Anaesth. Intens. Care (1977). 5.

284

BODY FLUID COMPARTMENTS

Roy W. PAIN*
Institute of Medical and Veterinary Science. Adelaide

SU:\c\lARY
. The terms mole. molality ..mola.ri~y. osmote. osm?lal!ty, osmolarity. osmolar gap and
anwn [!,ap are defined and thetr cllmcal usefulness mdlcated. The following body fluid
com:partments arc descrtbed: t.otal bOI('! water (TBW). extracellular fluid (ECF). intracellular
flmd (ICF), transc~llu!ar flltld ~TCI'). plas.ma vO~ltme, red cell volume and interstitial fluid
volume. Isotope-dlltttwn tcchmques are bnefly dIscussed and representatit,e normal c'alues
for the .various compart~ents a~co:di1:g t? s~x and ar.,e. are indicated. The Physiolo[!,ical
m~chamsm~ that 111~mtazn the dtstl11ctH'e W~tlC composltwns of the various fluid spaces are
brtefly. outhn.e4. ~\ ew concepts of the fu.nctlOlt of the gel matrix and of the lymph drainage
of the lIlterstztatm are presented. Opposzng models to the sodium-potassium membrane pump
are briefly described.

I~TRODUCTION gTams (g). A mole of any substance

This section deals with the anatomy of the
fluid compartments of the body and with the
physiological processes that keep the ionic
composition of these compartments so constant.
The anatomy of the body fluid spaces was
elucidated by radioisotope studies in the 1950's
and early 1960's and little has been added since.
Thus the publications of Gamble (1954), Edelman
and Leibman (1959) and Moore et al. (19t)3) are
still excellent reference works on this aspect.
Many of the physiological mechanisms that
regulate the fluxes of fluid and biological sub-
stances through these body compartments and,
at the same time, maintain their distinctive ionic
compositions, have likewise been known for some
time. However, ongoing research continues to
discover new regulatory mechanisms. The aim
of this introductory chapter is to define some of
the terms that appear on clinical chemistry
laboratory reports and to give a background
description of the fluid pools in which are dis-
solved the ions to be described in later sections.
1. Definitions of terms used in the description
of fluid and electrolyte status.
(i) Alole. The mole (mol) is the molecular
weight (MW) of a substance expressed in
* :\I.B.. B.S .• F.R.C.P.A. Senior Clinical Chemist
(Medical). Division of Clinical Chemistry.
Address for reprints: Dr. Roy W. Pain, Institute
of Medical and Veterinary Science, Frome Road.
Adelaide. South Australia 5000.
contains t)'061 x10 23 molecules (i.e. Avo-
gadro's number). A millimole (mmol)
is one thousandth of a mole. The Inter-
national System of units (SI units) uses
the mole unitage (where this is known)
rather than the "equivalent" (which is the
atomic or molecular weight divided by
valence).
(ii) Molality. The molality of a solution is
the number of moles of a solute per 1000
gTams of solvent (water in the case of
serum or plasma).
(iii) Molarity. The molarity of a solution is
the number of moles of solute per litre of
solution and is therefore, unlike the
molality, dependent on temperature. The
numerical difference between molal and
molar concentration is negligible in the
range of concentrations and temperature
of the body fluids.
(iv) Osmole. The osmotic effect of a sub-
stance in solution depends only on the
number of discrete particles dissolved and
is independent of the weight, electric
charge, valence or chemical formula.
One osmole (osm) of a substance yields,
in ideal solution, that number of particles
(Avogadro's number) which would de-
press the freezing point of the solvent
(water in the case of serum) by l·St)°C. A
milliosmole (m osm) lowers the freezing
point by l'St) x10- 3 °C. Each mole of

