You are on page 1of 88

Therapeutics

Intravenous Fluid Therapy


Edited by Andrew R Webb MD FRCP
Intravenous Fluid Therapy

Copyright© of this publication belongs to B. Braun Medical Ltd.


All rights are reserved and no part of this publication can be reproduced
or copied without the express permission in writing from B. Braun Medical Ltd.
Intravenous Fluid Therapy
Intravenous Fluid Therapy

Editorial

The subject of Intravenous Fluid Therapy is one that has been much
neglected by the medical professions and the pharmaceutical industry
in general.

There are many questions that have only partly been answered
regarding intravenous fluid therapy and the uses of the wide variety of
fluids currently in clinical use. New products are being developed by
the pharmaceutical industry and will no doubt find their places in the
clinician's armamentarium.

This edition of 'Therapeutics' provides a comprehensive review, by


leading experts in the field, of the current state of knowledge and
understanding about the clinical uses of these pharmaceuticals, and
points towards future developments.

I would like to take this opportunity to thank all the contributors for their
efforts in producing this review monograph and I hope that you, the
clinician find it a worthwhile and useful publication.

A Webb MD FRCP
Editor, 'Therapeutics'
September 2003

Therapeutics has been published with aid of an educational grant from


B. Braun Medical Ltd.
Intravenous Fluid Therapy Contributors

Joachim Boldt
Professor, Department of Anesthesiology and Intensive Care Medicine,
Klinikum der Stadt Ludwigshafen, Ludwigshafen, Germany
Edward Burdett
Clinical Fellow, Centre for Anaesthesia, University College London, London, UK
Peter Gosling
Consultant Clinical Scientist, University Hospital Birmingham,
Birmingham, UK
Hengo Haljamäe
Professor of Anaesthesiology and Intensive Care, Sahlgrenska
University Hospital, Göteborg, Sweden
Shumita Joseph
Staff Anaesthetist, The John Farman Intensive Care Unit,
Addenbrooke’s Hospital, Cambridge, UK
Roman Kocian
Department of Anesthesiology, University Hospital Lausanne, Lausanne,
Switzerland
Martin Kuper
Research Fellow Intensive Care, Chelsea and Westminster Hospital,
London, UK
Monty Mythen
Portex Professor of Anaesthesia and Critical Care, University College
London, London, UK
Peter Nightingale
Consultant in Anaesthesia & Intensive Care Medicine,
South Manchester University Hospital, Manchester, UK
Gilbert Park
Consultant in Anaesthesia, The John Farman Intensive Care Unit,
Addenbrooke’s Hospital, Cambridge, UK
Justin Roberts
SpR in Anaesthesia & Intensive Care Medicine,
South Manchester University Hospital, Manchester, UK
Neil Soni
Consultant ITU, Chelsea and Westminster Hospital, London, UK
Donat Spahn
Professor and Chairman, Department of Anesthesiology, University
Hospital Lausanne, Lausanne, Switzerland
Andrew R Webb
Medical Director and Consultant in Intensive Care,
University College London Hospitals, London, UK
Intravenous Fluid Therapy
Chapters Intravenous Fluid Therapy

Pages

History of intravenous fluid therapy 1 1-6


Andrew Webb

The Choice of intravenous fluid;


crystalloid or colloid 2 7 - 20
Martin Kuper and Neil Soni

Pathophysiology of capillary leak 3 21 - 32


Peter Gosling

Properties and use of crystalloids 4 33 - 38


Hengo Haljamäe

Properties and use of albumin 5 39 - 44


Shumita Joseph and Gilbert Park

Properties and use of gelatins 6 45 - 52


Justin Roberts and Peter Nightingale

Properties and use of


hydroxyethyl starches 7 53 - 62
Joachim Boldt

Properties and use of


artificial oxygen carriers 8 63 - 68
Donat Spahn and Roman Kocian

Clinical problems arising out


of intravenous fluid therapy 9 69 - 76
Edward Burdett and Monty Mythen

The importance of monitoring 10 77 - 82


Andrew Webb
Intravenous Fluid Therapy Chapter 1

History of intravenous fluid therapy


Andrew Webb

The first intravenous fluid infusion


Although blood transfusion had been attempted unsuccessfully in the
late 15th century the origins of intravenous fluid therapy can be traced
back to the British cholera epidemic of the 1830s. At that time there
was little understanding of the cause or physiology of cholera.
Treatment was reliant on blood letting as a result of the frequent
observation of black, thick blood. There were few that applied any
scientific thought to the problem until it was noted that the blood of a
cholera victim had no particular taste or smell and was depleted of the
water element.

William O’Shaughnessy, a young Edinburgh graduate with experience


of chemistry, wrote to The Lancet indicating that the blood had lost a
large proportion of its water, neutral saline, and was devoid of free
alkali. Where urine production was low he noted an increase in urea in
the blood. He later wrote to the Central Board of Health in 1832
indicating that cure of cholera was dependent on restoring the natural
specific gravity of the blood and restoring the deficient saline. He
stated the specific gravity could only be restored by imbibation or the
injection of aqueous fluid into the veins.

The technique of injecting drugs into veins had been known since the
17th century but Thomas Latta was the first to inject intravenous fluid
on hearing of William O’Shaughnessy’s analysis. He reported in The
Lancet in 1832 the miraculous but temporary recovery of an elderly
woman with cholera who was close to death, after injection of 6 pints
of fluid via the basilic vein. On witnessing the recovery, Thomas Latta
left the patient who deteriorated and died 5 hours later as vomiting and
diarrhoea returned. The recipe for the first infusion was a hypotonic
solution of saline and sodium bicarbonate infused at a temperature
higher than body temperature. Latta learned from the first case that
the diarrhoea returned as circulating volume was restored and the
treatment must be repeated as soon as the pulse fails. Thus, the need
for monitoring of intravenous fluid therapy was known from the
experience of the first infusion.

The use of intravenous fluid in cholera was slow to catch on. Many
believed the fluid infusions hastened death, failing to realise that the
diarrhoea and vomiting could not persist in the face of severe
dehydration. The return of diarrhoea and vomiting was therefore linked
to the intravenous fluid. Thomas Latta’s advice on repeating the
infusion when diarrhoea returned and the pulse weakened was not
heeded. In addition the unsterile preparation of the fluid increased the
risk of other infection (knowledge of infection did not exist at this time)
and the hypotonic nature of the infusion promoted haemolysis.

Laying the foundations for routine


intravenous fluid therapy
Florence Nightingale was dispatched to the Crimean War (1853-1856)
where mortality was higher due to disease rather than injury. Major
efforts went to improving sanitation and there was no use of
intravenous fluid. Research into the physiology of volume loss
1 continued in the late 19th century. It was not until some 30 years after
the end of the Crimean War that there was a recognition of the
importance of plasma volume rather than red cell content of blood in
maintaining life. Von Ott aus St Petersburg wrote in ‘Virchows Archives’
in 1883:
"The danger of blood losses of up to two thirds of the total blood
volume lies in the disproportion of vascular size and vascular content
caused by the loss, and the volume of fluid infused counteracts this
danger, no matter whether the fluid contains erythrocytes or not, so long
as it is not directly injurious, but rather is indifferent" (translated from the
original German).

The importance of albumin in maintaining the circulating volume in health


was noted by Starling in 1896. He wrote in the Journal of Physiology:
"I believe the explanation is to be found in a property on which much
stress was laid by the older physiologists and which they termed the
high endosmotic pressure of albumen.
…although the osmotic pressure of the proteids of the plasma is so
insignificant, it is of an order of magnitude comparable to that of the
capillary pressures; and whereas capillary pressure determines
transudation, the osmotic pressure of the proteids of the serum
determines absorption."

In 1899 Starling went on to note that haemodilution with saline caused a


diuresis which he believed was a result of dilution of albumin, i.e. a fall in
colloid osmotic pressure. Knowlton noted in 1911 that the diuresis
could be reduced by haemodilution with gelatin solution.

Development of intravenous
crystalloid solutions
The first intravenous crystalloid solution was an unsterile hypotonic
solution of saline and sodium bicarbonate approximately equivalent to
Sodium 58mmol/l, Chloride 49mmol/l and Bicarbonate 9mmol/l. The
appropriate formulation of salt solutions was developed from early
physiological experiments to support organ or tissue preparations in the
laboratory. The importance of infusing isotonic solutions to avoid
haemolysis was recognised in the late 19th century. At that time the use
of intravenous crystalloid remained rare and saline was the mainstay of
treatment. In 1896 Biedl and Krause successfully infused a glucose
solution intravenously.

While studying the action of inorganic salts on frog heart muscle Sidney
Ringer noted an improvement in muscle contraction when his laboratory
assistant substituted tap water for distilled water when making up saline
bath solutions. During the 1880s and as a result of these experiments
Ringer’s solution was developed as an isotonic solution of sodium,
potassium and calcium chloride. At this time the intravenous infusion of
balanced salt solution was not established practice.

During the early 20th century, intravenous fluid infusion remained rare,
being reserved for those who were the most extremely ill. The
introduction of the intravenous saline infusion routinely to support major
surgery is attributed to Rudolph Matas who use the technique in 1924.
Lactated Ringer’s solution was introduced by Alexis Hartmann in 1932
while studying metabolic disease in children.

Post operative salt water retention was recognised as part of the


metabolic response to surgery in 1944 leading to a reluctance amongst
surgeons and anaesthetists to administer crystalloids to their patients.
2
Intravenous Fluid Therapy Chapter 1

In 1964 Shires demonstrated an increased survival in animals bled and


then re-infused with the shed blood and crystalloid. The enthusiasm
for crystalloid infusion gained momentum during the late 1960s.

Development of intravenous colloid


solutions
Gum acacia was one of the first commercial colloid solutions but was
difficult to manufacture, was not metabolized by the body, agglutinated
red cells and caused serious reactions involving ‘cyanosis, pulmonary
oedema and death with acute atrophy of the liver.’ The first gelatins
tested in 1915 were abandoned since they were difficult to sterilize and
the gel point was above room temperature rendering them difficult to
keep in solution. Polyvinyl alcohol solutions were studied from the late
1930s but suffered from many of the problems associated with gum
acacia.

During the Second World War polyvinyl pyrolidone was used widely by
the German army. It was largely excreted unchanged in the urine but
larger molecules were stored indefinitely in reticuloendothelial system.

In 1940 Cohn developed a cold ethanol fractionation process to


separate plasma components. Albumin, gamma globulin and
fibrinogen were isolated and became available for clinical use. The use
of albumin infusion was first demonstrated by Elliott and used the
following year to treat shock in the victims of the Pearl Harbour attack.

Dextrans, produced by Leuconstoc mesenteroides from sucrose were


first recognised as potential plasma substitutes in 1944.
L. mesenteroides B512 was used to create the dextran molecule since
other strains produced dextrans with too many branching side chains.
The clinical products were created by partial hydrolysis and
fractionation of the dextran molecule. Dextran 75 was introduced in
1947 and Dextran 110 and Dextran 40 in 1966.

Further development of gelatins reduced the gel melting point by


hydrolysis with cross links introduced to maintain molecular size.
Oxypolygelatin was reported in 1951, formed by cross linking peptide
chains with dialdehyde glyoxal. Polygeline was reported in 1962 in
which urea bridges were created with hexamethyl di-isocyanate.
Succinylated gelatin was reported in 1955 and differed in that cross
links were not used; rather, a conformational change in the molecule
was achieved by replacing polypeptide amino groups with acid
carboxyl groups by reacting with succinic anhydride.
Hydroxyethyl starch was introduced in 1957 and the first hetastarch
introduced into clinical practice during the 1970s. Various molecular
modifications during the 1980s and 1990s leave us with the
pentastarches and low molecular weight starches in common use
today.

3
In 1957 Grönwall listed eight requirements of the ideal colloid solution:
"These requirements are primarily:
i. The colloid should have molecular dimensions which guarantee an
adequate colloid effect
ii. The solution should have a colloid osmotic pressure and viscosity
of the same order of magnitude as blood plasma
iii. The colloid substance should be as little foreign to the body as
possible and have no toxic properties. It should cause no injury by
being stored up in the body tissues or organs, but should be
eliminated from the body through metabolization and/or excretion
iv. The colloidal substance should remain in the body during a
sufficiently long time and at a concentration adequate to warrant a
therapeutic effect
v. The solution should not develop pyrogenic or allergic properties
vi. The solution must stand sterilisation by autoclaving and be free
from virus contamination
vii. It should be possible to produce the solution with constant and
clearly defined properties
viii. The solution should be stable enough to be stored for long periods
of time under varying temperatures".

Although the properties listed are uncontentious it is true to say we


have not found a solution that meets all these properties today.

Development of blood transfusion


Blood transfusion was first attempted in 1492 for Pope Innocent VIII.
Both the donors and the recipient died of the experience. This was
over 130 years before William Harvey described the circulation as we
know it today. The first successful blood transfusion was performed by
Jean Baptiste Denis in 1667. A small quantity of sheep blood was
transfused into a boy who had sustained multiple therapeutic bleedings
as a result of a fever. Human to human transfusions were first
performed successfully in the late 1820s by James Blundell for women
with severe peri-partum haemorrhage with a 50% success rate in his
first 10 cases. The technique was applied in haemophilia in 1840.

These early transfusions were not with matched blood groups.


Landsteiner discovered the ABO blood group system in 1901.
Ottenberg performed the first ABO cross matched blood transfusion in
1907.

Anticoagulation of blood to allow storage became possible in 1914 and


blood was stored in blood depots for use during World War I. In 1932
the first blood bank was established in a hospital in Leningrad. Blood
and plasma were widely used for the treatment of blood loss during
World War II.

The first recorded case of transmission of infection was a case of


transfusion related hepatitis in 1943. In the last 30 years concerns
have increased over the safety of blood for transfusion, particularly the
ability to transmit infection. During the same period oxygen carrying
blood substitutes have been developed with limited success.

4
Intravenous Fluid Therapy Chapter 1

Development of fluid infusion philosophies


Late into the 19th century it was a commonly held view amongst
physicians that a ‘useless excess’ of blood was a causative factor in all
disease. Venesection or bleeding with leeches was still a common
treatment for haemorrhage. Thus, intravenous fluid therapy, taken for
granted today, has seen all of its development (as far as treatment
strategy is concerned) during the 20th century.

The foundations laid by the physiological works of the late 19th century
lead to further understanding of the value of colloid solutions. The
problem of saline solution passing to the tissues and causing oedema
was well recognised by Bayliss. Writing in the ‘Proceedings of the
Royal Society of Medicine’ in 1917, he gave a clear view of the
effectiveness of colloid solutions in maintaining the circulation for longer
than crystalloid solutions.
“We see that by raising the viscosity of the injected fluid to that of
blood by the addition of gum or gelatin, the blood lost can be replaced
by an equal volume of the solution, with a return to its original height…
this height is maintained for an hour or so… Pure Ringer’s solution is
very inefficient in maintaining the blood pressure even at the height to
which it at first raises it”.

One year after this report Cannon wrote in ‘JAMA’:


“Injection of a fluid that will increase blood pressure has dangers in
itself. Haemorrhage in a case of shock may not have occurred to a
marked degree because pressure has been too low and the flow too
scant to overcome the obstacle offered by a clot. If the pressure is
raised before the surgeon is ready to check any bleeding that may take
place, blood that is sorely needed may be lost”.

Today we have still not resolved whether crystalloid or colloid infusion is


better in terms, of mortality. Questions have been raised as to what
targets we should use for blood transfusion. It remains clear that
Bayliss’ statement is true. More recently, Lyons has coined the term
Hospital Acquired Generalised Interstitial Edema (HAGIE) to describe
the problems of excessive crystalloid infusion to maintain the
circulation. We had to wait until 1994 for Bickell to demonstrate
scientifically that Cannon was right.

One of the problems of research in the field of fluid management is that


studies have relied on identifying differences in properties of various
fluids available, whether or not these have any relevance to the patient.
They have also relied on treatment endpoints that are not used today
and there is still argument over whether patients should be managed
wet or dry.

In 1943 Blalock wrote in ‘Surgery’:


“It is well known by those that are interested in this subject that the
blood volume and cardiac output are usually diminished in traumatic
shock before the arterial blood pressure declines significantly”.

Yet the ability to monitor the cardiac output and stroke volume in
routine clinical practice did not arrive until the 1970s. In the 19th
century, while developing a method for measuring cardiac output, Karl
Ludwig stated:
“the supply of adequate amounts of blood to the organs of the body is
the main purpose of the circulation... the pressures that are necessary
5 to achieve it are of secondary importance… the measurement of flow is
difficult while that of pressure is easy so that our knowledge of flow is
usually derivatory”.
Still today the practice of prescribing fluid according to formula or fixed
circulatory target (usually pressure based) is common. The practice of
titrating fluid to dynamic effect is comparatively recent. Now we know
how to use fluid we must revisit what we should use.

Bibliography

Cosnett JE. The origins of intravenous fluid therapy. Lancet 1989;768-


771.
von Ott aus St Petersburg. Über den EinFluß der kochsalzinfusion auf
den verbluteten Organismus im Vergleich mit anderen zur Transfusion
verwendeten Flüssigkeiten. Virch Arch 1883;93:114-168.
Starling EH. On the absorption of fluids from the connective tissue
spaces. J Physiol 1896;19:312-326.
Grönwall A. Dextran and its use in colloid infusion solutions. Almqvist
and Wiksell. Uppsala. 1957.
Bayliss WM. Methods of raising a low arterial pressure. Proc Roy Soc
Med 1917;89:380-393.
Cannon W, Fraser J, Cowell E. The Preventative Treatment of Wound
Shock. JAMA 1918:618-621.
Blalock A. A consideration of the present status of the shock problem.
Surgery 1943;14:487-508.

6
Intravenous Fluid Therapy Chapter 2

The Choice of intravenous fluid;


crystalloid or colloid
Martin Kuper and Neil Soni

Introduction
The problems associated with being a mammal inhabiting a non-liquid
milieu are as many and various as the benefits that accrue from that
environment. Our evolution from single cell to colonies of cells in an
aqueous milieu has provided a legacy in that our physiological systems
are largely fluid based. In particular, the means by which substrates are
taken up and moved to where they are needed, as well as the means
by which effluent is removed, is through fluid transport. Even the well-
being of each cell is dependent on the homeostasis of its fluid
components.

Of necessity there are sophisticated protective mechanisms to defend


this fluid milieu against both global and local fluctuations arising from
inhabiting a non-aqueous environment. At a global level there are thirst
mechanisms mediated both through the volume and osmolality of the
fluid compartments while, at a local level, simple mechanisms such as
vasoconstriction of vessels compensate for acute fluid losses as occurs
in trauma. Redistribution of fluids helps maintain fluid compartments in
a functional state until they can be replenished fully.

These mechanisms are efficient and effective in the animal kingdom but
when they fail the animal dies. Interventions in medicine go well
beyond what nature can cope with and these mechanisms are
inadequate when confronted with overwhelming injury, disease or major
surgery. In these new circumstances exogenous means of supporting
fluid homeostasis are necessary but the physiological defence
mechanisms still function and need to be considered when providing
external support.

Over the last century the physiological understanding of where, how


and why fluid is managed by the body in both health and disease has
increased significantly. Means have been developed to measure where
the fluid is and in what quantities. The concept of fluid management
has evolved based on both measurement methods to assess fluid
status and on trial and error. The idea that fluid distribution can be
manipulated has also developed.

It is important to emphasise the confidence with which fluid is managed


belies the relative paucity of information about which fluids to use, the
effective fluid status of our patients and the more subtle implications of
how our management impinges on the pathophysiology we are
treating. It is pertinent to define what fluids are used for and to give a
brief resume of the physiological concepts underlying fluid
management.

Physiology of fluid compartments


The fluid compartments of the body are most commonly described by
a three compartment model. This comprises the intracellular, interstitial
and intravascular compartments. A variation on that theme is a two
7 compartment model splitting the body into intra and extracellular
compartments. In disease states where the efficacy of the barrier
between the intravascular and interstitial space may be impaired it may
be a more relevant model especially when considering fluid
administration. The three compartments are described in Fig 2.1.

Fig 2.1

Intra- Interstitial Intracellular


vascular Figure 2.1
The three fluid compartments of
the body.

3L 12L 25L

The intracellular compartment has a metabolically active membrane


maintaining the electrolyte equilibrium between the intracellular space
and the extracellular space with active transport involved. This
compartment is very large, important in terms of cell function and
inaccessible to clinical measurement systems.

The interstitial space is a large space separated from the intracellular


space and with a different electrolyte balance, sodium being the
predominant cation. It has some proteins and other macromolecules.
It is separated from the intravascular space by membranes and pores.
The electrolyte composition is similar to the intravascular space but has
less protein and therefore a lower oncotic pressure. It is in dynamic
equilibrium with the intravascular compartment across a barrier of
endothelial cells that is loosely referred to as a 'membrane'. It is a
difficult space to measure and is usually assessed clinically by virtue of
the observation that this is where oedema is seen. It has an intrinsic
circulation being supplied with fluid that egresses the intravascular
compartment, bathes the cells and is drained by the lymphatics. It
transports substrates and metabolites between the circulation and the
cell and is a medium in which immunological mediators function. In
times of stress or injury the barrier between the intravascular and the
interstitium becomes leaky causing local oedema as seen following
direct injury. This is in part the mechanism by which white cells and
immunological mediators get access to the site of injury.

The relatively small intravascular compartment is similar in electrolyte


composition to the interstitial fluid but contains a higher concentration
of protein and all the blood components. This is the main transport
system of mammals carrying everything from oxygen to lactate around
the body. Its purpose is the supply of substrates, removal of
metabolites and delivery of immunological mediators. Physically, it is a
fluid based hydraulic system requiring a working pressure to generate
flow so it needs to be replete with fluid to function. That fluid is blood
8
Intravenous Fluid Therapy Chapter 2

which has two main components: red cells and protein-containing,


electrolyte-rich plasma. Oxygen carriage, one of the primary functions
of blood, is largely determined by the haemoglobin in red cells and this
function can, at present, only be sustained by that haemoglobin
although other oxygen carrying fluids are in development. There is a
huge redundancy in the ability to transport oxygen provided by the
potential of blood to bind oxygen and haemodynamic reserves that can
increase the rate of transport. Consequently red cell replacement only
becomes critical when the haemoglobin is very low, though this is
determined in large part by the fitness of the cardiovascular system. If
the red cell oxygen carriage capability is threatened, the natural choice
of fluid is blood. The issue of blood transfusion will not be considered
further.

Plasma is more complex. Its key role is to provide a fluid medium to


carry red blood cells and fill the intravascular space enabling the
hydraulic circulatory system to function in its delivery role. This is of
overwhelming importance as the primary function. Maintenance of
intravascular volume is vital to survival as loss of fluid from the
intravascular compartment threatens the pressure and the flow that
provides oxygen delivery and hence survival. A key issue is therefore
the physical ability of fluid to remain in the intravascular space.

The constituents of plasma have important functions, some of which


are readily identified, e.g. clotting. Replacement of large amounts of
the plasma with inert substitutes reduces these factors and can have a
clinical impact. The many other components presumably have
physiological roles. Teleologically, its components must be there for a
reason and should be useful. The detriment that arises from failing to
replace them is not yet apparent. Clearly, in fluid replacement a key
question should be "What is in plasma?"

