Clinical Science (1994) 86,
469-477
(Printed in Great Britain)
469
Glucose polymer in the fluid therapy of acute diarrhoea:
studies in a model of rotavirus infection in neonatal rats
A. V. THILLAINAYAGAM, J. A. DIAS, A. F. M. SALIM, F. H. MOURAD, M. L. CLARK
and M. J. G. FARTHING
Department of Gastroenterology, St Bartholomew's Hospital, London, U.K.
(Received 10 ]une/l4 September 1993; accepted 12 October 1993)
1. Unlike standard glucose-electrolyte oral rehydration solutions, solutions containing polymeric glucose as substrate can significantly reduce stool output, duration of diarrhoea and total oral rehydration
solution requirements. However, neither the underlying mechanisms nor the optimal size and concentration of glucose polymer has been defined.
2. W e have used a model of rotavirus diarrhoea in
neonatal rats to compare the effects on water and
solute absorption of varying the concentration of a
glucose polymer (mean chain length five glucose
residues) in experimental oral rehydration solutions.
Three polymer (P) solutions were compared with
solutions of identical electrolyte content (mrnol/l:
sodium, 60; potassium, 20; chloride, 60; citrate, 10)
containing equivalent amounts of free glucose (G) as
substrate by perfusion of the entire small intestine in
situ. The polymer (9, 18, 36mmol/I; 159, 168,
186mosmol/kg, respectively) and the monomer (45,
90, 180mmol/l; 195, 240 320 mosmol/kg) solutions
were perfused in normal and rotavirus-infected neonatal rats.
3. In normal intestine polymer solutions promoted
greater water absorption [P9, mean 291.4 (SEM
16.4); P18,331.9 (13.1);P36,284.3 (11.8)pl min- g-'1
than their equivalent monomer solutions [G45,, 220.8
(8.4); G90, 240 (21); G180,79.4 (14.5) plmin'-' g-';
P<0.021. In rotavirus-infected intestine, water absorption from all solutions declined, but the fall was
much less pronounced from the polymer solutions
[P9, 232.8 (6); P18, 277.2 (20.5); P36, 166
(18.2) pl min-' g-'1 than from their monomeric
counterparts [G45, 116.7 (25.5); G90, 68.7 (12.4);
G180, 21 (11.6)plmin-' g-'; P<0.005].
4. In both the normal net absorptive state and the net
secretory state induced by rotavirus infection, there
was a striking inverse correlation between net water
absorption and perfusate osmolality ( r = 0.94 and
r= -0.88, respectively; P< 0.05). In rotavirus-infected intestine, increasing the polymer concentration
from 18 to 36mmol/l resulted in a relative fall in
water absorption (P<0.01). The hypertonic solution
G180 was associated with the lowest water absorption
(P<O.Ol). None of the solutions was able to reverse
'
-
rotavirus-induced net secretion of sodium, which was
similar from all solutions, whether polymer- or
monomer-based.
5. These results (i) emphasize the pre-eminence of
hypotonicity among the factors promoting water
absorption from polymer-based oral rehydration solutions in acute diarrhoea, (ii) confirm the adverse
consequence of raising substrate concentration
(whether polymer or monomer) beyond certain limits
and (iii) indicate that the concentration of this glucose polymer yielding the optimum compromise
between substrate availability and low osmolality
may be approximately 9-18 mmol/l.
INTRODUCTION
Acute infectious diarrhoea continues to be a
major cause of morbidity and mortality worldwide.
Rotavirus (RV) is the most commonly isolated
enteropathogen in pre-school children [11, and
accounts for 50% of the episodes of serious diarrhoea and dyhydration. In the U.S.A. each year RV
causes an estimated 2 million episodes of diarrhoea
in children under 2 years of age; nearly 200000 of
these children seek medical attention and 22000 are
hospitalized [2]. RV infection is also becoming the
most important cause of infantile diarrhoea in developing countries [3]. Oral rehydration therapy with
glucose-electrolyte solutions has been the cornerstone of management of acute infectious diarrhoea
for over two decades. It has saved many millions of
lives and has been declared the most important
therapeutic breakthrough of this century [4].
Despite the undisputed therapeutic benefits of
standard oral rehydration solutions (ORS) containing glucose, they do not reduce stool output or
duration of diarrhoea and may actually increase the
purging rate and stool volumes [S, 61.
Cereal-based ORS diminish the severity of the
diarrhoea (total stool output and duration of illness)
and also reduce the overall ORS requirements [7161. The staple cereals, such as rice powder, wheat
and lentils, most commonly used in cereal-based
Key words: acute diarrhoea, fluid therapy, oral rehydration. rotavirus.
Abbreviations: G, free glucose: ORS, oral rehydration solution; P, polymer: PEG, polyethylene glycol; RV, rotavirus.
Correspondence: Professor M.J.G. Farthing, Department of Gastroenterology, S t Bartholomew's Hospital Medical College, Charterhouse Square, London ECI M 6BQ, U.K.
470
A. V. Thillainayagam et al.
ORS, are easily accessible, but the preparation of
such ORS in the home requires time, consumes
energy resources and may be incorrectly performed
[17, 181. Moreover, the thicker consistency of a
cereal-based solution may not be palatable to a
dehydrated child.
