrhage (SAH).

2. Describe the time course of these effects, including time to maximum effect.

3. Describe the magnitude of these effects.

4. Describe the changes in serum sodium concentration, and how serum sodium is related

to ICP during the study period.

5. Describe haemodynamic effects.

6. Describe how mean ICP and mean ICP wave amplitude are related.

7. Compare the effect of HSS on mean ICP with the effect on mean ICP wave amplitude

(intracranial pulse pressure).

8. Compare the effect on intracranial pulse pressure with the effect on systemic arterial

pulse pressure.

29

METHODS All data in the three studies have been collected from patients suffering from acute, non-

traumatic subarachnoid haemorrhage (SAH) admitted to the Intensive Care Unit at

Rikshospitalet University Hospital.

All patients included were unconscious due to the nature of their disease and could not

give their informed consent to participation. Possible risks and benefits were therefore given

thorough evaluation. The protocols were approved by the Regional Ethics Committee for

Medical Research and the Norwegian Medicines Agency. Close relatives were given oral and

written information in accordance with the terms for approval by the Ethics Committee.

Study design

A common denominator of the studies was the effort taken to ensure as laboratory-like

conditions as possible, given a clinical setting. The first measure to achieve this was of course

to only include SAH patients, the second measure was to only include patients which were

relatively stable before intervention, and the third was to make sure that nothing was done

with the patient during the study period unless dictated by the rescue treatment protocol.

In both Study I and II we looked at effects of HSS on mean ICP, CPP, and

haemodynamics during a 210 minutes study period. In Study I relevant patients with mean

ICP > 20 mmHg were included. This being an observational study, Study II was given a

randomised, single-blinded, placebo-controlled design including patients with mean ICP < 20

mmHg. Judged together, these two studies would then give sound evidence of an effect (Study

II), magnitude of effect (Study I), and time course of the effect (Study I and II). Study III

utilized data collected in a still ongoing prospective study looking at different ICP variables

and their association to outcome. Instances where known doses of HSS were administered at

30

known times were selected from a large database of pressure recordings, and before and after

values were analysed.

All data were recorded prospectively. In Study I and II they were analysed according

to the predetermined plan drawn up in the protocols. In Study III, the data were prospectively

collected as part of a different study still ongoing. The protocol for Study III was designed

after data collection had started.

Data collection

Intracranial pressure was measured via a solid ICP sensor (Codman MicroSensor™, Codman,

Raynham, MA) coupled to a Codman® pressure transducer (Codman ICP Express™,

Codman, Raynham, MA). For the continuous arterial blood pressure (ABP) monitoring, an

arterial cannula was placed in a radial or femoral artery and connected to a Baxter fluid sensor

(Baxter, Deerfield, IL). Arterial blood pressure was zeroed at the level of the heart. Both

signals together with heart rate (HR), central venous pressure (CVP), and peripheral oxygen

saturation (SpO2) were coupled to a vital signs Siemens 9000XL Series Monitor (Siemens

AG, Munich, Germany). Cardiac index (CI), intrathoracic blood volume index (ITBI), and

extravascular lung water index (ELWI) were measured by use of the PiCCO system (Pulsion

Medical Systems, Munich, Germany). Arterial blood gases, pH, haemoglobin, and sodium

were analysed with the ABL 725 (Radiometer, Denmark).

In study I and II, values from the Siemens monitor and the PiCCO system were

registered electronically on a bed side computer every 30 seconds (LabView, National

Instruments, Austin, TX). False blood pressure values due to blood sampling were removed

manually before analysis.

In Study III a different setup for electronic data harvest was used. By means of the

Siemens Infinity Gateway Software, the continuous ICP and ABP signals were transferred on

31

line via the hospital network to a computer server and stored as raw data files with sampling

rates of 100 Hz. These data were analysed retrospectively using an algorithm implemented in

the software SM NeuroWave Version 2.0 (Eide, 2006a). Information on the time and amount

of hypertonic saline administered was collected from the patient charts.

Statistical methods

All patients included in the studies were reported according to the principle of intention-to-

treat. Study I and II are very similar in design. Repeated measurements of continuous data

were collected from different individuals. While Study I was an observational study of ten

interventions in seven patients, Study II was a placebo-controlled comparison of two groups,

total 22 patients. Mean values of measured variables for each patient from a 5 min period just

prior to infusion of study drug served as baseline. For data acquired electronically every 30

seconds, the mean value of all 5 minute periods throughout the 210 minute study period were

calculated and used for analysis.

In Study I comparisons of values at the time of maximum effect and at the end of the

study were compared with baseline values applying Wilcoxon’s test for paired samples. Even

though the changes were highly significant, we chose to display the graphs of mean ICP and

CPP for the ten individual cases. Since the number of cases was so low, we found this to be

the best way to convey the results truthfully to the reader.

In Study II baseline values were also calculated as the mean of a 5-minute period prior

to infusion of HSS, and mean values for each 5-minute interval throughout the study period

were used for analysis like in Study I. Differences between the groups were assessed by un-

paired Student’s t-test with Welch correction when there were unequal variance. For analysis

of demographic data, Mann-Whitney and chi-square tests were also used. The number of

patients was equal to the number of cases. In this study we did not only compare maximum

32

effect to baseline levels, we calculated the area under the curve for the different variables and

standardized by the length of the study to get the mean change for the whole period. This is a

recommended approach to statistical analysis of serial measurements (Matthews et al., 1990).

In Study III we had a total of 52 registrations from 20 patients, ranging from one to six

per patient. Due to possible bias because of repeated measurements, only the first registration

in each patient was included in the main results. This is in accordance with the statistical

recommendations that the sampling unit, or the “unit of analysis”, must be the patient in

studies like this one (Altman and Bland, 1997). We did however include information also

about calculations based on the total material of 52 registrations. A paired t-test was applied

comparing values before and after intervention for each variable, correlation was calculated

according to Pearson, and standard linear and multiple regression analyses were applied. P-

values <0.05/number of comparisons (Bonferroni correction) were considered significant.

33

SYNOPSIS OF RESULTS

Paper I

This study was designed to document the short term effects of 7.2% (72 mg/mL) saline in 6%

(60 mg/mL) hydroxyethyl starch 200/0.5 solution (HyperHAES®) (HSS) on intracranial

hypertension after subarachnoid haemorrhage (SAH). Primary outcome variables were

changes in mean intracranial pressure (ICP), changes in cerebral perfusion pressure (CPP),

and number of treatment failures, here defined as not reaching mean ICP < 20 mmHg and

CPP > 60 mmHg during the study period. Secondary outcome variables were changes in

cardiac index (CI), intrathoracic blood volume index (ITBI), and extravascular lung water

index (ELWI).

Ten interventions in seven patients were included. All patients suffered from

spontaneous SAH, their source of bleeding had been secured, and they had raised mean ICP >

20 mmHg. A bolus of 2 mL/kg of HSS was infused during 20 minutes, and registrations

continued for a total of 210 minutes. A laboratory like setting was pursued.

All interventions resulted in decreased mean ICP < 20 mmHg and elevation of CPP >

60 mmHg, i.e. zero treatment failures. The mean value for maximum mean ICP decrease in

percent of baseline was 58%, a mean decrease of 14.3 mmHg from a baseline of 25 mmHg.

Maximum effect was reached at 40 minutes after start of infusion. The mean peak increase in

CPP was 26%. In the eight cases recorded for 210 minutes, the mean ICP was still 35% below

baseline values at the end of the study. The time course of changes in ICP and CPP is

displayed in Figure 1B (page 38). We could also demonstrate an inverse relationship between

mean ICP and serum sodium levels. The maximum mean serum sodium increase was 6.6

mmol/L. There were no statistically significant changes to the secondary outcome variables

(measured in only five cases), only a trend towards increase in CI (p = 0.06) and ITBI (p =

0.06).

34

We concluded that the infusion of 2 mL/kg of HSS had a predictable and clinically

significant beneficial effect on mean ICP and CPP.

Paper II

Study I being an observational study, this study was designed to strengthen those results. A

group given 2 mL/kg of HSS was compared with a placebo group receiving 2 mL/kg of

normal saline (NS). In order to do that, the inclusion criterion as to intracranial pressure was

mean ICP between 10 and 20 mmHg. The design apart from this was very similar to Study I

with almost the same outcome variables. Changes in mean ICP and CPP were however

primarily analysed as the changes in mean values for the entire study period, not only

maximum change and change at the end of the study period.

In the placebo-group, there were no changes in any variables indicating that the study

design was valid. In the HSS group we found a significant reduction in mean ICP and increase

in CPP. Mean difference between the groups (HSS – NS) in average �ICP was -3.0 mmHg

and �CPP was 5.4 mmHg. Mean maximal change in mean ICP was -5.6 mmHg in the HSS

group. No rebound increase in ICP was found during the study period. The temporal change

was very similar to the one found in Study I (Figure 2, page 42). In addition, we demonstrated

a significant increase in cardiac output in the HSS group. The changes in serum sodium found

in the HSS group was quite similar to the findings in Study I, mean maximum increase of 5.6

mmol/L measured at 30 minutes, and a level 3.3 mmol/L above baseline at the end of the

study.