Anaesthesia and Intensive Care. Vol. V. No. 4. November, 1977

 BODY FLUID COMPARTMENTS
an unionized substance equals one osmole.
Since each molecule of sodium chloride in
ideal solution dissociates into a sodium
and a chloride ion, sodium chloride
contributes twice as many osmotically
active particles as a non-ionized sub-
stance, e.g. glucose. Thus one milli-
mole of sodium chloride gives two millios-
moles. However, body fluids are not
ideal solutions and although the dissocia-
tion of strong electrolytes is complete,
the number of particles free to exert an
osmotic effect is reduced due to interac-
tions between the ions and thus one refers
to the "effective" concentration or
" activity". Bearing these restrictions
in mind, it is easy to show that sodium
chloride is 2,300 times as osmotically
active as albumin, for one gram of
albumin (MW 68,000) theoretically yields
0·015 milliosmoles while 1 gram of sodium
chloride theoretically yields 34·3 millios-
moles.
(v) Osmolality. The osmolality of a solution
is the number of osmoles of solute per
kilogram (kg) of solvent (water in the
case of serum or plasma). In clinical
chemistry osmolality is usually expressed
in milliosmoles/kg. The osmolality of
serum or urine is measured directly by an
instrument called an osmometer on the
principle that each milliosmole will
depress the freezing point of the solvent
water by 0 ·00186°C. The normal os-
molality of serum or plasma is 285-295
mosm/kg.
(vi) Osmolarity. The osmolarity of a solution
is the number of osmoles of solute per
litre of solution. This is usually express-
ed in milliosmoles/litre in clinical
chemistry. This parameter is calculated
by adding the total osmolar concentra-
tions of directly measured substances in
serum or plasma, i.e. by adding the
osmolar concentrations of electrolytes,
glucose and urea. At least seventeen
different formulae for making this cal-
culation have been described by Dorwart
and Chalmers (1975) and Weisberg (1975).
In our laboratory the formula used is
"1·86 (Na+K) + urea + glucose"
with all measurements expressed in 2. Total Body Water (TBW)
mmol/I. The factor of "1·86" makes
allowance for interaction between ions.
Our 95 percentile reference range is
272-283 m osm/I.
Anaesthesia and Intensive Care, Vol. V, No. 4. November. 1977
285
(vii) "Osmolar Gap". The numerical differ-
ence between measured osmolality and
calculated osmolarity is called the osmolar
gap and it is normally 0-24 milliosmoles/
kg in value. Certain alcohols, e.g.
ethanol, depress the freezing point of
plasma water and hence cause a falsely
elevated osmolality and thus an elevated
osmolar gap. This fact has been made
use of, to quickly estimate the plasma
ethanol concentration, by Robinson and
Loeb (1971), Champion et al. (1975) and
Phillips and Pain (1975). The normal
concentration of solids (protein and
lipids) in serum is 7% leaving 93% of
serum as water. With hyperproteinae-
mic states (e.g. multiple myeloma) or
hyperlipoproteinaemia the percent of
serum water decreases. Although the
concentration of electrolytes in the serum
water (and hence the osmolality) remains
normal the concentration of electrolytes in
serum (and hence the calculated osmo-
larity) decreasE$. This is another cause
of an elevated osmolar gap (Weisberg
1975).
(viii) "Anion Gap ". Anions are negatively
charged ions (e.g. Cl-, HC0 3 -, P04 ~,
proteins, lactate and ketones) and cations
are positively charged ions (e.g. Na+,
K+, Ca++, Mg++). The number of anions
must equal the number of cations to
ensure electrical neutrality. The anion
gap is usually calculated from the formula
(Na++K+)-(Cl-+HC0 3-) and the 95
percentile reference range in our labora-
tory is 11-19 mmol/l (Thomas, Pain and
Duncan 1973). In other words sodium
and potassium account for a greater
percentage of the cations than chloride
and bicarbonate do of the anions. Most
of the normally unmeasured anions are
due to protein. If the anion gap is
raised it is usually due to phosphate and
sulphate, etc. (in renal failure), ketones
(in diabetic ketoacidosis) or lactate (in
lactic acidosis). For an excellent review
of the clinical usefulness of the anion gap
see Emmett and Narins (1977), Lancet
Editorial (1977) and Narins and Emmett
(1977).
(i) Measurement
Although chemical analysis has been
carried out on some eight cadavers
(Edelman and Leibman 1959, Krzywicki
286 Roy W. PAIN

et al. 1974) only one subject, an period following a 60 tJ.Ci oral dose of
accident victim, could have been con- THO. Antipyrene is another tracer that
sidered normal. Thus, the estimation of has been used.
the fluid compartments of the body must (ii) Reference Values
always involve uncertainties since only
indirect methods of analvses can be used. Table 1 gives representative values for
Since different tracer mefhods give similar TBW throughout life. The reader is
results for total body water (TBW) both referred to the monumental work of
in humans and in animals and, in the Moore et al. (1963) where regression
latter, these results agree favourably equations are given for all the body fluid
with experimental data obtained by tissue compartments tailored to gender, age and
sampling and desiccation, this has lent body weight. These formulae have been
credibility to the results obtained by derived from parametric equations to
dilutional tracer measurements. compensate for extremes in body size.
Day to day fluctuations in TBW in
Isotope-dilution techniques normal man are very small amounting to
Volumes of distribution can be calcu- TABLE 1
lated for any substance that, having been Representative Normal Values for TBW, ECF and ICF
injected, will mix throughout a body as % of Total Body Weight
compartment, providing the concentra-
tion in the body fluids and the amount Age % Total Body
removed by excretion or metabolism can in Years Unless Weight
Otherwise --- --- --- ECF/ICF
be accurately measured. These tech- Stated TI3W ECF ICF
niques assume that the subject is in a ---
steady metabolic state (diet, water and Premature ·. 80 - - -
electrolytes, etc.) during the experi- Birth ·. ·. 76 42 33 1.27
1 month · . ·. 70 32 38 0.84
mental period and this may not always 1 .. ·. ·. 65 26 39 0.67
be so. The isotopes of water, deuterium 10 . . ·. ·. 62 26 36 0.72
oxide (D 20) and tritium oxide (THO) are Males 25 · . ·. 60 27 33 0.82
the tracers most frequently used for TBW 45 · . ·. 53 24 29 0.83
65 · . ·. 54 26 28 0.93
(Deuterium is 2H and tritium 3H). They
are small molecules that penetrate all
85 · . ·. 51 26 25 1. 04
Females 25 ·. 51 23 28 0.82
phases of body water rapidly with good 45 ·. 48 23 25 0.92
mixing and are not significantly metabo- 65 ·. 44 22 22 1.00
lized over the 3-4 hour equilibration 85 ·. 43 22 21 1.05
period. There is only 1-2% exchange
(Adapted from Edelman and Leibman 1959, Moore
with labile hydrogen atoms of the solid et al. 1963.)
constituents of the body, mainly proteins,
and less than 3 % appears in the urine per approximately 0'2% of body weight
24 hours. D 2 0 is 10% heavier than (Kleeman 1972). The large variation in
water, is non radioactive and is measured TBW observed between individuals is due
by difficult densimetric or mass-spectro- mainly to the reciprocal relationship
metric methods. between body water and body fat
Tritium is a weak beta emitter and the content since there is very little water in
use of THO is the isotope method of adipose tissue. With increasing degrees
choice. The radioactive half life is 12·4 of obesity the percent of body weight
years but the biologic half-life is only 10 which is water approaches 50% whereas
days and with the usual small doses of with increasing degrees of leanness it
0·5 mCi the total bodv radiation dose is approaches 70%. Estimates of body
well below acceptable limits. Equilibra- water based on fat-free body weight (the
tion occurs within 4 hours, perhaps 6-8 "lean body mass") are more constant
hours in obese or oedematous patients or than those based on body weight not
in those with ascites. Thereproducibility corrected for fat. However, although
of results is usually ±2%. Eberstadt there may be such an entity as the "lean
(1974) has described a rapid THO method body mass" in healthy young adult
where urine is collected over a two hour males, there is unlikely to be so in the

Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977

 BODY FLUID COMPARTMENTS 287

heterogeneous population as a whole. of whole blood. Similar considerations

This is especially so in the sick where
there is loss of muscle and fat tissue and
where metabolically inactive bone and
dense connective tissues assume much
bigger proportions of the total weight.
Thus prediction formulae based upon
mean ideal standards, apart from not
allowing for possible individual variation,
do not necessarily apply to the critically
ill patient.
3. Individual Compartments
(i) Extracellular Fluid (ECF)
By definition this includes all body
water which is external to the cells and,
in a physiological rather than an ana-
tomical sense, may be divided into the
five phases shown in Table 2. Whereas
the ECF was formerly considered as
composing 33 % of TBW and 20% of total
body weight the masterly review of
Edelman and Leibman (1959) indicates
that more accurate figures are 45 % of
TBW and 27% of total body weight. In
addition, Swan and Nelson (1973) in a
thought-provoking article, conclude that
red blood cell water is an expansion of the
ECF because 3-5% of injected radio-
chloride enters the erythrocytes within
five minutes. Since red cells are 70%
water then for every 100 ml of blood,
82·5 ml is water, the red cell water con-
tributing 38% of the total water content
with respect to the electrolyte composi-
tion of red cells should be kept in mind,
the authors claim, when deciding what
replacement fluids to use in cases where
whole blood has been lost (Swan and
Nelson 1973). With rare exceptions
chronic disease is associated with a fall
in intracellular fluid (ICF) volume and
an expansion of the ECF volume so that
the ECF may be equal to or even exceed
the ICF volume (Moore et al. 1963,
Elwyn, Bryan-Brown and Shoemaker
1975). Any major operation results in
catabolism of body tissue high in fat
content thus releasing 600-1000 ml of
sodium-free intracellular and metabolic
water a day, some of which becomes
extracellular (Moore 1959).
(ii) Measurement of ECF
(a) Plasma Volume
Plasma volume can be measured using
dyes that bind to plasma albumin, e.g.
Evans Blue (T-1824) or radioiodine-
labelled serum albumin (RISA). 7-10%
of iodinated albumin escapes per hour
from the vessel compartment into the
interstitial fluid and lymph causing an
overestimation of plasma volume in
normals and an even greater error in
patients with trauma, burns, cancer,
ascites, nephrosis and cardiopulmonary
disorders (Valeri and Cooper 1973). To

TABLE 2
Extent of Body Fluid Compartments (for a 70 kg young adult male) *

Percent Percent Volume

Compartment Body Weight TBW in litres ml/kgwt
Plasma · . ·. ·. ·. ·. ·. 4·5 7·5 3·2 45
Interstitial-lymph ** ·. ·. ·. ·. 12·0 20·0 8·4 120
Dense connective tissuet ·. ·. ·. 4·5 7·5 3·2 45
Bonet ·. ·. ·. ·. ·. ·. 4·5 7·5 3·2 45
Transcellular .. ·. .. ·. ·. 1·5 2·5 1·0 15
ECF (Total) ·. ·. ·. ·. ·. 27 45 19 270
Red Cellst ·. ·. ·. ·. ·. 2·3 3·8 1·6 23
ICF (Total) ·. ·. ·. ·. ·. 33 55 23 330
TBW ·. .. ·. ·. ·. ·. 60 100 42 600

* The body weight of the average young adult male is made up of 18% protein, 15% fat. 7% bone mineral
and 60% water. For the average young adult female whose weight is 61 kg, TBW is 51 %. ECF is 23%, fat is 25%.
** This includes 25% of connective tissue water and 10% of bone water.
t These compartments are .. inaccessible ". They comprise 75% of conneC'tive tissue water and 90% of bone
water.
t Some would include this compartment in the ECF.
(Adapted from Edelman and Leibman 1959).