Commercially available starch hydrolysates of
defined short-chain polymer composition may be an
alternative substrate [191, particularly in view of
their greater solubility in water. The mechanisms by
which complex carbohydrate ORS enhance water
and electrolyte absorption have not been clearly
defined, but our preliminary work with complex
carbohydrate substrate suggests that low osmolality
is of primary importance [20]. However, neither the
optimum polymer chain length nor the polymer
concentration that would yield the optimal balance
between total glucose availability and hypotonicity,
has been determined. Before ORS based on defined
glucose polymers can be recommended for routine
clinical use, randomized controlled clinical trials are
required. However, this is not a feasible method for
screening possible formulations of polymer ORS
because of the many variables involved. Suitable
animal and human models have been developed for
the preliminary screening of new ORS formulations
and these can be applied to the evaluation of
glucose polymer ORS before clinical trial [21].
RV diarrhoea is associated with lower stool
sodium losses than in cholera [22] and the pathophysiology of the diarrhoea has an osmotic component (due to disaccharide malabsorption) and a
secretory component (due to impaired sodium
chloride absorption). We have investigated the effect
of varying the concentration of a defined glucose
polymer (mean chain length of five glucose molecules) on water and solute absorption in an intestinal perfusion model of RV diarrhoea in the
neonatal rat in vivo. Three concentrations of glucose
polymer (P) were compared with monomer ORS of
identical electrolyte composition containing equivalent concentrations of free glucose (G) as substrate.
METHODS
Materials
Rat RV for our studies was kindly supplied by Dr
S. L. Vonderfecht (Division of Comparative
Medicine, Department of Pathology, Johns Hopkins
University Medical School, Baltimore, MD, U.S.A.).
It was a group B RV known as the infectious
diarrhoea of infant rat virus [23] and is morphologically almost identical with the classical group A
RV responsible for most human RV disease, differing in only one common group antigen, VP6 [24].
‘Maxijul’, a widely available commercial preparation of maltodextrins (Scientific Hospital Supplies), provided a range of short-chain (mean length
five glucose molecules) glucose polymers. Radiolabelled polyethylene glycol ([ 14C]PEG 4000) was
used as a volume marker and was obtained from
Amersham International. RPMI-1640 culture
medium without glutamine was obtained from
Gibco Ltd. Pentobarbitone sodium (veterinary
grade) (Sagatal; 60 mg/ml) was supplied by May and
Baker. All other chemicals were supplied by BDH
Chemicals or Sigma Chemical Co., and were of
Analar R grade.
Preparation of virus inocula
The inocula used to infect the neonatal rats were
prepared by the modified methods of Vonderfecht et
al. [23]. Inocula (0.25ml each) made up from 2 0 4
of intestinal homogenate (containing approximately
107-108 infectious diarrhoea of infant rat virus
particles/ml) mixed with 0.23 ml of water were given
by gavage to 8-day-old neonatal Wistar rats (1215g). The rats were returned to their mother and
allowed to suckle. At 24h after infection, when all
animals had developed diarrhoea, the rats were
killed and the whole intestinal tract from duodenum
to sigmoid colon was removed, pooled and ground
in a tissue grinder for at least 20min. The ground
tissue was mixed with RMPI-164 culture medium
without glutamine to produce a 10% (w/v) suspension. The suspension was centrifuged at 13 OOOg for
5min in a MSE Superspeed 50 centrifuge at room
temperature. After 5 min the tubes were removed
and the supernatant was collected in separate sterile
tubes before being passed through a filter-membrane
(Sartorius; pore size 0.22 pm) using 10 ml sterile
plastic syringes. This material was examined by
electron microscopy (Phillips 300) with negative
staining to confirm the presence of RV-like particles.
The material was then stored in 5OOpl aliquots at
-30°C for later use in infecting subsequent batches
of rats.
Inoculation of the rats
A 5 0 0 ~ 1aliquot of RV suspension was thawed
and made up to 6.25ml with water. Boluses of
0.25ml were then used to inoculate each rat intragastrically. After inoculation neonatal rats were
returned to the mother so that they could suckle for
the next 48 h ad libitum. During the characterization
of this model is was shown that the animals would
begin to lose weight and have profuse watery
diarrhoea within 24h and RV particles could be
detected in the faeces [25, 261.
Intestinal perfusion model of RV diarrhoea
Two hours before starting the intestinal perfusion,
neonatal rats were separated from their mother and
kept in a separate cage. Rats were anaesthetized
with sodium pentobarbitone (4.8 mg/k body weight).
Anaesthesia was maintained as necessary by interval
doses of intramuscular sodium pentobarbitone solution (1.2 mg/kg). Body temperature was maintained
Fluid absorption in a model of rotavirus diarrhoea
through the use of an underlying heat pad. After
mid-line laparotomy the proximal duodenum was
ligated and a polyethlene cannula (internal diameter
0.76 mm) was inserted into the second part of the
duodenum and advanced to the duodeno-jejunal
junction. Another polyethylene tube (internal diameter 1.4mm) was inserted into the distal ileum just
proximal to the ileo-caecal junction. Gentle lavage
with warm isotonic saline (37°C) and air was used
to clear small intestinal contents. The proximal
catheter was then used to infuse the test ORS (37°C)
into the small intestine at a rate of 0.25ml/min.
About 10ml of the test solution was perfused
through the small intestine to remove the saline
used for the lavage procedure, after which there was
an equilibration period of 30 min to ensure that
steady-state conditions were established. After the
equilibrated period, three successive 10min collections were made from the distal cannula for analysis. At the end of the experiment the rats were killed
by exsanguination after an intracardiac injection of
sodium pentobarbitone. The perfused segment of
small intestine was removed and dessicated in a hot
air oven at 100°C for 18h to obtain the dry weight.
Control rats of the same age, which had not been
inoculated with RV were perfused at the same time
(48 h) after sham inoculation.