We concluded that HSS reduce mean ICP and increase CPP in SAH patients, and that

positive haemodynamic effects can be found. Judged together with Study I we concluded that

HSS attenuates increased mean ICP in SAH patients with a maximum effect reached at 20-30

35

minutes after the completition of the infusion. The magnitude of the effect would be expected

to be in the range demonstrated in Study I.

Paper III

In this study we compared the effect of HSS on mean ICP with the effect on mean ICP wave

amplitude in SAH patients. Mean ICP wave amplitude is the intracranial pulse pressure

recorded according to the algorithm developed by Eide; each cardiac beat-induced single

pressure wave is automatically identified, and the pressure variables are calculated for

consecutive 6-second time windows (Eide, 2006a). We also compared the effect on mean ICP

wave amplitude with the effect on systemic arterial pulse pressure.

The study included 52 interventions with different amounts of HSS in 20 SAH

patients. Only the first intervention in each patient was included in the main results due to

statistical reasons. It was found that both mean ICP and mean ICP wave amplitude were

attenuated by HSS. A correlation was found between change in mean ICP and mean ICP wave

amplitude, but no correlation was found between baseline values of mean ICP and mean ICP

wave amplitude. There was also a difference as to whether the treatment goal was reached.

The goals were defined as mean ICP < 15 mmHg and mean ICP wave amplitude < 5 mmHg.

The goal was reached in 65% of cases for mean ICP, while only in 30% of cases for mean ICP

wave amplitude. The study also documented that there was no correlation between systemic

arterial pulse pressure and intracranial pulse pressure, indicating that the intracranial pulse

pressure is no reflection of systemic pressure changes but represents the response to the

intracranial volume changes with each heartbeat.

In conclusion, the study documents for the first time that a drug, in this case HSS, has

an attenuating effect on pulsatile ICP. The low percentage of target achievement for mean ICP

wave amplitude indicates, however, that the intracranial compliance state was less favourable

36

than expected in a majority of these SAH patients with mean ICP and mean CPP values

within normal ranges. The results also showed that the value of mean ICP wave amplitude

could not be deduced from the value of mean ICP or vice versa in any given individual

patient.

37

DISCUSSION

All three studies included in this thesis are about the effect of 7.2% (72 mg/mL) saline in 6%

(60 mg/mL) hydroxyethyl starch 200/0.5 solution (HSS) on intracranial pressure (ICP) in

patients suffering from acute, spontaneous, subarachnoid haemorrhage (SAH). There are,

however, thematically two different trails that are relevant to separate. Study I and II on one

hand, are documenting effects on the well known and much investigated variables mean ICP

and cerebral perfusion pressure (CPP). Study III on the other hand, is more on the edge to less

well described territory, contributing a piece of evidence concerning the newly described

intracranial pressure variable; mean ICP wave amplitude.

To put the perspective right, the findings in this thesis can tell nothing about whether

the use of HSS had any bearing on neurological outcome in the patients studied. But being

based on clinical studies, the results can, judged together with available literature, have direct

implications for our current practice treating these patients. I hope to supply the fundament for

an educated answer to whether clinical practice can be modified based on the findings in this

thesis.

Before continuing, I would like to make a small pause, and close a 85 years long loop

between the initial findings in 1919 by Capt. Lewis H. Weed and 1st Lt. Paul S. McKibben at

The Army Neuro-Surgical Laboratory, Johns Hopkins Medical School, their extensive

research in cats, and our own findings in patients admitted to our intensive care unit at the

beginning of the following century. The point is to visually illustrate the similarities between

an ICP curve from one of their publications (Weed and McKibben, 1919b) (Figure 1A, page

38) with our ICP curve from Study I (Figure 1B, page 38).

During the time I have worked with this thesis, it has remained a mystery to me why

hypertonic saline was not re-discovered before, as an osmotic agent in the context of

38

intracranial hypertension. HS is cheap and has a much higher blood brain barrier (BBB)

reflection coefficient than the other drugs studied throughout the 20th century.

A B

Figure 1. A: ICP curve from cat no. 1271 after IV infusion of 12 mL of 30% saline (Weed and McKibben, 1919b). B: Mean change in ICP and CPP after IV infusion of 2 mL/kg of 7.2% saline in 6% hydroxyethyl starch 200/0.5 solution in SAH patients, n = 10 (Paper I). ICP, intracranial pressure; CPP, cerebral perfusion pressure; IV, intravenous. Figure 1A used with the permission from The American Physiological Society.

The relevance of ICP and CPP

A fundamental assumption for the use of osmotherapy to reduce ICP is that raised ICP

correlates with bad outcome. Even though there is no single study that beyond doubt proves

that monitoring of ICP improves outcome as opposed to not monitoring ICP, there are strong

indications that support its value.

With data from 428 traumatic brain injury (TBI) patients, Marmarou et al. found that

the proportion of hourly ICP readings greater than 20 mmHg was highly significant in

explaining outcome (Marmarou et al., 1991). Several other authors have also demonstrated

worsened outcome in patients with increased ICP (Marshall et al., 1979; Narayan et al., 1981;

Fearnside et al., 1993). However, Resnick et al. failed to demonstrate such a correlation for

ICP > 20 mmHg measured more than 96 hours after injury (Resnick et al., 1997). In 1999,

Robertson et al. published a randomised study in 189 patients with severe head injury,

39

demonstrating that secondary ischaemic insults could be prevented with the implementation

of a target management protocol based on ICP and CPP (Robertson et al., 1999). Patel and co-

workers have published an abstract showing a mortality rate twice as high among patients

treated in non-neurosurgical units compared with patients treated in neurosurgical units (Patel

et al., 2003). The availability of ICP monitoring is of course only one out of many factors

differing between the groups in this study. All of these studies refer to TBI patients, and

Treggiari et al. have summarised evidence in a recent review (Treggiari et al., 2007).

Citerio et al. have published a survey looking at the treatment of 350 SAH patients in

Italian neurosurgical centres, and found high intracranial pressure to be an independent factor

related to unfavourable outcome (Citerio et al., 2007). Two groups have found outcome to be

related to mean ICP and pulsatility in SAH patients (Eide and Sorteberg, 2006; Soehle et al.,

2007). Finally, Zanier et al. have documented the clinical advantages of monitoring ICP via a

computerised system over manual recordings (Zanier et al., 2007). Based on all this I find it

safe to conclude that monitoring of ICP and CPP is valuable to the patients suffering from

severe head injury and severe SAH.

Different ICP variables

As pointed out earlier, the first two papers are concerned with mean ICP, and the last also

with mean ICP wave amplitude. We have in the last paper referred to the two ICP variables

respectively as static and pulsatile ICP. The pulsatile ICP reflects the difference between

systolic and diastolic ICP, the static ICP (mean ICP) however, does not carry any information

about the shape of the ICP curve. The static ICP is influenced by a number of factors besides

pathological processes within the skull. Methodological factors such as the point of

measurement and baseline pressure level determined by zero calibration and drift, greatly

impacts the final measurement and has been discussed previously.

40

Pulsatile ICP (mean ICP wave amplitude) or intracranial pulse pressure is however in

many respects a more robust variable. It describes the dynamic pressure response evoked by

each cardiac beat. It does not need a correct zero calibration and is not influenced by drift of

the zero level (Eide, 2006b). Further, it is not influenced by alterations in head position like

described for mean ICP. In the study by Eide, while mean ICP differed considerably between

two sensors, the differences in single pressure wave amplitude were marginal (Eide, 2006b).

A difference of > 1 mmHg was found in only 0.8% of the time windows registered. Also, the

intracranial compliance is better described by the pulsatile than the static ICP (Avezaat et al.,

1979; Gonzalez-Darder and Barcia-Salorio, 1989; Eide et al., 2007a). With the availability of

pulsatile ICP at the bed-side in real time, I find this variable worth investigating as a

supplement to current standard neuromonitoring.

Effects of HSS on mean ICP and CPP in SAH patients

In Table 3 (page 23) I have listed the clinical studies looking at different concentrations of HS

used with the intention of controlling raised ICP. There are few SAH patients included in

these studies, and there are no studies with only SAH patients. There is also little information

about the time course of the effect. Inspired by this lack of knowledge, the two first studies

were designed to determine the short time effects on mean ICP and CPP of a HSS infusion in

SAH patients only. The reason that there is two and not one study is that I wanted to carry out

this research in a fashion as controlled as possible given a clinical setting, including a placebo

control. A placebo control group could not be justified including patients with mean ICP > 20

mmHg, which is the group of patients that is relevant to look at. Two quite similar studies

were therefore designed, one observational where all patients had mean ICP > 20 mmHg, and

one placebo-controlled where the patients had mean ICP 10 – 20 mmHg. An observational

study can never prove that an intervention was the cause of changes documented. The changes

41

could be due to natural courses or the effect of other interventions even if everything is kept

unaltered. The thought was that if the findings were similar in the two studies, the placebo-

controlled study would strengthen the findings from the observational study. The magnitude

of the effect of HSS would be reflected in the observational study where the inclusion criteria

for osmotherapy were relevant. In both studies we paid much attention to keeping all other

things that might affect ICP constant during the study period. No sudden increase in ICP

should lead to inclusion and a five minute period prior to infusion of HSS supplied baseline

values. During the 210 minute study period, the patients were not stimulated, their position

were unchanged, likewise ventilation, infusion rates for vasopressors, sedatives, and fluids.