Other components include the cellular elements such as white cells,


platelets, clotting factors and proteins. The latter encompass a vast
array of proteins with a multiplicity of roles. Proteins and albumin
provide a colloid osmotic pressure and also bind molecules and are
involved in membrane integrity as well as clotting. Immunoglobulins,
acute phase proteins and many others presumably have defined
purposes and will not be replaced in a synthetic fluid. Is this
important? All are produced within the body and the amount in the
intravascular space represents variable proportions of the total
reserves, so they may not be lost immediately. Nevertheless, when
fluid is being replaced in large quantities with inert fluid the implication
of this dilution should be considered. The intravascular space also
supplies fluid that sustains both the interstitial and most importantly the
intracellular space.

Unlike the preceding two spaces the intravascular compartment is


amenable to measurement. There are a vast array of measurement
systems. Many have been developed because it was relatively easy to
do so rather than the cause they measure appropriate things. They are
frequently surrogates of the measurements that we actually require but
which are unavailable. The question is "what to measure and what
relevance that measurement has?'" Ideally, measurements should
provide information that can be utilised to understand what is
happening but, more importantly, to guide interventions and permit
manipulation of the fluid spaces. To take this further needs elaboration
of the way fluids are perceived to move.

9
Fluid movement
The basic rule of thumb for fluid movement after its administration is:
• water moves across all spaces
• sodium will equilibrate across the intravascular and interstitial space
• large molecules will stay, for the major part, in the intravascular
compartment

This means that a largely water comprised fluid such as 5% glucose


will, after the glucose has been metabolised, equilibrate rapidly across
all compartments approximately in proportion to their size. This is
good for replacing water in patients who are dehydrated but will have a
very transient effect on intravascular volume.

Salt containing solutions will distribute the sodium to equilibrium across


the interstitial and intravascular compartments. As the interstitial
compartment is 3-4 times the size of the intravascular compartment
most of the administered isotonic salt solution (3/4) will go to the
interstitial space after a relatively short period of time.

Protein containing fluids and cells will stay longer intravascularly.


Movement of the molecules and hence the accompanying fluid will be
inhibited by the endothelial cells. The size of the junctions between
endothelial cells is in the order of 6.5nm and, in theory, molecules the
size of proteins do not cross these gaps while sodium does. Therefore
these fluids should sustain the intravascular compartment since most
of the fluid given should stay in the appropriate space.

This simple rule of thumb needs to be moderated by consideration of


factors that influence distribution. These include:
• hydrostatic pressure gradients
• oncotic pressure gradients, although pressure gradients always try
to equilibrate and electrostatic forces influence the distribution of
larger charged molecules
• the shape of the molecules
• the size of molecules
• the condition of the endothelial barrier
• the time for redistribution (half-life) in the intravascular space

The equation that governs fluid movement out of the capillary into the
tissues or from intravascular compartment to the interstitial space is
seen in Fig 2.2.

10
Intravenous Fluid Therapy Chapter 2

Fig 2.2

(Capillary c)
Figure 2.2
Schematic diagram of the
Starling equation governing
fluid movement between Hydrostatic Oncotic
intravascular and interstitial Pressure Pressure
Pc
compartments. c

Capillary
Endothelium

Pi i Lymphatic
Return
Interstitium i

Starling equation: Qf = K1.[(Pc - Pi) - !("c - "i)]


Net fluid flux, Qf
K1 fluid filtration coefficient
Pc capillary hydrostatic pressure
Pi Interstitial hydrostatic pressure
! osmotic reflection pressure
c capillary oncotic pressure
i interstitial oncotic pressure

Given these considerations it should be easy to choose fluids for


various situations. If a patient is dehydrated, giving fluid that will
replenish all compartments will be of benefit and water is able to reach
all parts. If the primary consideration is the intravascular space then a
fluid that will preferentially stay and sustain that space would be ideal.

The other factor that is important is time. Any fluid infused rapidly into
one space will take a finite time to redistribute to equilibrium. This time
may be more than adequate to fulfil a primary goal. For example, in a
trauma patient who has lost a considerable proportion of the circulating
volume, a bolus of fluid will fill the space. If it solves the immediate
problem and, by the time it redistributes the bleeding is curtailed, no
further fluid is required. Redistribution is not an issue. In particular, in
this scenario the volume is lost rapidly from the intravascular space and
it is the only space affected. If however the blood loss is ongoing then
several factors are involved:

• fluid continues to be lost


• fluid shifts out of the interstitium and even the cells in an attempt to
stabilise the intravascular compartment
• redistribution will reduce the efficacy of administered fluids

All these factors will necessitate more fluid being given or, alternatively,
a fluid that is more likely to stay in the intravascular space will be
deemed more efficient. There is also the problem that redistributed
fluid will be in inappropriate places such as the interstitium and will
need to be removed later. In the interim it may cause problems as a
consequence of oedema. The distribution time of the fluid therefore
becomes relevant.
11
The argument may be summarised as - it is better to use a fluid that is
less likely to redistribute in order to:
• fix the problem faster and keep it fixed
• use less fluid
• cause less oedema

This assumes oedema is pathological in its own right. On the basis of


the discussions above it can be argued that a fluid with larger
molecules is less likely to redistribute or, if it does, will do so more
slowly. This is the basis of the argument for the use of colloids in
resuscitation as colloids will fill the intravascular space faster and for
longer.

What are colloids and crystalloids?


A crystalloid is a solution of water containing electrolytes or substances
that are easily metabolised such as glucose. The molecules are small
and the osmolality is variable. When glucose is metabolised it leaves
water, so it is a hypo-osmolar solution. Isotonic saline is iso-osmotic.
Hypertonic saline is also used and is osmotically active. The osmolar
effect is produced by the concentration of molecules not their size.
Distribution is determined by hydrostatic and osmotic pressures across
semi permeable membranes.

A colloid is a solution containing molecules that are large enough to be


impeded by endothelial linings. They exert a pressure, the colloid
osmotic pressure, by virtue of the number of large molecules on one
side of the barrier that cannot move so cannot equilibrate across it.
This oncotic pressure holds or draws water to it. Distribution is
determined by hydrostatic pressure, to a small degree but largely by
the physical characteristics of the molecules, including the size of the
molecule.

In practice colloids vary in size from 20,000 to 450,000 Dalton. Size,


shape and charge all influence the ability of those molecules to move
out of the intravascular space and clearly the larger the molecule the
less likely it is to move out. Therefore, even under normal
circumstances, there is a spectrum of intravascular effects of colloid
solutions.

Oedema
One of the key arguments in the choice of fluid is the creation of
oedema. Oedema is fluid in the interstitial space and occurs as a
consequence of redistribution. The nature of the oedema is influenced
by the material that is sequestered. Therefore saline will result in a
crystalloid oedema while a colloid may produce an oedema containing
that particular colloid molecule. The amount of oedema is governed by
many factors but included in these will be the propensity of the fluid to
move across semi permeable membranes. Hence, in theory, colloids
produce less oedema for two reasons:
1. Less fluid is given
2. Less fluid moves out into the interstitial space

12
Intravenous Fluid Therapy Chapter 2

The nature of the oedema may influence how it is cleared. There is


very little information on clearance, on how different colloids affect
lymphatic flow or what role the reticuloendothelial system plays. It is
frequently hypothesised that colloid derived oedema may be more
resilient to removal and may have a prolonged effect. This has never
been substantiated and the clinical impact of oedema is poorly
characterised so the suggestion that different oedema types have
different associated morbidity is entirely theoretical.

The whole issue of whether oedema is intrinsically a bad phenomenon


is also contentious. Oedema is a sign that a patient is ill and may be a
sign that various organs are not functioning well but is that the same as
it being a deleterious phenomenon in its own right? There is no doubt
that localised oedema can cause physical problems. Laryngeal
oedema, cerebral oedema and severe pulmonary oedema are all life
threatening. The former two are due to anatomical constraints, the
latter through impeding gas exchange but usually in company with
ventricular dysfunction or lung injury. Is it cause or effect? Is all
oedema bad?

Oedema is part of the response to injury. The circulation transports


substrates, immunological agents, repair material and defences to the
site of injury by virtue of the same mechanisms that produce oedema.
If oedema is the observed result of activation of a fluid medium for
delivery then teleologically it must have merits. Excess oedema may
impede the transport of oxygen in tissues but the normal mechanism
by which oxygen gets transported in tissues is via a fluid medium.
Furthermore the detrimental effects on oxygen transport at tissue level,
are hard to identify. So maybe it is excess oedema that is the problem.
The disappearance of oedema often coincides with correction of the
pathological circumstances that caused it. Is it cause or effect? It is
likely that, as with other things in nature, a small amount is beneficial
but it is hazardous in excess. So how much is enough?

The immediate relevance is that one of the most persistent arguments


in the choice of fluids is the absolute requirement to minimise oedema
based on the assumption that all oedema is bad. This assumption
bears further examination as it is not as secure as it seems.
Regrettably, the very modest literature on oedema is relatively unhelpful
and appears to demonstrate that detrimental effects to oxygen
transport at tissue level are hard to identify, hence confounding the
assumption.

Principles underlying fluid administration


Electrolyte replacement

In discussing the choice of fluids, crystalloid or colloid, we must


encompass electrolyte replacement. It is not intended to discuss
maintenance fluids in depth but rather to draw attention to the fact that
various body fluids may be lost that need replacement. While
protection of the intravascular compartment must be a primary
consideration, losses should be considered and appropriately replaced.
The composition of some commonly lost fluids is shown in Table 2.1.

In general the administration of electrolytes is either to replace ordinary


daily losses or, in some circumstances, to replace abnormally excessive
losses. The key issue is to replenish the spaces appropriately. The
sodium losses can be evaluated for clinical assessment and from
13 measurement of the serum sodium. The value will be determined by
the overall state of hydration, literally the water deficit, and this must be
considered across all three spaces. Sodium concentration reflects the
concentration across the intravascular and interstitial space.

Potassium deficit is more complex as the plasma value does not reflect
the total body potassium very effectively and is influenced by, and
influences the acid base status.

It is obvious that in replacing electrolytes it is important to use balanced


solutions that take into account electrolyte losses. It is an anomaly that
the colloid solutions frequently used in large volume are usually not
electrolyte balanced solutions. However, colloid solutions in a balanced
electrolyte solvent are beginning to come to market.

Table 2.1. Electrolyte content of some body fluids

Fluid Vol L/24hrs Sodium mmol/l Potassium mmol/l Chloride mmol/l Bicarbonate mmol/l
saliva 0.75 - 1.5 30 20 30 15
gastric 2 50 8 145 -
pancreas 1 140 4.5 30 100
bile 1 140 5 150 30
small intestine 4 100 5 100 25
diarrhoea - 70 25 60 25
sweat 1-3 50 10 50 -

Benefits and problems of crystalloid and colloid

We need to consider the specific benefits accrued from the use of a


specific fluid. Given that in resuscitation, filling the space is of
paramount importance and that other benefits are secondary it may be
hard to identify specific benefits from the components of a fluid. While
subtle but common benefits may be difficult to identify let alone
evaluate some complications may be rare but easy to identify so that a
balanced assessment may be difficult.

There are at least two areas to consider when focusing on problems:


those problems relating directly to the fluid and its contents and those
relating to how it is handled (Table 2.2).

Table 2.2. Some problems associated with fluid administration


Problem Nature
Volume of fluid given Electrolyte imbalance
Acid base
Dilutional coagulopathy
Distribution of fluid Failure to maintain intravascular volume
Sequestration and consequences
Elimination of sequestered fluid
Anaphylaxis Incidence
Rheological / Coagulation Coagulation effects
direct and dilution
Immunology Immunosuppresion
Activation
Diuresis Oncotic diuresis
None specific Itching
Cost Monetary definition easy, clinical impossible
Long term use Outcome

14
Intravenous Fluid Therapy Chapter 2

Volume and distribution

Any fluid is given for a reason and if it fails in that objective then this is
a potential problem. If a sustained intravascular effect is required a
rapidly redistributing fluid such as glucose is inappropriate and that is,
in itself, a problem. Saline distributes to a lesser degree and has a
longer period of efficacy as an intravascular compartment expander.
As the molecular size increases there is an unpredictable relationship
between molecular size and half-life in the intravascular compartment
mediated, in part, by the range of molecular sizes seen in many
'standard' colloid solutions.

Another aspect of this phenomenon is what happens to the fluid that


redistributes. With the crystalloids an accumulation of crystalloid in the
interstitium may result in a large expansion of that space, seen as
oedema and with implications in terms of pathological effect. Colloids
may result in less movement into the space and a smaller expansion of
that space, with less oedema. The merits of preventing oedema are
much discussed but there is little hard data. There is a secondary
problem intrinsic to the effect of the fluid in the interstitium:
• Is there a difference between a crystalloid and a colloid in the
interstitium?
• Will a crystalloid clear more quickly than a colloid?

The answers to these questions are not yet clear. The issue of
clearance of these fluids may be significant. Ideally a colloid might be
inert, administered for a purely physical effect in terms of presence and
is excreted either completely metabolised or unchanged. Crystalloids
are passed as urine eventually as are smaller colloids such as the
gelatins. The issue of a colloid osmotic diuresis has been raised in
burns resuscitation. Although there is some evidence that it occurs
and a colloid osmotic pressure can be measured in the urine, it is a
minor effect in clinical terms. Some have even advocated the use of
gelatins as being 'good for renal function' which may be in part be this
minor diuretic effect especially seen in those fluid replete. Starches are
less easily cleared. Some colloid molecules find their way into the
reticuloendothelial system where they may remain long term. There is a
postulated association between starch molecules in the
reticuloendothelial system and itching. This constitutes a practical
clinical problem in the recovery period. There is little information on the
impact of loading the reticuloendothelial system with starch and
consequent effects on the immune system.

Another aspect of the volume of fluid given is the effect on acid base
balance. Attention has been drawn to the association of large volumes
of saline with the development of acidosis. This phenomenon could,
theoretically, be seen in large volume colloid administration where there
may also be a large chloride load from an unbalanced electrolyte carrier
fluid.

The most immediate complication of giving a fluid in practice may be


anaphylaxis or reactions to the fluid or its contents. Most of the colloids
have an incidence of reactions variously quoted in the literature. Not all
are anaphylactoid. The incidence of these complications may be low
but the benefits of colloid use must be demonstrable to outweigh such
a side effect, albeit rare. Crystalloids do not constitute a problem from
this perspective.

15
Immunology

There is a considerable amount of data on the effect of colloids on


immune function in vitro but little in the way of convincing conclusions.
This may be because, in clinical practice, the subtle immunological
effects are overwhelmed by the pathology itself or by other aspects of
treatment.

Coagulation

Again there is a wealth of literature on the potential for many agents to


influence coagulation. This is either through physical dilution or through
direct effects on the coagulation pathways. In clinical terms the impact
may be negligible but nevertheless a consideration.

Long-term effects

Since the effect of colloids is difficult to delineate in terms of specific


clinical problems then one approach has been to try to assess their
overall effect in the longer term. No one doubts that assessing the
long term effects of fluids is extremely difficult. There is a wealth of
literature on the short term effects of fluids and direct comparison in
terms of efficacy on cardiovascular indices. This is relatively easy
focused research in the mould of 'cause and effect' using easily
recognised measurement systems. Beyond the immediate
cardiovascular and respiratory effects the position becomes more
difficult. It sounds easy to assess the intravascular half life of a fluid
but, in practice, it is difficult, even in volunteers. In clinical practice it
becomes very difficult and there are few studies and extremely limited
information on the intravascular half life of these agents in disease.
Extrapolation from volunteer studies is almost certainly misleading and
does not fit either with clinical observation or with the limited studies
available for scrutiny. However, the difficulties seen in these
physiological studies are minor when compared to trying to evaluate
the impact of a fluid on outcome in the longer term.

This is complex. When a patient is resuscitated the use of fluid may be


life saving and the relative benefit of one fluid over another is
overwhelmed by the obvious benefit of any fluid rather than none.
Following resuscitation the influence of the fluid used is difficult to
assess as it is only one of several supportive measures taking place.
Easily recognised noxious effects or blatant direct complications will
identify the fluid as causing greater morbidity than other fluids. Benefits
are likely to be less obvious. If no single event can be described then
the effect of the fluid has to be viewed in the non-specific context of
outcome in company with all the other variables likely to be involved.
This will include the initial pathology and its natural history as well as
any nosocomial problems. The populations studied have tended to be
critically ill which are homogenous in name alone. The actual
heterogeneity of patients in ICU is staggering and matched only by the
heterogeneity of the disease and the treatments. None of the studies
looking at outcome have been convincing except in suggesting that
there is little to choose between fluids. It is extremely hard to
determine a relative detriment between fluids used in supportive
treatment in a heterogenous population. Attempts to do so with very
limited data in an ill defined population have depended upon
sophisticated statistics, the results of which have been published. They
suggested that there were differences between colloids and crystalloids
and also that albumin is dangerous. The conclusions were imaginative
extrapolations from inconclusive statistics.
16
Intravenous Fluid Therapy Chapter 2

The real achievement of these analyses has been to highlight the topic and
identify how little information is available about a daily aspect of critical care
management.

How do these principles impact on choice?

Things to consider when choosing which fluid to use:


• Identify what the fluid is to be used for.
- Rapid expansion of the intravascular space - resuscitation.
- Slower maintenance of the intravascular space -
intraoperative replacement or top up of the space in the
critically ill.
- Replacement of water or electrolytes - either normal fluid
maintenance or correction of electrolyte deficit.
- Generating a diuresis - this may be where there is an
abnormality in electrolyte composition, e.g. hypercalcaemia,
or where there is toxicity from an agent interchangeable with
an electrolyte, e.g. lithium.
- Correcting acid base - the use of agents such as
bicarbonate.
- Replacing haemoglobin.
- Oncotic gradient - is it being used in an attempt to
redistribute fluid.
• Maintaining oncotic pressure.
• Consider the time scale of the intervention:
- Short lived intravascular expansion while the problem
is corrected. The half-life in the intravascular space is
irrelevant in this context.
- Longer term on going resuscitation - for example a patient
with severe sepsis in the ICU where intravascular repletion
may continue for a long time. Logically an agent with a
longer intravascular half-life would be more efficient.
• Consider the volume:
- Will it be a relatively small volume?
- Might it become a large volume over several days? In
these cases the cumulative effects of the fluid need to be
considered. Obvious examples are hypertonic saline, which
may be useful short term, but may impose metabolic
problems used excessively or over a long period.
• Are there immediate side effects?
- Anaphylaxis.
- Other reactions - for example bradykinins.
- Reactions to blood.
• Are there other detrimental effects?
- Rheology.
- Coagulopathy.
- Immunological problems.
17
• Does the fluid cause any degree of immunosuppression?
• Conversely is it helpful and does it somehow suppress
immunological triggers as has been suggested for starches?
• Is the fluid associated with effects on other organs such as the
kidney or the lungs?
• If the agent will redistribute does this have any negative
connotations?
- It was suggested that sequestered albumin may be
difficult to eliminate from the lungs and may contribute
to poor lung function. This has not been
corroborated.
• Oncotic effects.
- Can the fluid cause oncotic diuresis and if so is that a
problem?

• Is there a cost implication?


- Colloids are more expensive than crystalloids.

• Is there an infective risk?


- Human derived proteins still constitute a very small but
finite risk.

• Are there any specific benefits from the agent that would be helpful?
- Albumin has many physiological associations but none
have been shown to have clinical relevance.

The final consideration is never mentioned but always relevant. There


is huge intrinsic safety in doing what is routine and using the fluid that
one is used to using. The phenomenon of acute changes in practice
has not been examined but wisdom decrees it would cause morbidity.
In the absence of overwhelming benefit stay with what you are used to.

There is no one answer as to which fluid to use. It should be


appropriate to the requirements of the patient and geared to the
problems with which they present. The way in which this might be
addressed is illustrated in a few examples below.

How do these considerations


work in practice?
In trauma with major blood loss the primary problem is replacement of
the intravascular volume and, in a dire emergency, any fluid (crystalloid
or colloid) would do. The purpose is to maintain perfusion and oxygen
delivery to vital organs. Paradoxically, the fluid administration may
increase bleeding by increasing pressure and flow, so, if it is a very
short time to definitive treatment then fluid replacement should be
restrained. In more general terms adequate perfusion of vital organs is
fundamental to resuscitation so fluid must be given. Most commence
with crystalloid. Any fluid will do and there are many that, with a major
ongoing problem, would use colloids. The rationale is faster and better
sustained resuscitation with less fluid. There are also advocates of
hypertonic saline who suggest that the loss is predominantly from the
intravascular compartment and therefore endogenous fluid can be
mobilised from the interstitium. The rationale is the same as with
colloids in terms of fluid volume. This in theory removes the anxieties 18
Intravenous Fluid Therapy Chapter 2

about sequestered fluid and its late implications in terms of organ


function. It is without doubt best to use what one is used to using.

In contrast we can consider free flaps in plastic surgery. This is lengthy


major surgery with major blood and fluid loss. The free flaps are
perceived to be of increased risk of developing oedema, partly due to
lack of lymphatic drainage and therefore a decreased ability to
reabsorb excessive interstitial fluid. Some then suggest that while
limited crystalloid is useful for insensible losses, other losses should be
met with colloids or even with hypertonic saline. Haemodilution can be
used to improve blood flow.

In neurological injury a series of factors pertain that should be


considered. Unlike the larger junctions in endothelial cells the junctions
in the brain are very small, 0.4-0.7nm, which is too small even to allow
sodium to cross. The implications of the physics of this barrier are
that, under physiological circumstances, a reduction in plasma
osmolality will allow water ingress and hence oedema so that osmolality
is the major driving force in fluid distribution in the normal brain.
Isotonic solutions or colloids are therefore preferable if oedema is to be
avoided.

In anaesthesia there are some interesting situations where fluids are


used to protect the intravascular space when regional blocks are used.
This provides a model for comparing the efficacy of the two types of
fluid in repleting the intravascular space in normal physiology. Most
studies seem to suggest that the use of a smaller volume of colloid
was not only as effective but also more consistent in efficacy than the
crystalloid. None of the studies have looked at detrimental effects.
The issues of sustainable effect and sequestration of the colloid have
questioned the rationale of fluid preloading in these circumstances.

Sepsis is a condition frequently treated in the ICU requiring ongoing


support of the intravascular compartment. The pathophysiology of
sepsis involves changes in the nature of the semi-permeable barrier, in
effect, increasing the apparent ease with which fluids, including those
with large molecules, can redistribute. Oedema seems to be a feature
of this condition irrespective of the fluid used. Advocates of colloid
suggest there is less oedema than with crystalloid. Advocates of
crystalloid draw attention to the theoretical risks of colloid sequestration
in the lungs. There are few effective long term studies trying to
evaluate the effects of these different agents. Recently attention was
drawn to the fact that most colloids are in relatively non-physiological
solutions but the impact of this on the patient in the long term is hard
to evaluate. The theoretical benefits of using a physiological solution
such as albumin appear robust in terms of oncotic pressure, rheology,
transport mechanisms and its metabolism but despite a plethora of
potential benefits few can be demonstrated at all in practice. The
choice then becomes a personal preference and the relative
importance of cost increases against the apparent irrelevance of the
unsubstantiated benefits.