Rats were perfused at 48 h after inoculation with
RV because previous perfusion experiments using
plasma electrolyte solution (at 12, 18, 24, 48, 72 and
96 h after infection) had shown the secretory state
for water (- 15.2 f7.3 pl min-’ g- mean fSEM) to
be maximal at that time. The mean recovery of
[14C]PEG 4000 was 100.47f0.58%. The PEG concentrations of three successive collections of effluent
always varied <5% about the mean, confirming the
existence of steady-state conditions.
’;
H.p.1.c. of glucose polymers
H.p.1.c. analysis of the glucose polymer was performed by applying 2 0 0 ~ 1aliquots of a 0.26%
solution of glucose polymer (Maxijul) by injection
to a Dextropak (Waters Radial-PAK) column with
water as the mobile phase. Detection of the various
oligomers (Knauer Differential Refractometer) was
performed by a Shimadzu C-RIB integrator which
measured peak area under the elution pea.ks [27].
Relative concentrations of oligomers were quantified
as a percentage of the total glucose polymer measured. Samples were analysed in duplicate, and
quality control was ensured by regular analysis of a
standard solution (0.13mg/ml) of glucose monohydrate BP (BN P1071). The h.p.1.c. analysis showed
that the component oligomers varied between G1
and G9, with 81% comprising G2-G6 (Fig. 1).
GI
G2
47 I
G3
G4
G5
G6 G l
G8
G9 GI0
Oligomer chain length
Fig. 1. Chain length distribution within the glucose polymer mixture (‘Maxijul’) by h.p.1.c. analysis. Values are shown as meansf SEM.
Table 1. Composition of oral ORS perfused
Solute
(mmol/l)
Glucose monomer.. .G45
Sodium
Potassium
Chloride
Citrate
Glucose
Calculated osmolality
(mormol/kg)
Solute
(mmol/l)
G90
GI80
20
60
20
60
60
60
20
60
10
45
10
90
10
180
I95
240
330
PI8
36
60
Glucose polymer.. .P9
Sodium
Potassium
Chloride
Citrate
Polymer*
Calculated osmolality
(mosmol/kg)
60
9
60
20
60
10
18
I59
I68
I86
60
20
60
10
20
60
10
36
*Glucose polymer of mean chain length five glucose molecules.
studied, 9 mmol/l (P9), 18mmol/l (P 18) and
36mmol/l (P36), and each was matched by one of
three solutions (G45, G90 and G180) containing the
equivalent amount of free glucose as substrate
(Table 1).
Laboratory analysis of perfusate
The effluent sodium and potassium concentrations were measured by flame photometry
(Instrument Laboratories 943), the chloride concentration using a Corning Chloride Analyser 925,
the bicarbonate concentration using a Corning CO,
meter, and the glucose concentration with a
Beckman Glucose Analyser 2. Osmolality was analysed by the vapour pressure technique using a
Wescor 5500 osmometer. [14C]PEG concentrations
were measured in triplicate by liquid scintillation
spctroscopy in a LKB Wallac 1219 Rackbeta liquid
scintillation counter.
Composition of experimental ORS
Hydrolysis of glucose polymer
All ORS studied were of identical electrolyte
composition. Three polymer concentrations were
The polymer remaining in the intestinal effluent
was hydrolysed by combining 15Opl aliquots of
472
A. V. Thillainayagarn et al.
eMuent with 9 0 0 ~ 1of 2mol/l HCl and boiling in a
water bath for 2 h before quenching the reaction
with 4 5 0 ~ 1of KOH solution. Acid hydrolysis is a
well-established biochemical method of carbohydrate hydrolysis [28]. We validated it against an
enzymic method employing glucoamylase [29] and
found our method of acid hydrolysis to be effective
at 2 h in producing complete hydrolysis of the
glucose polymer. The extent of substrate hydrolysis
was determined by expressing the luminal disappearance rate as a percentage of the substrate
infusion rate and was calculated according to a
formula appropriate for a single-pass perfusion
system [29].
Calculating net water and solute movement
U
G45
P9
G90
PI8
GI80 P36
Experimental ORS
Fig. 2. Net water movement in normal intestine. Data are presented
as means & SEM. Statistical significance: *P=O.Ol, **P <0.005.
r4O01
M
Water or solute movement was expressed as the
mean result of three eMuent samples in pl
mi n - l g - ' or mmol m i n - l g - ' dry weight of small
intestine respectively. A negative value indicates net
secretion and a positive value indicates net absorption. Standard formulae for this single-pass
perfusion method were used to caluclate solute and
water transport [30].
Experimental ORS
Statistical methods
For each of the variables the following pairs of
perfusates were compared separately for normal and
RV-infected intestine: (i) P9 and G45, (ii) P18 and
G90, (iii) P36 and G180. Normally distributed variables were analysed using an unpaired t-test and
variables which proved to be non-parametric using
the Wilcoxon two-sample test. For each variable
and for each of the six perfusates, the results for the
normal and RV-infected intestine were compared
using unpaired t-tests for normally distributed variables and the Wilcoxon two-sample test otherwise.
Finally, the three glucose polymer formulations (P9,
P18 and P36) within each intestinal state were
compared using one-way analysis of variance for
normally distributed variables and the KruskalWallis test for the others. Where a significant overall difference was encountered, more sensitive pairwise comparisons were made using Duncan's multiple range test for the normally distributed
variables and the Wilcoxon two-sample test for the
others.
RESULTS
Fig. 3. Net water movement in RV-infected intestine. Data are presented as means +SEM. Statistical significance: *P =0.05, **P=O.OOOI.