Anaesthesia and Intensive Care, Vol. V, No. 4. November. 1977

288
reduce such errors the volume of distribu-
tion is determined after only 10-15
minutes equilibration or better one can
back extrapolate to zero time from
multiple readings. The above authors
improved the accuracy of plasma volume
measurement by using a labelled macro-
globulin. Representative normal values
according to age and sex are shown in
Table 3. It should be noted that the
expanded blood volume observed in
cirrhosis of the liver is a real phenomenon
and is not an artefact due to ascites
(Lieberman and Reynolds 1967, J\Iaddrey
et al. 1969).
(b) Red Cell Volume
This can be calculated from the plasma
volume and haematocrit values. However,
the whole body haematocrit is only 85-
92% of the large vein haematocrit.
Thus red cell mass calculated from
plasma volume and venous haematocrit
is overestimated by about 7%. The
better way to estimate red cell volume is
by re-injecting erythrocytes tagged with
radio chromium (5lCr). See Table 3 for
representative normal values. This is
the preferred way for testing the extent
of haemorrhage because the plasma
volume is quickly restored from the
interstitial fluid pool.
(c) Extracellular Fluid Volume (ECF)
Values ranging from 15-27 % of the
total body weight have been obtained for
ECF according to the tracer used to
measure it. Tracers fall into two main
categories, crystalloids (e.g. inulin, man-
nitol and sucrose) or ionic substances
(e.g. 82bromide, chloride isotopes and
Representative Normal Values for Plasma and Red Cell Volumes
Plasma \'olume
Age in Years
In Litres ml/kg Body Weight
Male 2.5 ·. ·. 3·3
45 · . ·. 3·2
65 · . ·. 2·9
86 · . ·. 2·7
------
Female 2tj ·. 2·7
4ii ·. 2·7
6ii ·. 2·5
85 ·. 2·4
I
I I
(Adapted from Moore et al. 1(63)
Roy W. PAIN
35 sulphate). The crystalloids being larger,
less diffusible molecules do not penetrate
cells but they also fail to penetrate the
entire ECF within a reasonable time.
The smaller, more diffusible ionic sub-
stances penetrate all parts of the ECF
quickly but also partly penetrate cell
membranes. Ionic substances thus give
larger values than crystalloids. No
material is known which is entirely
confined to the ECF. Inulin is slowly
metabolized. Isotopes of chloride have
half-lives either too long (400,000 years)
or two short (37 minutes). 82Br has a
half-life of 36 hours and is not metabo-
lized and only slowly excreted. It
rapidly, but non-progressively, enters red
blood cells (a proof, some conclude, that
red blood cell water is part of the ECF-
see earlier) and other cells more slowly.
Bromide equilibrates in 20-24 hours
during which time 3-5% is lost into the
urine at a fairly constant rate. Expo-
nential regression of plasma levels to
zero time allows for losses into urine and
elsewhere (Spears et al. 1974). The
variability of results can be up to ±5%,
but is usually ±2%. Note that it is
more correct to talk of the "20 hour
bromide space" than the " ECF space" ,
i.e. the tracer and the equilibration time
should be designated. The disappearance
of injected radiobromide from the plasma
is the result of two rates of transfer: (i)
to a rapidly equilibrating pool (20
minutes) which is in dynamic equilibrium
with the plasma and is referred to as the
"functional ECF". It is readily per-
meated by both crystalloids and ions and
is some 20% of TBW and 8·4 litres in
TABLE 3
Red CeIl Volume
In Litres ml/kg Body Weight
47 2·0 29
43 1·9 27
42 1·8 26
39 1·7 25
44 1·5 25
43 1·4 22
40 1·3 21
38 1·2 19
I
Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977
 BODY FLUID COMPARTMENTS 289

volume. It is a reservoir from which cellular membranes are crossed allowing

water and sodium can be mobilized into
the circulation or which will accept large
amounts of these constituents when
circumstances favour extravascular filtra-
tion. This pool includes 25% of dense con-
nective tissue and 10% of bone water. (ii)
to a slowly equilibrating pool (up to 24
hours) which includes the remaining
water of dense connective tissue (cartilage,
ligaments and tendons) and bone. Each
of these two latter phases are 7 ,5% of
TBW and thus equal in volume to the
plasma water. They were originally
thought to be part of the intracellular
fluid volume but are now considered part
of the ECF because their chloride
concentration per kg of tissue water is
similar to the plasma concentration. Like
total exchangeable sodium, the ECF
volume (as a percent of total body weight)
shows very little change with age during
adulthood so that most of the decrease in
TBW is in the intracellular fluid (Forbes
and Reina 1970).
(d) Interstitial SPace Volume
This cannot be measured directly as it
is difficult to sample and no tracer
equilibrates with this fluid alone. It is
calculated as the difference between ECF
volume and plasma volume.
Jenin et al. (1975) consider the body as
consisting of conductors (liquids) and
insulating materials (cellular membranes,
fat) the latter forming a barrier to AC
current. Using impedance measurements,
ECF only will be conductive at low
frequency currents (1-5 kHz). At high
frequency currents (100 kHz-1 MHz)
TABLE
Mean Ionic Composition of Fluid Compartments (in mmol/l)
a measurement of TBW. Only the ratio
of TBWIECF can be found by this
technique and is normally 1·50±0·05.
If either the absolute TBW or ECF space
is then measured by standard metho-
dology, the other space can be simply
calculated, e.g. with a TBW of 42 litres,
the ECF would be 28 litres. Although
this experimental method does not at
present yield results identical with estab-
lished methods it would appear to be
worthy of further investigation because
of its simplicity and potential for serial
measurements.
(iii) Intracellular Fluid (ICF)
The volume of the intracellular fluid
(ICF) cannot be determined by isotope
dilution directly but must be calculated
by the difference between TBW and ECF,
i.e. including transcellular fluid as part
of ECF. Thus the error of the ICF
calculation will be the sum of the errors
of the TBW and ECF estimations.
Allowing the ECF for a young adult male
to be 27% of total body weight, the ICF
is then 33% of body weight or 55% of
TBW (formerly considered as 67% of
TBW). Thus in adults the ratio of
ICFIECF is probably 1 ·2 rather than 2·0
as formerly believed (Edelman and
Leibman 1959). The ICF declines with
age (Forbes and Reina 1970) and accounts
for most of the age related decline in
TBW. The ionic composition of ICF
will vary from tissue to tissue and also
with the degree of activity of the tissue
(Table 4 gives mean values). As would be
expected there is a good correlation be-
4

Plasma Intracellular
Substance Interstitial Fluid
Whole Water Fluid Water

Sodium .. .. ·. ·. ·. 141 152 144 10

Potassium ·. .. ·. ·. 3·7 4·0 3·8 156
Chloride ·. .. ·. ·. 102 110 115 3·0
Bicarbonate ·. .. ·. ·. 28 30 30 10
Anion Gap ·. ·. .. ·. 15 - - -
·. .. ·. - - -
..· .
Calcium .. 2·4
Magnesium .. .. ·. 0·8 - - 11
Phosphate ·. .. .. ·. 1·1 - - 31
Protein .. ·. .. .. ·. 16 - 10 55

(Adapted from Hays 1972, plus our own measurements and calculations)