The level of resistance towards CSF drainage was kept unaltered.

The results in the group that received HSS in Study II did closely mimic those of

Study I with a fall in mean ICP and rise in CPP, though with less absolute change as expected,

as the patients included in Study II had far lower ICP at inclusion. There was no change in the

placebo group (Figure 2, page 42). This indicated that the study design was valid, and that we

managed to control other variables that might have affected ICP. We therefore concluded that

HSS attenuates mean ICP and increases CPP in SAH patients.

Temporal pattern of change

The curves in Figure 2 (page 42) and Figure 1B (page 38) illustrates another important aspect;

i.e. the temporal pattern of the effect. Maximum effect is not reached immediately, as it is

when arterial CO2 partial pressure is reduced by hyperventilation. In our studies it took about

twice the infusion time before the maximum effect was reached. In Weed and McKibben’s

study it took 25-30 minutes after the end of the bolus (given during five minutes) (Figure 1A,

page 38). I have found only two clinical studies that contain such information. The first is a

report of two TBI patients given 250 and 100 mmol of NaCl during 10 minutes respectively

42

Figure 2. Changes in intracranial pressure (ICP) (top panel), and cerebral perfusion pressure (CPP) (bottom panel), after infusion of 2 mL/kg of study solution. HSS, 7.2% saline in 6% hydroxyethyl starch; NS, normal saline.

43

(Worthley et al., 1988). Maximum effect was reached about 20 minutes after the end of the

infusions deduced from graphs included in the paper. The second report is in a study by

Battison et al. including six TBI and three SAH patients, where HS was infused during 5

minutes. The paper includes a time-series plot from one patient, where the time to maximum

effect seems to be 20 – 30 minutes (Battison et al., 2005). In our studies I and II, average time

to maximum effect from start of infusion was 40 and 60 minutes respectively, the infusion

times being 20 and 30 minutes.

This is important information both for the practitioner and the scientist. For the

scientist, this indicates that the major effect of HS in the brain is due to removal of water, not

mainly due to haemodynamic effects that would be expected to display an immediate effect. It

may also say something about the properties of Aquaporine-4, if it is correct that these

channels represent a rate-limiting step for water flux across the blood-brain-barrier (BBB).

For the practitioner using HS in clinical practice, it is very important to know when maximum

effect of the intervention can be expected.

Study solution

Our choice of HS solution should be commented on. We chose a commercially available

solution (HyperHAES®) containing 7.2% (72 mg/mL) saline in 6% (60 mg/mL) hydroxyethyl

starch 200/0.5 (HSS). Converted to mmol, the sodium content is 1.23 mmol/mL. This solution

was chosen for several reasons. The combination of 7.2 to 7.5% saline in either hydroxyethyl

starch or dextrane has been extensively used in clinical research (Holcroft et al., 1987; Vassar

et al., 1991; Mattox et al., 1991; Vassar et al., 1993a; Vassar et al., 1993b), and is as such

known to be safe. We wanted our findings to be readily transferable to clinical practice, and

using a solution that is available seemed sensible. 3% saline is quite frequently used in

publications from the United States, but is seldom in stock in Norwegian hospitals. For the

44

reasons mentioned, and for issues related to the use of new drugs in medical research, we did

not want to prepare our own solution for these studies. All our patients were sedated and had a

central venous catheter in place, so the hypertonicity was not a problem. In acute, pre-hospital

settings, the solution can also be given in a peripheral vein. Using a less concentrated solution

would also have yielded a higher volume load which could become a problem in some

patients with cardiac and/or renal impairments.

As mentioned previously under Introduction (page 19), the addition of 6%

hydroxyethyl starch or dextrane to HS has been shown to prolong a positive haemodynamic

effect, but not influence the effect on ICP. I therefore find it justified to attribute the

intracranial pressure effects documented in our three studies to the HS component of the HSS

solution.

Dose and practical use

The dose used in the first two studies was 2 mL/kg of 7.2% saline in 6% hydroxyethyl starch

(HSS) (2.46 mmol Na+/kg). This is half of the conventional dose used in small-volume

haemodynamic resuscitation research, but we had found 2 mL/kg to be sufficient in pilots

prior to the studies. The idea was to not use a higher dose than necessary and thereby not

increase serum sodium more than necessary, so that the treatment could be repeated if the

clinical state indicates that. As discussed in Paper II, the findings from these studies indicate

that a standard dose might be lowered from 2 mL/kg, as the maximum reduction in mean ICP

in Paper I was > 50%. We now recommend a dose of 1 – 2 mL/kg. In practical use this has

evolved into a standard dose of 100 mL of HSS during 10-15 minutes in the average 60 – 90

kg adult. However, a proper dose-finding study should be made.

Are bolus doses the optimal way of administering HS as osmotherapy? There is no

definite answer to that. Most of the clinical studies listed in Table 3 (page 23) report only

45

single or repeated boluses, but some have used continuous infusion. Simma et al. randomised

children with TBI to an infusion of lactated Ringer’s solution (sodium 131 mmol/L) or HS

(sodium 268 mmol/L) with the goal of a serum sodium of 145-150 mmol/L during three days

in the HS group. An increase in serum sodium correlated with lower ICP and higher CPP.

They also reported fewer interventions, fewer complications, and shorter ICU stay in the HS

group (Simma et al., 1998). Qureshi et al. on the other hand reported increased in-hospital

mortality in a group of TBI-patients receiving a 3% HS-infusion as compared to normal saline

in a retrospective study (Qureshi et al., 1999). The infusion was targeted at a serum sodium

level of 145 – 155 mmol/L. There was, however, more penetrating injury in the HS group and

the infusion was continued even when mean ICP was normal. There are two studies on

children with TBI from San Diego applying a different approach (Khanna et al., 2000;

Peterson et al., 2000). In these studies, HS is infused on a sliding scale to achieve a target

serum sodium level that would maintain mean ICP < 20 mmHg. The results in these studies

were promising with good effect on ICP and CPP, and they also reported very high serum

sodium levels without adverse effects, mean highest level in the study by Khanna et al. was

170.7 mEq/L (range, 157 – 187 mEq/L).

I find the sliding scale approach logic, that the dose of HS should be dictated by the

need, here defined by the level of mean ICP, instead of targeting a predefined level for serum

sodium. It also underlines the need to pay attention to the serum sodium level at all times, not

allowing it to drift. However, in clinical practice there is often a need for a bolus intervention

in order to treat an increased ICP. Mere adjustments of a continuous infusion would not take

effect rapidly enough. I would therefore recommend to give a bolus of HS as described

previously when ICP > 20 mmHg and osmotherapy is indicated, followed by a continuous

infusion to maintain the serum sodium level achieved after the bolus. When needed, a new

46

bolus can be administered, and the infusion continued to maintain the new sodium level. This

can be pictured as a “stair-case approach” with respect to the serum sodium level.

Adverse effects

The classic adverse effect described in patients with rapidly increased serum sodium is central

pontine myelinolysis (CPM). This has typically been described after rapid correction of pre-

existing chronic hyponatraemia (Sterns et al., 1986; Vassar et al., 1990; Qureshi and Suarez,

2000; Verbalis, 2006). CPM has never been reported after elevation of a normal serum

sodium level (Qureshi and Suarez, 2000; Himmelseher, 2007), but breakdown of the BBB

followed by myelinolysis can occur if serum osmolality reaches 370 – 380 mOsm/kg (Soupart

et al., 1996). Intracranial haemorrhages have been reported in small children and cats after

abrupt changes in serum sodium (Finberg, 1967), but does not seem to be an issue in the

context of adult neurointensive care. However, we have in our institution implemented safety

precautions implying that one should avoid increasing serum sodium more than 15

mmol/L/day. We found in Study I and II mean maximum increase in serum sodium well

within this limit; 6.6 and 5.6 mmol/L after 2 mL/kg of HSS.

Even though the reflection coefficient (�) across intact BBB for NaCl is 1.0, one has to

be aware of possible rebound increase in ICP. Since one can imagine all degrees of damage to

the BBB, NaCl might leek into brain tissue in injured areas, later contributing to rebound

increase in ICP. One has also to be aware of the transport of organic osmolytes, e.g.

myoinositol, glutamine, glutamate, taurine, and urea into the brain, gradually compensating

for the acute hyperosmolar state created intravascularely after HS. (Lien et al., 1990; Sterns

and Silver, 2006; Verbalis, 2006). This means, that after culmination of the intracerebral

hypertension, one has to carefully and in a controlled manner reduce the serum sodium level

back toward normal levels, allowing the brain to lose the accumulated organic osmolytes

47

again. Fortunately, the loss of osmolytes is more rapid than the uptake (Verbalis, 2006), but

we recommend a reduction of serum sodium of no more than 10 mmol/L/day. I think that

caution in tapering of the hypernatremic state can eliminate some problems that can be

misconceived as rebound phenomena.