C45
P9
G90
PIE
GI80
P36
Experimental ORS
Fig. 4. Net glucose movement in normal intestine, Data are presented
as means &SEM. Statistical significance: *P=O.OOl.
both produced greater water absorption than
(P<0.007). In both normal (r=-0.94) and
infected intestine ( r = - 0.88) water absorption
inversely correlated with osmolality ( P = 0.006
P = 0.02, respectively).
P36
RVwas
and
Water transport
Glucose transport
In normal and RV-infected intestine all three
polymer solutions produced more net water absorption ( P <0.01) than their equivalent monomer solutions (Figs. 2 and 3). Water absorption from the
three polymer solutions was similar in normal intestine. In RV-infected intestine, although water
absorption was similar from P9 and P18 (P=0.54),
In normal and RV-infected intestine, glucose
absorption was greater from P9 ( P < 0.03) than from
its equivalent monomer solution G45 (Figs. 4 and
5). In normal intestine, glucose absorption from P18
was similar to that from its equivalent monomer
solution, but in RV-infected intestine there was
substantially greater glucose absorption from P18
Fluid absorption in li model of rotavirus diarrhoea
473
Potassium transport
T
G45
P9
G90 PI8
Experimental ORS
GI80
P36
Fig. 5. Net glucose movement in RV-infected intestine. Data are
presented as means fSEM. Statistical significance: *P=0.02, **P=O.W.
There was net potassium absorption with all
solutions in normal intestine, but absorption was
reduced in RV-infected intestine. For P9 and G45
potassium absorption was similar in normal and
RV-infected intestine. In normal intestine P18 and
G90 led to similar potassium absorption, but RVinfected intestine P18 was associated with higher
potassium absorption than G90 ( P < 0.02). In both
normal and RV-infected intestine P36 produced
greater potassium absorption than its equivalent
monomer ORS G180 (P<O.O5). The results for net
potassium movement are illustrated in Table 2.
Chloride transport
j
$ -30
I
G45
P9
G90
PI8
Experimental ORS
GI80
P36
Fig. 6. Net sodium movement in normal intestine. Data are presented
as means f SEM.
For P9 and P18, chloride movement in both
normal and RV-infected intestine was similar from
each of their equivalent monomer solutions (G45
and G90, respectively). In normal intestine G180
produced modest net secretion of chloride in contrast to the substantial net chloride absorption from
its equivalent polymer ORS P36 (P<O.Ol). In secreting intestine, however, both P36 and G180 produced substantial net chloride secretion of a similar
level (Table 2).
Effluent osmolality
Ef€luent osmolality increased as substrate concentration was increased whether polymer or
monomer. The osmolality remained hypertonic with
G180 in normal and RV-infected intestine, but
remained hypotonic when the solution perfused was
hypotonic, whether it contained glucose polymer or
not.
35
1
-30
Glucose polymer hydrolysis
than from G90 (P=O.O4). P36 and its equivalent
monomer solution G180 produced similar glucose
absoprtion in both normal and RV-infected intestine.
Table 3 summarizes the extent to which the
glucose polymer was hydrolysed at the three different concentrations. In normal and RV-infected
intestine glucose polymer was hydrolysed to a
greater extent in P9 than P18 or P36 (P<0.003). In
normal but not in RV-infected intestine the percentage hydrolysis of polymer was greater from P18
than P36 (P<O.Ol). The extent of polymer hydrolysis was greater in normal intestine than in RVinfected intestine for P9 (P=0.006), but was similar
in both situations for P18 and P36.
Sodium transport
DISCUSSION
There was modest net sodium secretion from all
the ORS in normal intestine. In RV-infected intestine there was a substantial increase in net
sodium secretion from all solutions, whether
monomer or polymer-based, although there was no
difference between matched G and P solutions in
either state (Figs. 6 and 7).
The mucosal damage which follows RV infection
of the small intestine results in pathophysiological
changes which are reflected clinically by acute
watery diarrhoea. There is villous shortening and a
compensatory rise in the crypt cell proliferation rate
at the base of the villous and in the crypt [31]. The
model of RV infection used in our experiments was
G45
P9
G90
PI8
Experimental ORS
GI80
P36
Fig. 7. Net sodium movement in RV-infected intestine. Data are
presented as means f SEM.
A. V. Thillainayagam et al.
474
Table 2. Net electrolyte movement and effluent osmolality. Values are mean (SEM). Statistical significance: *P ~0.05,
t P < 0.01
compared with equivalent glucose ORS. Abbreviation: ND, not determined.
ORS
Sodium
movement
(mmol min- ' g -
Normal intestine
Polymer-ORS
P9
PI8
P36
Glucos~ORS
G45
G90
GI80
I)
Potassium
movement
(mmol min - I g-
I)
Chloride
movement
(mmol min- ' g - ')
Effluent
osmolality
(mosmol/kg)
-4.7 (1.8)
- 3.5 (2.4)
- I .6 (0.7)
5.3 (0.8)
8.0 (0.3)
6.8 (0.7)*
3.6(1.5)
4.8 (I.5)
I I .4 (I.O)t
201 (2)
204(2)*
208 (3) t
- 2. I (0.9)
O.l(l.1)
- 5.4(2.5)
6.9 (0. I)
8.0 (0.3)
3.7 (I3)
2.4 (I.2)
6.l(l.2)
- 1.8( I .9)
234 (3)
270 (I)
318(4)
-21.1 (2.9)
- 15.2(2.8)
-15.9(1.7)
S.Z(O.3)
4.7 (O.S)*
3.7(0.4)t
-6.3(1.3)
- 2.0( I . I )
- 5.3 (I .O)
ND
ND
232 (2)t
- 15.8(3.2)
3.l(l.O)
1.7(0.7)
1.6(0.9)
-5.2( I .2)
0.E(0.5)
-7.9( I.3)
227 (4)
ND
316(2)
RV intestine infected
Polymer ORS
P9
PIE
P36
Glucose ORS
G45
G90
GI80
- 9.6 (I .4)
- I5.8(2.5)
Table 3. Extent of glucose polymer hydrolysis. Values are mean (SEM).
