Osmolarity
Chemical
of Human Serum and of
Solutions of Biologic
Importance
Edward B. Hendry
With the use of the Fiske Osmometer, the mean total osmolarity of normal human
serum was found to be 289 mOsM (S.D., 4), which is equivalent to a mean freezing point of -0.537#{176}.
The isosmotic concentrations of some important biologic
solutions were determined. It was also found that M/15 solutions of disodium
hydrogen phosphate and of potassium dihydrogen phosphate are very hypotonic,
and that 3.8% sodiumcitrate is hypertonic. Hemolysisof erythrocytes in isosmotic
ammonium chloride solution can be considerablydelayed by the addition of 3.0%
glucose to the solution. Isosmotic concentrations of disodium hydrogen phosphate
and sodium dihydrogen phosphate were precisely determined, as were pH levels
of buffer solutionsmade from these two salts. The cause of the slight changes in
osmolarity that occur when these two isosmotic solutions are mixed is discussed.
“THE INThODUCTION
of the Fiske Osmometer* has eliminated most of
the technical difficulties of the measurement
of freezing-point
depressions.
Stirring is mechanical and can be regulated
to any predetermined
speed. The rate of cooling of the specimen can be reduced as the sample reaches the actual freezing temperature.
Freezing is induced mechanically
by means of a vibrator rod so that the
degree of supercooling
is precisely controlled and is constant in any
determination,
whether of pure chemical solutions or of biologic
fluids: This eliminates not only errors due to variation of the actual
temperature
at which freezing is induced (the greatest
source of
error in the conventional
Beckmann thermometer
method) but also
those introduced
by “seeding”
the sample with a crystal of frozen
solute. On any given solution or fluid, a series of readings may be
made in rapid succession, and the manipulations
involved can be reFrom
the
Department
Received
for publication
*Advanced
Instruments,
of
Biochemistry,
The
Western
July 22, 1960.
Inc., Newton
Highlands,
156
Mass.
Infirmary,
Glasgow,
W.1,
Scotland.
Vo’. 7, No. 2, 1961
OSMOLARITY OF HUMAN SERUM
157
peated identically-an
important
advantage
in this type of work.
Finally, the instrument
may be standardized
rapidly at any point on
the scale by means of standard solutions.
The accuracy of the Osmometer is ± 1 mOsM,* which is equivalent
to a freezing point depression
of 0.00186#{176}.
Since the total osmotic
pressure of normal body fluids is in the region of 300 mOsM, the percentage error is very small. The standard
solutions used for calibration contain sodium chloride prepared
on the basis of the accepted data of the International
Critical Tables (10) or on the basis
of the precision data of Scratchard
and Prentiss
(1). Calibration
is
carried out by means of molal (w/w) solutions;
are invariably
in
biologic work, molar (w/v) solutions used. The difference is very
slight, since 1 kg. of pure water at 20#{176}
occupies 1001.77 ml., and any
error introduced by interchanging
molal and molar solutions is less
than the probable error of the instrument.
Serum is the most readily available of all body fluids and is generally accepted as representative
in terms of total osmotic pressure,
of all the body fluid compartments.
This is based on the assumption
that there are no barriers to the movement of water between these
compartments,
and that movement
of water alone maintains
the
same total osmotic pressure
on both sides of any membrane.
Although perhaps not strictly accurate, this statement
can at least be
accepted as a first approximation.
The difference in total osmotic
pressure between serum and native plasma must be very small: The
only major chemical difference is the occurrence in plasma of a low
concentration
of fibrinogell (molecular weight, about 400,000), which
cannot appreciably
affect the total osmotic pressure.
Most modern textbooks quote -0.56#{176}
as the freezing point of normal human serum, but this figure is too low. The mean values and
ranges published in recent years are listed in Table 1. In instances
in which the vapor-pressure
method was used, the results are expressed in the literature
as concentrations
of sodium chloride; such
values have been converted to milliosmols for purposes of coinparison.
An osmolarity
of 288 mOsM is equivalent to a freezing point of
-0.535#{176};
in the investigation
reported in this article, the mean value
obtained using the Fiske Osmometer
(289 mOsM) was taken as tile
standard figure for normal human serum. The work reported deals
with the osmolarity of normal human serum and of various chemical
*Per
kg. water
understood.