We do also advise caution in patients with unsecured aneurysms. A substantial

intracranial volume and pressure reduction might precipitate re-bleeding. I have however

never seen figures relating re-bleeding to the application of osmotherapy, but the general risk

of very early re-bleeding (first 24 hours) is as high as 15% when no antifibrinolytic drugs are

given (Hillman et al., 2002).

When using HS one has also always to be aware of the acute volume load that is

introduced to the circulation, paying special attention to patients with pre-existing heart

failure. We have, however, not encountered problems with this using 1 – 2 mL/kg of HSS.

Neither have we been able to attribute special clinical problems to the inevitable

hyperchloremic acidosis induced by the continuous and/or repeated administration of NaCl.

A significant risk of acute renal failure exists when mannitol is administered in large

doses, especially when serum osmolarity is > 320 mmol/L (Bullock, 1995) and the patient is

hypovolemic. This does not seem to be a problem to the same extent with HS, and a serum

osmolality > 320 mmol/L is not considered a contraindication to its use (Qureshi and Suarez,

2000). And as opposed to mannitol, repeated administration of HS would contribute to

correction of a hypovolemic state, rather than contributing to it.

Effects of HSS on mean ICP wave amplitude in SAH patients

Bed-side measurement of single wave intracranial pulse pressure is available based on a

method developed by Per Kristian Eide at our institution (Eide, 2006a). This is an interesting

variable, given the methodological problems with mean ICP discussed previously, and the

48

emerging documentation that mean ICP wave amplitude (intracranial pulse pressure) is better

correlated with outcome than mean ICP (Eide and Sorteberg, 2006; Eide, 2006c; Eide et al.,

2007b).

The measurement of intracranial pulse pressure is not a new thing. The fundamental

difference in the new algorithm is that mean ICP wave amplitude is calculated based on exact

measurements of single waves generated by each cardiac beat. Previous technologies have

either only measured maximum and minimum ICP during a given time frame, e.g. 15 – 20

seconds, the results thus “contaminated” by e.g. the effects of the respiratory cycle on the ICP

curve, or calculated on the basis of Fast Fourier transformation.

One important finding in Study III is that there is no correlation between the amplitude

of ICP and the amplitude of systemic arterial blood pressure (ABP). We can therefore

conclude that mean ICP wave amplitude is an independent variable reflecting the response to

intracranial volume changes induced by each cardiac beat, i.e. a reflection of the intracranial

compliance.

The main finding of Study III was that mean ICP wave amplitude was attenuated by

HSS. This has to my knowledge not been demonstrated before. However, the effect on mean

ICP wave amplitude was not directly comparable to the effect on mean ICP. First, there was

not a good agreement between the two measurements, signifying that the value of one of them

can not be deduced from the other. Secondly, there was a great discrepancy in target

achievement between the two variables. “Normal” mean ICP wave amplitude was only

achieved in a minority of cases while “normal” mean ICP was achieved in a majority of cases.

This suggests that the intracranial compliance state was less favourable than expected from

mean ICP in a majority of these patients. This is in concordance with the clinical perception

we have from our clinical day to day practice in the ICU, that a mean ICP of < 15 mmHg is

not “good enough” for a number of our SAH patients. For example, the mean ICP may be

49

normal, but the patient does not wake up. Only upon further reduction in mean ICP e.g by

increasing CSF drainage, does the patient wake up. This is illustrated by a case report by Eide

et al. where a SAH patient with normal mean ICP still was not waking up after 44 days (Eide

et al., 2007c). Only after changing to an ICP-wave guided management which brought about

increased CSF drainage, did the patient wake up. The notion that increased mean ICP wave

amplitude is indicative of an unfavourable state, is also supported by another case report by

Eide et al. that showed how increased amplitude was associated with increased levels of brain

metabolites indicative of ischemia (Eide et al., 2007a). All this challenges the thought that

“normal” mean ICP has the same implications in the injured and non-injured brain. These are

interesting aspects deserving further attention in the future.

A positive spin-off effect of measuring mean ICP wave amplitude with an online

display of the actual ICP curve drawn on the basis of measurements at 100 Hz, is that one has

a useful tool for judging whether the ICP measurement is valid. The diagnosis of a partly

dislodged catheter can readily be made just by looking at the curve. And in addition the

system will classify the measurements as invalid. This is information not supplied by mere

mean ICP.

Future clinical research questions

Reporting short time effects of hypertonic saline (HS) on intracranial pressure (ICP) variables

in subarachnoid haemorrhage (SAH) patients is important, but to link the use of HS or

measurements of increased mean ICP wave amplitude to effects on outcome; that is the real

challenge.

• Further studies are needed to better enlighten many aspects related to the use of

hypertonic saline in SAH patients with intracranial hypertension:

50

o Optimal dose

o Bolus, continuous infusion, or a combination

o Rebound phenomenon

o Timing related to time since injury

o Safety issues related to serum sodium levels and adverse effects

• Interesting and valuable knowledge would presumably emerge from studies

combining metabolic changes measured with micro dialysis, cerebral water content

and cerebral blood flow derived from MRI-investigations, and intracranial pressure

variables during administration of hypertonic saline.

• There is an immediate need for more data relating mean ICP and mean ICP wave

amplitude to neurological outcome.

51

MAIN CONCLUSIONS

All conclusions relate to patients suffering from acute spontaneous subarachnoid haemorrhage

(SAH) investigated in the intensive care setting.

1. Intravenous bolus infusion of 7.2% (72 mg/mL) saline in 6% (60 mg/mL)

hydroxyethyl starch 200/0.5 solution (HSS) during 20 – 30 minutes reduced

intracranial pressure (ICP), both mean ICP and mean ICP wave amplitude, and

increased cerebral perfusion pressure (CPP).

2. Intravenous bolus infusion of normal saline (single blind placebo) had no effect on

mean ICP or CPP, verifying stable study conditions.

3. When HSS was infused as a bolus during 20 to 30 minutes, the reduction in mean ICP

started almost immediately, but reached its maximum 20 to 30 minutes after the end of

the infusion, thereafter tapering off, but without a rebound within 3 hours.

4. When mean ICP was above 20 mmHg, 2 mL/kg of HSS (2.46 mmol Na+/kg) reduced

mean ICP in average by more than 50% at time of maximum effect, and mean ICP

remained below baseline three hours later.

5. Changes in mean ICP were inversely related to changes in serum sodium levels, and

mean maximum increase in serum sodium was about 6 mmol/L after 2 mL/kg of HSS .

6. A bolus infusion of 2 mL/kg of HSS increased cardiac output in our patients.

7. There was poor agreement between mean ICP and mean ICP wave amplitude, so the

level of one variable cannot be deduced from the other for any given individual

patient.

8. When the target level for mean ICP was reached after HSS infusion, the target for

mean ICP wave amplitude was reached only in a few patients. This suggests that the

52

intracranial compliance may be less favourable than expected in many SAH patients

with normal mean ICP and CPP.

9. There was no correlation between mean ICP wave amplitude (intracranial pulse

pressure) and systemic arterial pulse pressure. This supports the view that mean ICP

wave amplitude is reflecting the pressure response to intracranial volume changes

induced by each cardiac beat, i.e. intracranial compliance.

53

IMPLICATIONS FOR CLINICAL PRACTICE

The studies included in this thesis have documented that intravenous infusion of HSS has a

reliable effect on increased mean ICP in SAH patients. Mean ICP is lowered and as a

consequence of that, CPP is increased. Based on the findings of this thesis, which are

supported by other published studies, I now find there is enough evidence to recommend the

use of HS for osmotherapy in SAH patients, as an alternative to or a supplement to mannitol.

In clinical practice, the mode of administration could preferably be modified compared

with the regime presented in this thesis. The findings from study I indicates that a bolus dose

could be reduced from the 2 mL/kg of HSS used in the study, and in clinical practice, an

infusion time of 20 to 30 minutes would be perceived as too long. Further, our studies show

how the positive effects on the ICP variables are gradually reduced as serum sodium returns

to pre-infusion levels. I would therefore recommend to give a bolus of HS, e.g. NaCl 1 – 2

mmol/kg during 5 – 15 minutes, when ICP > 20 mmHg and osmotherapy is indicated,

followed by a continuous infusion to maintain the serum sodium level achieved after the

bolus. When needed, a new bolus can be administered, and the infusion continued to maintain

the new sodium level. Since serum sodium in our studies increased by about 6 mmol/L after

2.5 mmol/kg, there is room for repeated boluses per day without increasing serum sodium

more than 15 mmol/L/day. The inverse relationship between serum sodium and mean ICP,

also teaches us that a careful tapering of a hypernatremic state is important and necessary to

avoid rebound increase in mean ICP.

In our patients, cardiac output increased after HSS. This is in most cases clearly

beneficial, as many of these patients require vasopressors to achieve a satisfactory

haemodynamic situation. HSS may therefore reduce the need for such drugs.

54

It was not within the scope of this thesis to define the role of mean ICP wave

amplitude in clinical practice. More research is needed, but the findings from Study III have

taught us that the intracranial compliance situation is not necessarily good even though mean

ICP and CPP is normal. So at least in the cases where the radiology findings are fine, mean

ICP and CPP are fine, but the clinical state of the patient is not fine, these results lend support

to actively try to improve intracranial compliance e.g by infusion of HS or by increasing CSF

drainage.