Statistical significance: *P <0.05 for P9 compared PI8 and P36; tP<O.OI
for normal intestine compared with RV-infected intestine; $P<O.OI for PI8
compared with P36.
Extent of hydrolysis (%)
Polymer ORS
Normal intestine
RV-infected intestine
21 8(1 3)*t
II 8(1 0):
8 5 (0 6)
I S 9(0 4)*
94(04)
96(04)
~~
P9
PI8
P36
developed with the feasibility of intestinal perfusion
in situ in mind [25, 261. As well as offering the
advantage that the intestine is perfused with its
blood supply intact, the model closely mimics
human R V infection in that maximal villous shortening is present by 24h with progressive recovery
occurring over the ensuing 48-72 h.
In the present study the net secretory state for
water and sodium was reversed by all the ORS
whether they were based on monomeric or polymeric glucose. However, in both normal and RV-infected intestine, polymer solutions consistently promoted more water absorption than their monomeric
equivalents. Increasing glucose concentration in
ORS beyond certain limits is known to have
adverse consequences with respect to water absorption [32]. Not unexpectedly, therefore, the highsubstrate monomeric solution G 180 was associated
with a substantial fall in water absorption in both
normal and RV-infected intestine. More surprising
were the findings for P36 in RV-infected intestine.
Although in normal intestine water absorption
among the polymer solutions was similar, in RV-
infected intestine increasing the concentration of
glucose polymer from 18mmol/l to 36mmol/l led to
a fall in water absorption. Glucose absorption from
P36 was also lower than from P18 in RV-infected
intestine where sodium-glucose-transporting capacity is known to be reduced and this may be the
explanation for the concomitant fall in water absorption. Another possibility might be the small rise
in osmolality attendant on increasing the glucose
polymer concentration, but this explanation would
seem less likely. The stimulation of sodium absorption in uitro [33] and in uiuo [30] by glucose is
known to be a saturable process, making sodiumglucose co-transport the rate-limiting step in
absorption of glucose from the bulk phase [34].
Our findings with regard to P36 certainly highlight
the differences inherent in ORS handling by normal
and diseased intestine. They are also less surprising
when viewed in the light of an uncontrolled clinical
trial, which showed that increasing the concentration of glucose polymer in ORS unduly can have
devastating reuslts [35]. The polymer ORS used
yielded 730 mmol/l free glucose on complete hydrolysis. Some of the infants in the trial suffered
severe osmotic diarrhoea and serious hypernatraemia. In that study, as in most other field trials
of cereal-based ORS, the principal rationale for the
use of complex carbohydrate as substrate was to
maximize the total glucose residues available for cotransport without incurring an osmotic penalty. Our
data suggest that the ability to increase total substrate availability is not foremost in importance in
determining the efficacy of glucose polymer in promoting water absorption from ORS.
In normal intestine P9 was associated with
Fluid absorption in a model of rotavirus diarrhoea
greater water glucose absorption than its monomeric counterpart G45, but with the other substrate
concentrations glucose absorption was similar
between equivalent monomer- and polymer-based
solutions. In RV-infected intestine both P9 and P18
were associated with higher glucose absorption than
their monomeric equivalents, but P36 and G180
produced similar glucose absorption. One of the
factors that may be important in explaining the
increased efficacy of polymer-based ORS is the
phenomenon of ‘kinetic advantage’. However, we
were not able to demonstrate a consistent kinetic
advantage for glucose polymer over free glucose.
The term kinetic advantage is now used to describe
the increased glucose absorption from polymers
than from free glucose and this has been shown best
in man by perfusion of normal human jejunum in
uiuo by Jones et al. [29, 361. The explanation for
this phenomenon is thought to depend on the
spatial arrangement of the hydrolase enzymes a.nd
glucose transporters. The proximity of these proteins would result in glucose accumulating in much
greater concentrations around the glucose transporter as a result of mucosal polymer hydrolysis
than might occur by simple diffusion of free glucose
from the bulk phase, thereby facilitating an ‘efficient
capture mechanism’ for glucose molecules by the
transporter proteins. The kinetic advantage conferred by glucose polymer has also been demonstrated in children with chronic diarrhoea [37] and
increased rates of glucose absorption resulting in
higher portal venous glycaemic responses have also
been shown in rats after intraduodenal infusion of
hypotonic, short-chain, rice-derived glucose polymer
[38, 391. There is evidence, however, against the
phenomenon of kinetic advantage. In amylase-free
canine and porcine Thiry-Vella fistulae, it was
shown that the rate of disappearance of glucose
polymer was less than that of free glucose and
dependent on chain length, absorption of shorter
glucose polymers being more rapid [40, 411. A
rabbit jejunum in uitro study also showed that the
rate of polymer-derived glucose transport was
slower than that of free glucose and it was concluded that the requirement for hydrolysis limited
glucose polymer assimilation [42].
Why we were not able to show that glucose
polymer conferred a kinetic advantage on glucose
absorption is not clear, but there are a number of
possible explanations. Our glucose polymer contained a mixture of chain lengths. The presence of
significant amounts of free glucose (approximately
3%) and polymers containing more than six glucose
residues (approximately 16%) may have been a
factor in view of the fact that Jones et al. [36] have
shown that oligomers containing between two and
six glucose residues were optimal for conferring
kinetic advantage. Although in normal intestine P36
was associated with higher glucose absorption than
P9 and P18, in RV-infected intestine increasing
polymer concentration from 18 to 36mmol/l was
415
associated with lower glucose absorption. These
findings probably relate to a quantitative reduction
in glucose transporters rather than impaired mucosal hydrolysis because P36 was hydrolysed to a
similar extent in normal and RV-infected intestine.