158
HENDRY
Table
No. of euaa.
1.
OsMoiARrrv
OP NORMAL. HUMAN
Oa.nojarffiit
Mean
M.Lkod*
Ch.mistry
Clinical
SERUM
(mOaM)
S.D.
Re/ir.ne.
30 M
VP
286
6
(2)
20 P
VP
287
6
(2)
21 M, P
28 M
VP
VP
289
288
6
4
(3)
(4)
B
291
8
289
4
75 M, F
50 M, F
P0
VP indicates vapor-pressure
method of Hifi (6);
P0, Fiske Osmometer method; M, male, F, female.
Mean of 6 values, 288 mOsM.
solutions
isosmotic
author.
B, Beckmann
(5)
Present
thermometer
work
method;
of biologic interest
and re-examines
the composition
of
sodium phosphate
solutions previously
described by the
Methods and Results
Serum was collected from convalescent
adult patients who could
reasonably
be regarded
as “normal”
and from members of staff.
The clotted blood could be kept at room temperature
(about 20#{176})
for at least 2 hours without in any way affecting the serum osmolarity. Once the serum had been separated,
it could be preserved at refrigerator
temperature
(4#{176})
for several hours without change in
osmolarity.
The mean value found in this series of 50 cases (25 male, 25 female) was 289 mOsM (S.D., 4). The osmolarity
was the same for
both sexes, and the means and standard
deviations
were identical.
The mean value corresponds
to a freezing point of serum equal to
-0.537#{176}.
This value is in good agreement with the results of other
workers using different methods (Table 1).
Since sodium and chloride ions are the main contributors
to the
total osmotic pressure
of normal serum (together,
they contribute
approximately
70 per cent), both were estimated in each specimen.
The mean serum sodium concentration
was 140 mEq./L. (S.D., 3.1),
and the mean chloride concentration,
104 mEq./L. (S.D., 3.3). The
serum urea concentration
was also determined
on all specimens to
exclude any possible case of undiagnosed
nitrogen
retention;
all
values were within normal limits.
Osmolarity of Standard Solutions
The chemicals used were of analytical
reagent quality, supplied
by British Drug Houses, Ltd. Each was recrystallized
once from
Vol. 7, No. 2, l9l
159
OSMOLARITY OF HUMAN SERUM
water and, if anhydrous,
dried in vacuo over phosphorus
pentoxide.
If the chemical was hydrated,
the procedure
described below was
used. All concentrations
were molar. For each compound, five solutions were prepared
in fresh glass-distilled
water, one near the assumed isosmotic concentration
and two on either side of it. The
osmolarity
of each solution was determined
five times and the mean
value for each was calculated.
The concentration
equivalent
to a
total osmotic pressure of 289 mOsM was then obtained with the use
of a large-scale
graph. Any error in the final result was not likely
to be greater than 1 part in 300. Typical results are shown in Table
2. These data are within 1 mOsM of the osmolarity calculated from
Table
2.
OSMOLAIUTY
os’ SoDIUM CHLORIDE SoLunoNs
Solution
Concentration
(grams per 100 ml.)
Osmolarity
(milliosmols)
0.960
305
0.921
292
0.898
286
0.870
277
0.850
270
the data given in the International
Critical Tables. Graphically
(or
by calculation),
the isosmotic concentration
of sodium chloride corresponding
to 289 mOsM was found to be 0.910 gm./100 ml. The
isosmotic concentrations
found for other useful biologic solutions
are shown in Table 3.
Table
3.
IsosMcYrIc
CONCENTRATIONS
OF 289-MOsM
Concentralion
(g,n./100
ml.)
