Clinical Science (1970) 39, 77-87.
T H E ELECTROLYTES O F ALKALINE H U M A N GASTRIC
JUICE
J. R. C. G A R D H A M
AND
M. HOBSLEY
Department of Surgical Studies, The Middlesex Hospital, London
(Received 9 July 1969)
SUMMARY
1. The electrolyteswere measured in alkaline gastric juice from eleven subjectswith
various gastric disorders; seven of these subjects had histamine-fast anacidity. The
aim was to investigate the relation between alkaline gastric aspirates and the alkaline
component predicted by the two-component hypothesis of gastric secretion.
2. The electrolyte concentrations Wered from one subject to another: decreasing
concentrations of sodium, chloride and alkali were correlated with increasing potassium concentrations.
3. The observed variations were completely explicable in terms of dilution of gastric
juice by saliva. It is therefore concluded that the uncontaminated gastric juice was
similar in all subjects.
4. The purest specimens of alkaline gastric juice closely resembled the alkaline
component predicted by the two-component hypothesis. It is unlikely that diffusion
in exchange for primary acid secretion could have produced the same results.
Two main alternatives have been advanced to explain the electrolyte composition of human
gastric juice. The two-component hypothesis (Hollander, 1932) proposed that gastric juice
consists of separate acid and alkaline components. A quantitative statement of this hypothesis
(Makhlouf, McManus & Card, 1966) defined the two components, and the validity of this
approach in a number of different human subjects has since been conlirmed (Hobsley &
Men, 1970). By contrast, it has been claimed that a primary acid secretion is altered by backdiffusion of hydrogen ions in exchange for sodium ions (Teorell, 1947), and it has been demonstrated that this process can occur under experimental conditions, (Davenport, Warner &
Code, 1964; Chapman, Werther & Janowitz, 1968).
The majority of previous work has been concerned with the electrolytes in acid gastric juice.
By contrast there is very little information about the electrolytes in alkaline human gastric
juice. Alkaline juice can always be aspirated from patients with pernicious anaemia (Callender,
Correspondence: Dr J. R. C. Gardham, Department of Surgical Studies, The Middlesex Hospital, London,
w.l.
77
78
J. R. C. Gardham and M. Hobsley
Retief & Witts, 1960), and can occasionally be aspirated, under resting circumstances, from
many other subjects.
It seemed that a study of the electrolytes in alkaline juice might provide new and direct
evidence of the existence and composition of the alkaline component. This paper reports data
from a small series of subjects who secreted alkaline gastric juice, and demonstrates that the
variations in electrolyte concentrations followed a distinct pattern. The interpretation is
discussed, and it is shown that such a pattern would be produced by a mixture of variable
proportions of saliva with an alkaline component of constant composition.
M A T E R I A L S A N D METHODS
The results of tests on eleven subjects are reported. The clinical diagnoses are indicated in
Table 1. Seven of the subjects had total anacidity throughout the tests : the remaining four had
basal anacidity, but did secrete acid in response to stimulation. The results reported in this
paper are only those from alkaline specimens (pH >7.0).
After overnight fasting, a nasogastric tube was passed into the stomach and the position
checked radiologically. With the subject in a semi-recumbent position, the stomach was emptied
and the initial aspirate discarded. Thereafter, aspiration was by continuous mechanical suction, aided by frequent air insufflation to clear the tube. Samples were collected at 15-min
intervals. Each test included a minimum of three consecutive alkaline samples without visible
bile staining. In eight tests intravenous histamine diphosphate (0.04 mg h-’ kg-’) was given
by a slow injection pump (C. H. Palmer Ltd.) into the tubing of a continuous infusion of 5%
Dextrose. In one test ametazole hydrochloride (‘Histalog’) 50 mg was given subcutaneously.
In the remaining two tests no gastric stimulant was given as the subjects were already known
to have histamine-fast anacidity, and gave their permission for the test to be conducted for
research purposes.
In all tests phenol red 0.15% w/v solution was instilled into the upper part of the stomach at
a rate of 4.8 ml/h by a Palmer pump via a fine vinyl tube, and the measurements of dye in the
samples were used to estimate the completeness of aspiration (Hobsley & Men, 1966, 1969).