19
Bibliography

Emerson TJ. Unique features of albumin: a brief review”. Crit Care Med
1989;17:690-694.
Lang K, Boldt J, Suttner S, Haisch G. Colloids versus crystalloids and
tissue oxygen tension in patients undergoing major abdominal surgery.
Anesth Analg 2001;93:405-409.
Mythen MG, Salmon, JB, Webb AR. The rational administration of
colloids. Blood Rev 1993;7:223-228.
Cochrane Injuries Group Albumin Reviewers. Human albumin
administration in critically ill patients: systematic review of randomised
controlled trials." BMJ 1998;317:235-240.
Velanovich V. Crystalloid versus colloid fluid resuscitation: a meta-
analysis of mortality." Surgery 1989;105:65-71.
Waikar SS. Chertow GM. Crystalloids versus colloids for resuscitation in
shock. Curr Opin Nephrol Hypertens 2000;9:501-504.

20
Intravenous Fluid Therapy Chapter 3

Pathophysiology of capillary leak


Peter Gosling

Introduction
The Roman physician Celus described the cardinal features of
inflammation as ‘rubor’ (redness), ‘calor’ (heat) ‘dolor’ (pain) and ‘tumor’
(swelling). The swelling is due to a localised increase in capillary
permeability allowing fluid rich in protein and electrolytes to enter the
interstitial compartment. This response probably evolved to allow
circulating antibodies to reach sites of injury and infection. The local
oedema may have provided some ‘splinting’ of the injured part or limb,
thereby speeding recovery. These inflammatory changes also produce
an increase in whole body capillary permeability to plasma proteins,
especially albumin. It is not clear that this systemic response carries
any survival benefit. Systemic capillary leak is a very early and
consistent feature of the acute inflammatory process, and forms part of
the systemic inflammatory response syndrome (SIRS), a ubiquitous
step in the development of multiple organ dysfunction, multiple organ
failure and death.

Fluid resuscitation is intended to preserve tissue oxygenation, for which


it is essential to have adequate capillary function. Failure to modify fluid
therapy according to the dynamic state of the systemic microcirculation
at best leads to sub-optimal care and at worse can precipitate organ
failures and death. Recent studies have demonstrated that the choice
of intravenous fluid can influence systemic capillary leak, the
inflammatory response and organ function. For these reasons it is
important to understand the pathophysiology of systemic capillary leak
and how it can be modified.

Clinical capillary leak syndrome


Systemic capillary leak syndrome or Clarkson’s disease, was first
described as a rare condition associated with an idiopathic increase in
capillary permeability. The characteristic features are body-wide
oedema due to a movement of water and protein from the vascular
space into the interstitium, with associated hypoalbuminaemia,
increased haematocrit, hypovolaemia and accumulation of interstitial
oedema fluid. Although Clarkson’s disease is a rare condition,
increased capillary permeability leading to fluid shifts from the vascular
to interstitial compartments occur in all patients undergoing an acute
systemic inflammatory episode.

Small changes in systemic capillary permeability occur in response to


even mild inflammatory stimuli such as intermittent claudication,
myocardial ischaemia or minor surgical procedures. However, the
increase in vascular permeability only lasts a few hours before returning
to normal and is usually clinically silent. In contrast, clinical capillary
leak syndrome can be defined as a syndrome characterised by a
prolonged and severe increase in capillary permeability, the features of
which include hypoalbuminaemia, clinical oedema and decreased
oxygen transport and consumption.

Large changes in systemic capillary permeability take place very rapidly


in response to acute inflammatory insults such as surgery, trauma,
ischaemia-reperfusion injury, pancreatitis or burns. An increase in
capillary permeability to radio-labelled albumin can be detected within 3
21
hours of surgery and this is the principal cause of post-operative
hypoalbuminaemia (Fig 3.1).

Fig 3.1

60

30
Figure 3.1
Transcapillary escape rate of
Albumin transcapillary escape rate (1% hr -1)

25 radiolabelled albumin (left panel)


50
and plasma albumin level (right

Plasma albumin (g l -1)


20 panel) in patients before and after
cardiac bypass surgery. (Taken
15
40
with permission from Fleck A,
Raines G, Hawker F, Trotter J,
10
Wallace PI, Ledingham I McA,
Calman KS. Increased vascular
5
30 permeability: a major cause of
hypoalbuminaemia in disease and
0
3 5 7
0
3 5 7 injury. Lancet 1985; I:781-784).
pre - op Hours after by-pass pre - op Hours after by-pass

Increased capillary permeability to radio-labelled albumin is also seen in


sepsis and malignancy and is again a major contributor to
hypoalbuminaemia in these conditions. Increased renal permeability to
albumin, as part of the systemic response, is one of the earliest
features of inflammation and occurs before changes in inflammatory
cytokines or rise in serum concentrations of acute phase proteins such
as c-reactive protein (Fig 3.2).
Fig 3.2

Figure 3.2
14

12
Stylised representation of the
typical changes in capillary
10 permeability (----), plasma
interleukin-6 (……) and serum
c-reactive protein ( ___ ) before
Multiples of normal

during and after major surgery.


6

Pre op surgery 2 4 6 8 10 12 18 24 36 48

Hours post surgery

Increased systemic capillary permeability occurs within minutes of the


inflammatory insult, is proportional to the severity of the initiating
condition (Fig 3.3) and is predictive of complications including
pulmonary dysfunction, multiple organ failure and death.

22
Intravenous Fluid Therapy Chapter 3

Fig 3.3

25
Figure 3.3 Hernia repair
Comparative changes in systemic Coronary bypass
grafting
capillary permeability assessed as
Bowel resection with
microalbuminuria, for hernia 20
complication
repair, uncomplicated coronary

Urine albumin creatinine ratio mg/mmol


by-pass grafting and bowel
resection for Crohns disease 15

requiring further surgery for


ischaemic bowel 10 hours after
the start of the first operation. 10
(Normal urinary albumin creatinine
ratio < 2.3 mg/mmol).

0
pre op 30 2h 3h 4h 10h 15h 20h 24h
mins
Time from start of operation

It is worth emphasising that, unlike Clarkson’s disease, clinical capillary


leak syndrome is common and can be found in many patients requiring
intensive care. Clinical capillary leak syndrome forms part of SIRS in
the development of multiple organ failure. Indeed, were the
cardiovascular system to be considered an organ, then capillary leak
syndrome could be considered as an organ failure.

Mechanisms for transcapillary transport

Capillary endothelial ‘pores’

The movement of fluids and solutes between the vascular and


interstitial space is controlled by the vascular endothelium lining the
capillary wall. The capillary endothelium secretes and metabolises
molecules which regulate vascular permeability, vascular tone and
coagulation (e.g. nitric oxide, histamine, von Willebrand factor,
fibronectin, P selectin, interleukins, thrombomodulin).

The vascular endothelium varies from organ to organ. Capillaries with


continuous endothelial barriers are found in the muscle, heart and
brain, fenestrated endothelium in the gastrointestinal tract and renal
glomerulus and discontinuous capillaries in the liver and spleen. The
variable permeability of endothelium to different sized molecules can be
considered in terms of ‘pores’ (Table 3.1). The high proportion of large
pores in liver and lung may relate to the need of these organs to
metabolise macromolecular precursors of lipoproteins and surfactant
respectively. This may also partly explain why the lungs, despite having
a high capacity lymphatic drainage system, become oedematous more
rapidly in response to increases in capillary permeability compared with
other organs.

23
Table 3.1. Relative size in nanometers (nm) and distribution of endothelial pores in different tissues
Tissue Small Large Ratio of large
Pores (nm) Pores (nm) to small pores

Subcutaneous 5.0 20.0 1:3000


Skeletal muscle 6.0 22.0 1:3600
Brain 0.4 - -
Intestine 4.6 20.0 1:6400

Liver 9.5 33.0 1:50


Lung 8.0 20.0 1:200

Molecular dimensions of some plasma constituents


NaCl/Glucose < 1.0nm lgG 5 x 24 nm
Albumin 4 x 14 nm Fibrinogen 2 x 70 nm

(Taken with permission from Zikiria BA, Oz MO and Carlson RW (eds) Reperfusion Injuries and Clinical
Capillary Leak Syndrome. Armonk, NY: Future Publishing Company, 1994)

Other mechanisms for transcapillary transport


The precise mechanism by which water, electrolytes and proteins cross
the vascular endothelium and how this process is controlled is still not
well understood. Although the ‘pore’ description above goes some
way to describe the experimental findings of variations in capillary
permeability of different tissues, other mechanisms have also been
proposed to explain the movement of molecules across the endothelial
barrier, based on electron microscopy and mathematical models
derived from permeability experiments. Recent evidence shows inter-
endothelial cell gaps forming within 1 minute of neurogenic
inflammation produced by substance P injection. These gaps are large
enough to explain passage of proteins such as albumin and
immunoglobulins across the capillary wall (Fig 3.4).

Fig 3.4

Figure 3.4
Scanning electronmicrographs of
rat capillary endothelial cell
junctions.
Panel A: control with
interendothelial cell junction
(arrowed),
Panel B: 1 minute after injection
of substance P into the circulation
showing interendothelial cell gaps.
(arrowed).
A B
B (Taken with permission from
McDonald DM, Thurston G, Baluk
P. Endothelial gaps as sites for
Vesicles which traffic between luminal and abluminal sides of the cell, plasma leakage in inflammation.
or coalesce to form a transendothelial channel have also been Microcirculation 1999; 6:7-22 ).
proposed to explain movement of large molecular weight proteins out
of the circulation.

Membrane water channels have been described through which water


can pass, yet which prevent the passage of ions and small molecules,
thereby creating an osmotic gradient. These have been termed
‘aquaporins’ and a series of 28,000 Dalton membrane proteins have
been identified in capillaries from a variety of tissues.

24
Intravenous Fluid Therapy Chapter 3

It would be wrong to consider transcapillary movement of fluids and


proteins as solely a pathological response. In normal subjects about
165 grams albumin (60% of total body albumin pool) is in the interstitial
space, and in-vivo radio-isotope studies have shown that about 6
grams albumin move from the vascular space to the interstitial space
every hour returning to the circulation via the lymphatics. Every gram
of albumin ‘binds’ about 18 grams water, so normal albumin flux is
theoretically associated with a fluid circulation of about 2.5 litres every
24 hours.

Capillary endothelium and colloid osmotic


pressure

What is plasma colloid osmotic pressure?

Plasma colloid osmotic pressure is exerted across the capillary wall due
to trapping of molecules within the lumen of the capillary, because they
are too large to cross the capillary wall. Colloid osmotic pressure
opposes the pressure within the capillary derived from the pumping
action of the heart to push fluid out. Classical Starling theory indicates
that, in the arteriolar end of the capillary, the pressure forces fluid out of
the circulation into the interstitial space against the colloid osmotic
pressure. As the capillary pressure falls towards the venous end of the
capillary, the colloid osmotic pressure is greater than the capillary
pressure and fluid is forced back into the circulation.

The plasma colloid osmotic pressure is governed by the number and


size of macromolecules within the circulation (e.g. albumin) and the
permeability of the capillary wall to those molecules. Thus, if the
permeability of the capillary wall to large molecules increases allowing
them to pass into the interstitial space, the colloid osmotic pressure
across the capillary wall will fall. Similarly, the colloid osmotic pressure
falls if the permeability of the capillary wall remains unaltered but the
number of macromolecules decreases. In reality, both of these
mechanisms contribute to the rapid fall in colloid osmotic pressure
seen following acute inflammatory stimuli.

Plasma colloid osmotic pressure measurement

It is possible to measure the plasma colloid osmotic pressure in the


laboratory using a colloid osmometer, but this is only an approximation
of the in-vivo colloid osmotic pressure. Table 3.1 illustrates how the
capillary permeability to macromolecules will vary from tissue to tissue
depending on the relative numbers of ‘large’ pores, which allow
macromolecules to pass and ‘small’ pores to which macromolecules
are excluded. For example, capillaries in the lung with 1 large pore for
every 200 small pores are relatively permeable to large molecules. The
colloid osmotic pressure across the capillary wall will be lower than that
across the capillaries of skeletal muscle, which are relatively permeable
to large molecules with 1 large pore for every 3600 small pores. An
artificial membrane is used in the laboratory colloid osmometer with a
typical ‘cut-off’ of 20,000 Dalton. This means that molecules with
molecular weight greater than 20,000 Dalton will be unable to cross the
membrane and will therefore contribute to the measured plasma colloid
osmotic pressure. Thus the plasma colloid osmotic pressure measured
in the laboratory is based on the assumption that, on average, the
systemic capillaries will exclude molecules of greater than 20,000
Dalton.
25
Despite this limitation serial measurements of plasma colloid osmotic
pressure give an indication of how rapidly it changes following acute
inflammatory stimuli such as surgery. Plasma colloid osmotic pressure
is normally between approximately 22 and 28 mmHg and typically falls
within 6 hours of major surgery to around 15 mmHg. The major cause
of this fall is increased systemic capillary permeability and movement of
plasma proteins and fluid into the interstitial space. As the patient
recovers over the subsequent few days, the colloid osmotic pressure
slowly returns to normal, as fluid and proteins are returned to the
circulation. Failure of post operative colloid osmotic pressure to return
to normal is associated with organ failures and death. This is
consistent with the view that a prolonged very low plasma colloid
osmotic pressure is due to a large transcapillary escape of plasma
proteins and severe interstitial oedema, persistence of which reflects
uncontrolled inflammation and a failure to restore fluid homeostasis.
There have been attempts to normalise persistently low colloid osmotic
pressure by infusion of colloids, but there is no evidence that this
improves outcome. Low colloid osmotic pressure seems to be a
reflection of the patient’s clinical state rather than a primary abnormality
which can be treated with resulting recovery.

Mediators of capillary permeability


The mediators of the rapid and proportional increase in capillary
permeability following injury or infection are interwoven with the acute
inflammatory response. In addition to their roles in co-ordinating the
inflammatory response, many different classes of mediator have been
shown to increase capillary permeability (Fig 3.5). To date, no single
mediator pathway has emerged that can be inhibited to achieve down-
regulation of either the systemic inflammatory response or capillary
leak.
Fig 3.5

Complement activation Anaphylotoxins (C3a C5a) Figure 3.5


Pathways linking injury and
Inflammatory insult, e.g. trauma, ischaemia

Increased systemic capillary permeability

Macrophage activation Tumour necrosis factor


activation of inflammatory
Platelet-activating factor
Interleukins 1, 6, 8
pathways with systemic capillary
leak.
reperfusion, infection

Endothelial cell activation Na/K pump failure


or dysfunction Nitric oxide

Platelet activation Thromboxane A2

Neutrophil activation Proteolytic enzymes


Leukotrienes
Free radicals
Interleukins 1, 6, 8

Mast cell activation Prostaglandins/prostacyclin

Cytokines and Leukocytes

Systemic capillary leak syndrome due to diffuse vascular injury has


been observed in patients treated with specific immune cancer therapy
using recombinant inflammatory cytokines such as interleukin-2 or
granulocyte-macrophage colony stimulating factor (GM-CSF). In-vitro
studies of vascular endothelial injury suggest that cytokines alone are 26
Intravenous Fluid Therapy Chapter 3

not vascular damaging but require the presence of polymorphonuclear


neutrophils or macrophages.

Polymorphonuclear neutrophil activation is mediated by specific agents


such as activated complement (C3a, C5a), interleukins 1, 6, and 8,
superoxide radicals, opsonised bacteria, tumour necrosis factor,
lipopolysaccharide (bacterial endotoxin), leukotriene B4 and platelet
activating factor. Mediation of neutrophil activation leads to expression
of adhesion molecules on the surface of both neutrophils and vascular
endothelium. Adhesion of polymorphonuclear neutrophils to vascular
endothelium is accompanied by release of a number of cytotoxic
agents such as oxygen derived free radicals and proteases as well as
additional inflammatory cytokines. There is evidence that activated
neutrophils have a role in capillary leak either directly or indirectly,
causing local release of permeability mediators. Recent clinical studies
have shown that the peri-operative increase in renal glomerular
permeability (microalbuminuria) and renal tubular damage associated
with coronary bypass surgery are significantly reduced in leukodepleted
patients suggesting that leukocyte activation is an important
component of the in-vivo changes in systemic vascular permeability
(Fig 3.6).

Fig 3.6

Figure 3.6 120

CPB
Changes in renal glomerular
permeability to albumin each 100 LEUKO

post-operative day (POD)


following coronary bypass surgery 80
randomised to standard treatment
(CPB n=22) and and
60
leukodepletion (Leuko n=22).
Urinary MA : Cr

Leukodepleted patients showed


significantly less glomerular 40

permeability. (Taken with


permission from Tang AT, Alexiou 20
C, Hsu J, Sheppard SV, Haw MP,
Ohir SK. Leukodepletion reduces
renal injury in coronary 0

revascularisation: a prospective Pre-op POD1 POD2 POD3 POD4 POD5

randomised study. Ann Thorac -20

Surg 2002;74:372-7).

Nitric oxide and prostacyclin

Under normal circumstances the vascular endothelium releases


prostacyclin and nitric oxide which prevent adhesion of both leucocytes
and platelets to the vessel wall. Nitric oxide is produced by the action
of nitric oxide synthase on L-arginine, an enzyme which can be broadly
classified into two main subgroups; constitutive and inducible.

Constitutive nitric oxide synthase produces nitric oxide continuously in


picomolar amounts causing muscle relaxation and, together with
prostacyclin, maintaining the thromboresistant lining of the endothelium
inhibiting platelet aggregation. In contrast, inducible nitric oxide
synthase, produces nitric oxide in 100 times higher concentrations in
27 response to endotoxin or specific cytokines. Over production of nitric
oxide (e.g. in septic shock) can damage tissues by reacting with free
radicals to form peroxynitrite or hydroxyl radicals. When the vascular
endothelium becomes injured, either as a result of direct trauma or due
to the effects of activation of polymorphonuclear neutrophils (in sepsis
for example), the prostacyclin and nitric oxide-mediated
thromboresistant lining of endothelium is lost and polymorphonuclear
neutrophils adhere to the damaged vessel wall. Enhanced platelet and
neutrophil adhesion to vascular endothelium is closely associated with
increased capillary permeability.

Complement and contact cascades

Activation of the contact and complement cascades increases vascular


permeability. Anaphylotoxins C3a and C5a increase vascular
permeability and treatment with C1 esterase-inhibitors reduce capillary
leak both in septic shock and interleukin-2 induced capillary leak
syndrome associated with immune cancer therapy described above. In
animal models C5a mediates neutrophil-dependent inflammatory lung
injury causing pulmonary leakage of radio-labelled albumin and
accumulation of myeloperoxidase; these effects can be attenuated in a
dose-dependent manner with antibodies to C5a.

Histamine and Bradykinin

The characteristic ‘weal and flare’ reaction to histamine is due to its


combined effects of increasing vascular permeability causing oedema
(weal) and vasodilatation producing a localised increase in blood flow
(flare). Histamine H1 receptors located on the surface of endothelial
cells, stimulate a change in the orientation of intracellular actin
microfilaments altering the cytoskeleton and producing a
conformational change in the cell. Histamine, bradykinin and thrombin
increase endothelial cell permeability within about 10 minutes and the
effect lasts for about 2 hours.

Other rapidly acting permeability mediators have been identified


including substance P (see Fig 3.4) and vascular endothelial growth
factor which has been noted to induce endothelial fenestration in post
capillary venules within 10 minutes of in vivo intradermal injection.

Endothelial cell damage

A mechanism by which endothelial cells allow plasma proteins to pass


into the interstitium during inflammatory episodes may also involve loss
of endothelial cell surface glycosaminoglycans and glycoproteins.
These compounds are responsible for maintaining the net negative
change on the endothelial membrane surface which normally repels
negatively charged albumin. Loss of endothelial glycosaminoglycans
and the negative charge (as a result of the action of membrane-
damaging free radicals and proteases released by activated
polymorphonuclear neutrophils) facilitates the passage of plasma
albumin and fluid into the interstitial space.

Another possible mechanism for increased permeability is endothelial


cell swelling, probably due to failure of the energy-dependent
sodium/potassium pump, allowing sodium and water to enter the cell
causing cellular distortion and widening of inter-endothelial gaps.

Some of the mediators known to increase vascular permeability are


summarised in Table 3.2. It can be seen that the mediators and 28
Intravenous Fluid Therapy Chapter 3

mechanisms that cause rapid changes in capillary endothelial


permeability to plasma proteins overlap and involve many components
of the inflammatory response.

Table 3.2 Some soluble permeability mediators released during acute inflammation

Histamine
Bradykinin
Thrombin
Vascular endothelial growth factor
Platelet Activating Factor
Proteases (released from polymorphonuclear neutrophils)
Oxygen derived free radicals (e.g. superoxide, hydroxyl and peroxynitrite radicals)
Tumour necrosis factor - alpha
Interleukins (e.g. IL-2, IL-6, IL-8 acting on leukocytes)
Activated complement (C3a, C5a)
Leukotriene - B4
Thromboxane - A2
Nitric Oxide (in high concentrations due to the activity of inducible nitric oxide synthase)

Patterns of capillary leak following acute


injury or insult
Systemic capillary permeability, measured as the transcapillary escape
rate of radio labelled albumin, has been shown to be closely related to
low level urine albumin excretion (microalbuminuria) in both normal
subjects and diabetic patients. Simultaneous measurement of urinary
albumin and plasma protein leak in the lungs in patients with acute
respiratory distress syndrome (ARDS) has shown that both reflect a
systemic increase in vascular permeability.

Assessment of microalbuminuria following trauma or burn injury


indicates systemic capillary permeability is increased on admission in
proportion to the severity of the injury. In uncomplicated cases capillary
permeability rapidly returns towards normal within 12 to 24 hours.
Capillary permeability also increases within 30 minutes of the start of
elective surgery and, again, the size and duration of the leak is
proportional to the magnitude of the procedure (Fig 3.3). Post-trauma
corrective surgery is associated with a further increase in capillary
permeability, presumably due to a restimulation of the inflammatory
response (Fig 3.7).

Fig 3.7

Case study : 40 year old female with 41% burns


Figure 3.7 250

Changes in capillary permeability


(microalbuminuria) in a patient
Albumin creatinine ratio mg/mmol (normal < 2.3)

200
14% excision
suffering 41% burns and who 10% donor site
underwent extensive grafting 72
hours later. Note how infusion of 150

hydroxyethyl starch (HES)


temporarily reduces capillary 100
HES Infusion
leak.
50

29
1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97

Hours Post Injury


The pattern of capillary permeability following injury or surgery in those
patients who subsequently develop complications such as pulmonary
dysfunction or ARDS is markedly different. Patients who develop
pulmonary complications 24 hours after aortic surgery show a
significantly greater capillary leak within 4 hours of the start of operation
(Fig 3.8).

Fig 3.8

Aortic Surgery Blunt Trauma


Ann Vasc Surg 1994;8:1-5 J Trauma 1997;47:1056-61
Figure 3.8
A = 10 patients with lung dys. at 24hrs A = 13 patients with later lung dys. or ARDS Differing patterns of capillary leak
B = 30 patients with no lung complications B = 19 patients with no lung complications (microalbuminuria) following elective
surgery and blunt trauma. Left panel:
180
600 Forty patients undergoing elective
microalbuminuria mg/mmol creatinine

160
abdominal aortic aneurysm repair, of
500
140 whom 13 developed pulmonary
A A
microalbuminuria ug/min

120 400
dysfunction 24 hours later. Right panel
100 32 patients suffering blunt trauma (injury
300 severity score > 18) of whom 13 went on
80
to develop pulmonary dysfunction or
60 200 acute respiratory distress syndrome.
40 (Median urine albumin excretion rate
100 B
20 B (95% confidence intervals) (Taken with
0 0
permission from: Smith CT, Gosling P,
0 2 4 6 8 10 12 4 8 12 18 24 Sanghera K, Green MA, Paterson IS,
hours post start of surgery hours post admission Shearman CP. 1994: Microproteinuria
predicts the severity of systemic effects
of reperfusion injury following infra renal
aortic aneurysm surgery. Annals of
Vascular Surgery :8:1-5 and Pallister I,
Similarly, blunt trauma victims who develop later pulmonary dysfunction
Gosling P, Alpar K, Bradley S. Prediction
or ARDS have a severe and prolonged capillary leak 8 hours after of post traumatic adult respiratory
admission (Fig 3.8). Several studies have shown that increased distress syndrome by albumin excretion
capillary permeability, assessed by microalbuminuria, is associated with rate eight hours after admission. Journal
sepsis, the development of multiple organ failure and non-survival of of Trauma 1997:42:1056-61).
both medical and surgical ICU patients (Fig 3.9).