Solution
Sodium
Potassium
SoLu’rIoNs
chloride
(Table
2)
0.910
1.169
0.836
chloride
Ammonium chloride
Disodium hydrogen phosphate
(anhydrous)
Sodium
(a#{241}hydrous)
dilaydrogen
phosphate
Sodium dihydrogen phosphate dihydrate
Sodium bicarbonate
Sodium citrate (Na3C6H507.2H20)
Glucose (anhydrous)
1.792
1.927
2.505
1.346
3.06
5.20
Dataon Salts Studied
Ammonium Chloride
Solutions of this salt must be given intravenously
only in those
rare cases in which the patient has severe hypochioremia
with alkalosis, where it is desirable to increase rapidly the concentration
of
160
HENDRY
Clinical Chemistry
chloride in the extracellular
fluids without, at the same time, increasing the sodium concentration.
Since the human erytlirocyte
is rapidly permeated
by ammonium salts, red cells hemolyze in any concentration of ammonium
chloride.
The phenomenon
has been extensively studied by M. H. Jacobs and his colleagues (7). When administered intravenously,
ammonium
chloride solution must be given
slowly, and it must never be used in cases of hepatic dysfunction,
since it rapidly induces coma.
The hemolytic action of ammonium chloride solution may be reduced by the addition of glucose (which enters the erythrocyte
relatively slowly), as shown by the following experiment.
Heparinized
venous blood was diluted 1:10 in isosmotic solutions of ammonium
chloride containing
increasing
concentrations
of glucose. The systems were incubated at 37#{176}
and the times taken to reach complete
hemolysis were measured.
The results were as noted in Table 4.
Table
4.
RATR
OF HEMOLYSIS
Syatem
Isosmotic
ammonium
No glucose added
Plus 1.0% glucose
Plus 2.0% glucose
Plus 3.0% glucose
IN
SOLUTIONS
Osinolariiy
OF AMMONIUM
(mOd!)
CHLORIDE
Time
for
PLUS
complete
GLUCOSE
hemolyei.
chloride
289
345
401
457
2 miii. 40 sec.
3 miii. 30 sec.
4 mm. 45 sec.
over 15 mm.
An isosmotic solution of ammonium chloride containing,
in addition, 3.0 gm. of glucose per 100 ml. will not cause intravascular
hemolysis.
Although the solution is markedly hypertonic
to plasma,
the shift of water between cells and extracellular
compartments
will
be minimal since the ammonium salt enters the cells.
Disodium Hydrogen Phosphate
This salt is marketed
either as the anhydrous
compound, which
may contain up to 4 per cent moisture,
or as the dodecahydrate,
Na2HPO4’12H20.
The latter is markedly efflorescent. It is stated in
several standard
textbooks that the dodecahydrate
loses water on
exposure to air and is converted to the dihydrate
(S#{216}rensen’s
salt).
In several attempts, the author was not able to confirm this statement. The only reliable procedure
to avoid the complication
of
variable water of crystallization,
is use of the anhydrous
salt prepared in the following way. After recrystallization
from water, the
hydrate is dried in vacuo at room temperature,
first over concen-
Vol. 7, No. 2
1961
OSMOLARITY OF HUMAN SERUM
161
trated sulfuric acid and then over phosphorus
pentoxide.
This takes
several days. When almost all the water of crystallization
has been
removed, the salt is finally dried in an electric oven at 120#{176},
with
occasional stirring to prevent caking. The salt is quite stable at this
temperature;
it does not begin to decompose until 230#{176}.
The anhydrous salt is slightly hygroscopic.
If the attempt to drive off water of crystallization
directly by
heating is made, the dodecahydrate
melts and loses water, forming
a hard glassy mass that is very difficult to handle.
The isosmotic concentration
was found to be 1.792 gm. Na2HPO4
per 100 ml., and the pH of this solution lies between 9.07 and 9.10.
The solubility of the salt decreases rapidly with fall of temperature,
and at 0#{176}
it is approximately
1.75 gm. of Na2HPO4 per 100 ml. As a
result, some difficulty was occasionally
encountered
in the measurement of the osmolarity
of the more concentrated
solutions used.
Solutions of this salt must be protected from the atmosphere.
Sodium Dihydrogen Phosphate
The common form of this salt is the dihydrate.
A good crystalline
specimen contains the theoretical percentage
of water, 23.1 per cent.
This hydrate is, however, slightly efflorescent, and the percentage
of water of crystallization
should be determined
for each specimen.