Four subjects were asked during the test to expectorate small samples of saliva for analysis,
but in no test was any attempt made to prevent swallowing of saliva. The reasons for this
decision are set out in the discussion section. In the interpretation of the results, the presence of
saliva has been taken into account.
Analyses were performed as follows: pH on a direct reading pH meter (Electronic Instruments Ltd., Model 27A, fitted with a Jena dual micro-electrode) and titratable alkalinity with
O.OIM-HCIto pH 7.0, or to the end point of phenol red. Sodium and potassium were measured
on a flame photometer (Evans Electroselenium Ltd., Mark 11) and chloride on a potentiometer
(Evans Electroselenium Ltd., chloride meter). Phenol red concentration was measured on a
spectrophotometer (Unicam Instruments Ltd., Model S.P. 600), with readings at wavelengths
of 550 and 410 mp incorporating a correction for extinction by blood in the samples (Crawford
& Hobsley, 1968).
Calculation of results
The concentration of sodium, potassium, chloride, and titratable alkali in a sample was
multiplied by the volume to give the quantity of each in mEq, in each single sample. The sum of
Electrolytes of alkaline gastric juice
79
the quantities in successive samples was then divided by the total volume, and the result expressed as the mean concentration in mEq/l for the test. The phenol red marker dye contributed
small amounts to the volume, sodium and alkali in the aspirated fluid, and therefore, after
estimating the volume of phenol red in each sample, appropriate subtractions were made
before the final calculation of mean concentration. In addition to the measured volumes of
aspirate, volumes corrected for incomplete aspiration by the observed recovery rate of phenol
red are also given. The validity of this correction is discussed below.
RESULTS
The results are summarized in Table 1.
Electrolyte concentrations
The subjects have been arranged for convenience by the mean sodium concentration of the
gastric juice, which ranged from 14648 mEq/l. The ranges of mean concentrations for the
other electrolytes were : potassium 26.7-1 1.1 mEq/l, chloride 147-56 mEq/l, and titratable
alkali 18-3-3.3 mEq/l. In individual subjects, successive specimens closely resembled one
another, and the use of mean concentrations does not conceal any appreciable variations.
Examination of correlations between any pair of the four electrolyte concentrations resulted
in six regression equations. Five of these six pairs had correlation coefficients which were
statistically significant: the regression equations for these pairs of electrolytes are given in Figs.
1-5, and the regression lines are drawn as solid lines.
“a’] vs.
[Alkali] vs.
[Alkali] vs.
[K’] vs.
“a+] vs.
[Cl-]
[Cl-]
“a’]
[Cl-]
[K+]
:r
:r
:r
:r
=
+0.970 : P<O.OOl (Fig. 1)
(Fig. 2)
(Fig. 3)
(Fig. 4)
(Fig. 5)
= +0.677 : P<0*05
= +0.610 : P<0.05
= -0.759 : P<O.O1
: r = -0.706 : P<0*05
The one remaining correlation was not statistically significant at conventional levels.
[Alkali] vs. [K’] : r
=
-0.557 : 0.05<P<O.l
The three correlations in which potassium was included were negative, the remainder were
positive.
Volume
The aspirated volumes varied from 0.5 to 15.8 m1/15 min, and the corrected volumes from
3.4 to 21.0 m1/15 min. The volumes did not correlate significantly with the electrolyte concentrations.
Response to histamine or ‘Histalog’
Nine subjects received histamine or ‘Histalog’ during these tests, and in five of these there
was no consequent change in the volume, pH, or electrolyte concentrations of the alkaline juice. The remaining four subjects, who had alkaline basal juice, later produced acid in
response to histamine. This paper reports only the results from the alkaline samples.
P
F
M
F
75
64
61
67
59
76
63
61
72
64
2
3
4
5
6
7
8
9
10
11
F
F
F
Pernicious
anaemia
Atrophic
gastritis
Pernicious
anaemia
Gastric
ulcer
Gastric
ulcer
Duodenal
ulcer
Pernicious
anaemia
Pernicious
anaemia
Atrophic
gastritis
Pernicious
anaemia
Pernicious
anaemia
Diagnosis
Mean values
M
F
M
F
M
73
Sex
1
No.