Fig 3.9

Figure 3.9
Capillary leak (microalbuminuria) in 140
1000 surgical and medical patients measured
within 15 minutes of admission to the
ICU. The median capillary leak in 25
Urine albumin creatinine ratio mg/mmol

non-survivors was significantly higher


(P<0.0002 Mann Whitney ) (Medians,
100 95% confidence intervals for medians,
50th centiles and 95% limits) ( Taken with
permission from: Gosling P, Brudney S,
McGrath L, Riseboro S, Manji M.
Mortality prediction at admission to
intensive care: a comparison of
10 microalbuminuria with acute physiology
scores after 24 hours. Critical Care
Medicine. 2003; 31:98-103.

0.1

Survivors Non-survivors
30
Intravenous Fluid Therapy Chapter 3

These studies illustrate three essential features:


1. There is an increase in systemic capillary permeability within minutes
of injury which is proportional to injury severity.
2. In uncomplicated cases capillary permeability rapidly returns to
normal within 12 to 24 hours.
3. A severe and prolonged capillary leak is associated with pulmonary
dysfunction and later organ failures.

Capillary leak and organ function


The association of systemic inflammatory response syndrome and
multiple organ failure with severe positive fluid balance is well
recognised. What is less widely appreciated is the impact systemic
capillary leak and resultant interstitial oedema has on organ function.

Hypovolaemia caused by blood loss due to surgery, trauma, or


vasodilatation, as in bacterial sepsis or pancreatitis, requires infusion of
red blood cells and fluid to preserve tissue oxygenation. However,
these conditions are also associated with a severe and prolonged
increase in systemic vascular permeability that allows plasma proteins
and fluid to leave the vascular compartment. Thus the combination of
an increase in systemic capillary permeability and aggressive fluid
administration places the patient at risk of severe interstitial oedema
and a positive fluid balance of many litres. Non-cardiogenic interstitial
oedema in the lungs is due to extravasation of plasma protein and
water due to capillary leakage and is a feature of ARDS.

The kidneys are vulnerable to permeability oedema producing both


intra-renal and extra-renal pressures compromising filtration. Interstitial
oedema within the kidney produces high intra-renal pressures because
of the tight non-expandable renal capsule which may be accentuated
by visceral oedema causing external renal compression. A high intra-
renal pressure opposes the filtration pressure derived from the pumping
action of the heart and reduces filtration. Thus, in addition to fluid and
electrolyte retention mediated by stress related release of vasopressin,
the renin-angiotensin-aldosterone axis and inhibition of atrial natriuretic
peptide, interstitial oedema of the kidney also opposes excretion of fluid
and sodium used for volume expansion.

Oedema of the intestinal mucosa may impede microvascular perfusion


and reduce the normal osmotic gradient between the lumen of the gut
and the mucosa. Both these mechanisms will compromise intestinal
absorption.

Implications of capillary leak for fluid


management
Organ dysfunction is associated with an excessive and prolonged
increase in capillary permeability and interstitial oedema. In comparison
with crystalloids, gelatins or albumin, infusion of medium molecular
weight hydroxyethyl starches during episodes of increased capillary
permeability have been shown to improve cardiac index, reduce
positive fluid balance, improve pulmonary gas exchange and reduce
inflammation. However, as yet there is no evidence that use of such
fluids reduces mortality.

31
Bibliography

Fleck A, Raines G, Hawker F, et al. Increased vascular permeability: a


major cause of hypoalbuminaemia in disease and injury. Lancet
1985;I:781-784.
Zikria BA, Oz MO, Carlson RW (Eds). Reperfusion Injuries and Clinical
Capillary Leak Syndrome. Armonk, NY: Futura Publishing Company,
1994.
Gosling P. Salt of the earth or a drop in the ocean? A
pathophysiological approach to fluid resuscitation Emerg Med J
2003:20:306-315.

32
Intravenous Fluid Therapy Chapter 4

Properties and use of crystalloids


Hengo Haljamäe

Introduction
Crystalloids are commonly used in clinical practice for intravenous fluid
therapy. The main indications are maintenance of adequate hydration
of patients not capable of managing their basal fluid intake and
correction of fluid derangement in connection with trauma, surgical
procedures, and in the treatment of critically ill patients in the intensive
care unit (ICU). The physiological responses to accidental trauma and
blood loss, as well as those occurring during major elective surgery or
in critical illness, include internal changes of the fluid homeostasis
between the different fluid spaces of the body. Abnormal intra- as well
as extracellular fluid volumes may also exist in patients with pre-existing
cardiovascular, renal and hepatic diseases. Therefore, in the fluid
resuscitation of critically ill patients and patients undergoing major
surgery, intravenous administration of a crystalloid solution at a rate
sufficient to maintain an hourly urine output between 0.5 and 1.0
ml/kg/h is generally recommended. It is considered that such an
infusion rate will decrease the incidence of haemodynamic instability
and acute renal failure. The maintenance of a relatively normal water
and sodium excretion by the kidney in most clinical situations is
consequently an important strategy for prevention of adverse events,
the risk of organ failure and unfavourable morbidity and mortality.

Composition and Properties


of Crystalloids
There are some, although often minor, differences in the composition of
crystalloid intravenous fluids used in different countries. The main
reason for such differences is the fact that within most countries there
is an amount of generally accepted national consensus on acceptable
composition of crystalloids to be used in different clinical situations.
Such consensus will consequently influence the composition and
commercial production of intravenous crystalloid fluids within the
country. Table 4.1 lists the main components of commonly available
crystalloid solutions.

The fluids listed in Table 4.1 represent crystalloid solutions for two main
clinical indications:
• fluids for maintenance (Ringer’s glucose, buffered glucose,
glucose/saline combination)
• fluids for replacement/plasma volume expansion (isotonic saline,
Ringer’s acetate/lactate)

There are in addition crystalloids to be used in specific clinical


situations:
• fluids for special purposes (sodium bicarbonate, hypertonic saline)

33
Table 4.1. Electrolyte composition, acetate, lactate, glucose content, osmolality and pH of crystalloids.

Fluid Na+ K+ Ca2+ Mg2+ CI- Acetate Lactate Glucose Osmolality pH


mmol/l mmol/l mmol/l mmol/l mmol/l mmol/l mmol/l g/l mosm/kgH20

Ringer’s glucose 73.5 2 1.15 - 77.8 - - 25 290


Buffered glucose 70 - - - 45 25 - 25-50 270-440
Isotonic saline 150-154 - - - 150-154 - - - 300-308 5
Ringer’s acetate 130 4 2 1 110 30 - - 270-290
Ringer’s lactate 130-131 4-5 2-3 - 109-111 - 28-29 - 273-280 6-7
5% glucose - - - - - - - 50 252 4
Saline/glucose 31 - - - 31 - - 40 262 4

Fluids for maintenance

Fluids for maintenance of fluid balance are designed to replace the daily
insensible and urinary fluid losses. Urine and insensible loss contain a
lower concentration of electrolytes than plasma. Therefore, fluids for
maintenance need to provide a sodium concentration just over half that
of plasma. The glucose content of the fluid provides a physiological
osmolality level of the solution despite the low electrolyte content. It is
metabolised leaving water. Thus, maintenance fluid may consist of
combinations of saline and glucose to provide 70-80 mmol sodium and
1.5-2.0 litres/day. It may be questioned to what extent a content of
minor amounts of K+, Ca2+ and Mg2+ in some of the fluid preparations
is of any physiological importance.

By addition of acetate or lactate the chloride load can be reduced thus


reducing the possibility of hyperchloraemic acidosis. Acetate and
lactate are metabolised to bicarbonate providing buffering capacity.
The choice of glucose concentration in maintenance fluids may vary
depending on the nutritional status and the estimated energy
requirements of the patient to be treated.

Fluids for replacement/plasma


volume expansion

Isotonic saline

Isotonic saline, containing sodium chloride at a physiological


concentration (0.9%), is still commonly used for replacement of fluid
deficits, as a plasma volume expander for fluid resuscitation in the
perioperative period and in connection with critical illness. Isotonic
0.9% saline may be considered physiological as a result of its
osmolality. It has a sodium concentration higher but not extremely
different from that of the extracellular fluid. However, the chloride
content of 150-154 mmol/l is considerably higher than that of the
extracellular fluid and is high enough to promote acidosis if a large
volume of isotonic saline is infused. It has been suggested that
hyperchloraemic metabolic acidosis may not only influence acid-base
balance but also induce haemostatic defects and impair urine output.
Furthermore, in clinical practice, it is important also to realise that many
of the available colloids are in saline solutions. Infusion of isotonic
saline as well as saline based colloids will therefore result in a high total
chloride load and increase the risk of clinically relevant hyperchloraemic
metabolic acidosis. Hyperchloraemic acidosis induced by isotonic
saline infusion may worsen any prevailing acidosis due to tissue
hypoperfusion prior to the fluid administration. 34
Intravenous Fluid Therapy Chapter 4

Ringer’s solutions

Ringer’s laboratory studies in the late 1800s defined the concentrations


of various inorganic salts necessary for the health of cells bathed in
crystalloid. These studies were the basis for the original formulation of
Ringer’s solution. Lactate was added in the 1930s by Hartmann in
order to counteract acidosis. The chloride load could be diminished
and when lactate was metabolised in the liver sodium was made
available to combine with other anions. The use of Ringer’s solutions,
having a lower chloride concentration (usually about 110 mmol//l), will
consequently prevent the occurrence of metabolic acidosis resultant
from hyperchloraemia and concomitant decrease in strong ion
difference.

The buffering capacity of Ringer’s solutions is partly due to their content


of either lactate or acetate since the metabolic breakdown of these
substances produces bicarbonate. Although lactate containing
Ringer’s solution is most commonly used globally, the choice of acetate
instead of lactate may be advantageous. The metabolism of lactate is
mainly dependent on the metabolic capacity of the two main lactate-
clearing organs, the liver and the kidney. Acetate, on the other hand,
can be metabolised by most tissue cells of the body. Therefore,
Ringer’s acetate rather than Ringer’s lactate will maintain a considerable
buffering capacity when used for fluid resuscitation of severe shock
when the capacity of the kidney and liver to metabolise lactate may be
severely reduced. In such situations the infusion of a lactate containing
solution may aggravate the already present high lactate load.

Fluids for special purposes

Buffers

Sodium bicarbonate is available for correction of metabolic acidosis.


Bicarbonate has mainly an extracellular buffering effect that is
dependent on formation and elimination of CO2. Indeed the CO2
formed may diffuse into cells (unlike the bicarbonate ion) contributing to
further intracellular acidosis. Bicarbonate used as a buffer causes
hypokalaemia, hypocalcaemia and represents a high sodium load. An
alternative solution with more direct intracellular buffering effect is
available. Its composition is: Na+ 195 mmol/l, HCO3- 155 mmol/l,
phosphate 20 mmol/l, acetate 200 mmol/l, and THAM (Tris-
hydroxymethyl aminomethane) 300 mmol/l; osmolality 800 mosm/kg.
By the use of this solution for correction of severe acidosis extra - as
well as intracellular buffering effects will be obtained followed by slow
prolonged buffering when bicarbonate is produced by the metabolism
of the acetate.

Hypertonic saline

Small volume (4 to 6 ml/kg) hypertonic saline (7.5% sodium chloride


solution; 2,400 mOsm/kg) resuscitation is considered advantageous in
several clinical situations since hypertonic saline has been shown
experimentally as well as clinically to increase systemic blood pressure,
cardiac output, peripheral tissue perfusion, and survival rates.
Concentrated sodium ions seem to play an important functional role
since, in spite of similar osmolality, hypertonic solutions having other
ionic compositions are not associated with similar beneficial effects on
outcome when used in the resuscitation of shock and trauma.
Beneficial effects of hypertonic saline have also been demonstrated in
35 connection with increased intracranial pressure, cardiac surgery,
vascular surgery and a number of other experimental and clinical
situations. However, most data indicate that fluid resuscitation
regimens using a combination of hypertonic and hyperoncotic fluid
components are superior to those based on hypertonic saline alone.
Hypertonic saline products are available containing either Dextran 70 or
hydroxyethyl starch (200/0.5) as colloids.

Advantages and Disadvantages


of Crystalloids
In clinical practice the advantages and disadvantages of the different
types of crystalloids have to be considered according to the choice of
therapeutic strategy.

Fluids for maintenance/rehydration

Fluids used for maintenance of the fluid homeostasis and basal


rehydration are expected to have a distribution volume corresponding
to the whole extracellular space and they will also, to some extent,
reach the intracellular compartment. Since hydration is an important
component of the maintenance of fluid homeostasis, the use of high
water - low electrolyte solutions (Ringer’s glucose, buffered glucose,
saline/glucose combinations) is to be preferred unless glucose
administration is contraindicated. If so, a Ringer’s solution should be
chosen instead. Isotonic saline may also be considered but only when
a rather limited volume is to be infused. Whenever infusion of a large
volume of crystalloid is indicated a Ringer’s solution is the most optimal
choice to avoid plasma electrolyte disturbances and the risk of
hyperchloraemic acidosis.

Fluids for rehydration/plasma volume expansion

Advantages and disadvantages of Ringer’s type of crystalloid fluid,


when used for rehydration and plasma volume expansion, are
summarised in Table 4.2.

Table 4.2. Advantages and disadvantages of Ringer’s crystalloid fluids

Advantages Disadvantages

Balanced electrolytes composition Poor plasma volume support


Buffering capacity (lactate/acetate) Large quantities needed
No risk of adverse reactions Reduced plasma COP
No disturbance of haemostasis Risk of overhydration
Promoting diuresis Risk of oedema formation
Inexpensive Risk of hypothermia

Advantages

Balanced electrolyte composition, buffering capacity, absence of risk of


adverse anaphylactoid type of reactions, and minimal influences on
haemostasis (other than effects caused by the haemodilution per se)
are all definite advantages associated with Ringer’s solution. A slight
procoagulant effect of crystalloid resulting in enhancement of
coagulation mechanisms has been suggested but the clinical relevance
of such an in vitro effect (as measured by thrombelastography, TEG) is
not evident. Additional beneficial characteristics of crystalloids are
diuresis promoting effects and low cost (compared to colloids).
36
Intravenous Fluid Therapy Chapter 4

Disadvantages

Salt solutions will freely cross capillary walls and equilibrate within the
whole extracellular compartment. The interstitial fluid compartment is
four times larger than the intravascular one. Therefore, within 20 to 30
minutes about 75% of the infused volume of a Ringer’s solution or
saline solution will leave the vascular compartment and lodge mainly in
the interstitial fluid space. Consequently, the plasma volume
supporting capacity of the crystalloid is poor and infusion of a relatively
large volume of fluid, about four times the estimated intravascular
volume deficit, is necessary in order to achieve normovolaemia and
haemodynamic stability. At the same time plasma colloid osmotic
pressure will be reduced and a new Starling equilibrium for
transcapillary fluid exchange will be established between the
intravascular and the interstitial compartments. Therefore, it is obvious
that crystalloid fluid resuscitation is associated with a potential risk of
development of tissue oedema. It is usually claimed that extravascular
fluid will accumulate mainly in tissues with a high compliance such as
skin and connective tissue but experimental studies indicate that the
fluid content is increased also in vital organs, e.g. in the lungs and the
gastrointestinal tract. The clinical experience is that a significant weight
gain as a result of fluid resuscitation may be associated with increased
need for ventilatory support, impaired wound healing and prolonged
intensive care unit stay. Therefore, it seems obvious that excessive
crystalloid fluid resuscitation is associated with considerable risk of
increased morbidity and even mortality.

Infusion of large quantities of cold fluids may induce hypothermia which


is known to be associated with a risk of cardiac arrhythmias, impaired
tissue perfusion, metabolic disturbances, and coagulopathy. Many of
the coagulation mechanisms function poorly at a temperature below
37°C and at the same time platelet function is impaired. Hypothermia
will furthermore reduce the hepatic synthesis of coagulation factors.
When a large volume of crystalloid is needed it is of importance to
prevent such deleterious temperature dependent effects by proper
warming of intravenously administered fluids.

Perioperative Crystalloid Fluid Therapy

In most perioperative situations intravenous fluid therapy is indicated


and usually infusion of a crystalloid is started prior to induction of
anaesthesia . At this stage the patient has usually been subjected to a
period of restricted fluid and food intake. Therefore, at least in the case
of major surgery, the choice of a solution compensating for the relative
water deficit and the basal fluid requirements seems the most
advantageous approach, i.e. infusion of a maintenance type of fluid
(Ringer’s glucose, buffered glucose, saline/glucose combinations). By
such an approach the risk of hypoglycaemia, a potentially serious
condition that could cause brain damage in the patient with severely
exhausted nutritional resources, is also reduced.

Intraoperatively, the surgical procedure will result in local fluid losses


from the wound area (evaporation, exudation) and blood loss. Ringer’s
acetate (lactate) is the most suitable fluid to compensate these
extracellular fluid losses. Considering the disadvantages of balanced
crystalloid solutions, particularly in major surgical procedures, it is
logical to add a colloid at a rather early stage before excessive
amounts of crystalloid has been infused. By combined infusion of
crystalloid and colloid it is assumed that the following goals of
37 intravenous fluid therapy are more optimally achieved:
• maintenance of normovolaemia and haemodynamic stability

• compensation of internal fluid fluxes between interstitial and


intracellular compartments

• maintenance of an acceptable plasma colloid osmotic pressure

• achievement of adequate urine output (0.5 and 1.0 ml/kg/h)

In the case of minor surgery the choice of crystalloid is less important.


When only one litre of fluid is to be infused during the surgical
procedure either isotonic saline or a maintenance type of fluid (Ringer’s
glucose, buffered glucose) may be chosen.

Conclusions
Crystalloids are commonly used for maintenance of the fluid balance
and correction of intra- as well as extravasacular fluid derangements in
connection with surgical procedures and in the treatment of critically ill
patients. Maintenance fluid (Ringer’s glucose, buffered glucose,
saline/glucose combination) is suitable for compensation of basal fluid
requirements while Ringer’s acetate/lactate is to be preferred for
rehydration and plasma volume expansion. Since excessive crystalloid
fluid administration may result in formation of tissue oedema and, in
case of isotonic saline infusion, the development of hyperchloraemic
acidosis, it is of importance to monitor the fluid therapy properly and
avoid excessive crystalloid fluid infusion.

Bibliography

Haljamäe H. Use of fluids in trauma. Int J Intensive Care 1999;6:20-30.


Haljamäe H. Crystalloids versus colloids: The controversy. In: NATA
Textbook. Paris: R & J Éditions Médicale, 1999:27-36.
Arieff AI. Fatal postoperative pulmonary edema: pathogenesis and
literature review. Chest 1999;115:1371-1377.
Waters JH, Gottlieb A, Schoenwald P, et al. Normal saline versus
lactated Ringer's solution for intraoperative fluid management in
patients undergoing abdominal aortic aneurysm repair: an outcome
study. Anesth Analg 2001;93:817-822.
Haljamäe H, McCunn M. Fluid resuscitation and circulatory support:
Fluids - when, what, and how much? In: Søreide E, Grande CM (Eds).
Prehospital Trauma Care. New York: Marcel Dekker Inc., 2001, pp 299-
322.

38
Intravenous Fluid Therapy Chapter 5

Properties and use of albumin


Shumita Joseph and Gilbert Park

Introduction
Albumin is the most abundant plasma protein. It performs a range of
important functions in health. Hypoalbuminaemia is associated with a
poor outcome in critical illness. It seems logical that replacement of
albumin in critical illness would improve outcome. Indeed, albumin has
been used as replacement fluid in conditions such as burns, shock and
hypoalbuminaemia. However, many studies have now shown that, in
the critically ill, correction of hypoalbuminaemia does not improve
outcome and, in one meta-analysis, was associated with a poorer
prognosis compared to alternative colloid solutions. Although this
meta-analysis was based on old studies and is not definitive in its
conclusion, the use of albumin has been questioned. The consensus
view after the publication of the meta-analysis is that stringent
randomised controlled trials need to be undertaken examining the
efficacy of albumin in critical illness. As a result, the use of albumin has
declined in favour of other colloid solutions. However, there still remain
a few specific clinical conditions where the use of albumin has been
shown to be of benefit.

Albumin in health

Structure

Plasma electrophoresis reveals five major plasma proteins. Albumin,


#1-, #2-, $ globulin and % globulin. Albumin is the most abundant
comprising 50 to 60% of all plasma proteins.

Albumin consists of a single polypeptide chain of 585 amino acids with


a molecular weight of 69,000 Dalton. It is one of the few plasma
proteins that contain no carbohydrate moieties. The single polypeptide
chain is folded into a series of # helices that form three structural
domains. These are held together by 17 disulphide bonds. The three
domains have polar outer walls and a hydrophobic core. In solution
they form an ellipsoid pattern. The albumin molecule is flexible and
strong. It is only denatured at non-physiological temperature and pH.

Metabolism

Total plasma protein concentration ranges from 60 to 80 grams per


litre. Albumin is present at a concentration of about 40 grams per litre.
It is synthesised only in the liver and has a half-life of approximately 20
days. After synthesis albumin is not stored but secreted into the blood
stream with 42% remaining in the intravascular compartment. It is
important to realise that the majority of albumin is found in the
extravascular compartment. Skin and muscle hold the majority of
extravascular albumin, a small amount being present in gut, liver and
subcutaneous tissue.

In health, synthesis matches metabolism. About 13 grams per day are


synthesised in the liver and catabolized in the muscle, skin, kidney, liver
and gut. The liver has a limited capacity to increase albumin synthesis.
It can only increase metabolism by twofold since, at rest, most of the
39 liver’s synthetic activity is already dedicated to albumin. Thus, in states
where there is increased catabolism or loss, such as renal loss in the
nephrotic syndrome or gastrointestinal loss in protein losing
enteropathy, the body is unable to maintain a normal plasma
concentration of albumin. Additionally, certain disease states such as
trauma, sepsis, liver disease and fasting result in reduced albumin
synthesis, and low plasma concentrations of albumin.

Function

Albumin has various functions in health. The main role it plays is


maintaining colloid osmotic pressure and in the transport of
endogenous and exogenous compounds.

Colloid osmotic pressure

The circulation can be divided in to muscular arterioles, which are


resistance vessels, capillaries consisting of a single layer of endothelial
cells and veins, which serve as capacitance vessels. It is at the level of
the capillaries, or microcirculation that albumin exerts its important
colloid osmotic effect.