Anaesthesia and Intensive Care, Vol. V, No. 4, November, 197'/

290 Roy W.
tween ICF and total exchangeable
potassium. Erythrocytes and leukocytes,
being specialized tissues, may not be
representative of cell composition as a
whole. There are also probably sub-
divisions of intracellular water and
electrolytes among the organelles within
the cell, c.g. during active respiration,
mitochondria contain very high levels of
calcium, magnesium and inorganic phos-
phate and the sodium concentration of the
nucleus is probably higher than in other
parts of the cell.
(iv) Transcellular Fluid (TCF)
Transcellular fluids have the common
property of being formed by the transport
activities of cells. They are extracellular
fluids and comprise 2 ,5% of TBW. TCF
includes the CSF and the fluids in the
lumen of the gastrointestinal tract; in the
urinary excretory passages; in the ducts
of glands of the pancreas, biliary tree,
skin, bronchial tree, gastrointestinal tract
including salivary glands; in the follicles
of endocrine glands; in the aqueous
humour of the eye; in the endolymph of
the inner ear; in joints and in the pleural,
pericardial and peritoneal cavities. The
electrolyte composition varies from site
to site and differs from that of a simple
ultrafiltrate of plasma. In the gastro-
intestinal tract, 7-12 litres of secretions
are secreted per day of which 98-99% are
reabsorbed. In obstruction of the lower
bowel, the total TCF can double or treble
at the expense of the remainder of the
ECF and 3-5 litres of "concealed" loss
of fluid can accumulate.
Secretion into an obstructed bowel
continues at a normal rate for 24, hours
and thereafter increases. Although there
is a wide range of concentration of various
ions in the gastrointestinal secretions, in
practice for purposes of replacement, one
can assume them to be isotonic with
plasma with a potassium concentration
about double that of plasma.
4. Physiological Regulation of Compartments
Certain physiological processes which play a
major role in maintaining the composition of the
fluid compartments of the body will now be
briefly described.
(i) Diffusion. Chemical gradients across
fluid compartmental barriers can lead to
shifts in substances by simple diffusion.
Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977
PAIN
Lipophilic substances cross membranes
(which are mainly lipid) at higher rates
than hydrophilic substances.
(ii) Pores in Cell ~1fembranes. These were
predicted because total volume flow across
membranes was 3-5 times that expected
from diffusion alone. There is evidence
that cell membranes are interrupted by
pores of varying sizes, the smallest being
approximately O· 7 nanometres (nm) in
diameter (Ganong ]976).
(iii) Hydrated Radius. Different substances
in the body are hydrated with varying
numbers of water molecules. It is the
hydrated radius that determines the
diffusibility of a subs tance rather than
the size of the non-hydrated ion or
molecule (Felgenhauer 1974, Felgenhauer
and Renner 1977).
The atomic weight (A W) of sodium is
23, its hydrated radius is 0·28 nm com-
pared to an A W of 39 and a hydrated
radius of 0·23 nm for potassium (Birch
1974). This is one reason why potassium
ions diffuse through membranes more
readily than sodium ions. It also
explains why lithium (A W =6, hydrated
radius =0 ·34, nm) in ways acts more like
magnesium (A W =30, hydrated radius=
0·47 nm) and calcium (AW =4,0, hydrated
radius =0 ·32 nm) than like sodium
(Birch 1974).
(iv) Osmosis. Osmosis is the movement of
solvent molecules across a membrane into
a region where there is a higher con-
centration of a solute to which the
membrane is impermeable. The os-
motic pressure" is that hydrostatic
pressure which would have to be applied
to the membrane to stop the movement
of solvent molecules. Since all cell
membranes and capillary walls are freely
permeable to water it follows that all
fluid compartments must be isotonic with
plasma except when there has been
insufficient time for equilibration to occur
after a sudden change in composition.
Sodium concentration is the prime con-
troller of extracellular osmotic pressure
and potassium of the intracellular osmotic
pressure. Hence, alteration of sodium
concentration in the ECF will be ac-
companied by alterations in ICF volume
and osmolality and vice versa. It follows
that acute hyponatraemia will always
 BODY FLUID COMPARTMENTS
cause an increase in ICF volume regardless
of the level of TBW and vice versa with
hypernatraemia. If the concentration of
plasma sodium is normal, the ICF volume
will likewise be normal. Since the ICF
volume is dependent on the ECF sodium
concentration it is possible to be oedema-
tous or dehydrated even though TBW is
normal. In actual fact, either intra-
cellular potassium concentration or
extracellular sodium concentration could
be the prime controller of the body's
osmolality. A portion of the intracellular
ions may be osmotically inactive, that is
bound to proteins and other cell con-
stituents. For example, magnesium
binds readily to a variety of cell
lipoproteins and nUcleoproteins, ribo-
nucleic acids and free A.T.P. and as
much as 30% of intracellular magnesium
may be bound. There is considerable
controversy regarding the extent to which
potassium is bound to the cytoplasm
(Ling and Ochsenfeld 1973). Neither is
all cell water osmotically active, for some