Subject Age
7.4
7.9
7.3
9.3
14.9
3.4
7.0
3.O
0.5
4
4
12
5.9
O*
0
-
* Negative response to histamine during previous test.
4
-
7.8
11.5
7.7
8
0
10.4
8.2
21 .o
1.9
4
O*
5.9
64.0
-
17.5
15.8
3
31.5
79.5
17.0
91.9
8.1
4.3
65.6
26.7
48.0
4.1
69.5
20.4
3.5
56.0
51.0
3.3
66.9
20.7
19.8
6.1
11.1
79.5
80.0
19.8
14.6
55.8
60.8
61.0
67.0
-
14.8
11.1
3
27
7.3
11.3
14.8 104
84.5
8.0
5.0
5.0
3
1
92.0
18.3
117
15.9
108
8.1
4.3
2.4
10
0
11.3
9.0
11.1 133
131
-
5.4
4.2
10.4
Titratable
alkali
6
0
0
11.9 147
C1
146
K
(mEq/l)
-
Na
Mean concentrations
6.8
Maximum
PH
6.4
Corrected
volume
(ml/l5 min)
8
Acid
secretion Number of Aspirated
after
alkaline
volume
histamine samples (m1/15min)
(mEq/h)
TABLE1. Composition of alkaline gastric juice in eleven patients
81
Electrolytes of alkaline gastric juice
160-
“a+]=
1.10 [Cl-] -20.9
[Cl-]= 234-l-0.859 [.NO+]
I40
120.
p Alkaline
Component
0
-
100
a
\
U
-$
0
00.
0
I
0
0
0
60.
I‘.
40.
20
l
l
l
l
20
0
l
l
40
l
l
l
80
60
l
l
l l l l l ~
120 140 I60
100
No+ ( mEq/l)
FIG.1. Regression of sodium and chloride concentrations in allcaline gastric aspirate (solid lines),
compared with a hypothetical mixture (interrupted line) between pure alkaline component of
gastric juice and pure saliva in varying proportions.
25-
24 -
[Cl-] -1.45
(2)[Alkali]=O.104
/
22 -
/
/
20 -
-
2
L
._
g
a
Alkaline
component
/
/
/
18-
16-
/
141210-
/
/
/
8-
/
6-
/
d
-
Saliva
2-
I
I
I
I
I
I
1
I
I
l
l
[
J. R. C. Gardham and M . Hobsley
82
'
,
/
(2) [Alkali] =1.48+0.083[Na+]
22
Alkaline
component
@
,
( 1 ) [No+] =43.4 t 4 . 5 0 [Alkali]
2
I
0
I
I
26 -
l
l
l
I
I
I
1
00
100
(2)
I
I
120
I
140
]
160
[ K'] -28.4-0.125 [Cl-]
\
\Y\*
22
-
-
-
l
60
@
(1)
i
24 -
-<w"
l
40
20
&
\
20
I6
1614-
E
12-
"y
10-
a\ \
0
\
8-
\
@Alkaline
component
642-
I
I
I
I
I
I
l
l
l
l
l
l
I
I
I
1
)
I
:
83
Electrolytes of alkaline gastric juice
These values are compatible with other reports of salivary composition at low rates of flow
(Thaysen, Thorn & Schwartz, 1954), and confirm the high potassium content of saliva.
Saliva
(I) [Not] = 162-4.86 [K’]
(2)[ K t ] =25-2-0403
[ No+]
(11
+
E
16
14
12
:
2
bAlkaline
component
20
40
60
80
100
120
140
I60
Nat(rnEq/l)
FIG. 5. Regression of potassium and sodium concentrations, compared with a hypothetical
of alkaline component and saliva.
DISCUSSION
The electrolyte concentrations in alkaline gastric contents varied from one subject to another,
and were different from those of saliva. A pattern of interrelation between the electrolytes
emerged in the six regression equations. It will be shown that this pattern closely resembled
that which would be expected if a pure alkaline gastric juice of constant composition were
mixed in varying proportions with saliva.