55

REFERENCES

Agre P, King L S, Yasui M, Guggino W B, Ottersen O P, Fujiyoshi Y, Engel A, Nielsen S. Aquaporin water channels--from atomic structure to clinical medicine. J Physiol 2002; 542: 3-16.

Albeck M J, Borgesen S E, Gjerris F, Schmidt J F, Sorensen P S. Intracranial pressure and cerebrospinal fluid outflow conductance in healthy subjects. J Neurosurg 1991; 74: 597-600.

Altman D G, Bland J M. Statistics notes. Units of analysis. B M J 1997; 314: 1874.

Anderson J T, Wisner D H, Sullivan P E, Matteucci M, Freshman S, Hildreth J, Wagner F C, Jr. Initial small-volume hypertonic resuscitation of shock and brain injury: short- and long-term effects. J Trauma 1997; 42: 592-600.

Arai T, Tsukahara I, Nitta K, Watanabe T. Effects of mannitol on cerebral circulation after transient complete cerebral ischemia in dogs. Crit Care Med 1986; 14: 634-637.

Auer L M, Haselsberger K. Effect of intravenous mannitol on cat pial arteries and veins during normal and elevated intracranial pressure. Neurosurgery 1987; 21: 142-146.

Avezaat C J, van Eijndhoven J H, Wyper D J. Cerebrospinal fluid pulse pressure and intracranial volume-pressure relationships. J Neurol Neurosurg Psychiatry 1979; 42: 687-700.

Bacher A, Wei J, Grafe M R, Quast M J, Zornow M H. Serial determinations of cerebral water content by magnetic resonance imaging after an infusion of hypertonic saline. Crit Care Med 1998; 26: 108-114.

Bakke S J, Lindegaard K F. [Subarachnoid haemorrhage--diagnosis and management]. Tidsskr Nor Laegeforen 2007; 127: 1074-1078.

Balestreri M, Czosnyka M, Steiner L A, Schmidt E, Smielewski P, Matta B, Pickard J D. Intracranial hypertension: what additional information can be derived from ICP waveform after head injury? Acta Neurochir (Wien) 2004; 146: 131-141.

Battison C, Andrews P J, Graham C, Petty T. Randomized, controlled trial on the effect of a 20% mannitol solution and a 7.5% saline/6% dextran solution on increased intracranial pressure after brain injury. Crit Care Med 2005; 33: 196-202.

Battistella F D, Wisner D H. Combined hemorrhagic shock and head injury: effects of hypertonic saline (7.5%) resuscitation. J Trauma 1991; 31: 182-188.

Bermueller C, Thal S C, Plesnila N, Schmid-Elsaesser R, Kreimeier U, Zausinger S. Hypertonic fluid resuscitation from subarachnoid hemorrhage in rats: A comparison between small volume resuscitation and mannitol. J Neurol Sci 2006; 241: 73-82.

Bhardwaj A. Osmotherapy in neurocritical care. Curr Neurol Neurosci Rep 2007; 7: 513-521.

56

Bloch O, Manley G T. The role of aquaporin-4 in cerebral water transport and edema. Neurosurg Focus 2007; 22: E3.

Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care, AANS/CNS. Guidelines for the management of severe traumatic brain injury. II. Hyperosmolar therapy. J Neurotrauma 2007a; 24 Suppl 1: S14-S20.

Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care, AANS/CNS. Guidelines for the management of severe traumatic brain injury. VIII. Intracranial pressure thresholds. J Neurotrauma 2007b; 24 Suppl 1: S55-S58.

Broderick J P, Brott T, Tomsick T, Huster G, Miller R. The risk of subarachnoid and intracerebral hemorrhages in blacks as compared with whites. N Engl J Med 1992; 326: 733-736.

Bullock R. Mannitol and other diuretics in severe neurotrauma. New Horiz 1995; 3: 448-452.

Burke A M, Quest D O, Chien S, Cerri C. The effects of mannitol on blood viscosity. J Neurosurg 1981; 55: 550-553.

Cahill J, Calvert J W, Zhang J H. Mechanisms of early brain injury after subarachnoid hemorrhage. J Cereb Blood Flow Metab 2006; 26: 1341-1353.

Cantore G, Guidetti B, Virno M. Oral glycerol for the reduction of intracranial pressure. J Neurosurg 1964; 21: 278-283.

Cascino T, Baglivo J, Conti J, Szewczykowski J, Posner J B, Rottenberg D A. Quantitative CT assessment of furosemide- and mannitol-induced changes in brain water content. Neurology 1983; 33: 898-903.

Chapman P H, Cosman E R, Arnold M A. The relationship between ventricular fluid pressure and body position in normal subjects and subjects with shunts: a telemetric study. Neurosurgery 1990; 26: 181-189.

Chen C H, Xue R, Zhang J, Li X, Mori S, Bhardwaj A. Effect of osmotherapy with hypertonic saline on regional cerebral edema following experimental stroke: a study utilizing magnetic resonance imaging. Neurocrit Care 2007; 7: 92-100.

Chen L W, Huang H L, Lee I T, Hsu C M, Lu P J. Hypertonic saline enhances host defense to bacterial challenge by augmenting Toll-like receptors. Crit Care Med 2006; 34: 1758-1768.

Christensen L, Borgesen S E. Single pulse pressure wave analysis by fast Fourier transformation. Neurol Res 1989; 11: 197-200.

Citerio G, Gaini S M, Tomei G, Stocchetti N. Management of 350 aneurysmal subarachnoid hemorrhages in 22 Italian neurosurgical centers. Intensive Care Med 2007; 33: 1580-1586.

Claassen J, Carhuapoma J R, Kreiter K T, Du E Y, Connolly E S, Mayer S A. Global cerebral edema after subarachnoid hemorrhage: frequency, predictors, and impact on outcome. Stroke 2002; 33: 1225-1232.

57

Coimbra R, Hoyt D B, Junger W G, Angle N, Wolf P, Loomis W, Evers M F. Hypertonic saline resuscitation decreases susceptibility to sepsis after hemorrhagic shock. J Trauma 1997; 42: 602-606.

Czosnyka M, Pickard J D. Monitoring and interpretation of intracranial pressure. J Neurol Neurosurg Psychiatry 2004; 75: 813-821.

de Felippe Jr J, Timoner J, Velasco I T, Lopes O U, Rocha-E-Silva. Treatment of refractory hypovolaemic shock by 7.5% sodium chloride injections. Lancet 1980; 2: 1002-1004.

De Vivo P, Del Gaudio A, Ciritella P, Puopolo M, Chiarotti F, Mastronardi E. Hypertonic saline solution: a safe alternative to mannitol 18% in neurosurgery. Minerva Anestesiol 2001; 67: 603-611.

DeWitt D S, Prough D S, Deal D D, Vines S M, Hoen H. Hypertonic saline does not improve cerebral oxygen delivery after head injury and mild hemorrhage in cats. Crit Care Med 1996; 24: 109-117.

Doyle J A, Davis D P, Hoyt D B. The use of hypertonic saline in the treatment of traumatic brain injury. J Trauma 2001; 50: 367-383.

Ducey J P, Mozingo D W, Lamiell J M, Okerburg C, Gueller G E. A comparison of the cerebral and cardiovascular effects of complete resuscitation with isotonic and hypertonic saline, hetastarch, and whole blood following hemorrhage. J Trauma 1989; 29: 1510-1518.

Eide P K. A new method for processing of continuous intracranial pressure signals. Med Eng Phys 2006a; 28: 579-587.

Eide P K. Comparison of simultaneous continuous intracranial pressure (ICP) signals from a Codman and a Camino ICP sensor. Med Eng Phys 2006b; 28: 542-549.

Eide P K. Intracranial pressure parameters in idiopathic normal pressure hydrocephalus patients treated with ventriculo-peritoneal shunts. Acta Neurochir (Wien ) 2006c; 148: 21-29.

Eide P K, Bentsen G, Stanisic M, Stubhaug A. Association between intracranial pulse pressure levels and brain energy metabolism in a patient with an aneurysmal subarachnoid haemorrhage. Acta Anaesthesiol Scand 2007a; 51: 1273-1276.

Eide P K, Czosnyka M, Sorteberg W, Pickard J D, Smielewski P. Association between intracranial, arterial pulse pressure amplitudes and cerebral autoregulation in head injury patients. Neurol Res 2007b; 29: 578-582.

Eide P K, Sorteberg W. Intracranial pressure levels and single wave amplitudes, Glasgow Coma Score and Glasgow Outcome Score after subarachnoid haemorrhage. Acta Neurochir (Wien) 2006; 148: 1267-1276.

Eide P K, Sorteberg W, Meling T, Jorum E, Hald J, Stubhaug A. From intracranial pressure to intracranial pressure wave-guided intensive care management of a patient with an aneurysmal subarachnoid haemorrhage. Acta Anaesthesiol Scand 2007c; 51: 501-504.

58

Einhaus S L, Croce M A, Watridge C B, Lowery R, Fabian T C. The use of hypertonic saline for the treatment of increased intracranial pressure. J Tenn Med Assoc 1996; 89: 81-82.