The thin walled capillaries are separated from the interstitium by only a
single layer of cells - the endothelium. This permits the rapid exchange
of water and solutes with the interstitium. Molecules pass from the
capillaries to the interstitium by three main pathways - diffusion,
filtration and pinocytosis. Albumin affects the rate of filtration. The
hydrostatic and colloid osmotic pressures inside and outside the
capillary govern the rate of filtration of water across the capillary wall
according to the Starling equation. An increase in hydrostatic pressure
inside the vessel favours the movement of fluid into the interstitium. An
increase in colloid osmotic pressure inside the capillary draws fluid in.
The opposite is true for pressures in the interstitium.

Albumin exerts a greater colloid osmotic pressure than would be


expected from the number of particles dissolved in the plasma. This is
because the molecule has a negative charge of 17. The negative
charge attracts cations such as Na+ into the intravascular
compartment. Since the product of diffusible ions must balance across
a semi-permeable membrane (the Gibbs-Donnan effect) the increase in
cations as a result of the charge on the albumin molecule is greater
than the decrease in diffusible anions. These ion shifts contribute an
increase in measured osmotic pressure.

Endothelial integrity

Albumin is also needed to maintain the integrity of the endothelial


membrane. Removing all plasma protein and replacing it with non-
protein colloids causes the filtration rate across the endothelium to
double as a result of an increase in endothelail permeability. The
addition of a small amount of albumin restores normal permeability to
the membrane.

Transport

Albumin is important in the transport of many endogenous and


exogenous compounds in the blood. Albumin carries normal
components of the plasma such as bilirubin and fatty acids. It binds
divalent cations such as calcium and magnesium. It is also a
secondary carrier for steroids that have their own specific binding 40
Intravenous Fluid Therapy Chapter 5

proteins like vitamin D and thyroxine. Although steroids have a low


affinity for albumin, a large amount is bound because of the high
concentration of albumin.

Many commonly used drugs bind to plasma proteins. Acidic drugs


tend to bind to albumin, and basic drugs to #1- acid glycoprotein.
Warfarin, furosemide, salicyclates, phenytoin, benzylpenicillin and
amiodarone are some of the many drugs that are transported by
albumin in the blood.

A decreased concentration of albumin can have important clinical


effects. Warfarin, for example is highly bound, with only 1% existing as
free drug. However, it is the free fraction of drug that is responsible for
its clinical effect. A decrease in albumin concentration will lead to an
increase in free drug and significant effects on coagulation. Similarly,
the simultaneous administration of another drug, such as amiodarone,
which will compete for binding sites, will displace warfarin and again
lead to an increase in the free fraction of drug and an increased
prothrombin time.

Other functions of albumin

Albumin performs a range of other functions in health. It is a plasma


buffer, metabolises endogenous and exogenous compounds and
scavenges oxygen free radicals.

Albumin in critical illness


Since albumin performs important functions in health it seems logical
that critical illness associated with hypoalbuminaemia should be treated
with albumin replacement therapy. Indeed, a low serum albumin can
be used as a prognostic marker in critical illness. However, the
Cochrane Injuries Group Albumin Reviewers questioned this view.
They found that albumin administration may increase mortality in
critically ill patients, although this has subsequently been contested.

Albumin as a prognostic marker

Albumin has been found to be a valuable prognostic marker in critical


illness, both in individual studies and in a recent meta-analysis. In
hospitalised patients low serum albumin is associated with increased
length of stay, higher complication rates and higher mortality. In
critically ill patients survivors have a higher serum albumin
concentration than non-survivors.

A recent meta-analysis examined 90 cohort studies with a total of


291,433 patients. Hypoalbuminaemia was found to be an independent
predictor of poor outcome. A 10 grams per litre decrease in serum
albumin increased the odds of mortality by 137%. Serum albumin
concentration at 24 to 48 hours is as accurate as the APACHE II
system in predicting mortality.

The value of albumin replacement

The value of albumin replacement is less clear. Albumin was licensed


for the treatment of hypovolaemic shock, burns, and clinical states
associated with hypoalbuminaemia. The Cochrane Report, published
in1998, analysed 30 randomised controlled trials in which
administration of albumin was compared with non-albumin fluids, in
41
these conditions. Mortality from any cause was taken as the main
outcome measure. For each disease state it was found that the risk of
death was increased in the albumin treated group compared to the
control group. Overall, it was suggested that there were 6 additional
deaths for every 100 patients treated with albumin.

The limitations of this meta-analysis were that it was based on relatively


small and old individual trials in which the total number of deaths was
small. The report nevertheless questioned the traditionally held view
that administration of albumin was of benefit in critical illness. It called
for rigorous, large, randomised controlled trials to assess the value of
albumin therapy.

Since the Cochrane Report, conflicting studies have been published.


One study analysed nine prospective controlled trials and examined the
complication rates after the administration of albumin. The study
suggested that complication rates were reduced when serum albumin
was maintained greater than 30g/l.

A further study examined 55 randomised controlled trials that used


albumin in the treatment of trauma, burns, hypoalbuminaemia, high-risk
neonates, and ascites. It was found that albumin did not have a
significant effect on mortality in any of these conditions.

Thus the effect of albumin on mortality in critical illness remains


debated. It should be borne in mind that a meta-analysis examines a
heterogenous group of studies and sometimes extracts data that the
studies were not initially designed for. There still remains the need for a
specifically designed, randomised control trial evaluating the effect of
albumin on mortality in specific disease states.

Many studies in critically ill adults have shown that albumin has no
advantages over other colloids in this group of patients, but is much
more expensive.

In children an improvement in the absorption of food has been seen


when hypoalbuminaemia has been corrected. This is thought to be
because of the reduction of bowel oedema. In neonates albumin has
been used to treat hypotension, metabolic acidosis and promote a
diuresis. Unfortunately, many of the studies are small and some of
these effects may also be achieved with other agents that are both
more effective and cheaper.

Possible reasons why albumin may not


work
A number of reasons as to why albumin therapy may not improve
outcome in the critically ill have been proposed.

Effects on colloid osmotic pressure

Starling’s principle is one of the main reasons why albumin therapy was
thought to be of benefit. In hypovolaemic shock increasing the colloid
osmotic pressure would encourage the reabsorption of fluid from the
interstitium. However, the effect is transient and may only amount to
an increase in the intravascular volume by 500 millilitres over 15
minutes.

42
Intravenous Fluid Therapy Chapter 5

In septic shock the release of inflammatory mediators has been


implicated in increasing the ‘leakiness’ of the vascular endothelium.
Albumin then escapes into the interstitial tissue, causing potentially life-
threatening conditions such as pulmonary oedema. This loss of
albumin is the main cause of hypoalbuminaemia in early critical illness.
The administration of exogenous albumin does not treat the damaged
endothelial membrane. Indeed, it may compound the problem by
adding to the interstitial oedema.

A small, randomised controlled study examined the effects of treating


acute lung injury patients needing mechanical ventilation with albumin
and furosemide or placebo. There was a significant diuresis, weight
loss, and improvement in haemodyanamic indices in the
albumin/diuresis group at 24 hours, but no significant difference at day
seven. Importantly, the difference may have been achieved by
furosemide therapy alone. Interestingly, a similar conclusion was
reached when the effects of albumin and furosemide were investigated
in hypoalbuminaemic patients with renal failure in an attempt to
promote a diuresis. The furosemide was found to be the key
component rather than the albumin.

Increased catabolism

In diseased states catabolism is increased. An increased transcapillary


flux of albumin is associated with increased catabolism. The structure
of albumin is also altered in disease. For example, in pancreatitis
bisalbumin is formed. In critically ill patients albumin behaves differently
to normal when it is examined by electrophoresis, migrating more
quickly than in controls. It is possible that altered albumin is more
susceptible to catabolism. Again the administration of exogenous
albumin does not treat the cause of the problem, and might therefore
not be of benefit.

Adaptation to hypoalbuminaemia

Humans may be able to adapt to low albumin states although the


mechanism remains unclear. There is a rare condition where individuals
have a serum albumin concentration of less than 1 gram per litre. It is
thought to be because of a spontaneous genetic mutation.
Interestingly, these people are able to live a relatively healthy life. They
do not suffer cardiorespiratory dysfunction such as shock or pulmonary
oedema, although their colloid osmotic pressure is reduced from 25 to
16mmHg. The main features are peripheral oedema, fatty deposition in
the lower limbs and fatigue.

The way in which these individuals adapt to a low albumin state is


unknown, but may be of relevance in the acute hypoalbuminaemia
seen in critical illness.

Specific Indications for the use of albumin


Despite the recent controversy surrounding the use of albumin there
remain a few clinical in indications for its use that are supported by
small studies.

Albumin infusion following paracentesis

Paracentesis is indicated for the treatment of tense ascites in cirrhotic


patients who are refractory to medical treatment. Patients in liver failure
43
have marked disorders of their cardiovascular system with a
hypodynamic, hypotensive circulation. Following large volume
paracentesis they are at risk of further paracentesis induced circulatory
dysfunction. In studies of patients having paracentesis those who
received albumin rather than saline as replacement fluid had a reduced
risk of circulatory dysfunction. However, use of artificial colloid
solutions may also reduce the risk of circulatory dysfunction.

Extracorporeal albumin dialysis (ECAD)

ECAD has been used in patients with liver cirrhosis and a


superimposed acute injury who had not responded to standard medical
therapy. It works on the principle that whole blood is dialysed against a
solution of 15% human serum albumin. Albumin bound molecules
transfer from blood to dialysate. Using this technique
hyperbilirubinaemia, 30-day survival, and encephalopathy improved
significantly. It was thought to be the removal of toxic compounds
such as ammonia that was of benefit.

After liver transplantation

In the early post operative period after liver transplantation there is a


degree of hepatic dysfunction and patients can become significantly
hypoalbuminaemic. This can increase the concentration of free drug
used for immunosupression therapy with resulting adverse effects. We
have shown that a low serum albumin leads to an increase in the free
fraction of tacrolimus and an increase in nephrotoxicity as evidenced by
an increase in serum creatinine. Conversely, treating patients with
tacrolimus and albumin may reduce adverse effects but may also
reduce the immunosuppressive activity of the drug, increasing rejection.

Conclusion
There is much that is unknown about the role of albumin in health and
disease. Although many of its functions are known in health there
remain questions, such as how analbuminaemic humans adapt to a
low serum albumin. Although albumin has been found to be a
prognostic marker in critical illness, how it functions in disease is not
clear. There is no clear evidence for the use if albumin, rather than
other colloids, as fluid replacement therapy in conditions such as
hypovolaemia, burns and hypoalbuminaemia. There is an urgent need
for randomised controlled trials seeking to answer the question whether
albumin administration is associated with excess mortality and whether
it should be used to treat hypoalbuminaemia.

Bibliography

Cochrane Injuries Group Albumin Reviewers. Human Albumin


Administration in Critically Ill Patients: systematic review of randomised
controlled trials, BMJ 1998;317:235-240.
Nicholson J, Wolmaris M, Park G. The Role of Albumin In Critical
Illness, Br J Anaesth 2000;85:599-610.
Nicholson J, Park G. The Need for Reappraisal of the Albumin
Molecule, Curr Opin Crit Care 1998;4:224-227.
Peters TJ. All about albumin. Biochemistry, genetics and medical
applications. San Diego: Academic Press, 1996.
Park G. Molecular Mechanisms of Drug Metabolism in the Critically Ill. 44
Br J Anaesth 1996;77:32-49.
Intravenous Fluid Therapy Chapter 6

Properties and use of gelatins


Justin Roberts and Peter Nightingale

Introduction
Gelatins are a group of large molecular weight proteins that have been
developed as plasma replacement solutions. When used as plasma
replacement fluids gelatins exert a colloid osmotic pressure across
capillary walls, transiently retaining fluid within the intravascular space.
They are not considered as significant plasma expanders.

Two types of gelatin are in common use:


• Succinylated gelatin (MFG - modified fluid gelatin)
• Urea cross-linked gelatin (polygeline)

A number of other gelatin solutions are available world-wide, e.g.


oxypolygelatin produced by dialdehyde glyoxal cross-linking, as well as
other modified fluid gelatins differing in concentration of the gelatin and
solute make up of the carrying solution.

Manufacture
Gelatin was first used as a colloid replacement fluid by Hogan in 1915.
The first solutions used had a high molecular weight (approximately
100,000 Dalton) with the advantage of exerting a significant colloid
osmotic effect. Unfortunately, the high molecular weight meant the
solutions also had a tendency to gel at low temperatures.
Modifications led to gelatins that exerted a moderate colloid osmotic
effect but had a low gel melting point, i.e. remained fluid at room
temperature.

The two gelatin types in common use are manufactured from bovine
collagen, a helical array of three polypeptide chains. During synthesis
collagen rich bovine tissue is denatured. The collagen is then allowed
to renature and, in doing so creates, new inter-chain bonds that give
gelatins their characteristic appearance and properties.

Succinylated gelatins are prepared by heating and chemical hydrolysis.


Succinic anhydride is used to replace free amino groups with acid
carboxyl groups resulting in a conformational change in the molecule.
Polygeline is produced by thermal degradation of the same raw
material but the peptides are cross-linked using hexamethylene di-
isocyanate to form urea bridges. At any one molecular weight the
‘open architecture’ succinylated molecule will be larger than the cross
linked polygeline molecule.

Fig 6.1

Figure 6.1 H2C CH2


o

Structure of succinylated gelatin. NH2


+ c
o
c
o
NH-C-CH2-CH2-C cation+
o o

Peptide Succinic anhydride Succinylated protein chain


e.g. (modified fluid gelatin)
negative charged chain due to
carboxyl-groups [Mw 30 000]
45 Modification of polypeptide fragments by use of succinic anhydride after hydrolysis of gelatin
Fig 6.2

Figure 6.2
NH-C-NH-R-NH-C-NH Structure of polygeline.
NH-C-NH-R-NH-C-NH O O

O O
N=C=O
NH2
+ R
N=C=O
NH-C-NH-R-NH-C-NH

Peptide Diiso - cyanate O O

Urea linked gelatin


[Mw 35 000]

Cross linking of polypeptide fragments by use of diiso-cyanate after hydrolysis of gelatin

The shelf life for succinylated gelatin is 2-3 years, depending on the
packaging, and 5 years for polygeline.

Pharmacology
Gelatins act as plasma substitutes and maintain intravascular volume
thus allowing re-establishment of normal haemodynamics in the
hypovolaemic patient. They do not draw fluid from the extravascular
compartment and initial increases in intravascular volume in the healthy
patient reflect the total infused volume. After intravenous administration
the clinical effect is dependent upon:

• Rate and volume of infusion


• Blood volume
• Their colloid osmotic pressure
• Their water-binding capacity (~ 15 ml/g)
• Capillary permeability

Pharmacokinetics
Gelatins are complex solutions of molecules of varying molecular
weight and their pharmacokinetics are influenced greatly by multiple
factors.

Following the infusion of 500 to 1000 ml of polygeline the apparent


volume of distribution has been calculated as 120 ml/kg; this equates
to a plasma volume expansion of 350-500 ml per litre of gelatin
infused. However, these results were obtained in healthy volunteers
and caution is needed in applying them to sicker patients. It is easy to
see how a patient with sepsis who is hypovolaemic and has an
increased capillary permeability would have a different volume of
distribution from the healthy volunteer.

Very little metabolism of gelatins occurs prior to elimination from the


body; studies using radio-labelled polygeline suggest 1-3%.
Metabolism occurs predominantly as a result of endogenous proteases
such as trypsin, plasmin and cathepsin. 46
Intravenous Fluid Therapy Chapter 6

Excretion of the largely unchanged gelatin is via the renal route. After
24 hours approximately 60-70% of an infused volume is present in the
urine, 12% remains in the plasma and 15-25% is in the interstitial
space. Normally, no gelatin is detectable in the blood after 48 hours
and by 96 hours approximately 75% has been excreted via the
kidneys.

The elimination half-life from plasma has been found to vary according
to the patient population studied but is estimated to be approximately
5-8 hours. Patients with end-stage chronic renal failure on
haemodialysis can still receive gelatins although a prolonged plasma
half-life has been observed; this does not appear to cause problems.
Gelatins are not stored in the reticulo-endothelial system or elsewhere
in the body.

Physicochemical properties
Both succinylated gelatin and polygeline are supplied as preservative-
free, sterile solutions in sodium chloride. Their physicochemical
properties are detailed in Table 6.1.

In solution gelatins have a range of molecular weights, i.e. they are


polydispersed with a variable number of molecules of each weight.

Table 6.1. The physicochemical properties and effects of succinylated gelatin and polygeline
Succinylated gelatin Polygeline

Colloid content (g/l) 40 35


Na+ (mmol/l) 154 145
Cl- (mmol/l) 120 145
K+ (mmol/l) 0.4 5.1
Ca++ (mmol/l) 0.4 6.25
Molecular weight
weight average (MWw) 30,000 35,000
number average (MWn) 23,200 24,500
Polydispersity ratio 1.29 1.43
pH 7.4±0.3 7.3±0.3
Iso-electric point (pH) 4.5±0.3 4.7±0.3
COP [colloid osmotic pressure] (mmHg) 33 28
COP50 / COP10 0.37 0.18
Osmolarity (mOsm/l) 274 300
Volume effect (%) >80 80
Duration of volume effect (hrs) 3-4 1-2

Therefore average figures must be quoted. The number-average


molecular weight (MWn) and weight-average molecular weight (MWw)
can be used.

total weight of all molecules


MWn=
total number of all molecules

& (weight of molecules of each weight x each weight)


MWw=
total weight of all molecules

The weight-average MW is always the larger figure and is usually


quoted by manufacturers. However, the number-average molecular
weight better reflects the mean osmotically active particle weight of a
colloid; it is this number that should be compared to the value of
69,000 Dalton for albumin. Since albumin is a monodispersed colloid
Mw = Mn whereas for artificial colloids the polydispersity ratio, Mw/Mn,
is always greater than one.

Although each manufacturer quotes an osmotic pressure for their


product the actual value will depend on a number of factors, including
47
the pore size of the membrane used. A colloid with a large number of
small molecules will have a short duration of action as these leak out of
the circulation. A colloid with a greater number of large molecules will
have a longer effect. The colloid with the most molecules will have the
greater osmotic activity.

A useful technique to allow comparison of these variables with albumin


was devised by Webb and colleagues. They measured the osmotic
pressure of colloids using a relatively permeable membrane (nominal
molecular weight cut-off 50,000 Dalton) and one with a smaller pore
size (nominal molecular weight cut-off 10,000 Dalton) to produce the
COP50 / COP10 ratio; the value for albumin was 0.37, comparable to
the value for succinylated gelatin. It can be seen that this value helps
one to choose a colloid of suitable duration of action.

Pharmaceutics
Succinylated gelatin is presented as a 4% w/v gelatin in ionic solution
supplied in 500 millilitre or 1000 millilitre plastic infusion bags. It
contains no additional electrolyte components and is thus compatible
with blood for infusion.

Polygeline is presented as a 3.5% gelatin solution which, in addition to


sodium and chloride, also contains potassium and calcium as detailed
in Table 6.1.

Pregnancy and Lactation

There are no human studies reporting on carcinogenicity, mutagenicity,


fertility impairment or passage into breast milk. Although gelatins have
been found to be embryotoxic when given in sufficient quantities to
pregnant rats they have been given many times to pregnant women
without apparent problem.

Indications for use

Gelatins are widely used for the following indications:

• Plasma volume substitute in hypovolaemic states


• Acute normovolaemic haemodilution
• Plasma exchange
• Extracorporeal circulation
• Isolated organ perfusion
• Carrier for other substances

It should be noted that some of these are not licensed indications.

Plasma volume substitutes

When red cells are not needed the choice of fluid for plasma volume
replacement and expansion is controversial. Recent meta-analyses
have been criticised because of the age and heterogeneity of the
studies included which did not reflect current practice in fluid
management. Proponents of colloid therapy suggest that individual
48
Intravenous Fluid Therapy Chapter 6

patient identification and correct use of the colloid are important in


determining outcome on a single patient basis.

The rate and volume of infusion will be determined by the state of the
patient. It is advised that on first exposure the first 20 to 30 millilitre
should be infused slowly in case there is an adverse reaction. Where
necessary, gelatins can be infused rapidly and in large volumes,
warming appropriately if possible. There is no upper limit to the
amount that can be infused but generally the haematocrit should be
maintained above 25%. Gelatins can rapidly increase intravascular fluid
volume in many clinical circumstances:
• Hypovolaemia which maybe absolute due to haemorrhage, trauma,
burns, etc. or relative due to anaesthesia, distributive shock, sepsis,
etc.

• Acute normovolaemic haemodilution - a technique whereby fresh


blood is taken from a patient immediately prior to surgery.
Intraoperative blood loss occurs at a lower haematocrit and so less
red cells are lost. The risks of transfusion with donor blood are
reduced and the beneficial effects of a fresh blood transfusion are
obtained at the end of the procedure.

• Volume pre-loading prior to regional anaesthesia - a technique to


counter the decrease in cardiac pre-load and consequent
hypotension due to peripheral vasodilatation following regional
anaesthesia.

There have been several recent randomised controlled trials examining


the clinical usefulness of gelatins. A comparison of a rapid 1000
millilitre infusion of either succinylated gelatin or Ringer’s lactate solution
for the emergency resuscitation of 34 patients with either traumatic or
neurogenic shock showed improved haemodynamics in the gelatin
group; there were significantly greater rises in central venous pressure
and pulmonary artery wedge pressure within the first hour of
resuscitation.

Plasma exchange

The presence of a gelatin may decrease the initial hypotensive


response. There appear to be no apparent effects on coagulation.

Extracorporeal Circulation

Gelatins may be used for extracorporeal cardiopulmonary bypass. A


beneficial effect is the decreased use post-operatively of diuretics for
oliguria.

A comparison of polygeline with Ringer’s solution for volume therapy


following coronary artery bypass grafting, using stringent protocols,
showed no significant difference in haemodynamic variables. However,
the total volume of gelatin infused was less. Interestingly, within 8
hours of surgery, total blood volume index and intrathoracic blood
volume index were higher for the gelatin group. A similar study of the
effects of gelatins and hetastarch solutions in the critically ill population
with hypovolaemia demonstrated improvement in haemodynamic
variables with increased oxygen delivery but without any deleterious
effect on respiratory function.

49
Carrier for other substances

Gelatins can be used as a carrier fluid for the intravenous infusion of


other substances. There are no known compatibility problems with
many of the water soluble drugs used in clinical practice.

Advantages

There are several advantages to the use of gelatins over crystalloids


and other colloids that make them attractive for fluid therapy. These
include:

• Rapidity of intravascular fluid replacement; haemodynamics


normalised more readily than crystalloid.
• Persist intravascularly for a short but clinically useful time
(2-4 hours).
• Smaller volumes required, compared to crystalloids, for an
equivalent expansion of intravascular volume.
• Maintain plasma colloid osmotic pressure compared to crystalloids.
• Equipotent total fluid volumes are less compared to crystalloids and
hence there is less risk of tissue oedema.
• Renal excretion of gelatins provides an osmotic diuresis and allows
room for blood to be transfused, if necessary, with less risk of
circulatory overload.
• No upper limit on the volume that can be infused in contrast to
starch solutions.
• Elimination of the gelatins is complete; there is no persistence in the
tissues.
• Cheaper and more readily available than plasma proteins.
• No infection risk.
• Long shelf-life and stability at room temperature.
Undesirable effects

The main risk is volume overload, especially in patients with congestive


cardiac failure and severe renal impairment.