Before discussing the results in detail, it is necessary to comment on possible inaccuracies in
the measurements of titratable alkali and of volume. Specimens were collected in aerobic
conditions which allowed evolution of COz. Any evolution of C 0 2 leaves in solution a
corresponding amount of hydroxyl ions, so that the measurement of titratable alkali corresponds with bicarbonate in the original fluid. The volumes available for titration were often
very small (e.g. 0.5 ml) and the measurement of titratable alkali was probably less accurate
than that of the other electrolytes. This lack of accuracy may be reflected in the lesser values
for those correlations involving alkali.
The volume of stomach contents aspirated is likely to be incomplete (Hobsley & Silen, 1969).
For this reason the phenol red marker dye technique was used to correct the volume of aspirate
to the true volume of gastric secretion. There is doubt as to whether phenol red is an adequate
marker in these circumstances. It has been shown (Bloom, Jacobson & Grossman, 1967), that
in an alkaline medium recovery of phenol red is not complete, but the proportion recovered in
their experiments was 90-95%. Therefore the error from this source is probably not gross. In
addition, adequate mixing may not have occurred in the very small volumes of gastric juice in
some of these tests.
J. R. C. Gardham and M . Hobsley
For these reasons, and because in some tests a very large correction for incomplete recovery
was necessary, there must be some doubt about the validity of the corrected volume results.
Correlations interpreted as the efect of saliva
The subjects were allowed to swallow saliva, because it was judged to be impracticable
entirely to prevent this. The most effective method of preventing salivary contamination is
probably that used by Makhlouf et al. (1966), namely the use of frequently changed dental
pledgets in the floor of the mouth. Whilst this procedure was suitable for Makhlouf et al.’s two
practised subjects, it was considered to be too arduous for elderly subjects in this study. Other
methods, such as expectoration, or use of a dental sucker, have been ineffective in our hands.
Comparison of our results with those of Lambling, Bernier, Badoz-Lambling & Dives, (1956)
supports the conclusion that such methods are not effective. Lambling and his colleagues
studied seven patients with pernicious anaemia and attempted to exclude saliva by expectoration, or by a dental sucker. They found the electrolyte concentration of the aspirates to be
variable, and concluded that alkaline gastric juice was not constant in composition. However
our findings were similar to those of Lambling et al. (1956) (Table 2) and this suggests that their
precautions against swallowing of saliva were not effective.
TABLE
2. Mean electrolyte concentrations in alkaline gastric juice. (The ranges of observed values are
given in parentheses for this series and for Lambling et al., 1956)
1
2
3
4
5
6
MAN
MAN
MAN
DOG
DOG
DOG
Altamirano
(1963)
greater
curvature
Hollander
(1963)
Heidenhain
pouches
Grossman
(1959)
antral
pouches
137
138
133
151
6
4
4
9
117
117
120
145
23
13
8
Lambling et Makhlouf et al.
Present series
al. (1956)
(1966)
eleven subjects seven subjects
calculated
Na
+
K+
CI HCO
80
(48-146)
17
(1 1-27)
92
(56-147)
8
(3-1 8)
94
(64-135)
20
(14-24)
83
(50-113)
18
(1 6 2 2 )
(25*)
Concentrationsin mEq/l (to nearest mEq). * Assumed value.
The varying electrolyte concentrations in the aspirate might be explained on the hypothesis
that the pure alkaline gastric juice was diluted by varying quantities of saliva in different
subjects. This hypothesis can be examinedwith reference to Figs. 1-5. In each of the figures the
two solid lines are the regression lines from our data. Two lines have been reproduced for each
pair of electrolytes because there was no reason to select one electrolyte as the independent
variable. In addition a hypothetical mixture of saliva and alkaline gastric juice is also represen-
Electrolytes of alkaline gastric juice
85
ted in each figure as an interrupted line. The points at either end of this line indicate the composition of pure saliva and of pure alkaline component of gastric juice. The data for saliva
are from the four subjects who expectorated specimens, and the data for the hypothetical pure
alkaline component are taken from Makhlouf et al. (1966). It is notable that those regressions
involving potassium are negative, but that the remainder are positive. This distinction corresponds with the fact that only potassium is present in greater concentration in saliva than in
gastric juice. Thus the concentration of potassium was high when salivary contamination was
maximal, but all other ions were diluted by salivary contamination. In both positive and
negative correlations the lines from the data are closely comparable to the lines from the hypothetical mixture of saliva and pure alkaline component. This resemblance is not a formal
proof that the observed results were due to such a mixture, but does form the basis of a working
hypothesis.