Fearnside M R, Cook R J, McDougall P, McNeil R J. The Westmead Head Injury Project outcome in severe head injury. A comparative analysis of pre-hospital, clinical and CT variables. Br J Neurosurg 1993; 7: 267-279.

Finberg L. Dangers to infants caused by changes in osmolal concentration. Pediatrics 1967; 40: 1031-1034.

Fisher B, Thomas D, Peterson B. Hypertonic Saline Lowers Raised Intracranial Pressure in Children After Head Trauma. J Neurosurg Anesthesiol 1992; 4: 4-10.

Fremont-Smith F, Forbes H S. Intraocular and intracranial pressure: An experimental study. Archives of Neurology and Psychiatry 1927; 18: 550-564.

Garcia-Sola R, Pulido P, Capilla P. The immediate and long-term effects of mannitol and glycerol. A comparative experimental study. Acta Neurochir (Wien ) 1991; 109: 114-121.

Ghajar J. Intracranial pressure monitoring techniques. New Horiz 1995; 3: 395-399.

Gonzalez-Darder J M, Barcia-Salorio J L. Pulse amplitude and volume-pressure relationships in experimental hydrocephalus. Acta Neurochir (Wien ) 1989; 97: 166-170.

Guillaume J, Janny P. [Continuous intracranial manometry; importance of the method and first results.]. Rev Neurol (Paris) 1951; 84: 131-142.

Gunnar W, Jonasson O, Merlotti G, Stone J, Barrett J. Head injury and hemorrhagic shock: studies of the blood brain barrier and intracranial pressure after resuscitation with normal saline solution, 3% saline solution, and dextran-40. Surgery 1988; 103: 398-407.

Gunnar W P, Merlotti G J, Barrett J, Jonasson O. Resuscitation from hemorrhagic shock. Alterations of the intracranial pressure after normal saline, 3% saline and dextran-40. Ann Surg 1986; 204: 686-692.

Hackett M L, Anderson C S. Health outcomes 1 year after subarachnoid hemorrhage: An international population-based study. The Australian Cooperative Research on Subarachnoid Hemorrhage Study Group. Neurology 2000; 55: 658-662.

Hamaguchi S, Ogata H. Does hypertonic saline have preventive effects against delayed neuronal death in gerbil hippocampus? Shock 1995; 3: 280-283.

Hamaguchi S, Okuda Y, Kitajima T, Masawa N. Five percent, 7.5% or 10% hypertonic saline prevents delayed neuronal death in gerbils. Can J Anaesth 2002; 49: 745-748.

Hartl R, Ghajar J, Hochleuthner H, Mauritz W. Hypertonic/hyperoncotic saline reliably reduces ICP in severely head-injured patients with intracranial hypertension. Acta Neurochir Suppl (Wien) 1997a; 70: 126-129.

Hartl R, Medary M B, Ruge M, Arfors K E, Ghahremani F, Ghajar J. Hypertonic/hyperoncotic saline attenuates microcirculatory disturbances after traumatic brain injury. J Trauma 1997b; 42: S41-S47.

59

Harutjunyan L, Holz C, Rieger A, Menzel M, Grond S, Soukup J. Efficiency of 7.2% hypertonic saline hydroxyethyl starch 200/0.5 versus mannitol 15% in the treatment of increased intracranial pressure in neurosurgical patients - a randomized clinical trial [ISRCTN62699180]. Crit Care 2005; 9: R530-R540.

Heimann A, Takeshima T, Alessandri B, Noppens R, Kempski O. Effects of hypertonic/hyperoncotic treatment after rat cortical vein occlusion. Crit Care Med 2003; 31: 2495-2501.

Hillman J, Fridriksson S, Nilsson O, Yu Z, Saveland H, Jakobsson K E. Immediate administration of tranexamic acid and reduced incidence of early rebleeding after aneurysmal subarachnoid hemorrhage: a prospective randomized study. J Neurosurg 2002; 97: 771-778.

Himmelseher S. Hypertonic saline solutions for treatment of intracranial hypertension. Curr Opin Anaesthesiol 2007; 20: 414-426.

Holcroft J W, Vassar M J, Turner J E, Derlet R W, Kramer G C. 3% NaCl and 7.5% NaCl/dextran 70 in the resuscitation of severely injured patients. Ann Surg 1987; 206: 279-288.

Hop J W, Rinkel G J, Algra A, van G J. Case-fatality rates and functional outcome after subarachnoid hemorrhage: a systematic review. Stroke 1997; 28: 660-664.

Horn P, Munch E, Vajkoczy P, Herrmann P, Quintel M, Schilling L, Schmiedek P, Schurer L. Hypertonic saline solution for control of elevated intracranial pressure in patients with exhausted response to mannitol and barbiturates. Neurol Res 1999; 21: 758-764.

Huang S J, Chang L, Han Y Y, Lee Y C, Tu Y K. Efficacy and safety of hypertonic saline solutions in the treatment of severe head injury. Surg Neurol 2006; 65: 539-546.

Hughes J, Mudd S, Strecker E A. Reduction of elevated intracranial pressure by concentrated solutions of human lyophile serum. Archives of Neurology and Psychiatry 1938; 12: 1277-1287.

Hutchinson P, Timofeev I, Kirkpatrick P. Surgery for brain edema. Neurosurg Focus 2007; 22: E14.

Jackowski A, Crockard A, Burnstock G, Russell R R, Kristek F. The time course of intracranial pathophysiological changes following experimental subarachnoid haemorrhage in the rat. J Cereb Blood Flow Metab 1990; 10: 835-849.

Javid M. Urea in intracranial surgery. A new method. J Neurosurg 1961; 18: 51-57.

Javid M, Settlage P. Effect of urea on cerebrospinal fluid pressure in human subjects; preliminary report. J Am Med Assoc 1956; 160: 943-949.

Johnston S C, Selvin S, Gress D R. The burden, trends, and demographics of mortality from subarachnoid hemorrhage. Neurology 1998; 50: 1413-1418.

Kaufmann A M, Cardoso E R. Aggravation of vasogenic cerebral edema by multiple-dose mannitol. J Neurosurg 1992; 77: 584-589.

60

Keyrouz S G, Diringer M N. Clinical review: Prevention and therapy of vasospasm in subarachnoid hemorrhage. Crit Care 2007; 11: 220.

Khanna S, Davis D, Peterson B, Fisher B, Tung H, O'Quigley J, Deutsch R. Use of hypertonic saline in the treatment of severe refractory posttraumatic intracranial hypertension in pediatric traumatic brain injury [see comments]. Crit Care Med 2000; 28: 1144-1151.

Koenig M A, Bryan M, Lewin J L, III, Mirski M A, Geocadin R G, Stevens R D. Reversal of transtentorial herniation with hypertonic saline. Neurology 2008; 70: 1023-1029.

Koffijberg H, Buskens E, Granath F, Adami J, Ekbom A, Rinkel G J, Blomqvist P. Subarachnoid haemorrhage in Sweden 1987-2002: regional incidence and case fatality rates. J Neurol Neurosurg Psychiatry 2008; 79: 294-299.

Kofke W A. Mannitol: potential for rebound intracranial hypertension? J Neurosurg Anesthesiol 1993; 5: 1-3.

Kolsen-Petersen J A. Immune effect of hypertonic saline: fact or fiction? Acta Anaesthesiologica Scandinavica 2004; 48: 667-678.

Kramer G C, Perron P R, Lindsey D C, Ho H S, Gunther R A, Boyle W A, Holcroft J W. Small-volume resuscitation with hypertonic saline dextran solution. Surgery 1986; 100: 239-247.

Lien Y H, Shapiro J I, Chan L. Effects of hypernatremia on organic brain osmoles. J Clin Invest 1990; 85: 1427-1435.

Linn F H, Rinkel G J, Algra A, van G J. Incidence of subarachnoid hemorrhage: role of region, year, and rate of computed tomography: a meta-analysis. Stroke 1996; 27: 625-629.

Lundberg N. Continuous recording and control of ventricular fluid pressure in neurosurgical practice. Acta Psychiatr Scand Suppl 1960; 36: 1-193.

Marmarou A, Anderson R L, Ward J D, Choi S C, Young H F. Impact of ICP instability and hypotension on outcome in patients with severe head trauma. J Neurosurg 1991; 75: s59-s66.

Marshall L F, Smith R W, Shapiro H M. The outcome with aggressive treatment in severe head injuries. Part I: the significance of intracranial pressure monitoring. J Neurosurg 1979; 50: 20-25.

Matthews J N, Altman D G, Campbell M J, Royston P. Analysis of serial measurements in medical research. B M J 1990; 300: 230-235.

Mattox K L, Maningas P A, Moore E E, Mateer J R, Marx J A, Aprahamian C, Burch J M, Pepe P E. Prehospital hypertonic saline/dextran infusion for post-traumatic hypotension. The U.S.A. Multicenter Trial. Ann Surg 1991; 213: 482-491.

Mayer S A, Kreiter K T, Copeland D, Bernardini G L, Bates J E, Peery S, Claassen J, Du Y E, Connolly E S, Jr. Global and domain-specific cognitive impairment and outcome after subarachnoid hemorrhage. Neurology 2002; 59: 1750-1758.