intracellular water will be bound to cell
proteins as water of hydration. This may
account for 16-40% of intracellular
water (Olmstead 1966).
Since urea diffuses freely across cell
membranes it does not contribute to the
" effective" osmotic pressure in vivo but
it does contribute to the in vitro os-
molality of plasma as measured in the
laboratory. Note too, that although a
5 % dextrose infusion is isotonic in the
bottle, in the body it is metabolized to
carbon dioxide and water so that the net
effect is of infusing water. In sudden
hyperglycaemia, the osmalility of the
ECF is increased, water is osmotically
drawn from the cells diluting the sub-
stances in the interstitial fluid and plasma.
Under such circumstances a simple rule
is that for every 3 mmol/l elevation of
plasma glucose there is a decrease of
1 mmol/l of plasma sodium due to osmotic
dilution (Katz 1973).
(v) Gibbs-Donnan Equilibrium. The presence
of higher concentrations of anionic pro-
tein molecules within cells and capillaries
than in the interstitial fluid causes (i)
the concentration of diffusible cations
(e.g. sodium or potassium) to be greater;
(ii) the concentration of diffusible anions
(e.g. chloride) to be less and (iii) the total
Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977
291
number of diffusible ions to be greater, in
the compartments containing the higher
concentrations of protein (Hays 1972).
It should be remembered that the
difference in actual number of diffusible
anions and cations between compartments
is extremely small relative to the total
number of anions and cations present
within those compartments. The result-
ant differences in ionic composition in
neighbouring compartments are examples
of the Gibbs-Donnan equilibrium.
(vi) Starling's Forces. The forces described
by Starling in 1896 cause a very large
turnover of water and diffusible solutes
between the intravascular and interstitial
compartments without a net change of
volume in either compartment. The
total filtering surface of the capillary bed
is about 6,300 M2 in the adult. Since
the osmotic pressure due to the plasma
proteins is 25 mm Hg throughout the
length of the capillary whereas the blood
pressure falls from 35 mm Hg at the
arterial end to 15 mm Hg at the venous
end of the capillary there will be free
flow of water and diffusible ions into and
out of the interstitium at these two ends
respectively. Most workers still believe
that the overall Starling model is valid
even though it has been shown (Aukland
1973) that (i) the venous end of the
capillary has a greater surface area and
is more permeable than the arterial end;
(ii) the albumin concentration of the
interstitial fluid is of the order of 20 g/l
so that the interstitial osmotic pressure
due to albumin is about 10 mm Hg.
This fluid also has Cl. hydrostatic pressure
of 0-5 mm Hg.
The interstitial fluid is normally held
tightly in place in the form of a gel. The
reticulum of this gel is composed mainly
of hyaluronic acid which is slightly cross-
linked with collagen fibres to form a
meshwork in which the interstitial fluid
is trapped. The presence of this gel
reduces markedly the mobility of inter-
stitial fluid (so that less than 1 % is
" free ") but it hardly effects the indi-
vidual molecular diffusion of small ions
and molecules (Guy ton et al. 1973)
I t has been estimated (Landis and
Poppenheimer 1963) that the bulk flow
of water through the interstitium is 20
litres per day in an adult. With an
292 Roy W.
interstitial fluid volume of 8· 4 litres this
would mean that interstitial water is
renewed about two and a half times daily.
c.;uyton, Granger and Taylor (1971)
claIm that the normal interstitial fluid
pressure (at least in subcutaneous tissues)
IS 6-8 mm Hg subatmospheric. Other
methods have not confirmed these find-
ing~, .e.g. needle pressure technique giyes
posltlYe pressures and the wick method
1-:2 mm Hg subatmospheric (Aukland
]973).
It is generally thought that 90-95°1 ;0 of
t Ile normal efflux from the arterial end of
the capillary is re-absorbed In- the
venous end of the capillary and tllat the
remainin~ 5-]0% flO\\:s il1to the lymph
vessels.. fhe Iyml?hatlcs are capable of
rhythmIc contractlons, the intensity of
which are related to the rate of Iyinph
fl?w m a regulatory fashion (Hall 1969).
1 he movement of skeletal muscles and
the pulsations transmitted from arteries
a:e not sufficient to explain lymph flow.
1 here are local variations in both struc-
ture and functior: of the lymphatic
vessels. The opemngs in the terminal
capillaries of the lymphatic system are
constructed in such a way' that the
endothelium cells overlap each other
~cting as flap val,:es to allow fluid entry
IIlt(~, b~t not eXIt out of, the lymph
caplllanes. The outer surfaces of the
lymphatic endothelial cells are held
t!ghtly to the surrounding interstitial
tlssues by anchoring filaments. There-
fore when the interstitial tissue expands
due to increased fluid entry, the l\'mph
capIllary also expands (being pulled open
by the surrounding tissue) and the cell
flaps open inwards allowing increased
amounts of fluid to enter (Guy ton et al.
197?). A~~ordingto.Guytonctal. (1973),
as mtershtlal flUld mcreases, the inter-
stitial gel will absorb the excess water and
~he interstitial hydrostatic pressure will
mc:ease sharply to zero, thus effectively
reslstmg oedema formation. After the