Effects on coagulation

Haemodilution of clotting factors will occur as with any fluid when given
in sufficient volume. There are mixed data on the direct effects of
gelatins on coagulation, demonstrating no effect, enhanced
coagulability or diminished coagulability. Thromboelastography
suggests that, at low dilutions, gelatins may enhance coagulation. This
has been challenged by some studies that suggest this could be an
environmental effect since thromboelastography values may vary from
baseline in patients presenting for surgery and receiving no fluid. A
further study has demonstrated a 1.7 fold increase in bleeding time,
with saline having no effect. This was due to a 32% reduction in von
Willebrand factor and a 45% reduction in thrombin-antithrombin
complexes. Ristocetin induced thrombocyte aggregation may be
impaired and scanning electron microscopy has shown that fibrin forms
a less extensive mesh in the presence of gelatins compared with
crystalloids. However, in contradiction to this, other studies and case
reports show no evidence of a significant disturbance of haemostasis 50
Intravenous Fluid Therapy Chapter 6

even with large volume gelatin infusions. Polygeline contains calcium


which may activate the clotting cascade when mixed with citrated
blood or fresh-frozen plasma.

Other undesirable effects

The erythrocyte sedimentation rate may increase.


The specific gravity of the urine will be increased by gelatins and hence
should not be used as an index of renal function.
Significant proteinuria has been reported. One study found an increase
in $2-microglobulin and #1-microglobulin in intensive care patient
patients but without significant renal tubular damage.
The hyperoncotic syndrome is unlikely with gelatins although
theoretically possible with succinylated gelatin in renal failure.
Gelatins form complexes with fibronectin in-vivo inhibiting opsonisation
which is necessary for phagocytosis.
Polygeline has been shown to activate complement C3a more than
hydroxyethyl starch although there were no clinical manifestations.

Anaphylactic and anaphylactoid reactions may occur. Histamine


release may lead to urticarial rash, hypotension, tachycardia and
bronchospasm although not all features always occur. Polygeline given
by inappropriately rapid infusion, especially to normovolaemic patients,
may cause release of vasoactive substances. Cross-reaction can
occur in a patient known to be susceptible to one gelatin when a
different gelatin is used. Most reactions cause minor clinical upset
although there have been some fatalities; in a major prospective study
the incidence of reaction to gelatins was 1 in 2,500 but serious
reactions are much less common, around 1 in 10,000. Reactions can
occur on first clinical use without previous exposure. The infusion must
be stopped; further management is as for any anaphylactic reaction.
Gelatins are more expensive than crystalloid solutions but cheaper than
hydroxyethyl starch and protein solutions.

Differences of significance between


succinylated gelatin and polygeline
Due to the methods of manufacture, polygeline is presented in plastic
containers containing up to 45 ml of air; this has resulted in venous air
embolism when infused using pressure bags. Polygeline contains
calcium ions and increases in serum calcium concentration have been
reported in the initial hours following large volume resuscitation; there
were no reported clinical sequelae. It would seem prudent to avoid
polygeline if there is hypercalcaemia or where this may lead to a drug
interaction, e.g. in patients on cardiac glycosides. Conversely, in
hypocalcaemic patients circulatory status may improve with polygeline.
Polygeline also contains potassium ions; this may be beneficial to those
patients who are hypokalaemic.

Succinylated gelatin has lower chloride content which may reduce the
incidence of hyperchloraemic metabolic acidosis. Succinylated gelatin
should not be frozen or stored above 25°C whereas there are no
special storage precautions for polygeline. Succinylated gelatin has an
overall net negative charge which contributes to intravascular retention
by repulsion by the negatively charged capillary wall and increases the
colloid osmotic pressure (by the Gibbs Donnan effect).
51
Conclusions
Gelatins are useful intravenous solutions that fill the gap between
crystalloids and blood products. They are easy to store, relatively short
acting, maintain the colloid osmotic pressure, probably do not affect
coagulation and are safe to use.

Bibliography

Webb AR, Barclay SA, Bennett ED. In vitro colloid osmotic pressure of
commonly used plasma expanders and substitutes: a study of the
diffusibility of colloid molecules. Intensive Care Med 1989;15:116-120.
De Jonge E, Levi M, Berends F, et al. Impaired haemostasis by
intravenous administration of a gelatin-based plasma expander in
human subjects. Thromb Haemost 1998;79:286-290.
Kohler H, Fuchs P, Stalder K, Distler A. Elimination of hexamethylene
diisocyanate cross-linked polypeptides in patients with normal or
impaired renal function. Eur J Clin Pharmacol 1978;14:405-412.
Wu JJ, Huang MS, Tang GJ, et al. Hemodynamic response of modified
gelatin compared with lactated ringer’s solution for volume expansion in
emergency resuscitation of hypovolemic shock patients: preliminary
report of a prospective randomized trial. World J Surg 2001;25:598-
602.
Wahba A, Sendtner E, Strotzer M, et al. Fluid therapy with Ringer’s
solution versus Haemaccel following coronary artery bypass surgery.
Acta Anaesthesiol Scand 1996;40:1227-1233.
Tait AR, Larson LO. Resuscitation fluids for the treatment of
hemorrhagic shock in dogs: effects on myocardial blood flow and
oxygen transport. Crit Care Med 1991;19:1561-1565.
Beards SC, Watt T, Edwards JD, et al. Comparison of the
hemodynamic and oxygen transport responses to modified gelatin and
hetastarch in critically ill patients: a prospective, randomized trial. Crit
Care Med 1994;22:600-605.

52
Intravenous Fluid Therapy Chapter 7

Properties and use of hydroxyethyl


starches
Joachim Boldt

Introduction
The choice of fluid to restore volume in hypovolaemic shock and to
guarantee stable macro- and micro-haemodynamics while avoiding
excessive fluid accumulation in the interstitium engenders much
controversy. There is still a dispute over the beneficial and adverse
effects of each fluid type. Hydroxyethyl starch (HES) appears to be a
very effective plasma substitute and is widely used to treat the
hypovolaemic patient.

Pharmacology of Hydroxyethyl Starch


HES is a derivative of amylopectin, which is a highly branched starch
compound (Fig 7.1). In humans and animals, amylopectin is rapidly
hydrolysed by #-amylase and renally excreted. In order to slow down
the metabolic degradation, anhydroglucose residues of the amylopectin
are substituted with hydroxyethyl groups. Substituting hydroxyethyl for
hydroxyl groups results in highly increased solubility and retards
hydrolysis of the compound by amylase, thereby delaying its
breakdown and elimination from the blood. The hydroxyethyl groups
can be introduced mainly at positions C2, C3, and C6 of the
anhydroglucose residues.

Fig 7.1

HES Structure
Fig 7.1 1,4 linkage (fixed)

Structure of hydroxyethyl starch


(HES).

1,6 linkage (flexible)

Substitution of hydroxy group with


further hydroxyethyl group

53
HES preparations differ widely with regard to their physico-chemical
properties.

They are characterized by:


• concentration
• molar substitution (MS: the molar ratio of the total number of
hydroxyethyl groups to the total number of glucose units)
• degree of substitution (DS: the ratio of substituted glucose units to
the total number of glucose molecules)
• molecular weight described as weight averaged mean molecular
weight (MWw) and

& (weight of molecules of each weight x each weight)


MWw=
total weight of all molecules

total weight of all molecules


MWn=
total number of all molecules

• number average molecular weight (MWn)

Three of these characteristics are given by the manufacturers


(Table 7.1):

• concentration: 3%,6%,10%,
• weight average molecular weight (MWw):
low-molecular weight [LMW]-HES: 70,000 Dalton
medium-molecular weight [MMW]-HES: ranging from 130,000 to
270,000 Dalton
high-molecular weight [HMW]-HES: >450,000 Dalton
• degree of substitution (DS):
low DS: 0.4 and 0.5
high DS: 0.62; 0.7

There is convincing evidence that the ability of #-amylase to hydrolyse


the HES molecule depends on the position of the hydroxyethyl groups
on the glucose molecule (C2,C3,C6). The ratio of C2:C6
hydroxyethylation appears to be an important factor in the
pharmacokinetic behaviour of HES and possibly also in its side-effects
(e.g. accumulation). The C2:C6 ratio, however, is not yet shown by the
different manufacturers.

The water-binding capacity of HES ranges between 20 and 30 millilitres


per gram. Therefore, HES solutions have a good plasma volume
expanding capacity. Following the infusion of HES there is initially a
rapid amylase-dependent breakdown and renal excretion of up to 50%
of the administered dose within 24 hours. The hydroxyethyl residues,
especially when bound to the C2 carbon position of glucose, hinder the
plasma amylase, hence increasing the intravascular half-life of the HES
solution. A higher molecular weight range and a more extensive
degree of substitution result in slower elimination. Smaller HES
molecules (<50,000 to 60,000 Dalton) are eliminated rapidly by 54
Intravenous Fluid Therapy Chapter 7

glomerular filtration. Renal elimination by filtration continues as larger


HES molecules are hydrolysed to smaller molecules. A small amount
of the administered dose passes to the interstitial space for later
redistribution and elimination. Another fraction is taken up by the
reticuloendothelial system (RES) where the starch is slowly broken
down. Thus, trace amounts of the preparations can be detected for
several weeks after administration.

Table 7.1. Characteristics of different HES solutions


HES 70/0.5 130/0.4 200/0.5 200/0.5 200/0.62 450/0.7
Concentration (%) 6 6 6 10 6 6

Volume efficacy (%) 100 100 100 130 100 100


Volume effect (hours) 1-2 2-3 3-4 3-4 5-6 5-6
MWw (Dalton) 70,000 130,000 200,000 200,000 200,000 450,000

Degree of substitution 0.5 0.4 0.5 0.5 0.62 0.7


C2/C6 ratio 4.1 9.1 6.1 6.1 9.1 4.6:1

In Europe, numerous types of HES preparations with different


combinations of concentration, weight-average molecular weight, and
hydroxyethylation pattern are available (Table 7.1). The development of
HES solutions has resulted in lower and medium molecular weight
solutions (Fig 7.2).

Fig 7.2

Fig 7.2:
Development of hydroxyethyl
starch (HES) preparations.

Another modification of HES preparations has involved the


development of a physiologically "balanced" first generation HMW-HES
preparation (molar substitution: 0.7; weight average molecular weight:
approximately 450,000 Dalton) in a balanced electrolyte solution. The
balanced electrolyte solution has been reported to eliminate the
negative side-effects of standard HMW-HES.

55
Clinical Use of Hydroxyethyl Starch

Haemodynamic effects

Stabilisation of systemic haemodynamics has been well established


with all kind of HES preparations. The duration and the initial effect of
the specific HES solution, however, is dependent on the physico-
chemical characteristics of the HES preparation (Table 7.1). Long-
acting HES solutions (e.g. HES 450/0.7) appear to be less favourable
because they are associated with several undesired side-effects.
Short-acting HES solutions (e.g. HES 70/0.5) have only a very limited
duration of hemodynamic stabilisation and have to be reinfused every 1
to 2 hours thus increasing costs considerably. Medium-acting HES
solutions (HES 130/0.4, HES 200/0.5) appear to be best suitable for
volume deficit restoration because they have beneficial haemodynamic
effects and few side-effects.

Microcirculatory effects

Although macro-haemodynamics may be restored to a similar degree in


crystalloid and colloid treated subjects, even massive crystalloid
resuscitation is less likely to achieve adequate restoration of
microcirculatory blood flow. In septic animal experiments, greater
capillary luminal area associated with less endothelial swelling and less
parenchymal injury has been shown using MMW-HES (HES 270/0.5)
than with Ringer`s lactate. In patients undergoing major abdominal
surgery third-generation MMW-HES (HES 130/0.4) increased tissue
PO2 compared to patients who received exclusively saline solution for
volume replacement. This was despite similar systemic
haemodynamics and systemic oxygenation data in both volume
groups. Thus, MMW-HES improves the microcirculation compared to
crystalloid.

Dose-limitations

Recommended maximum doses per day from the manufacturers varies


with regard to the specific HES preparation (Table 7.2). There is no
specific limitation for the speed of infusion, whereas, for albumin, it is
recommended not to administer more than 3-5 ml/min. Maximum
cumulative doses over time is not definitely known, but no harm was
reported when more than 5 litres of a MMW-HES with a low DS (HES
200/0.5) was administered over 5 days in the intensive care patient.
With the third generation HES (6% HES 130/0.4) 70ml/kg have been
given in patients without severe side-effects.

Table 7.2. Recommended maximum doses of hydroxyethyl starch (HES)


1st generation 20ml/kg 6% HES 450/0.7
nd
2 generation 33ml/kg 6% HES 200/0.5
33ml/kg 6% HES 70/0.5
20ml/kg 10% HES 200/0.5
3rd generation 50ml/kg 6% HES 130/0.4

56
Intravenous Fluid Therapy Chapter 7

Side-Effects of Hydroxyethyl Starch


As with all plasma substitutes, HES solutions used to restore
hypovolaemia have their merits and limitations.

Coagulation

Abnormal coagulation has been reported for the first-generation, high-


molecular weight (HMW) HES preparation (mean molecular weight
[MWw] 450,000 Dalton, degree of substitution [DS] 0.7) or other HES-
preparations with a high DS (0.62). Infusion of HMW-HES may result in
a type I von Willebrand-like syndrome with reduced factor VIII
coagulant activity, decreased von Willebrand`s factor antigen and factor
VIII-related ristocitin cofactor. HES with lower molecular weight (LMW-
or MMW-HES) influences concentrations of VIIIR:Ag and VIIIR:RCo
significantly less than HMW-HES and appears to be safe with regard to
haemostasis.

The weight average mean molecular weight (MWw) of the HES


preparation seems to play a role in the influence on coagulation; when
either 500ml of HES with a MWw of 70,000 Dalton or HES with a
MWw of 200,000 Dalton was given over 30min, von Willebrand factor,
prothrombin time, and maximum amplitude measured by
thrombelastography were altered more by HES 200/0.5 than by HES
70/0.5.

Highly-substituted HES preparations are associated with considerable


negative effects on blood clotting and subsequently increased
postoperative bleeding tendency. This has been confirmed in patients
scheduled for minor elective surgery in whom either 15 ml/kg saline
solution or 15 ml/kg of a HES preparation with a medium MWw
(200,000 Dalton) but a high DS (0.62) was given; a reduced GPIIb-IIIa
expression and a reduced maximum amplitude measured by
thrombelastography were seen only in the HES group. A meta-analysis
on postoperative bleeding in cardiac surgery patients in whom either
albumin or different HES preparations were given was performed by
Wilkes et al. Both solutions were used either before or after
cardiopulmonary bypass or as an addition to the priming of the
extracorporeal circuit. When HMW, highly substituted HES (HES
450/0.7) was compared with albumin postoperative bleeding was
significantly higher in the HMW-HES patients. When HES with a lower
MWw (200,000 Dalton) was compared with albumin there was no
significant difference in postoperative bleeding.

New HES preparations have shed new light on the problem of HES
and coagulation. HMW-HES in balanced electrolyte solution and a
third- generation, low-molecular weight/low-substituted HES (MWw
130,000 Dalton; DS 0.4) have been developed to avoid side-effects of
HES on haemostasis. HMW-HES in balanced electrolyte solution has
been reported to affect coagulation significantly less than standard
HMW-HES although this has not been confirmed in other studies. The
third-generation HES (HES 130/0.4) shows favourable effects on
coagulation; in 50 healthy patients undergoing minor elective surgery
10ml/kg of saline or 10 ml/kg of 4 different HES preparations including
the new HES 130/0.4 were given. After infusion of HES 130/0.4,
platelet function remained unaffected, whereas the other HES
preparations (HES 70/0.5; HES 200/0.5; HES 450/0.7) resulted in
considerable alterations in platelet function. Using activated
thrombelastography (TEG), it was shown that infusion of approximately
57
2,500ml of HES 130/0.4 in patients undergoing major abdominal
surgery was associated with almost no negative effects on
haemostasis. This new generation of HES did not show significant
negative effects on coagulation time, clot formation time, and maximum
clot firmness measured by TEG. In studies using HES 130/0.4 in
orthopaedic and cardiac surgery patients, less bleeding and less use of
blood/blood products than with a conventional HES 200/0.5
preparation were reported.

Renal

Elimination of HES varies mostly with molecular weight (MWw) and


degree of substitution (DS). Large HES molecules are split by
hydrolytic cleavage by alpha-amylase. Smaller HES molecules are
eliminated by glomerular filtration. Histological studies have shown
reversible swelling of tubular cells of the kidneys after administration of
HES, which appears to be due to reabsorption of macromolecules.
Swelling of tubular cells may result in tubular obstruction, medullary
ischaemia and in the development of renal failure. Glomerular filtration
of colloid molecules causes a hyperviscous urine and stasis of tubular
flow resulting in obstruction of tubular lumen. This kind of renal failure
can be induced by almost all colloids.

The dehydrated patient who receives considerable amounts of


hyperoncotic colloid is at particular risk of developing hyperoncotic
acute renal failure. Thus colloids have to be administered in addition
rather than in lieu of crystalloids.

In a retrospective study of brain-dead kidney donors in whom HES with


a high DS (0.62) was infused, osmotic-nephrosis-like lesions were
shown. However, there was no relation between these lesions and
graft function or serum creatinine levels 3 and 6 months after
transplantation. A multicentre study in 129 intensive care patients with
sepsis or septic shock reported a higher incidence of renal failure in
patients treated with MMW-HES with a high DS (200/0.62) compared
to gelatin. However, there was no difference in mortality, length of
hospital stay or need for renal replacement therapy. Compared to a
control group in whom albumin was administered, long-term use over 5
days of HES 200/0.5 in critically ill intensive care patients was without
negative effects on renal function (serum creatinine levels). Promising
results have been shown with the third-generation, rapidly-degradable
HES preparations in patients showing moderate-to-severe kidney
dysfunction.

Accumulation, storage, and pruritus

Accumulation

The degree of accumulation is highly dependent on the type of HES


used for volume replacement. HES with a high DS is associated with
considerable accumulation, particularly after multiple dosing (Figs 7.3
and 7.4). For HES with a high DS caution is required over the total
dose administered.

58
Intravenous Fluid Therapy Chapter 7

Fig 7.3

Fig 7.3 Multiple dosage of


Plasma levels of HES after 6% HES 450/0.7
repetitive doses of HES- (500 ml/day for 3 days)
preparations with a high degree of
substitution (DS).

Multiple dosage 6%
HES 200/0.6
(500 ml/day for 5 days)

Fig 7.4

Fig 7.4 Multiple dosage


Plasma levels after repetitive 10% HES 200/0.5
doses of HES-preparation with a (500 ml/day for 5 days)
medium degree of substitution
(DS) but different mean molecular
weight (Mw).
Multiple dosage
6% HES 70/0.5
(500 ml/day for 5 days)

Storage
Depending on the characteristics of the HES preparation, a varying
degree of the infused HES leaves the vascular space and is taken up
by the reticulo-endothelial system (RES). The effects of storage of HES
on the RES are not well clarified although HES storage appears to be
without detrimental consequences. The new third- generation HES
(HES 130/0.4) is associated with less tissue storage than other HES
preparations (Fig 7.5).

Fig 7.5

14
Fig 7.5 Total body C-HES in % of infused dosage
%
Storage of HES 130/0.4 vs 8
HES 200/0.5. 6% HES 130/0.4
6% HES 200/0.5
6 -50%

*p<0.01
4 *
-50%
-50%
2 * -75%
*
*
0

3 10 24 52
Days after final application
59
Itching

Itching after administration of HES has been demonstrated in some


reports. Special features of HES-induced pruritus include long latency
of onset and persistence. A dose-dependent uptake of HES was first
detected in macrophages and, thereafter, in endothelial and epithelial
cells. Patients suffering from pruritus consistently showed additional
deposition of HES in small peripheral nerves. Most of these reports
originated from patients treated for sudden deafness. These patients
received a considerable amount of HES (up to 20 litres) over a long
period (10 to 20 days) - mostly HMW-HES or HES with a high DS. The
incidence of pruritus after surgery is not clearly known because it may
occur weeks or even months after administration. Occasional reports
of pruritus have been published after single use of 2 litres of HES.
However, in a questionnaire covering more than 700 patients no
increased incidence of pruritus after infusion of two different HES
preparations (LMW- and MMW-HES) in comparison with lactated
Ringer`s solution was shown.

Anaphylactic reactions

All colloidal plasma substitutes, including human albumin, present a risk


of inducing anaphylactic reactions. In a large clinical trial of
approximately 20,000 patients it was demonstrated that HES-
preparations were associated with a very small incidence of
anaphylactic reactions.

Conclusions
Because of the important physico-chemical differences between the
varying HES preparations, it appears inappropriate to summarise all
HES solutions in a "HES group". The ideal volume replacement
regimen should have beneficial effects on organ blood flow and
microcirculation and should also be free of negative side-effects. Thus
merits and limitations of each HES preparation have to be considered
carefully including the effects on haemostasis, renal function, and on
other organ systems. The first-generation HES solutions with a high
MWw and a high DS (hetastarch) have unfavourable physico-chemical
characteristics. Whether changing the solvent of HMW-HES is able to
eliminate the negative effects of this plasma substitute remains to be
confirmed.

No specific recommendation can be made at present with regard to


outcome regarding the principle "crystalloids versus colloids" for
volume replacement. This also applies to the different HES solutions.
Nor can recommendations be made with respect to the different HES
solutions. Most studies with HES were not primarily focused on
outcome, outcome was not even shown in several studies. The
different physico-chemical properties of the various HES solutions have
mostly been neglected in the different meta-analyses. Using modern
(second and especially third-generation) HES specifications we have
very effective measures to treat hypovolaemia. There are still open
questions when using them in certain kind of patients, such as young
children or burns victims.

60
Intravenous Fluid Therapy Chapter 7

Bibliography

Taylor RJ, Pearl R. Crystalloid versus colloid: all colloids are not created
equal. Anesth Analg 1996;83:209-212.
Funk W, Baldinger V. Microcirculatory perfusion during volume therapy.
Anesthesiology 1995;82:975-982.
Boldt J, Müller M, Mentges D, et al. Volume therapy in the critically ill:
is there a difference? Intensive Care Med 1998;24:28-36.
Schortgen F, Lacherade JC, Bruneel F, et al. Effects of
hydroxyethylstarch and gelatin on renal function in severe sepsis: a
multicentre randomised study. Lancet 2001;357:911-916.
Bothner U, Georgieff M, Vogt NH. Assessment of the safety and
tolerance of 6% hydroxyethyl starch (200/0.5) solution: a randomized,
controlled epidemiology study. Anesth Analg 1998;86:850-856.

61
62
Intravenous Fluid Therapy Chapter 8

Properties and use of artificial


oxygen carriers
Donat Spahn and Roman Kocian

Introduction
Artificial oxygen carriers aim at replacing the oxygen carrying capacity
of the blood. Other functions, such as blood coagulation and immune
response are not replaced by these agents. Artificial oxygen carriers
thus may be used as an alternative to allogenic blood transfusions or to
improve tissue oxygenation and function of organs with marginal
oxygen supply. Modified haemoglobin solutions and perfluorocarbon
emulsions have undergone extensive clinical testing. In addition,
haemoglobin containing liposomes and allosteric modification of
haemoglobin based oxygen delivery may represent future options to
decrease the need for allogenic blood transfusions. Data of
approximately 500 - 1000 patients dosed with these compounds have
been published but there is still a significant amount of non-published
data which renders the overall assessment difficult.

Artificial oxygen carriers are grouped into:

• modified haemoglobin solutions


• perfluorocarbon (PFC) emulsions

The haemoglobin may originate from outdated human blood, be of


bovine origin or be genetically engineered. The native human
haemoglobin molecule needs to be modified in order to decrease
oxygen affinity in order to improve tissue oxygen off-loading and to
prevent rapid dissociation of the native #2-$2 tetramer into #-$ dimers.