61

Mazzoni M C, Borgstrom P, Intaglietta M, Arfors K E. Capillary narrowing in hemorrhagic shock is rectified by hyperosmotic saline-dextran reinfusion. Circ Shock 1990; 31: 407-418.

Menghini V V, Brown R D, Jr., Sicks J D, O'Fallon W M, Wiebers D O. Incidence and prevalence of intracranial aneurysms and hemorrhage in Olmsted County, Minnesota, 1965 to 1995. Neurology 1998; 51: 405-411.

Mocco J, Prickett C S, Komotar R J, Connolly E S, Mayer S A. Potential mechanisms and clinical significance of global cerebral edema following aneurysmal subarachnoid hemorrhage. Neurosurg Focus 2007; 22: E7.

Morgalla M H, Krasznai L, Dietz K, Mettenleiter H, Deininger M, Grote E H. Methods of experimental and clinical assessment of the relative measurement accuracy of an intracranial pressure transducer. Technical note. J Neurosurg 2001; 95: 529-532.

Muizelaar J P, Lutz H A, III, Becker D P. Effect of mannitol on ICP and CBF and correlation with pressure autoregulation in severely head-injured patients. J Neurosurg 1984; 61: 700-706.

Muizelaar J P, Wei E P, Kontos H A, Becker D P. Mannitol causes compensatory cerebral vasoconstriction and vasodilation in response to blood viscosity changes. J Neurosurg 1983; 59: 822-828.

Munar F, Ferrer A M, de Nadal M, Poca M A, Pedraza S, Sahuquillo J, Garnacho A. Cerebral hemodynamic effects of 7.2% hypertonic saline in patients with head injury and raised intracranial pressure. J Neurotrauma 2000; 17: 41-51.

Nakayama S, Kramer G C, Carlsen R C, Holcroft J W. Infusion of very hypertonic saline to bled rats: membrane potentials and fluid shifts. J Surg Res 1985; 38: 180-186.

Narayan R K, Greenberg R P, Miller J D, Enas G G, Choi S C, Kishore P R, Selhorst J B, Lutz H A, III, Becker D P. Improved confidence of outcome prediction in severe head injury. A comparative analysis of the clinical examination, multimodality evoked potentials, CT scanning, and intracranial pressure. J Neurosurg 1981; 54: 751-762.

Ohkuma H, Tsurutani H, Suzuki S. Incidence and significance of early aneurysmal rebleeding before neurosurgical or neurological management. Stroke 2001; 32: 1176-1180.

Paczynski R P. Osmotherapy. Basic concepts and controversies. Crit Care Clin 1997; 13: 105-129.

Patel H C, Bouamra O, King A, Lecky F. Mortality After Head Injury: Effect of Neurosurgical Care. Euroacademia Multidisciplinaria Neurotraumatologica. Graz, 2003 [Abstract]. Acta Neurochir (Wien) 2003; 145: 1135-1148.

Peterson B, Khanna S, Fisher B, Marshall L. Prolonged hypernatremia controls elevated intracranial pressure in head-injured pediatric patients [see comments]. Crit Care Med 2000; 28: 1136-1143.

Piepgras A, Thome C, Schmiedek P. Characterization of an anterior circulation rat subarachnoid hemorrhage model. Stroke 1995; 26: 2347-2352.

62

Piper I, Barnes A, Smith D, Dunn L. The Camino intracranial pressure sensor: is it optimal technology? An internal audit with a review of current intracranial pressure monitoring technologies. Neurosurgery 2001; 49: 1158-1164.

Piper I, Spiegelberg A, Whittle I, Signorini D, Mascia L. A comparative study of the Spiegelberg compliance device with a manual volume-injection method: a clinical evaluation in patients with hydrocephalus. Br J Neurosurg 1999; 13: 581-586.

Polderman K H, van de K G, Dixon J M, Vandertop W P, Girbes A R. Increases in spinal fluid osmolarity induced by mannitol. Crit Care Med 2003; 31: 584-590.

Preston G M, Carroll T P, Guggino W B, Agre P. Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science 1992; 256: 385-387.

Prough D S, Johnson J C, Poole G V, Stullken E H, Johnston W E, Royster R. Effects on intracranial pressure of resuscitation from hemorrhagic shock with hypertonic saline versus lactated Ringer's solution. Crit Care Med 1985; 13: 407-411.

Prough D S, Johnson J C, Stump D A, Stullken E H, Poole G V, Howard G. Effects of hypertonic saline versus lactated Ringer's solution on cerebral oxygen transport during resuscitation from hemorrhagic shock. J Neurosurg 1986; 64: 627-632.

Prough D S, Whitley J M, Taylor C L, Deal D D, DeWitt D S. Regional cerebral blood flow following resuscitation from hemorrhagic shock with hypertonic saline. Influence of a subdural mass. Anesthesiology 1991; 75: 319-327.

Prough D S, Whitley J M, Taylor C L, Deal D D, DeWitt D S. Rebound intracranial hypertension in dogs after resuscitation with hypertonic solutions from hemorrhagic shock accompanied by an intracranial mass lesion. J Neurosurg Anesthesiol 1999; 11: 102-111.

Prunell G F, Mathiesen T, Svendgaard N A. Experimental subarachnoid hemorrhage: cerebral blood flow and brain metabolism during the acute phase in three different models in the rat. Neurosurgery 2004; 54: 426-436.

Qureshi A I, Suarez J I. Use of hypertonic saline solutions in treatment of cerebral edema and intracranial hypertension. Crit Care Med 2000; 28: 3301-3313.

Qureshi A I, Suarez J I, Bhardwaj A, Mirski M, Schnitzer M S, Hanley D F, Ulatowski J A. Use of hypertonic (3%) saline/acetate infusion in the treatment of cerebral edema: Effect on intracranial pressure and lateral displacement of the brain [see comments]. Crit Care Med 1998; 26: 440-446.

Qureshi A I, Suarez J I, Castro A, Bhardwaj A. Use of hypertonic saline/acetate infusion in treatment of cerebral edema in patients with head trauma: experience at a single center. J Trauma 1999; 47: 659-665.

Raichle M E, Eichling J O, Straatmann M G, Welch M J, Larson K B, Ter-Pogossian M M. Blood-brain barrier permeability of 11C-labeled alcohols and 15O-labeled water. Am J Physiol 1976; 230: 543-552.

Rapoport S I. Transport in cells and tissues. In: Blood-Brain Barrier in Physilogy and Medicine. (Ed.Rapoport SI). New York: Raven Press, 1976; 17-42.

63

Ravussin P, Archer D P, Meyer E, bou-Madi M, Yamamoto L, Trop D. The effects of rapid infusions of saline and mannitol on cerebral blood volume and intracranial pressure in dogs. Can Anaesth Soc J 1985; 32: 506-515.

Resnick D K, Marion D W, Carlier P. Outcome analysis of patients with severe head injuries and prolonged intracranial hypertension. J Trauma 1997; 42: 1108-1111.

Robertson C S, Valadka A B, Hannay H J, Contant C F, Gopinath S P, Cormio M, Uzura M, Grossman R G. Prevention of secondary ischemic insults after severe head injury. Crit Care Med 1999; 27: 2086-2095.

Rudehill A, Gordon E, Ohman G, Lindqvist C, Andersson P. Pharmacokinetics and effects of mannitol on hemodynamics, blood and cerebrospinal fluid electrolytes, and osmolality during intracranial surgery. J Neurosurg Anesthesiol 1993; 5: 4-12.

Saltarini M, Massarutti D, Baldassarre M, Nardi G, De Colle C, Fabris G. Determination of cerebral water content by magnetic resonance imaging after small volume infusion of 18% hypertonic saline solution in a patient with refractory intracranial hypertension. Eur J Emerg Med 2002; 9: 262-265.

Schatzmann C, Heissler H E, Konig K, Klinge-Xhemajli P, Rickels E, Muhling M, Borschel M, Samii M. Treatment of elevated intracranial pressure by infusions of 10% saline in severely head injured patients. Acta Neurochir Suppl (Wien) 1998; 71: 31-33.

Schirmer C M, Hoit D A, Malek A M. Decompressive hemicraniectomy for the treatment of intractable intracranial hypertension after aneurysmal subarachnoid hemorrhage. Stroke 2007; 38: 987-992.

Schmoker J D, Zhuang J, Shackford S R. Hypertonic fluid resuscitation improves cerebral oxygen delivery and reduces intracranial pressure after hemorrhagic shock. J Trauma 1991; 31: 1607-1613.

Schubert G A, Thome C. Cerebral blood flow changes in acute subarachnoid hemorrhage. Front Biosci 2008; 13: 1594-1603.

Schwarz S, Schwab S, Bertram M, Aschoff A, Hacke W. Effects of hypertonic saline hydroxyethyl starch solution and mannitol in patients with increased intracranial pressure after stroke. Stroke 1998; 29: 1550-1555.

Shackford S R, Bourguignon P R, Wald S L, Rogers F B, Osler T M, Clark D E. Hypertonic saline resuscitation of patients with head injury: a prospective, randomized clinical trial. J Trauma 1998; 44: 50-58.