The composition of pure alkaline gastric juice
If the variations in concentrations were due to dilution by saliva, and if the composition of
saliva did not vary, then the pure gastric juice must also have been of approximately the same
composition in the different patients. The assumption of constant composition for saliva
140-
m
m
:::::.:::'
:::<.
@&
...............
...............
.
:
j..l'
..l:
....
::.:I:.:,:
120100 -
-
00-
'
60-
\
s
No
Cl
No
40 20 -
Mokhlouf
(calculated 1
Ave
Da
ge of
'nts
1,2 and3
FIG.6. Comparison of the electrolyte concentrations in the alkaline component of acid gastric
juice (Makhloufet ul., 1966) with the mean of the measured concentrations in subjects 1,2 and 3 of
this series.
secreted at low rates is reasonable (Thaysen et al., 1954). So, despite the variety of pathological
states in the subjects studied, it appears that a characteristic alkaline gastric juice was present
in all subjects.
The exact electrolyte concentrations in pure alkaline gastricjuice cannot be determined from
our results. In Table 2 the data from the present series are compared with other data from the
literature. In our series and that of Lambling et al. (1956) salivary contamination probably
lowered the mean sodium and chloride concentrations,and raised the potassium concentration.
The third column of Table 2 shows the alkaline component calculated by Makhlouf et al.
86
J. R. C. Gardham and M . Hobsley
(1966) for one normal subject tested repeatedly with strict exclusion of saliva. There is a close
resemblance between this estimate and those of our samples which appeared to be least contaminated with saliva (Fig. 6). Makhlouf and his colleagues based their calculations on an
assumed bicarbonate concentration of 25 mEq/l. The difference between this assumption and
the maximum titratable alkali of 18 mEq/l in our series accounts for the divergence from the
predicted regressions in Figs. 2 and 3. The possibility that samples were contaminated with
minute amounts of acid cannot be denied.
Three authors (Altamirano, 1963; Hollander, 1963; Grossman, 1959) have published
results from dogs, and these results are also reproduced in Table 2. In each instance the method
ensured that the alkaline gastric juice was entirely free of saliva.
Alkaline component or diflusion
Do these observations indicate whether the two component hypothesis or the diffusion
theory is the correct physiological concept of gastric secretion?
In favour of the two component hypothesis is the finding that, after allowing for salivary
contamination, the aspirated alkaline gastric juice appeared to be of constant composition, and
resembled the alkaline component which has been predicted from a theoretical basis. The alkaline gastric juice from our subjects, like that from dog pouches, resembled extracellular fluid,
and may have been a simple transudate, as suggested by Hollander (1963). Alternatively this
fluid might have been the result of back-diffusion of a primary acid secretion in exchange for
sodium. Space does not allow a detailed discussion of the theory of Teorell (1947), but his
paper indicated that a number of factors would influence the diffusion process.
These factors include the primary acid concentration, the rate of secretion, the area of
diffusion and the thickness of the membrane. Chapman et al. (1968) have recently shown that
200 ml of acid instilled via a nasogastric tube into the stomachs of four subjects with pernicious
anaemia resulted in diffusion exchange for sodium of approximately 10% of the instilled
acid within 15 min. If diffusion in exchange for acid had been the main factor in our subjects,
it would be necessary to postulate that the interaction of several variable factors combined in
each subject to produce the same composition in the alkaline juice. Therefore, if a primary
acid secretion did occur at very low rates, the results could be explained if diffusion proceeded
to equilibrium in all cases.