The oxygen transport characteristics of modified haemoglobin solutions


and PFC emulsions are fundamentally different. Most haemoglobin
solutions exhibit a sigmoidal oxygen dissociation curve similar to blood.
In contrast, the PFC emulsions are characterised by a linear relationship
between oxygen partial pressure and oxygen content. Most
haemoglobin solutions thus provide oxygen transport and unloading
capacity relatively similar to blood. This means that at a relatively low
arterial oxygen partial pressure substantial amounts of oxygen are
being transported. In contrast, relatively high arterial oxygen partial
pressures are necessary to maximise the oxygen transport capacity of
PFC emulsions. Therefore, PFC emulsions are generally used in
conjunction with pure oxygen ventilation.

Haemoglobin solutions
Efficacy of haemoglobin solutions to transport and unload oxygen has
been shown in a variety of animal models including extreme
haemodilution. In a rat model of haemorrhage and surgical trauma, it
has been demonstrated that treatment with #-# Diaspirin cross-linked
haemoglobin improved wound healing, enhanced hepatic cell
proliferation and decreased splanchnic bacterial translocation when
compared with transfusion of fresh autologous blood. In septic,
oxygen supply dependent rats #-# Diaspirin cross-linked haemoglobin
increased oxygen uptake similarly to transfusion of fresh (<6 days old)
63 red blood cells whereas animals treated by stored red blood cells
exhibited a high mortality. Furthermore, #-# Diaspirin cross-linked
haemoglobin enabled extreme, virtually red blood cell free,
haemodilution in pigs with absence of subendocardial ischaemia at a
haematocrit of 1%. In a similar model but with a critical coronary
stenosis, pigs resuscitated with #-# Diaspirin cross-linked haemoglobin
survived experimental haemorrhagic shock more frequently than
animals resuscitated with albumin.

Modified haemoglobin solutions therefore improve oxygen transport


and tissue oxygenation. Without the need for cross matching, these
solutions may be alternatives to allogenic blood transfusions and in pre-
hospital resuscitation of trauma victims or in specific situations in
intensive care medicine.

Since the breakdown of the native #2-$2 haemoglobin tetramer into


#-$-dimers is largely prevented by genetic or chemical modification,
nephrotoxicity is no longer a potential side effect of these solutions.
However, there are still side effects such as vasoconstriction resulting in
an increase in systemic and pulmonary artery pressures which has
been observed with most modified haemoglobin solutions. The
mechanisms involved include nitric oxide scavenging, endothelin
release and a sensitisation of peripheral #-adrenergic receptors.
Distinct vasoconstriction is viewed as a major limitation in the
development of haemoglobin based oxygen carriers since any increase
in blood pressure may aggravate blood loss in trauma victims and
compromise survival. Indeed, a study in trauma victims was terminated
prematurely due to an increased mortality in patients treated with #-#
Diaspirin cross-linked haemoglobin, and the development of this
haemoglobin solution was stopped.

Allogenic blood transfusion may be reduced with the use of #-#


Diaspirin cross-linked haemoglobin in patients undergoing cardiac
surgery. In a prospective randomized multicentre study 209 patients
were allocated to receive either packed allogenic red blood cells or up
to 750 ml of a 10% #-# Diaspirin cross-linked haemoglobin solution
when reaching a defined transfusion trigger following cardiopulmonary
bypass. In the #-# Diaspirin cross-linked haemoglobin group, 59% of
patients avoided allogenic blood transfusions until the first post-
operative day while all patients randomised to the control group had
received packed allogenic red blood cells. At hospital discharge, 19%
of patients in the #-# Diaspirin cross-linked haemoglobin group still
avoided any allogenic transfusion as compared to none in the control
group. In emergency surgery the amount of allogenic blood
transfusions was reduced by the use of a haemoglobin solution as well.

The results of several phase III trials have been published recently.
O-raffinose cross-linked human haemoglobin in conjunction with
intraoperative autologous donation reduced the need for allogenic
blood transfusions in 299 patients undergoing coronary artery bypass
surgery. Reported side effects included a 10% elevation of arterial
blood pressure, a higher incidence of episodes of hypertension and a
transient elevation of bilirubin due to haemoglobin metabolism. In 693
patients undergoing major orthopaedic surgery, HBOC-201, a bovine
derived haemoglobin based oxygen carrier, increased the percentage of
patients avoiding any allogenic blood transfusion from 0% (patients
were randomised at the first perioperative transfusion decision) to 59%
for the entire study period of 6 weeks. In this preliminary report the
analysis of adverse and serious adverse events were not reported,
rendering safety assessment difficult at present time. However, the
side effect profile of HBOC-201 was published recently. In this study 64
Intravenous Fluid Therapy Chapter 8

escalating doses of 0.6 to 2.5 grams per kilogram of HBOC-201


(corresponding to infusion volumes of 380±87 to 1384±309 millilitres)
were given to 42 patients and data were compared to 26 control
patients receiving lactated Ringer’s solution. Blood pressure was
slightly but significantly higher in HBOC-201 dosed patients, a trend
towards elevated postoperative lipase levels and late
methaemoglobinaemia (peak at the 3rd postoperative day) were
observed and IgG-antiHBOC-201 antibody was detected at follow-up.
In contrast, renal function, platelet count, blood coagulation indices and
general clinical laboratory values were similar in both groups.

The use of HBOC-201 also allowed a decrease in transfusion exposure


in 98 patients undergoing cardiac surgery by 34%. However, a lower
haematocrit was observed in the HBOC-201 group during the first 3
postoperative days. It is, therefore, not entirely clear whether the use
of HBOC-201 or simply a more restrictive transfusion regimen in the
early postoperative period resulted in a lower transfusion exposure.

A very interesting study focused on 171 patients with blood loss from
trauma or surgery. In a dose escalation study, patients received up to
20 units of human polymerised haemoglobin (one unit is 50 grams
haemoglobin in 500 millilitres). The nadir red blood cell haemoglobin
was less than 3 grams per decilitre (mean 1.5±0.7 grams per decilitre)
in 40 patients but their total haemoglobin was still maintained at
6.8±1.2 grams per decilitre. In these 40 patients mortality was 25%
(10 of 40) as compared with 65% (20 of 31) in an historic control
group.

PFC emulsions
PFCs are carbon-fluorine compounds characterised by a high gas
dissolving capacity (oxygen, carbon dioxide and other gases), low
viscosity, and chemical and biological inertness. PFCs are virtually
immiscible with water. Only manufacturing an emulsion with specific
characteristics (size of droplets of approximately 0.16 mm diameter)
rendered them biocompatible with only few side effects. With the
development of a stable 60% (58% perfluorooctyl bromide and 2%
perfluorodecyl bromide) emulsion there is now a relatively highly
concentrated emulsion which is clinically well tolerated.

After intravenous infusion, the emulsion is taken up by the


reticuloendothelial system. This uptake into the reticuloendothelial
system determines intravascular half life which is dose dependent and
approximately 10 hours after a 1.8 grams per kilogram PFC emulsion
dose. After the initial uptake of the PFC emulsion into the
reticuloendothelial system, the droplets of the emulsion are slowly
broken down, the PFC molecules are taken up in the blood again
(bound to blood lipids) and transported to the lungs, where the
unaltered PFC molecules are finally excreted via exhalation. At present
time, no metabolism of PFC molecules is known in humans.

PFC emulsion was assessed following acute normovolaemic


haemodilution at a haematocrit of 10 % and a massive rise in mixed
venous oxygen partial pressure and mixed venous saturation was
observed. The percentage of metabolised oxygen originating from
endogenous haemoglobin decreased with the infusion of PFC
emulsion, indicating that the oxygen transported by PFC emulsion is
preferentially metabolised, most likely due to the excellent oxygen
unloading characteristics.
65
Furthermore, Holman et al. tested PFC emulsion in severely
haemodiluted dogs undergoing cardiopulmonary bypass. Without
using catecholamines, dogs treated with increasing doses of PFC
emulsion survived cardiopulmonary bypass progressively better than
control animals.

In addition, Habler at al. found that mixed-venous oxygen partial


pressure was higher in PFC emulsion treated animals after
haemodilution to a haemoglobin of 7 grams per decilitre than in control
animals. Measures of left ventricular contractility were found to be
improved after PFC emulsion administration at a haemoglobin level of 3
grams per decilitre. This might be explained by an augmented oxygen
delivery through very narrow capillaries where PFC emulsion particles
(<0.2 mm in diameter) may penetrate better than the relatively large red
blood cell (7-8 mm in diameter) and thereby increase local tissue
oxygenation including in the myocardium.

PFC emulsion has also been used in humans. In one of the


prospective randomised multicentre studies, patients were
haemodiluted preoperatively to a haemoglobin level of 9 grams per
decilitre. After the patients had reached a predefined transfusion
trigger, they were randomized into 4 groups: (A) standard of care
(retransfusion of 450 ml of autologous blood at an unchanged FiO2 of
0.4), (B, C), PFC emulsion (0.9 or 1.8 grams per kilogram) with colloid
to a total 450 millilitre and ventilation with an FiO2 of 1.0 and (D),
infusion of 450 millilitre of colloid with ventilation with an FiO2 of 1.0.
PFC emulsion (1.8 grams per kilogram) was most successful in
reversing transfusion triggers in 97% of patients as compared to 60%
in the control group. Transfusion trigger reversal lasted longer (80 min)
in the PFC emulsion 1.8 g/kg group than in the control (55 minute) and
colloid (30 minute) groups. Thus, physiological transfusion triggers may
be treated with PFC emulsion. This illustrates the remarkable potency
of PFC emulsion to deliver readily available oxygen to those areas in the
body where additional oxygen is most needed.

PFC emulsions have side effects. Volunteers experienced mild flu-like


symptoms with myalgia and light fever and an approximately 15%
decrease in platelet count 3 days post-dosing returning to normal by
day 7. Traditional coagulation tests, bleeding time and platelet
aggregation, however, were unaffected by PFC emulsion. Finally,
enrolment in a phase III study in cardiac surgery was voluntarily
suspended in 2001 due to an apparent imbalance in adverse
neurological outcome. Experts, however, agree that these events were
not directly related to the PFC emulsion itself but rather to the rapid
blood harvesting procedure early in cardiopulmonary bypass. In fact, in
the same study it was observed in a subset of patients monitored with
gastric tonometry that gastric mucosal pH was higher in PFC treated
patients resulting in earlier postoperative bowel movement.

Augmented-acute normovolaemic haemodilution with PFC emulsion


has recently been shown to reduce the need for allogenic blood
transfusion in 492 patients undergoing major non-cardiac surgery.
PFC-treated patients first underwent acute normovolaemic
haemodilution to a haemoglobin of 8.0 grams per decilitre, followed by
dosing with PFC emulsion. In the intent-to-treat population transfusion
avoidance at 24h post-operatively was higher than in the control group
(53% versus 43%) and less allogenic blood was transfused (median 0
versus 1 unit). In the protocol defined subgroup of patients with major
blood loss (> 20 millilitres per kilogram), transfusion avoidance at 24
hours post-operatively was higher (32% versus 13%) and less blood
66
Intravenous Fluid Therapy Chapter 8

was transfused (median 2 versus 4 units). These transfusion benefits


remained until hospital discharge. As in previous studies, platelet
counts decreased approximately 20% more postoperatively in PFC
treated patients but this neither resulted in exaggerated post-operative
bleeding nor in more postoperative blood transfusions. More serious
adverse effects were observed in the PFC group (32% versus 21%) but
no organ system was particularly involved. This difference may be
explained by the fact that several participating centres did have a
limited experience with acute normovolaemic haemodilution and low
intra-operative haemoglobin levels. In addition, a trend towards under-
dosing of crystalloids and colloids in the PFC group was observed with
consequent hypovolaemia which, in itself, may have resulted in more
post-operative complications.

Haemoglobin containing Liposomes


The development of haemoglobin containing liposomes, also referred to
as "neo red cells", is still in the pre-clinical stage. Haemoglobin
containing liposomes may have a number of theoretical advantages as
compared to haemoglobin solutions and PFC emulsions: a relatively
long intra-vascular half life, a lesser pressor effect and optional co-
encapsulaton of 2,3-DPG and methaemoglobin reductase within the
liposomes to optimize the oxidation status of the haemoglobin and its
oxygen affinity. Although fascinating, the development of haemoglobin
containing liposomes may prove complex in its final realisation.

Allosteric Modifier
RSR13, (chemical formula C20H22NO4Na) modifies haemoglobin
oxygen affinity, thereby mimicking the effect of 2,3-diphosphoglycerate.
Therefore, the oxygen dissociation curve is shifted to the right and
oxygen off-loading in the tissue is enhanced. RSR13 was initially
developed to augment tumour oxygenation in order to achieve radio-
sensitisation. However, it is also conceivable to view such a drug as
being used to enhance oxygen off-loading and thereby avoiding
allogenic blood transfusions.

Future Uses of Artificial Oxygen Carriers


Future use of haemoglobin solutions and PFC emulsions may include a
combination of acute normovolaemic haemodilution with infusion of an
artificial oxygen carrier during surgery, a procedure termed augmented-
acute normovolaemic haemodilution. Augmented-acute
normovolaemic haemodilution is a concept in which patients undergo
acute normovolaemic haemodilution to relatively low haematocrit levels
prior to surgical blood loss. Acute normovolaemic haemodilution may
thus be performed preoperatively or during the first part of surgery prior
to the phase of major blood loss. In the phase of major surgical blood
loss, colloids and crystalloids are administered to avoid hypovolaemia
and artificial oxygen carriers are co-administered to maintain tissue
oxygenation. As a consequence, lower haematocrit levels can be
tolerated safely. Towards the end of surgery, the autologous blood
harvested during acute normovolaemic haemodilution will be re-
transfused. This will increase postoperative haematocrit levels and
oxygen delivery will again be provided by endogenous red blood cells.
Therefore, greatly elevated arterial PO2 values are not necessary in the
postoperative period and the relatively short half life of all artificial
oxygen carriers (< 24 hours) will not compromise their success in
67 reducing perioperative allogenic blood transfusion requirement.
Outlook
Progress in the development of artificial oxygen carriers has been
achieved in recent years but no artificial oxygen carrier has achieved
market approval yet in the US, Canada or Europe. Achieving market
approval is obviously the first goal. Today, we should think of the
necessary education of healthcare professionals to familiarise them with
these new concepts, physiological aspects and therapeutic
compounds. Only in the hands of the experienced physicians will
artificial oxygen carriers be used to the benefit of patients.

Bibliography

Winslow RM. Blood substitutes. Adv Drug Deliv Revs 2000;40:131-


142.
Riess JG. Oxygen carriers ("blood substitutes")--raison d'etre,
chemistry, and some physiology. Chem Rev 2001;101:2797-920.
Stowell CP, Levin J, Spiess BD, Winslow RM. Progress in the
development of RBC substitutes. Transfusion 2001;41:287-299.
Gould SA, Moore EE, Hoyt DB, et al. The life-sustaining capacity of
human polymerized hemoglobin when red cells might be unavailable. J
Am Coll Surg 2002;195:445-452.
Spahn DR, Waschke K, Standl T, et al. Use of perflubron emulsion to
decrease allogenic blood transfusion in high-blood loss non-cardiac
surgery: results of a European phase 3 study. Anesthesiology
2002;97:1338-1349.

68
Intravenous Fluid Therapy Chapter 9

Clinical problems arising out of


intravenous fluid therapy
Edward Burdett and Monty Mythen

Introduction
Clinical problems arising out of intravenous fluid therapy can broadly be
categorised into side effects of the fluids themselves or adverse events
related to the way the fluids are used. The side effects can further be
categorised as predictable (i.e. dose dependent) and unpredictable or
idiosyncratic. Adverse effects may be the inevitable consequence of
poor fluid choice (e.g. hyponatraemia as a result of injudicious use of
5% glucose). However, we will focus on the adverse events that may
arise despite choosing the right fluid for the job. We accept that the
edges between the two are blurred. Most of these side effects will
relate to colloids. One exception is the relative contribution of the
formulation of the carrier solution as the issues here apply equally well
to both colloids and crystalloids.

Crystalloids
Electrolyte and acid-base disturbance

Normal saline (0.9% sodium chloride solution) and colloids suspended


in normal saline are often infused because they are easily available and
are isotonic with plasma.

Saline-based fluids are non-physiological in three ways. Firstly, the level


of chloride is significantly above that of plasma; secondly, they lack
several electrolytes present in normal plasma, including potassium,
calcium, glucose, and magnesium; thirdly, they lack the bicarbonate or
bicarbonate precursor buffer necessary to maintain plasma pH within
normal limits. Each of these may be responsible for homeostatic
disruption.

As seen in Table 9.1, there are significant differences in the electrolyte


content of 0.9% sodium chloride (saline, saline based fluids and most
colloid preparations) compared to balanced electrolyte fluid
preparations (e.g. Hartmann’s solution, Lactated Ringer’s) and human
plasma.

Table 9.1: Relative concentrations of two commonly used crystalloid solutions, and normal human plasma.
All values quoted are electrolyte concentration in mmol/l

Electrolyte Hartmann’s Solution 0.9% NaCl Plasma


Concentration
Sodium 131 154 135 - 146

Chloride 111 154 100 - 106

Potassium 5 0 3.5 - 5.0

Calcium 2 0 2.20 - 2.67

Magnesium 0 0 0.7 - 1.1

Phosphate 0 0 0.8 - 1.5

Bicarbonate/lactate 29 0 22 - 30

69
Hyperchloraemic metabolic acidosis

Normal saline infusion causes metabolic acidosis. The term "dilutional


acidosis" has been used to describe this effect implying that plasma
expansion and dilutional reduction of plasma bicarbonate are the
underlying mechanisms. This is not the whole story. Recent work
emphasises the importance of hyperchloraemia in the aetiology of this
metabolic acidosis.

How does a change in chloride concentration bring about an alteration


in acid-base equilibrium? The answer can be explained by Stewart's
method of analysis of quantitative acid and base chemistry. Stewart
introduces the concept of the strong ion difference (SID). This principle
rests on the chemistry of electrolytes in the body, which are almost
completely ionised in aqueous solutions. The most notable of these
are Na+, K+, Ca2+, Mg2+, Cl-, and Lactate-.

In plasma, when all the positive and negative charges of strong ions
mentioned above are added together, the result does not add up to
zero. This accounts for the strong ion difference. Weak acids, for
example, have not been taken into account in the equation. Stewart
originally described the equation as follows:

(Na+ + K+) - (Cl- + Lactate-) = SID

The SID can be affected by changes in these electrolyte concentrations


in the plasma. It exerts enough electrochemical force to determine
what the plasma hydrogen ion concentration will be. In other words,
the forces generated by this SID determine the amount of H+ ions
required to "balance" plasma charges. A decrease in the SID will
create an increase in the H+ concentration. An increased plasma
chloride ion concentration will reduce the SID, increasing the hydrogen
ion concentration, and causing acidosis.

The clinical implications of hyperchloraemic metabolic acidosis are not


clear. We know that metabolic acidosis, whatever the aetiology, can
depress myocardial function, reduce cardiac output, and reduce renal
and intestinal perfusion. Acidaemia can inactivate membrane calcium
channels, and inhibit the release of norepinephrine from sympathetic
nerve fibres. This may result in the redistribution of cardiac output
away from internal organs.

While this may have little effect in fit patients undergoing minor surgery,
the concern is the effect of severe hyperchloraemic acidosis from
aggressive fluid resuscitation in acutely ill patients during major surgery,
or following trauma or burns. For example, after tourniquet release,
lactate and carbonic acid load may be superimposed on the iatrogenic
hyperchloraemic acidosis towards the end of the procedure.

In addition, there is the concern of incorrect diagnosis of the acidaemia.


If an intra-operative metabolic acidosis becomes apparent during fluid
replacement, a hyperchloraemic metabolic acidosis may not be
differentiated from lactic acidosis by the inexperienced clinician. The
clinician may be misled into believing that the patient is hypovolaemic,
or has a surgical cause for their acidosis. Further administration of
non-balanced fluids will exacerbate rather than relieve the problem.

Hyperchloraemic metabolic acidosis affects a variety of homeostatic


mechanisms. Coagulation, as with any other physiological system, has
optimal pH and electrolyte levels at which it functions most effectively.
70
Intravenous Fluid Therapy Chapter 9

Systematic review and meta-analysis of the available data of trials


investigating balanced electrolyte versus saline-based fluids has shown
a significant reduction in blood loss in patients treated with balanced
fluids.

Animal studies suggest that hyperchloraemia causes renal


vasoconstriction, and a decrease in glomerular filtration rate and renal
blood flow that is related to a tubular reabsorption mechanism involving
chloride. A number of clinical studies have now shown that renal
indices such as creatinine clearance and urine output are perturbed by
normal saline infusion in the peri-operative setting. Trials in this area
are small, and many are not powered for patient outcome, but the data
reinforce the animal models; hyperchloraemic metabolic acidosis
reduces renal function.

Other electrolyte abnormalities

Over-infusion of normal saline will dilute electrolytes normally present in


small quantities in plasma. Calcium, for example, is required for
myocardial and cerebral stability and haemostasis. In-vitro data have
shown that normal saline infusion can cause impairment of whole blood
coagulation by virtue of hypocalcaemia.

Conversely, some intravenous fluid preparations contain supra-


physiological levels of calcium. There is concern that these fluids may
cause coagulation of citrated blood if mixed, and their use through the
same port as blood products is contra-indicated.

Intravenous fluids that are not balanced in their composition affect


other electrolytes present in normal plasma. Some of these
electrolytes, such as potassium and magnesium, are tightly controlled
physiologically. Derangement of their concentrations in the plasma
could have serious adverse effects, such as the predisposition to
arrhythmias and convulsions.

Balanced electrolyte formulations contain a buffer, most often a


bicarbonate precursor, which is converted into bicarbonate in the liver.
This buffer is substituted instead of chloride in these fluids, thereby
reducing acidosis by two strategies. Since bicarbonate is not stable in
standard isotonic crystalloid formulations, it is more practical to use
lactate or acetate. In certain circumstances, however, such as in
severe hepatic impairment, lactate is not broken down correctly, and
can build up. Paradoxically, therefore, buffered fluids can under certain
circumstances increase the lactate load, and create a lactataemia. This
will only worsen the clinical effects outlined above, and contribute to
difficulty in diagnosing acidosis.

Colloids

Predictable (dose related) effects

ABO serum grouping can be compromised due to the action of dextran


on inducing rouleaux formation. This can produce problems in
detecting weaker antibodies when cross matching. Clinicians are
advised to take blood samples for cross match before starting a
dextran infusion. Several laboratory tests may also be compromised by
concurrent use of dextran. Troponin I titres and glucose measurement
may be considered inaccurate in the presence of dextran.
71
Coagulopathy

As mentioned above, intravenous fluids can adversely affect


coagulation by inducing acidosis, and by altering plasma
concentrations of electrolytes. Colloid molecules themselves have also
been shown to reduce haemostasis. Dextrans interfere with clotting by
reducing blood viscosity and by lowering platelet and erythrocyte
aggregation. They also negatively influence haemostatis by reducing
von Willebrand factor. This may be beneficial in reducing deep venous
thrombosis and micro-anastomosis failure post-operatively, but carries
the increased risk of bleeding following trauma and orthopaedic
surgery. Hydroxyethyl starch has similar properties.