In the study of this salt, all solutions were made up from the anhydrous salt prepared
in exactly the same way as in the method
described above for the disodium salt. Sodium dihydrogen
phosphate does not begin to decompose until 190#{176}.
If attempts to dry the
dihydrate directly by heating are made, the salt decrepitates,
melts,
loses water, and also forms a hard glassy mass. The anhydrous
salt
is not noticeably
hygroscopic.
The isosmotic
concentration
was
found to be 1.927 gm. of NaH2PO4 per 100 ml., which is equivalent to
2.505 gm. of the dihydrate
per 100 ml. The pH of this solution lies
between 4.33 and 4.35.
Sodium Citrate
The common sodium citrate, the trisodium salt, crystallizes
as tile
dihydrate, Na3C6H5O7.2H20. This hydrate is exceedingly stable. Indeed, in order to determine the percentage of water of crystallization
(12.2 per cent), the dihydrate
must be heated to 150#{176}.
There is no
tendency to effioresce.
The isosmotic concentration
of this salt was found to be 3.06 gm.
of Na3C5H5O1.2H20 per 100 ml., and the pH, 8.60. This concentration is different from the usually quoted “3.8% sodium citrate.”
Clinical Chemistry
HENDRY
162
Apart from the dihydrate,
sodium citrate also exists as Na3C6H5O7
51120 and as 2Na3C6H5O7.111120 (or Na3C6H5O7.5#{189}H2O),and the
conventional
“3.8% sodium citrate” presumably
refers to one or the
other of these higher hydrates.
A 3.80% solution of the dihydratethe one commonly used-is
hypertonic
and has an osmolarity of 352
mOsM. The tonicity of this solution is 1.22 (normal human serum,
1.00).
Phosphate Buffer Solutions
One of the most commonly used buffer solutions consists of varying mixtures of M/15 Na2HPO4.2112O and M/15 KH2PO4. This buffer
has several disadvantages:
(1) it is difficult to prepare Na2HPO4
21120 containing
the correct percentage
(20.2) of water of crystallization (mentioned
above); (2) in systems containing
mammalian
cells it is not advisable to have a high extracellular
concentration
of
potassium,
and M/15 potassium dihydrogen
phosphate contains 66.7
mEq. of potassium
per liter; (3) neither of the components
of the
solution is isosmotic with human plasma, and, indeed, both are very
hypotonic (Table 5); (4) the components have differ1ent osmolarities,
and consequently,
osmolarity of mixtures of them sll vary from one
buffer to another; and (5) the weights of these two salts required to
make up 1 L. of M/15 solution are frequently
misquoted
in the
literature.
The osmolarities
of M/15 solutions of these salts are shown in
Table 5.
Table
5. OSMOLARITY or M/15
PHOSPHATE
Concentration
(gm./L.)
Norm al human serum
M/15 Na2HPO4
M/15 KH2PO4
‘Equivalent
Oem ohwiiy
(mOaM)
289
163
124
9.465*
9.073
to 11.867 gm. of Na2HPO4.2H20
SOLUTIONS
Ton icily
1.00
0.565
0.43
per liter.
Sodium Phosphate Buffers
In a previous publication,
I have shown that when dilute solutions
of disodium hydrogen phosphate
and sodium dihydrogen
phosphate
of the same osmotic pressure
are mixed in any proportion,
the
osmotic pressure remains constant while the pH varies from 4.3 to
9.1 (8). The solutions used at that time were hypotonic with respect
to plasma since the object of the experiments
was to study the varia-
Vol. 7, No. 2, 1961
163
OSMOLARITY OF HUMAN SERUM
tion in hemolysis of erythrocytes
at different pH values, but without
any osmotic complications.
The solutions actually used had a tonicity
in the region of 0.35 (9). A tentative effort was made to determine
the concentrations
of the two sodium phosphates
that would be
isosmotic with human plasma, but this had to be done by extrapolation, and it was made on the erroneous
assumption
that solutions
isosmotic with plasma would have the same freezing point as a solution containing
0.85 gm. of sodium chloride per 100 ml.-a
concentration that is much too low. The calculated results were not very
accurate.