in~e:stitial gel has absorbed 30-50% of its
ongmal volume, free fluid (i.e. oedema)
occurs. \Yhen this happens the lym-
phatic return, which is normally less than
2 ml per minute from the entire body
increases by at least 20 fold (see earlier f;;;
the. mechanism for this). This causes
rapId loss of the protein-rich fluid from
Anaesthesia atld Intensive Care, T·ol. V, No. 4, November, 1977
PAIN
the interstitial space so that the inter-
stitial osmotic pressure decreases and
wa.ter .is. drawn into the capillaries by the
unmlllblted plasma osmotic pressure.
An infusion of a concentrated solution
of albumin or of some other large mole-
cular species will initially expand plasma
volume at the expense of the interstitial
fluid volume though significant amounts
of albumin will pass into the interstitial
space within 24 hours. The plasma
volume cannot be specifically increased
unless the administered fluid contains a
colloi.dal agent. For example, if saline
SolUtlOIl alone is administered to a sub-
ject who has lost blood this will re-
expand the extra-cellular fluid volume
but virtually all the expansion will be
confined to the interstitial compartment.
(Yii) Sodium-potassium I'u1l1p. The intra-
cellular content of soluble but non-
diffus.ihle macro~olecules (largely
protems and orgamc phosphates of net
negative electrical charge) tend to draw
extracellular fluid into the cell by osmotic
forces. Such uncompensated movement
of .wat~r could cause cellular swelling
whlCh, 1Il the case of the brain, could be
disatrous. This tendency to swollen
cells is offset by "pumps ,; located in the
outer plasma membranes of all cells.
These pUIllJlS continuously and actively
extrude sodium from the {CF to the ECl'
as rapidly as the sodium enters the cell
by diffusion from high extracellular to low
intracellular concentrations. Since they
simultaneously "pump" potassium into
cells against a chemical gradient they are
called "Xa-K pumps". As a result of
their pumping of ions, the pumps main-
tain a potential difference of from 10-100
millivolts across cell membranes in differ-
ent tissues (nO-80 millivolts for muscle
and nerve cells) with the inside negative
in relation to the outside. This electrical
gradient tends to repel negatively charged
chloride ions back into the ECF. The
bal~nce of d~stribution of osmotically
actIve solute IS such as to stabilize cell
volume. It is obvious that the body
thus does not Exist at equilibrium but
rather in a steady state away from
equilibrium.
The activity of the pump depends on
an adequate supply of adenosine tri-
 BODY FLUID COMPARTMENTS
phosphate (A TP) which is produced by
the metabolic processes of the cell and
which is hydrolysed by adenosine tri-
phosphatase (Na-K-ATP'ase) to adeno-
sine diphosphate (ADP) plus energy. It
is estimated that more than 1/3 of the
energy produced by resting muscle cells
is used to maintain the Na-K pumps
(Leaf 1970). The activity of Na-K-
ATP'ase is proportional to the concentra-
tion of potassium in the ECF and of
sodium in the ICF. The pump mechan-
ism is markedly temperature dependent
and is inhibited by ouabain and related
cardiac glycosides and by metabolic
poisons which prevent the formation of
ATP (Skou 1965, Schwartz, Lindenmayer
and AlIen 1975). Inhibition of the Na-K
pump causes entry of sodium into and
potassium out of cells with decline in
membrane potential and swelling of the
cell (" sick cell syndrome ").
Leaf (1970) in a thought-provoking
editorial discusses how failure of the
Na-K pump may be important in the
genesis of such diseases as cerebro-
vascular accident and myocardial in-
farction.
The Na-K pump or "membrane"
hypothesis is the generally held
mechanism for maintaining the difference
in ionic concentrations between the ICF
and ECF. Two other hypotheses are
worthy of note, if only to realize that
such exist. Ling (Ling 1962, Ling 1965,
Ling and Ochsenfeld 1973, Ling and
Walton 1976) advances the "association-
induction" hypothesis by which he
believes it is the cytoplasm rather than
the plasma membrane plus Na-K pump
which determines the composition of ICF.
He believes that cell water is not free but
exists almost entirely in multi-layer
absorption on protein molecules which
are kept in a necessary extended state by
the action of ATP. DeHaven and
Shapiro (1968) point out that at least 50
separate membrane carriers would need
to be postulated in order to support the
" membrane" theory in toto and that
few if any such carriers have been
positively identified. These authors
suggest that modern solution theory using
ionic strengths, dielectric constants and
dielectric increments could explain per
se the ionic differences between ICF and
Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977
B
293
ECF. No further work has been pub-
lished on this latter seemingly plausible
mechanism since it was first advanced.
REFERENCES
Aukland, K. (1973): "Autoregulation of Interstitial
Fluid Volume. Edema Preventing Mechanisms ",
Scand. ]. Clin. Lab. Invest., 31, 247.
Birch, N. (1974): "Lithium and Magnesium-Depen-
dent Enzymes ", Lancet, 2, 965.
Champion, H., Caplan, Y., Baker, S., Long, W.,
Benner, C., Cowley, R, Fisher, R, Gill, W. (1975) :
"Alcohol Intoxication and Serum Osmolality",
Lancet, 1, 1402.
De Haven, ]., Shapiro, N. (1968): "Speculations on
Physicochemical Fluid Properties in Physiological
Regulation ", Persp. BioI. Med., 12, 31.