No consistent pattern emerges from our data to
support the existence of kinetic advantage and
whatever the explanation, these results argue
strongly against the kinetic advantage factor being
of central significance in determining the enhanced
efficacy of polymer ORS.
None of the experimental solutions perfused was
able to effect net sodium absorption even in normal
intestine, nor were they able to prevent the substantial increase in net sodium secretion in RVinfected intestine. These findings are not surprising
as sodium absorption in the small intestine
is concentration-dependent. Although nutrientdependent sodium absorption via the glucose cotransport mechanism has achieved clinical importance as the rationale for oral rehydration therapy,
passive processes of sodium absorption, such as
electroneutral sodium chloride absorption and solvent drag, are quantitatively of more importance in
the normal physiological state [43].
A linear relationship between net sodium movement and sodium concentration in the perfusate has
been shown to exist in normal and secreting rat
intestine by intestinal perfusion in situ [44, 451. Net
sodium absorption in normal intestine only
occurred when the sodium concentration was
60mmol/l or above [45]. Experiments in normal
human jejunum have confirmed this relationship,
although net sodium absorption occurred only when
the sodium concentration exceeded 80-90 mmol/l
[46]. Rolston et al. [44] in a rat model of secretory
diarrhoea, showed that the degree of sodium secretion was primarily influenced by the sodium concentration and net sodium absorption only took
place when the sodium concentration was
120mmol/l or above.
The pattern of potassium absorption in the
present study is broadly similar to the results for
water absorption consistent with the view that
potassium absorption is passive [47]. Given the
uniformity of potassium concentration in the experimental solutions any differences in potassium
absorption can only be explained by the effects of
solvent drag and reflect differences in water
absorption.
In previous experiments in a rat model of secretory diarrhoea using cholera toxin we have studied
electrolyte-containing complex carbohydrate as substrate in the form of rice-starch [20] or a defined
glucose polymer [48]. We showed that water
absorption from hypotonic solutions (whether they
contained monomer or polymeric glucose) was significantly greater than from standard hypertonic
ORS such as the WHO/UNICEF solution [20] and
that increasing the polymer concentration beyond
certain limits would result in a fall in net water
476
A. V. Thillainayagam et al.
absorption [49]. In both studies we showed an
inverse relationship between osmolality and water
absorption for both normal and secreting intestine
and we concluded that the low osmolality of
polymer-based solutions plays a dominant role in
promoting enhanced water absorption from these
ORS. Our conclusions accord well with those of
other workers who examined the relationship
between osmolality, glucose concentration and
sodium concentration in ORS in normal rat intestine [49, 501. An inverse relationship between net
water absorption and perfusate osmolality has also
been demonstrated in normal human jejunum by
intestinal perfusion in uiuo with solutions of
250 mosmol/kg or greater [43], while other workers
have suggested that below 250 mosmol/kg this relationship may disintegrate [Sl].
Our results show a striking inverse correlation
between water absorption and osmolality, thereby
emphasizing that osmolality is pre-eminent among
the factors influencing water absorption from
polymer ORS. The findings also demonstrate that
increasing polymer concentration beyond certain
limits can compromise the increase in water absorption conferred by hypotonicity, even if the increase
in osmolality is only modest. In field studies where
cereal-based polymer ORS were used, the concentrations of glucose polymer were much higher
than we have used in our study, because the cerealbased ORS were formulated on the basis of maximizing substrate availability without unduly increasing ORS osmolality. This rationale was founded on
the idea that increasing total glucose available for
co-transport would result in increased glucose and
sodium absorption and subsequently greater secondary water absorption. What is the middle ground
between the benefits afforded by increased glucose
availability and low osmolality respectively? Our
findings indicate that in an ORS of this electrolyte
composition the ideal concentration of this glucose
polymer would be approximately 9-18 mmol/l. Solutions containing 9 or lSmmol/l of this glucose
polymer, in this study, produced significantly greater
water absorption than their monomer-based equivalents which we called G45 and G90. The latter
monomer ORS, which is itself hypotonic, has
recently been recommended by the European
Society for Paediatric Gastroenterology and
Nutrition as being the optimal formulation for
European children with acute diarrhoea [52].
A limitation of our model of RV diarrhoea is that
the colon is excluded, but it does allow an assessment of potential advances in ORS formulation in a
situation which mimics RV infection in humans.
Although the relevance of the behaviour of ORS in
this model to that in human intestine during acute
diarrhoea might be questioned, we have demonstrated close similarities in the handling of ORS in
the rat models of enterotoxin-mediated and RVinduced diarrhoea and normal human jejunum [53].
Although it still remains to be established whether
observations made in this model can predict the
efficacy of an ORS in acute RV diarrhoea, a recent
clinical study has confirmed the ability of a hypotonic monomer ORS (almost identical in composition
with G90) to reduce stool volumes, duration of
diarrhoea and hospital stay compared with conventional hypertonic ORS [54]. This trial supports
the validity of our model of RV infection for
evaluating potential advances in anti-diarrhoea1
therapy and our conclusions from the present study.
Nevertheless, caution should be used in extrapolating findings from any animal model to human
disease. The final arbiter of the therapeutic potential
of the proposed hypotonic, low-substrate (glucose
polymer 9-18 mmol/l) polymer ORS will, of course,
be the controlled clinical trial.