In seven of our subjects the pH of the aspirate never fell below 7.0, and in these subjects
there was no overt evidence of ability to produce a primary acid secretion. Teorell also considered the possibility of absence of primary acid secretion and derived an equation for simple
diffusion of NaCl. In the situation of total anacidity this concept does not differ from that of
Hollander (1963).
Therefore, the possible alternatives are either that back-diffusion to the same equilibrium
occurred in each subject, or that the results were due to simple diffusion, or secretion, of an
alkaline juice of k e d composition. It has already been indicated that back diffusion would
depend on the rate of secretion of acid. Calculations of the apparent alkaline component at
various rates of acid secretion (Makhlouf et al., 1966; Hobsley & Silen, 1970) may be considered in conjunction with the present data. The similar compositions of the observed alkaline
juice in the absence of acid, and the derived alkaline component in the presence of acid secreted
at different rates, support the concept of an alkaline component which is not dependent on the
rate of acid secretion.
Electrolytes of alkaline gastric juice
87
ACKNOWLEDGMENTS
We are grateful to Dr G. D. Hadley for permission to study patients under his care, and to
Miss G. Crawford for technical assistance.
REFERENCES
ALTAMIRANO,
M. (1963) Alkaline secretion produced by intra-arterial acetylcholine.Journal of Physiology, 168,
787-803.
BLOOM,
D.S., JACOBSON,
E.D. & GROSSMAN,
M.I. (1967) Validation of dilution indicators. Gastroenterology, 52,
205-210.
CALLENDER,
S.T., RETIEF,F.P. & W m s , L.J. (1960) The augmented histamine test with special reference to
achlorhydria. Gut, 1, 32C336.
CHAPMAN,
M.A., WERTHER,
J.L. & JANOWITZ,
H.D. (1968) Response of the normal and pathological human
gastric mucosa to an instilled acid load. Gastroenterology, 55, 3, 344353.
G. & HOBSLEY,
M. (1968) Spectrophotometric estimation of phenol red in gastric juice in the
CRAWFORD,
presence of blood. Biochemical Journal, 1 0 7 , 2 6 ~ .
DAVENPORT,
H.W., WARNER,
H.A. & CODE,C.F. (1964) Functional significance of gastric mucosal barrier to
sodium. Gastroenterology, 47, 142-1 52.
GROSSMAN,
M.I. (1959) The secretion of the pyloric glands of the dog. Znternationul Congress of Physiological
Scientists, 21st. Buenos Aires Symposium Special Lectures, 226-228.
HOBSLEY,
M. & SILEN,W. (1966) Improved method for studying the rate of human gastric secretory activity
under varying conditions of acid-base balance. Surgical Forum, 27, 324-326.
HOBSLEY,
M. & SILEN,W. (1969) Use of an inert marker (phenol red) to improve accuracy in gastric secretion
studies. Gut, 10, 787-795.
HOBSLEY,
M. & SILEN,W. (1970) The relation between the rate of production of gastric juice and its electrolyte
concentrations. Clinical Science, 39, 61-75.
HOLLANDER,
F. (1932) Studies in gastric secretion. IV. Variations in the chlorine content of gastric juice and
their significance. Journal of Biological Chemistry, 97,2, 585-604.
HOLLANDER,
F. (1963) The electrolyte pattern of gastric mucinous secretions: its implication for cystic fibrosis.
Annals of the New York Academy of Sciences, 106,2, 757-766.
LAMBLING,
A., BERNIER,
J.J., BADOZ-LAMBLING,
J. &DIVES,C. (1956) Contribution a 1’6tude de la physiopathologie de la secretion gastrique histaminique. Semaine des hipituux de Paris, 32, 1677-1744.
MAKHLOUF,
G.M., MCMANUS,
J.P.A. & CARD, W.I. (1966) A quantitative statement of the two-component
hypothesis of gastric secretion. Gastroenterology, 51, 2, 149-171.
TEORELL,
T. (1947) Electrolyte diffusion in relation to the acidity regulation of the gastric juice. Gastroenterology,
9,425-443.
THAYSEN,
J.H., THORN,H.A. & SCHWARTZ,
I.L. (1954) Excretion of sodium, potassium, chloride and carbon
dioxide in human parotid saliva. American Journal of Physiology, 178, 155-159.