Dorwart, W., Chalmers, L. (1975): "Comparison of
Methods for Calculating Serum Osmolality from
Chemical Concentrations, and the Prognostic
Value of Such Calculations ", Clin. Chem., 21, 190.
Eberstadt, P. (1974): "Total-body Water Determina-
tion with a Practical Internal Standard Method ",
Int. J. Nucl. Med. BioI., 1, 191.
Edelman, 1., Leibman, ]. (1959): "Anatomy of Body
Water and Electrolytes ", Amer. J. Med., 27, 256.
Editorial (1977): "The Anion Gap ", Lancet, 1, 785.
Elwyn, D., Bryan-Brown, c., Shoemaker, W. (1975) :
" Nutritional Aspects of Body Water Dislocations
in Post Operative and Depleted Patients ", Ann.
Surg., 182, 76.
Emmett, M., Narins, R (1977): "Clinical Use of the
Anion Gap ", Medicine, 56, 38.
Felgenhauer, K. (1974): "Protein Size and Cere-
brospinal Fluid Composition ", Klin. Wschr., 52,
1158.
Felgenhauer, K., Renner, E. (1977): "Hydrodynamic
Radii versus Molecular Weights in Clearance
Studies of Urine and Cerebrospinal Fluid ", Ann.
Clin. Biochem., 14, 100.
Forbes, G., Reina, ]. (1970): "Adult Lean Body Mass
Declines with Age: Some Longitudinal Observa-
tions ", Metabolism, 19,653.
Gamble, ]. (1954): "Chemical Anatomy, Physiology
and Pathology of Extracellular Fluid ", Harvard
University Press, Cambridge, Mass.
Ganong, W. F. (1973): " A Review of Medical Physi-
ology ",6th Ed. page 12, Lange Medical Publishers
California.
Guyton, A., Granger, H., Taylor, A. (1971): "Inter-
stitial Fluid Pressure ", Physiol. Rev., 51, 527.
Guyton, A., Taylor, A., Granger, H., Gibson, W. (1973):
"Regulation of Interstitial Fluid Volume and
Pressure ", Adv. Exp. Med. Biol., 33, 111.
Hall, ]. (1969): "The Flow of Lymph ", New Engl.
J. Med., 281, 720.
Hays, R (1972): "Dynamics of Body Water and
Electrolytes" in "Clinical Disorders of Fluid and
Electrolyte Metabolism" by Maxwell, M. H. and
Kleeman, C. R, 2nd Ed., page 15, McGraw-Hill,
New York.
lenin, P., Lenoir, ]., Roullet, C., Thomasset, A. L.,
Ducrot, H. (1975): "Determination of Body
Fluid Compartments by Electrical Impedance
Measurements ", Aviat. SPace Environ. Med., 46,
152.
Katz, M. (1973): "Hyperglycemia-induced hypo-
natremia-Calculation of Expected Serum Sodium
Depression ", New Engl. ]. Med., 289, 843.
294 Roy W.
Kleeman, C. R. (1972): ., Water :Metabolism" in
"Clinical Disorders of Fluid and Electrolyte
Metabolism" by 1\laxwell, 1\L H. and l<]epman,
C. R, 2nd edn., page 216, McGraw Hill, New York.
Krzywicki, H., \Vard, G., Rahman, D., Xelson, R.,
Consolazio, C. (1974): "A Comparison of Methods
for Estimating Human Eody Composition ",
Anzer. j. Clin. Nutr., 27, 1380.
Landis, E., Pappenheimer, j. (l!J63): "Exchange of
Substances Through the Capillary \Valls ", pages
961-1034 in Hamilton, \\'. F. and Dow, 1'. (Eds.)
"Handbook of Physiology, Section 2: Circulation
Yol. H." Amer. Physiol. Soc., \\'ashington, D.C.
Leaf, A. (1970): "Regulation of Intracellular Fluid
\'olume and Disease ", Amer. j. Xfed., 49, 2!1l.
Lieberman, F., Reynolcls, T. (1967): "Plasma Volume
in Cirrhosis of the Liver: Its Relation to Portal
Hypertension, Ascites, and Renal Failure ", j.
Clin. Invest., 46, 12\l7.
Ling, G. ~. (l!l62): "A Physical Theory of the Living
State ", Blaisdell, \\'altham, :\lass.
Ling, G. ~. (1065) : "The Membrane Theory and Other
VIews for Solute Permeability, Distribution, and
Transport in Li\'ing Cells ", Persp. Bioi. i1led.,
9,87.
Ling, G., Ochsenfeld, 1\I. (1973): "1\Tobility of Potassi-
um Ion in Frog :\Iuscle Cells, both Living and
Dead ", Science, 181, 78.
Ling, G., \Ya!ton, C. (1076): "What Retains \Yater in
Li\'ing Cells ", Science, 191, 293.
:\laddrey, \V., Boyer, j., Sen, N., Thomas, j., Basu, A.,
lber, F. (106ll): "Plasma Volume Expansion in
Portal Hypertension ", Hopkins Med. j., 125, 17l.
:\loore, F. (1959): "Metabolic Care of the Surgical
Patient ", page 284. W. B. Saunders, Phila-
delphia.
Anaesthesia and Intensive Care, Vol. V, No. 4, November, 1977
PAIN
l\Ioore, F., Olesen, K., :VTc:Vlurrey, j., Parker, H., Ball,
1\1., Boyden, C. (1963): "The Body Cell :\Iass and
its Supporting Environment. Body Composition
in Health and Disease ", \V. B. Saunders,
Philadelphia.
~arins, G., Emmett, M. (1977): "The Anion Gap ",
Lancet, 1, 1304.
Olmstead, E. (Ill66): "Mammalian Cell Water ",
London: Henry Kimpton.
Phillips, P., Pain, R (Ill7!;): "Blood-alcohol and
Serum Osmolality", Lancet, 2, 232.
Robinson, A., Loeb, j. (Ul7I): " Ethanol Ingestion
Commonest Cause of Elevated Plasma Osmolality",
New Engl. j. Med., 284, 1253.
Schwartz, A., Lindenmayer, G., Alien, J. (1975):
" The Sodiulll-potassiulll Adenosine Triphos-
phatase : Pharmacological, Physiological and
Biochemical Aspects ", Pharmacal. Rev., 27, 3.
Skou, J. (In6;';): "Enzymatic Basis for Active Tran-
sport of Na+ and K+ Across Cell ;\lembrane ",
Physiol. Rev., 45. 596.
Spears, C., Hyatt, K., Yogel, J., Langfitts, S. (l!l74) :
" t:nified Method for Serial Study of Body Fluid
Compartments ", Aerospace J1ed., 45, 274.
Swan, H., ::'\elson, A. (1973) : "The Chemical Anatomy
of Extracellular \Yater: Contribution of Erythro-
cyte \Vater ", j. Cardiovasc. Surg., 14,615.
Thomas, D., Pain, K, Duncan, B. (1973): "The
Anion Gap ", Lancet, 2, 848.
Yaleri, c., Cooper, A. (J!J73): "Limitations of
:\lcasuring Blood \' olumc with Iodinated p25
Serum Albumin ", Arch. Intern. 'lIed., 132, 534.
V,'eisberg, H. (1975) : "Osmolality-Calculated, 'delta',
and More Formulas", Clin. Chem., 21, 1182.