ACKNOWLEDGMENTS
This study was supported by a grant from Rorer
Healthcare Ltd. M.J.G.F. gratefully acknowledges
financial support from the Wellcome Trust.
REFERENCES
I . Guandalini S. Overview of childhood acute diarrhoea in Europe: implications
for oral rehydration therapy. Acta Paediatr Scand 1989 Suppl. 361: 5-12.
2. Rodriguez WJ, Kim HW, Brandt CD, et al. Longitudinal study of rotavirus
infection and gastroenteritis in families served by a pediatric medical practice:
clinical and epidemiological observations. Pediatr Infect Dis J 1987; 6 1706.
3. Haffejee IE. The pathophysiology, clinical features and management of
rotavirus diarrhoea. Q J Med 1991; 7 9 289-99.
4. Editorial. Water with sugar and salt. Lancet 1978; ii: 3M).
5. Field M. New strategies for treating watery diarrhoea. N Engl J Med 1977;
191: 1121-2.
6. Farthing MJG. History and rationale of oral rehydration and recent
development for formulating an optimal solution. Adv Oral Rehydration
Drugs 1988; 36 (Suppl. 4): 80-90.
7. Molla AM, Sarker SA, Hossain M, Molla A, Greenough WB. Rice-powder
electrolyte solution as oral therapy in diarrhoea due to Vibrio choleroe and
Escherichio coli. Lancet 1982; i: 1317-19.
8. Patra FC, Mahalanabis D, Jalan KN, Sen A, Banerjee P. Is oral rice electrolyte
solution superior to glucose electrolyte solutions in infantile diarrhoea? Arch
Dis Child 1982; 57: 910-12.
9. Molla AM, Ahmed SM, Greenough WB. Rice-based oral rehydration solution
decreases the stool volume in acute diarrhoea. Bull WHO 1985; 6 3 751-6.
10. Mehta MN, Subramaniam S. Comparison of rice water, rice electrolyte
solution, and glucose electrolyte solution in the management of infantile
diarrhoea. Lancet 1986; i: 843-5.
11. Alam AM, Sarker SA, Molla AM, Rahaman MM, Greenough WB. Hydrolysed
wheat based oral rehydration solution for acute diarrhoea. Arch Dir Child
1987; 62: 4404.
12. Bhan MK, Ghai OP, Khoshoo V, et al. Efficacy of Mung bean (lentil) and pop
rice based rehydration solutions in comparison with the standard glucose
electrolyte solution. J Pediatr Gastroenterol Nutr 1987; 6 392-9.
13. Murtaza A, Zulfiquar I, Khan S-R, Lindblad BS, Sahlgren, BA, Aperia A. The
benefits of the very early introduction of powder rice and dried edible seeds
(dal moong) in the oral rehydration solution during the treatment of acute
infectious diarrhoea. Acta Paediatr Scand 1987; 76 861-864.
14. Molla AM, Molla A, Nath SK, Khatun M. Food-based oral rehydration salt
solution for acute childhood diarrhoea. Lancet 1989 ii: 429-31.
15. Santosham M, Fayad IM, Hashem M, Goepp JG, Refat M, Sack RB. A
comparison of rice-based oral rehydration solution and 'early feeding' for the
treatment of acute diarrhoea in infants. J Pediatr 1930; 116 868-75.
16. Pizarro D, Posada G, Sandi L, Moran MBC, Moran JR, Rice-based oral
electrolyte solutions for the management of infantile diarrhoea. N EnglJ Med
1991; 314 517-21.
17 Rahman ASMM. Bari A, Molla AM, Greenough WB. Mothers can prepare and
use rice-salt oral rehydration solution in rural Bangladesh. Lancet 1985; ii:
539-40.
Fluid absorption in a model of rotavirus diarrhoea
18. Sathanakrishnan BR, Sankaranarayanan VS. Rice water solution in diarrhoea1
dehydration. Ind J Pediat 1985; 52: 479-82.
19. Rabbani GH, Lu R-V, Horvath K, Lebethal E. Short-chain glucose polymer and
20.
21.
22.
23.
anthracind-carboxylic acid inhibit water and electrolyte secretion induced by
dibutyryl cyclic AMP in the small intestine. Gastroenterology 1991; 101:
1046-53.
Thillainayagam AV, Dias ]A, Carnaby S, Clark ML, Farthing MJG. Evidence of
a dominant role for low ormolality on the efficacy of cereaCbased oral
rehydration solutions: studies in a model of secretoq diarrhoea. Gut 1994
(In press).
Hunt JB, Elliott EJ, Farthing MJG. Comparison of rat and human intestinal
perfusion models for assessing efficacy of oral rehydration solutions. Aliment
Pharmacol Ther 1991; 5: 49-59.
Molla AM, Rahman M, Sarker SA, Sack DA, Holla A. Stool electrolyte
content and purging rates in diarrhoea caused by rotavirus, enterotoxigenic E.
coli and V. cholerae in children. J Pediatr 1981; 48:8358.
Vonderfecht SL, Huber AC, Eiden J, Hader LC. Yolken RH. Infectious
diarrhea of infants and rats produced by a rotavirus-like agent. J Virol 1984,
52: 9 4 4 .
2.4. Bridger j€. Novel rotaviruses in animals and man. Ciba Found Symp 1987;
Izs: 5-u.
25. Salim AFM, Walker-Smith ]A, Farthing MJG. Characterisation of experimental
rotayirus (RY) diarrhoea in rat. Gut 1988; ZP: A1433.