Unpredictable or Idiosyncratic effects

Itching

In several recent studies and case reports, dermatologists have


highlighted a problem peculiar to hydroxyethyl starches. The starch
molecules are absorbed and deposited by phagocytes in various
organs, including skin, where they remain long-term.

Up to a third of patients receiving solutions containing HES


(hydroxyethyl starch) will suffer pruritis. The itching can be persistent,
refractory to treatment, and potentially debilitating. The pruritis tends to
be generalised (although may be localised to the anogenital region and
trunk), unrelated to atopy or age, and often in the absence of
identifiable skin lesions. A histamine-independent pathway for the
itching is assumed because of the ineffectiveness of antihistamines in
treatment.

In the largest study of HES-related pruritis, the incidence of this


adverse effect was related both to the type of HES (higher
concentration and molecular weight caused greater incidence) and the
total dose received. In this study, 32% of patients developed pruritis
with a mean duration of nearly 9 weeks. Most resolved by 50 weeks
although, in some the itching continued over 2 years later.

Anaphylaxis and anaphylactoid reactions

These rare but potentially fatal consequences of any drug


administration occur with some colloid groups. Although uncommon
across all colloid groups, adverse reactions occur more frequently with
gelatins, ranging from Grade I (skin manifestations such as flush and
urticaria) to Grade IV (cardiac and/or respiratory arrest).

An anaphylactic reaction is an exaggerated response to a substance to


which an individual has become sensitised in which histamine,
serotonin, tryptase and other vasoactive substances are released in an
immunoglobulin E (IgE) mediated response. Anaphylactoid reactions are
clinically indistinguishable from anaphylaxis but are not IgE mediated.

In a large, prospective study of nearly 20,000 patients in France


between 1991 and 1992, the frequency of severe reactions (Grade III &
IV) to colloid was calculated as 0.2%. The breakdown showed a
frequency of 0.35% for gelatins, 0.27% for dextrans and 0.06% for
HES. Multivariate analysis revealed four distinct risk factors: giving
gelatins, giving dextrans, being male, and a history of drug allergy.

72
Intravenous Fluid Therapy Chapter 9

The underlying mechanisms for these are clear for gelatins (histamine
release) and dextrans (antibody mediated), but the mechanisms for
HES reactions are less clear. Antibodies to HES do exist but do not
seem to be clinically relevant. HES is the least antigenic of the plasma
expanders and this may be in part due to its manufacture from
amylopectin, a substance very similar to glycogen.

Despite the differences between the colloid groups it should be noted


that anaphylactic and anaphylactoid reactions to colloids are rare.

Side effects from mode of administration

Volume overload: overdose of intravenous fluids

Cardiac output is the volume of blood pumped by the left ventricle into
the aorta each minute and is possibly the most important factor when
considering the circulation, for it is responsible for delivering and
removing blood from the tissues; on average the cardiac output is
nearly 5 litres a minute.

The function of the heart as a pump was described by Starling in ‘The


Linacre Lecture on the Law of the Heart’ in 1918:

• From the after-load side: the more the heart muscle is stretched, the
greater the force of contraction and hence the greater the volume of
blood is pumped into the aorta.

• From the pre-load side: within physiological limits, the heart pumps
all blood that comes into it without allowing excessive damming in
the veins.

At the cellular filament level, a larger filling pressure increases the force
of contraction.

Cardiac output cannot increase indefinitely with increasing


requirements. In a number of situations cardiac output can actually
reduce with increasing requirements.

Increasing atrial pressure, usually by intravenous fluid, will increase


cardiac output up to a point. Beyond this, cardiac output will
decrease, resulting in reduced arterial pressure and worsening tissue
hypoxia due to pump failure. In addition, oedema as a result of fluid
excess compresses capillaries reducing the perfusion of tissues.

Contamination

In 1996 a hospital in Brazil reported 35 newborn infants had died due


to intravenous fluid contaminated with endotoxins, produced by a local
company later closed for inadequate quality-control. There are other
case reports of contaminated intravenous fluid, but in the Western
hemisphere at least, this contamination tends to be due to migration
through air vents, dirty drip sets or additives. Contamination is
appreciably more common with high concentration glucose solutions.

73
Technical problems

As with any drug prescribed, physicians need to obey the maxim "First,
do no harm…..". The technical problems associated with fluid
administration are common, serious, and often overlooked.

Peripheral cannula insertion

The adverse effects of cannula insertion need to be balanced against


the therapeutic effects.

Thrombophlebitis can manifest as pain, swelling and redness, and


occurs due to several factors, including the nature of the fluid being
administered, the patient’s condition, and the cannula (size, length,
duration). Extravasation can occur when a peripheral cannula exits the
vein and fluid is forced into surrounding tissues. A pneumatic pressure
bag may compound the problem. This may present difficulties intra-
operatively in knowing how much of a drug has been administered via
a peripheral cannula which is often inaccessible to the anaesthetist.

As with all intravenous access, thrombosis of the vein can occur, with a
chance of embolus formation. Contamination of the site or specific
patient circumstances (e.g. diabetes mellitus) may result in infection
and abscess formation.

Invasive monitoring

Goal-directed or fluid therapy needs accurate and time-sensitive


monitoring to assess the management rationale. The use of central
venous, pulmonary artery and arterial monitoring is common practice to
optimise the acutely ill patient. Invasive monitoring has its own risks
and benefits; adverse effects are a significant factor in iatrogenic
morbidity and mortality.

Potential complications of inserting central venous catheters are


pneumothorax (in 5% of subclavian vein insertions), and arterial
puncture (commonest with the internal jugular approach at up to 15%
incidence). The most serious complication with pulmonary artery
catheters is perforation of the pulmonary artery.

Nosocomial infections via intravenous and intra-arterial catheters are


an enormous problem with significantly higher rates of infection for
femoral catheters: nosocomial sepsis is one of the biggest killers in the
modern intensive care unit.

Other complications include nerve injury (brachial plexus or phrenic


nerve), thrombosis, haemothorax, great vessel damage, haematoma,
air embolism, malposition, cardiac tamponade and arrhythmias. In
order to reduce the incidence of catheter or cannula related problems
clinicians must use a strict aseptic technique and, if operators are
inexperienced, consider the use of ultrasound guided methods.

74
Intravenous Fluid Therapy Chapter 9

Other problems

Intravenous fluids and oedema

The lack of colloid osmotic pressure support with crystalloids causes


them to leak very rapidly out of the vascular space into the interstitial
space. An isotonic crystalloid infusion will increase vascular volume by
only one-quarter of the volume infused. Crystalloid fluid infusion is
therefore linked to tissue oedema. This can be unsightly and
uncomfortable for the patient if dependent extremities are involved, and
will result in decreased mobility. Other, less visible tissues will become
oedematous too, and will cause adverse clinical effects. In particular,
there is now evidence that crystalloid infusion causes small bowel
oedema, and that this may be responsible for impaired gastrointestinal
function, including intolerance of enteral feeding, and post-operative
nausea and vomiting. Increased cerebral oedema may worsen
prognosis after brain surgery or head trauma.

Hypothermia

It is essential to be aware of the consequences of infusing intravenous


fluids with a temperature differential of up to 15 C from body
O

temperature, as core body temperature can be depressed after only


one to two litres of fluid. Hypothermia has profound effects on many
organ systems - it produces a coagulopathy (potentially significantly
affecting outcome in major trauma cases), is arrhythmogenic (impairing
oxygen delivery) and impairs the action of some drugs. Tachycardia
and an increased metabolic rate in the early stages of mild hypothermia
can increase oxygen consumption at a potentially critical time.
Warmed intravenous fluids are now standard practice intra-operatively
for major surgical cases.

Intravenous additives

Drug errors can occur with intravenous fluids as easily as with any
other drug administration. A recent study in the UK gave a rate of
intravenous drug errors of 49% with about a third resulting in potential
harm. Because intravenous fluids are seen as relatively innocuous,
sufficient care is often not taken over their prescription and
administration. A recent cause for concern has been the inadvertent
intravenous infusion of potassium chloride. Several cases of potassium
overdose have led to stricter controls over their use on general wards.

Conclusion
In clinical practice, and especially in the field of anaesthesia and peri-
operative medicine, the prescription and administration of intravenous
fluids can have deleterious clinical consequences. It has now been
firmly established that electrolyte disturbances, in particular
hyperchloraemic metabolic acidosis are a predicable consequence of
saline-based intravenous fluid administration.

The colloid components of some intravenous fluid products, while


having theoretical advantages, are responsible for a variety of unwanted
immunomodulatory effects. In addition, the physical and technical
consequences of all that intravenous fluid administration entails can
have considerable iatrogenic impact. While undoubtedly intravenous
fluid administration is a vital part of hospital practice, we must always
73
75 be aware of the risks involved, and consider the safest fluid to deliver.
Bibliography

Stewart PA. How to understand acid-base. Elsevier, New York, 1981.


Kellum JA. Saline-induced hyperchloraemic metabolic acidosis. Crit
Care Med 2002;30:259-261.
Wilkes NJ, Woolf R, Mutch M, et al. The effects of balanced versus
saline-based hetastarch and crystalloid solutions on acid-base and
electrolyte status and gastric mucosal perfusion in elderly surgical
patients. Anesth Analg 2001;93:811-816.
Roche AM, James MF, Mythen MG. In vitro addition of calcium to
saline-based intravenous fluids only partially compensates for TEG®
pictures observed during haemodilution. Anesth Analg 2002;94:S73.
Murphy M, Carmichael AJ, Lawler PG, et al. The incidence of
hydroxyethyl starch-associated pruritis. Br J Derm 2001;144:973-976.
Laxenaire MC, Charpentier C, Feldman L. Anaphylactoid reactions to
colloid plasma substitutes: frequency, risk factors, mechanisms. Ann Fr
Anaesth Reanim 1994;13:301-310.

74
76
Intravenous Fluid Therapy Chapter 10

The importance of monitoring


Andrew Webb

Introduction
As can be seen from the preceding chapters there is much knowledge
of the properties and limitations of the various solutions available for
intravenous fluid therapy. However, it has become clear that the
differences in properties of fluids make little difference to the outcome
of patients given them. Much more important is ensuring the right
volume of fluid is titrated appropriately. Identifying the hypovolaemic
patient requires appropriate monitoring since, in clinical shock,
inadequate tissue perfusion may be overt (with obvious clinical signs) or
covert (with few clinical signs). Both may lead to organ dysfunction
and failure. Recognising covert shock requires a high degree of
suspicion, early recognition of hypovolaemia and tools for assessment
of tissue perfusion. Adequate correction of circulation volume cannot
occur until tissue perfusion has been restored, or has reached the best
level possible with fluid alone.

Managing shock follows the basic principles of resuscitation followed


by attention to the cause. After ensuring adequate oxygenation and
appropriate haemodynamic monitoring, resuscitation requires a
functional approach to the circulation. If the circulation is thought of as
a system of conduits carrying oxygen and nutrient containing fluid
(blood) to tissues and driven by a pump (the heart) it is clear there are
four main components which may be targeted with treatment (the
blood, the heart, the blood vessels and the tissues).

Usually the circulating volume will be targeted first. Apart from


attention to the oxygen supply, correction of volume deficit is the most
critical intervention in shock and may even modulate the immune
response. Few would disagree that early administration of fluid often
corrects features of shock, such as hypotension or oliguria.
Hypovolaemia must be treated urgently and before clinical signs of
organ hypoperfusion are obvious to avoid the serious complication of
organ dysfunction.

Diagnosing hypovolaemia

The gold standard

The gold standard for the diagnosis of hypovolaemia is the


measurement of blood volume. Techniques available rely on the
principles of indicator dilution, usually involving radio-isotopes as the
indicators. Most techniques do not lend themselves to rapid, bedside
estimation and, therefore, preclude rapid intervention. Furthermore,
true estimation of blood volume requires an indicator that is detectable
before it distributes outside the circulation. The only indicator that can
currently be used in this fashion, mixing with the whole circulation
before any loss from the circulation, is radio-chromium labelled red
blood cells. The development of methods based on carbon monoxide
labelled red cells is promising since labelling and measurement can be
completed at the bedside. Even if we can develop a reliable method of
quick and accurate blood volume estimation at the bedside, there
remain significant problems of interpretation. Normal blood volume is a
77 poor indicator of physiological requirement in the shock state and is
dependent on body composition, being lower in obese patients. In
sicker patients or those with diminished cardiac reserve a higher than
normal blood volume is required to maintain the circulation.
Understanding how much blood volume is required to optimise the
circulation is dependent on measurement of the functional response of
the circulation to various blood volume levels, the blood volume itself
being of secondary importance. We therefore rely on surrogate
markers of volume status.

Clinical signs

Clinical signs of hypovolaemia (reduced skin turgor, oliguria, tachycardia


and hypotension) are late indicators. The presence of these signs
signifies hypovolaemia of a degree that requires urgent intervention
since there is a likelihood that organ perfusion is already compromised
with the risk of subsequent organ dysfunction. Since the blood
pressure is maintained by homeostatic sympathetic vasoconstriction a
low blood pressure as a result of hypovolaemia implies up to 25% loss
of blood volume. The absence of these clinical signs does not exclude
hypovolaemia. Indeed paradoxical bradycardia has been described in
severe haemorrhagic shock, polyuria may be a cause of hypovolaemia
and the presence of oedema may hide a deficit in circulating volume in
states of capillary leak. The diagnosis of lesser degrees of
hypovolaemia which require treatment for maintenance of tissue
perfusion and avoidance of the sequelae of organ dysfunction is difficult
clinically. Clinical assessment is dependent on patient position, there
being an increase in plasma volume, and therefore a minimisation of
clinical signs, associated with supine positioning. Where hypovolaemia
is suspected in a supine patient, lifting the legs and watching for an
improvement in the circulation is a useful indicator. This however,
requires that hypovolaemia has already affected the circulation and
restoration of blood volume should already have begun.

The central venous pressure

The central venous pressure (CVP) is the most popular surrogate


marker of volume status. Its popularity is based on ease of
measurement but there are a number of pitfalls. CVP is dependent on
venous return to the heart, right ventricular compliance, peripheral
venous tone and posture. Thus CVP is not only an index of right
ventricular filling pressure. Many use the CVP in a static fashion to
determine the need for fluid, often having in mind a fixed level of CVP
as a target. However, a normal CVP does not exclude hypovolaemia
and the CVP is particularly unreliable in pulmonary vascular disease,
right ventricular disease, isolated left ventricular failure and valvular
heart disease. In patients with an intact sympathetic response to
hypovolaemia the CVP may fall in response to fluid. In ventilated
patients the increase in mean intrathoracic pressure will be partially
transmitted to the central venous system thus elevating the CVP at any
one level of circulating volume. Thus we cannot aim for magic
numbers when dealing with the CVP since, like the blood volume, we
have no idea what the appropriate target is for the individual.

Pulmonary artery wedge pressure


and cardiac output

The pulmonary artery catheter became a popular technique for


measurement of pulmonary artery wedge pressure (PAWP) and cardiac
output (CO). The PAWP provides similar information regarding fluid
78
Intravenous Fluid Therapy Chapter 10

status as the CVP and is subject to similar constraints. As with CVP


the absolute level of PAWP does not confirm or exclude hypovolaemia.
Left ventricular disease may increase the level of PAWP required to
ensure an adequate circulating volume. Interpretation of PAWP
requires caution in ventilated patients. PAWP is a useful indicator of
relative hypovolaemia where CVP is high and PAWP is significantly
lower, e.g. in selective right ventricular dysfunction or chronic airflow
limitation where the CVP is unreliable.

The pulmonary artery catheter allows measurement of cardiac output.


The original techniques was based on the indicator dilution method
using a thermal indicator. The thermal indicator was a bolus of cool
(compared to blood temperature) fluid. Some modern catheters
equipped with fast response thermistors and thermal filaments allow
detection of a heat pulse rather than the change in temperature
induced by a cooled fluid bolus. Such catheters allow semi-continuous
monitoring of cardiac output. Since the purpose of fluid resuscitation is
to provide optimal blood flow for tissue perfusion, measurement of
PAWP without assessment of CO cannot be justified. As with other
surrogate markers of volume status knowledge of the absolute CO
does not confirm or refute hypovolaemia. More important is the
response of CO to fluid therapy and whether CO is adequate depends
on whether tissue perfusion is adequate.

Pulmonary artery catheterisation is known to be associated with


significant complications and has even been suggested as a cause of
increased mortality. CO can now be measured by non-invasive
techniques such as thoracic bioimpedence and oesophageal Doppler
ultrasound. Both techniques have been used to assess body fluid
status.

Tissue perfusion

Irrespective of the volume status adequate tissue perfusion does not


require treatment. Global assessment of tissue perfusion is based on
demonstration of the absence of anaerobic metabolism, i.e. no lactic or
metabolic acidosis. The presence of lactic acidosis does not
necessarily indicate an inadequate circulation, e.g. liver dysfunction,
and the absence of lactic acidosis does not guarantee adequate
perfusion of all tissues.

The gut mucosa is one of the earliest tissues to be compromised in


hypovolaemia and tonometry provides a simple, non-invasive method
of assessment of the adequacy of perfusion. Refinement of the
method of gastric tonometry has produced a semi-continuous,
automated system. Although the technique may be used to monitor
tissue perfusion, correction of which may be the ultimate goal of fluid
resuscitation, it must be remembered that alternative treatments may
be required to correct tissue perfusion. Global haemodynamic
assessment in response to fluid therapy is still required, although the
tonometer may help with the difficult diagnosis of otherwise covert
hypovolaemia.

The fluid challenge

The fluid challenge is a method of safely restoring circulating volume.


Rather than using fixed haemodynamic endpoints fluid is given in small
aliquots to produce a known increment in circulating volume with
assessment of the dynamic haemodynamic response to each aliquot.
79 No fixed haemodynamic endpoint is assumed and the technique
provides a diagnostic test of hypovolaemia (via an appropriate positive
response of the circulation to fluid) and a method of titrating the optimal
dose of fluid to the individual’s requirement.

Assessing the response to a fluid challenge

The response of CVP or stroke volume and PAWP, should be


monitored during a fluid challenge. Fluid challenges should be
repeated while the response suggests continuing hypovolaemia.

CVP or PAWP response

The change in CVP or PAWP after a 200 millilitre increment in


circulating volume depends on the starting circulating volume (Fig
10.1).
Fig 10.1

Fig 10.1.
Stroke volume PAWP/CVP The response of Stroke Volume,
CVP or PAWP to a 200 ml
increment of blood volume. In the
hypovolaemic patient no
Underfilled
Well Filled
significant rise in CVP or PAWP
Overfilled would be expected but an
3mmHg
increase in stroke volume would
be expected. In the optimally
<3mmHg filled patient a rise in CVP or
Blood Volume Blood volume PAWP with no significant rise in
stroke volume would be
expected.

A 3mmHg rise in CVP or PAWP represents a significant increase and is


probably indicative of an adequate circulating volume. It is important to
assess the clinical response and the adequacy of tissue perfusion in
addition; if inadequate it is appropriate to monitor stroke volume before
further fluid challenges or considering further circulatory support.

Stroke volume response

In the inadequately filled left ventricle a fluid challenge will increase the
stroke volume (Fig 10.1).

Failure to increase the stroke volume with a fluid challenge may


represent an inadequate challenge, particularly if the PAWP or CVP fails
to rise significantly (by at least 3mmHg). This indicates that cardiac
filling was inadequate as the increment in circulating volume fills the
depleted peripheral vascular space. In this case the fluid challenge
should be repeated. It is important to monitor stroke volume rather
than cardiac output during a fluid challenge. If the heart rate falls
appropriately in response to a fluid challenge the cardiac output may
not increase despite an increase in stroke volume.

Choice of fluid for a fluid challenge

An immediate transudation to the interstitial space of three quarters of


the volume of saline infused implies that oedema is a necessary side 80
Intravenous Fluid Therapy Chapter 10

effect of any increase in intravascular volume and therefore any


correction of the shock state with crystalloid resuscitation. There is no
difference in the process of oedema formation whether the cause is an
increase in capillary hydrostatic pressure or capillary leak. In the lung
this process follows the sequence of interstitial fluid accumulation,
followed by distension of the alveolar capillary barrier and finally alveolar
flooding. Thus, it is the intravascular Starling forces that are most
relevant and any increase in capillary hydrostatic pressure through
resuscitation will contribute to an increase in interstitial volume.

Recently, the crystalloid-colloid controversy has been fuelled by several


meta-analyses demonstrating no benefit to mortality with colloid
resuscitation. It is important to note that the research on which these
meta-analyses were based did not titrate fluid requirements by the
dynamic fluid challenge method. In many cases titration was not to a
haemodynamic endpoint at all. The aim of a fluid challenge is to
produce a small but significant (200 millilitre) and rapid increase in
circulating volume. Colloid fluids are ideal in that they give a reliable
increase in plasma volume. Crystalloids are rapidly lost from the
circulation and larger volumes would be required to achieve an
increase of 200 millilitre in circulating volume.

A variety of colloid fluids are available for fluid resuscitation but none
has been demonstrated to be superior to others in terms of outcome.
In a study of retention of colloid across a standard leaking ultrafiltration
membrane hydroxyethyl starch was superior to albumin, dextrans and
gelatins. Whereas gelatins should suffice for resuscitation where
capillary leak is not an issue, it would seem logical to use hydroxyethyl
starch in cases of capillary leak.

Conclusion
Despite the plethora of physiological monitoring commonly available in
the intensive care unit there remain some challenges in the avoidance
of hypovolaemia. Interpretation of physiological data assumes a
reference point of normality or acceptability. Normality is a statistical
concept from which acceptable cannot be derived for the individual.
Avoiding hypovolaemia requires a high degree of suspicion and a "try it
and see" approach with corrective treatment. The fluid challenge
approach provides a safe method of confirming or refuting the
diagnosis while titrating appropriate volume replacement.

While there are many fluids available for support of the circulation and
choice of fluid may depend on certain properties of relevance to the
clinical situation, it is clear that fluid choice is of secondary importance
to the monitoring techniques used to support fluid delivery. It is of
critical importance to choose the goals of fluid therapy first and adopt
physiological monitoring tools to ensure the goals are achieved.
Without these the greatest danger in intravenous fluid therapy is that
covert hypovolaemia persists.

81
Bibliography

Weil M, Shubin H, Rosoff L. Fluid repletion in circulatory shock - central


venous pressure and other practical guides. JAMA 1965;192:668-674.
Hagan R, Diaz F, Horvath S. Plasma volume changes with movement
to supine and standing positions. J Appl Physiol 1978;45:414-418.
Baek S, Makabali G, Byron-Brown C, et al. Plasma expansion in
surgical patients with high central venous pressure; the relationship of
blood volume to hematocrit, CVP, pulmonary wedge pressure, and
cardiorespiratory changes. Surgery 1975;78:304-315.
Singer M, Allen M, Webb A, et al. Effects of alterations in left ventricular
filling, contractility, and systemic vascular resistance on the ascending
aortic blood velocity waveform of normal subjects. Crit Care Med
1991;19:1138-1145.
Webb A. The fluid challenge. In Webb A, Shapiro M, Singer M, Suter
P (Eds) Oxford textbook of Critical Care, Oxford University Press,
Oxford, 1999, pp32-34.

82
Therapeutics Intravenous Fluid Therapy

ACADEMIA

Published by Aesculap Academia.


Thorncliffe Park, Sheffield, S35 2PW • Tel (0114) 225 9000 • Fax (0114) 225 9111

You might also like