Shackford S R, Zhuang J, Schmoker J. Intravenous fluid tonicity: effect on intracranial pressure, cerebral blood flow, and cerebral oxygen delivery in focal brain injury. J Neurosurg 1992; 76: 91-98.

Shields C J, O'Sullivan A W, Wang J H, Winter D C, Kirwan W O, Redmond H P. Hypertonic saline enhances host response to bacterial challenge by augmenting receptor-independent neutrophil intracellular superoxide formation. Ann Surg 2003; 238: 249-257.

64

Simma B, Burger R, Falk M, Sacher P, Fanconi S. A prospective, randomized, and controlled study of fluid management in children with severe head injury: lactated Ringer's solution versus hypertonic saline. Crit Care Med 1998; 26: 1265-1270.

Smith G J, Kramer G C, Perron P, Nakayama S, Gunther R A, Holcroft J W. A comparison of several hypertonic solutions for resuscitation of bled sheep. J Surg Res 1985; 39: 517-528.

Soehle M, Chatfield D A, Czosnyka M, Kirkpatrick P J. Predictive value of initial clinical status, intracranial pressure and transcranial Doppler pulsatility after subarachnoid haemorrhage. Acta Neurochir (Wien) 2007; 149: 575-583.

Soupart A, Penninckx R, Namias B, Stenuit A, Perier O, Decaux G. Brain myelinolysis following hypernatremia in rats. J Neuropathol Exp Neurol 1996; 55: 106-113.

Stegmayr B, Eriksson M, Asplund K. Declining mortality from subarachnoid hemorrhage: changes in incidence and case fatality from 1985 through 2000. Stroke 2004; 35: 2059-2063.

Steiner L A, Andrews P J. Monitoring the injured brain: ICP and CBF. Br J Anaesth 2006; 97: 26-38.

Steiner L A, Czosnyka M, Piechnik S K, Smielewski P, Chatfield D, Menon D K, Pickard J D. Continuous monitoring of cerebrovascular pressure reactivity allows determination of optimal cerebral perfusion pressure in patients with traumatic brain injury. Crit Care Med 2002; 30: 733-738.

Sterns R H, Riggs J E, Schochet S S, Jr. Osmotic demyelination syndrome following correction of hyponatremia. N Engl J Med 1986; 314: 1535-1542.

Sterns R H, Silver S M. Brain volume regulation in response to hypo-osmolality and its correction. Am J Med 2006; 119: S12-S16.

Suarez J I, Qureshi A I, Bhardwaj A, Williams M A, Schnitzer M S, Mirski M, Hanley D F, Ulatowski J A. Treatment of refractory intracranial hypertension with 23.4% saline [see comments]. Crit Care Med 1998; 26: 1118-1122.

Suarez J I, Qureshi A I, Parekh P D, Razumovsky A, Tamargo R J, Bhardwaj A, Ulatowski J A. Administration of hypertonic (3%) sodium chloride/acetate in hyponatremic patients with symptomatic vasospasm following subarachnoid hemorrhage. J Neurosurg Anesthesiol 1999; 11: 178-184.

Sudlow C L, Warlow C P. Comparable studies of the incidence of stroke and its pathological types: results from an international collaboration. International Stroke Incidence Collaboration. Stroke 1997; 28: 491-499.

Talacchi A. Sequential measurements of cerebral blood flow in the acute phase of subarachnoid hemorrhage. J Neurosurg Sci 1993; 37: 9-18.

Taylor G, Myers S, Kurth C D, Duhaime A C, Yu M, McKernan M, Gallagher P, O'Neill J, Jr., Templeton J, Jr. Hypertonic saline improves brain resuscitation in a pediatric model of head injury and hemorrhagic shock. J Pediatr Surg 1996; 31: 65-70.

65

Teunissen L L, Rinkel G J, Algra A, van G J. Risk factors for subarachnoid hemorrhage: a systematic review. Stroke 1996; 27: 544-549.

Thiex R, Tsirka S E. Brain edema after intracerebral hemorrhage: mechanisms, treatment options, management strategies, and operative indications. Neurosurg Focus 2007; 22: E6.

Thomale U W, Griebenow M, Kroppenstedt S N, Unterberg A W, Stover J F. Small volume resuscitation with HyperHaes improves pericontusional perfusion and reduces lesion volume following controlled cortical impact injury in rats. J Neurotrauma 2004; 21: 1737-1746.

Todd M M, Tommasino C, Moore S. Cerebral effects of isovolemic hemodilution with a hypertonic saline solution. J Neurosurg 1985; 63: 944-948.

Toung T J, Chen C H, Lin C, Bhardwaj A. Osmotherapy with hypertonic saline attenuates water content in brain and extracerebral organs. Crit Care Med 2007; 35: 526-531.

Treggiari M M, Schutz N, Yanez N D, Romand J A. Role of intracranial pressure values and patterns in predicting outcome in traumatic brain injury: a systematic review. Neurocrit Care 2007; 6: 104-112.

van Gijn J, Rinkel G J. Subarachnoid haemorrhage: diagnosis, causes and management. Brain 2001; 124: 249-278.

Vassar M J, Fischer R P, O'Brien P E, Bachulis B L, Chambers J A, Hoyt D B, Holcroft J W. A multicenter trial for resuscitation of injured patients with 7.5% sodium chloride. The effect of added dextran 70. The Multicenter Group for the Study of Hypertonic Saline in Trauma Patients. Arch Surg 1993a; 128: 1003-1011.

Vassar M J, Perry C A, Gannaway W L, Holcroft J W. 7.5% sodium chloride/dextran for resuscitation of trauma patients undergoing helicopter transport. Arch Surg 1991; 126: 1065-1072.

Vassar M J, Perry C A, Holcroft J W. Analysis of potential risks associated with 7.5% sodium chloride resuscitation of traumatic shock [published erratum appears in Arch Surg 1991 Jan;126(1):43]. Arch Surg 1990; 125: 1309-1315.

Vassar M J, Perry C A, Holcroft J W. Prehospital resuscitation of hypotensive trauma patients with 7.5% NaCl versus 7.5% NaCl with added dextran: a controlled trial. J Trauma 1993b; 34: 622-632.

Velasco I T, Pontieri V, Rocha, Lopes O U. Hyperosmotic NaCl and severe hemorrhagic shock. Am J Physiol 1980; 239: H664-H673.

Velasco I T, Rocha, Oliveira M A, Silva R I. Hypertonic and hyperoncotic resuscitation from severe hemorrhagic shock in dogs: a comparative study. Crit Care Med 1989; 17: 261-264.

Verbalis J G. Control of brain volume during hypoosmolality and hyperosmolality. Adv Exp Med Biol 2006; 576: 113-129.

Vialet R, Albanese J, Thomachot L, Antonini F, Bourgouin A, Alliez B, Martin C. Isovolume hypertonic solutes (sodium chloride or mannitol) in the treatment of refractory

66

posttraumatic intracranial hypertension: 2 mL/kg 7.5% saline is more effective than 2 mL/kg 20% mannitol. Crit Care Med 2003; 31: 1683-1687.

Walsh J C, Zhuang J, Shackford S R. A comparison of hypertonic to isotonic fluid in the resuscitation of brain injury and hemorrhagic shock. J Surg Res 1991; 50: 284-292.

Ware M L, Nemani V M, Meeker M, Lee C, Morabito D J, Manley G T. Effects of 23.4% sodium chloride solution in reducing intracranial pressure in patients with traumatic brain injury: a preliminary study. Neurosurgery 2005; 57: 727-736.

Weed L H, McKibben P S. Experimental alteration of brain bulk. Am J Physiol 1919a; 48: 531-558.

Weed L H, McKibben P S. Pressure changes in the cerebrospinal fluid following intravenous injection of solutions of various concentrations. Am J Physiol 1919b; 48: 512-530.

Worthley L I, Cooper D J, Jones N. Treatment of resistant intracranial hypertension with hypertonic saline. Report of two cases. J Neurosurg 1988; 68: 478-481.

Yau Y H, Piper I R, Contant C F, Dunn L T, Whittle I R. Clinical experience in the use of the Spiegelberg automated compliance device in the assessment of patients with hydrocephalus. Acta Neurochir Suppl 2002; 81: 171-172.

Zanier E R, Ortolano F, Ghisoni L, Colombo A, Losappio S, Stocchetti N. Intracranial pressure monitoring in intensive care: clinical advantages of a computerized system over manual recording. Crit Care 2007; 11: R7.

Zausinger S, Thal S C, Kreimeier U, Messmer K, Schmid-Elsaesser R. Hypertonic fluid resuscitation from subarachnoid hemorrhage in rats. Neurosurgery 2004; 55: 679-686.

Ziai W C, Toung T J, Bhardwaj A. Hypertonic saline: first-line therapy for cerebral edema? J Neurol Sci 2007; 261: 157-166.

Zornow M H. Hypertonic saline as a safe and efficacious treatment of intracranial hypertension. J Neurosurg Anesthesiol 1996; 8: 175-177.

Zornow M H, Scheller M S, Shackford S R. Effect of a hypertonic lactated Ringer's solution on intracranial pressure and cerebral water content in a model of traumatic brain injury. J Trauma 1989; 29: 484-488.