26. Salim AFM, Phillips AD, Walker-Smith JA, Farthing MJG. Pathogenesis of
rotayirus (RV) diarrhoea: studies in a nwnatal rat model. Gastroenterology
989; 9R: A436.
27. Cheetham W, Sirimanne P, Day W. HPLC separation of carbohydrate
pligomers. J Chromatogr 1981; WE 43944.
28. Roberts JE, Stewart R, Caseiro MC. Organic chemistry. New York: WA
r i n Inc, 1971.
29. onps BJM, Brown BE, Loran IS, et al. Glucose absorption from starch
ydrolysates in the human jejunum. Gut 1983; U: I152-60.
30. Sladen GE, DawsoQ AM. Interrelationships between the absorption of glucose,
so4ium and water by the normal human jejunum. Clin Sci 1969; 3 6 I19.-32.
31. Hall GA. Mechanisms of mucosal injury: animal studies. In: Farthing MJG. ed.
Viruses and the gut: Proceedings of the Ninth BSG S K I F International
Workshop. Welwyn Garden City: SmithKline and French Ltd, 1988: 27-9.
32. Mepwisse GW. High sugar worse than high sodium in oral rehydration
solutions. Acfa Paediatr Scand 1983; 72: 161-6.
33. Hellier MD, Thirumalai C, Holdsvorth CD. The effect d amino acids and
dipepfidp on sodium and water absorption in man. Gut 1973; It 41-5.
34. Grq GM. Carbohydrate digestion and absorption. Gastroenterology 1970 5&
96-197.
35. Sandhu BK, Jones BJM, Brook CGD, Silk DBA. Oral rehydration in acute
infantile diarrhoea with a glucose polymer-electrdyte solution. Arch Dir
Child 1982; 8: 152-4.
36. Jones BJM, Higgins BE, Silk DBA. Glucose absorption from maltofrime and
glucose oligomers in the human jejunum. Clin Sci 1987 72 409-14.
477
37. Sloven DG. Jirapinyo P, Lebenthal E. Hydrolysis and absorption of glucose
polymers from rice compared with corn in chronic diarrhea of infancy. J
Pediatr 1990; 116 876-881.
38. Azad MAK, Lebenthal E. Role of rat intestinal glucoamylase in glucose
polymer hydrolysis and absorption, Pediatr Res 1990; UI: 166-70.
39. Zheng BY, Khin-Maung-U, Lu R-B, Maiese RL, Maiese S, Lebenthal E.
Absorption of glucose polymers from rice in oral rehydration solutions by rat
small intestine. Gastroenterology 1993; IM 81-5.
40. Kenner B, Sloan HR, Haase G, McClung HJ, Ailabouni AH. The jejunal
absorption of glucose polymers in the absence of pancreatic enzymes. Pediatr
Res 1981; IS: 250-3,
41. Kerzner 8, Sloan HR, McClung HI, Chidi C, Ailabouni AH, Seckel C.
Absorption of glucose polymers from canine jejunum deprived of pancreatic
amylase. Am J Physiol 1986; G824-9.
42. Heitlinger LA, Sloane HR, Oevore DR, Lee D-C, Lebenthal E, Duffey ME.
Transport of glucose polymerderived glucose by rabbit jejunum.
Gastroenterology 1992; Ip1: 443-7.
43. Fordtran IS, Rector FC, Carter NW. The mechanisms of sodium absorption in
the human small intestine. J Clin Invest 1968; 47: E84-900.
44. Rolston DDK, Barado MM, Kelly MI, Dawson AM, Farthing MJG. Efficacy of
oral rehydration solutions in a rat model of secretory diarrhea. J Pediatr
Gastroenterol Nutr 1%7; 6 624-30.
45. Elliot EJ, Watson AJ, Walker-Smith ]A, Farthing MJG. Search for the ideal oral
rehydration solution: studies in a model of secretory diarrhoea. Gut 1991; 32:
I3 1420.
46. Spiller RC, Jones BJM, Silk DBA. Jejunal water and electrolyte absorption
from two propriety enteral feeds in man and importance of sodium content.
Gut 1987; uk 681-7.
47. Turnberg LA. Potassium transport in the human small bowel. Gut 1971; 12:
81 1-18.
48. Thillainayagam AV, Carnaby S, Clark ML, Farthing MJG. Water absorption
from glucose polymer-electrolyte solutions in secretory diarrhoea: glucose
load, kinetlc advantage or ormolality! Gastroenterology 19% 111: A558.
49. Saunders OR, Slllery JK. Absorption of carbohydrate-elettrolyte solutions In
rat duodenojelunum: impllcations for the composltlon of otal electtolyte
solutions in man. Dig Dir Sci 1985; 30: 154-60.
50. Lifshitr F, Wapnit RA. Oral hydration solutions: experimental optimization of
water and sodium absorption. J Pediatr 1985; 1w: 383-9.
51. Soergel KH, Whalen GE, Harris ]A. Passive movement of water and sodium
across the human m a l l intestinal mucosa. J Appl Physiol 1968; 24: 40-9.
52. Report of an ESPGAN Working Group. Recommendations for composition of
oral rehydration wluiioni for the children of Europe. J Pediatr Gastroenterol
Nutr 1992 14 113-15.
53. Hunt JB, Thillainayagam AV, Carnaby S, Fairclough PD, Clark HL, Farthing
MJG. Water absorptim from a hypotonic oral rehydration solution (06):
evaluation in human and animal perfusion models. Gut 1992; 33: 1625-9.
54. Rautanen T, El-Radhi S, Vesikari T. Clinical experience with a hypotonic oral
rehydratian solution in acute diarrhoea. Acta Paediatr Scand 1993; 82: 52-4.