However, such a problem can be tackled easily and accurately by means of the Fiske Osmometer.
The results observed by
use of this means are shown in Table 6.
The graph obtained when the volume of one of these solutions is
plotted against the corresponding
pH is the typical sigmoid curve;
from it, a buffer of any pH between 4.4 and 9.0 can be made up.
The deviations
in the observed osmolarity
will be discussed below.
Table
Volume
(ml.)
NoR ,P04
6.
SODIUM
PHOSPHATE
BUFFER SOLUTIONS*
Oamolarity
NaHPO4
0.00
5.00
10.00
20.00
(mOd!)
pif
100.00
95.00
90.00
80.00
9.09
7.89
7.56
7.20
289
288
288
287
30.00
70.00
6.91
40.00
60.00
6.79
50.00
50.00
6.61
286
60.00
70.00
40.00
30.00
75.00
80.00
90.00
100.00
25.00
20.00
10.00
0.00
6.42
6.22
6.10
5.97
5.60
4.34
286
286
287
288
289
289
*Stock
isosmotie
solutions:
19.27
gm. of NaH9PO4
per liter,
286
-
and 17.92
286
gm. of Na,HPO4
per liter.
Discussion
In cases in which chemical solutions must be administered
intravenously,
it is preferable
that they be given in a concentration
isosmotic with body fluids since this will cause minimal osmotic disturbance to the blood and tissue cells. However, small variations
in
the osmolarity
are of little account. For in vitro work, it is important to cause as little osmotic disturbance
as possible in experimental cells.
164
HENDRY
Clinical
Chemistry
The variation in osmolarity when isosmotic solutions of disodium
hydrogen
phosphate
and sodium dihydrogen
phosphate
are mixed
(Table 6) is interesting.
It will have been observed that the change
is regular in pattern.
In a study reported previously
(8), when very
hypotonic solutions were being used, this change was not observed
when the osmotic pressure was measured
either by the Beckmann
thermometer
or by the electrical conductivity.
Its occurrence in the
present investigation
is presumably
related to the increased concentrations
employed.
The most probable explanation
is as follows.
Sodium dihydrogen
phosphate ionizes according to the equation:
NaH2PO4
with the further
ionization,
H2P04-
When disodium hydrogen
sodium ion concentration
disodium salt ionizes:
Na2HPO4
Na+
+ H2P04-
to a smaller
H
)
extent:
+
HPO4=
phosphate
is added to this solution,
will be increased,
relatively,
since
‘2
Na
(1)
+ HPO4=
(2)
the
the
(3)
The total sodium concentration
of isosmotic sodium dihydrogen phosphate solution is 160 mEq./L., and of isosmotic disodium hydrogen
phosphate,
252 mEq./L.
The ionic sodium concentrations
will be in
approximately
the same ratio. On addition of the disodium salt, the
increase in sodium ion concentration
will force reversal of Eq. 1,
thus causing a decrease in, the total ionic concentration
and a slight
drop in the observed osmolarity.
The combination
of hydrogen and
hydroxyl ions to form water will have the same effect, but quantitatively this reaction cannot account for the observed change.
The
drop in osmolarity
of 3 mOsM introduces
a maximal error of 1
per cent.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Scratchard,
G., and Prentiss, S. S., J. Am. Chem. Soc. 55, 4355 (1933).
Culbert, B. W., J. Biol. Chem. 109, 547 (1935>.
Benham, 0. H., Duke-Elder, W. S., and Hodgson,
T. H., J. Physiol.
92, 355 (1938).
Lifeon, N., J. BIOZ. Cheni. 152, 659 (1944).
Olmstead, E. G., and Roth, D. A., Am. J. Med. Sci. 233, 392 (1957).
Hill, A. V., Proc. Eojai Soc. London, Series A 127, 9 (1930).
Jacobs, M. H., and Stewart, D. R., /. Cellular Corny. Physiol. 7, 351 (1936).
Hendry, E. B., Edinburgh.
Med.
J. 55, 142 (1948).
Hendry, E. B., Edinburgh
Med.
J. 55, 427 (1948).
International
Critical Tables, 4, 258 (1928).