Use of Oral Gentamicin, Metronidazole, and

Cholestyramine in the Treatment of Severe

Persistent Diarrhea in Infants

Ivor D. Hill, MD, FCP, Michael D. Mann, M Med, PhD,

Keith C. Househam, FCP, and Malcolm D. Bowie, MD, FRCP

From the Institute of Child Health, Red Cross War Memorial Children’s Hospital,
Rondebosch, Republic of South Africa

ABSTRACT. Oral gentamicin, metronidazole, and choles- Because the losses of nutrients from the bowel are
tyramine were given either as single agents or in various directly related to stool weight, maintenance of
combinations to infants who still required treatment after
nutrition by the oral route is usually not possible
seven days in the hospital for persistent diarrhea. The
effect of these drugs and the interactions between them in infants with fecal losses exceeding this value.2
were assessed by comparing daily stool output during Termination of the diarrhea is crucial in preventing
treatment with that in the pretreatment period. The the vicious cycle of gastroenteritis and malnutri-
effect of the drugs on apparent nitrogen and fat absorp- tion.
tion was also studied. On the first day of treatment the
A previous, uncontrolled trial demonstrated that
presence of cholestyramine was associated with a signif-
icantly greater decrease in stool output. This effect ap- a combination of oral gentamicin, metronidazole,
peared to be largely due to an interaction with gentami- and cholestyramine therapy was effective in stop-
cm. Thereafter, only gentamicin produced a significantly ping persistent diarrhea in the majority of cases.”3
greater decrease in stool weight. At no stage was metro- The present study was undertaken to test this in a
nidazole of benefit. Gentamicin and cholestyramine also
controlled fashion and to determine which of the
indirectly improved apparent nitrogen and fat absorption
by reducing stool output. The combination of oral gen- drugs used singly or in combination was most effec-
tamicin and cholestyramine is recommended as a safe tive. The effect of individual drugs and combina-
and effective way of treating infants with severe persist- tions of drugs was assessed by comparing the de-
ent diarrhea following acute gastroenteritis. Pediatrics crease in stool output following the institution of
1986;77:477-481; persistent diarrhea, gentamicin, choles-
treatment. The effect of individual drugs on appar-
tyramine.
ent nitrogen and fat absorption during treatment
was also evaluated.

PATIENTS AND METHODS

Most children admitted to Red Cross War Me-
mona! Children’s Hospital in Cape Town with gas- Infants between the ages of 6 weeks and 1 year
troenteritis have a self-limiting illness, and the requiring admission to the hospital for dehydrating
diarrhea diminishes after a few days. In 10% to diarrhea were assessed. Only boys were studied
15% of patients, severe diarrhea continues and re- because of the ease in separating urine and stool
quires therapy for more than seven days.’ Persist- collections. Infants with a history of diarrhea for
ent diarrhea is defined as a continuing daily stool more than 72 hours and any with a history of having
output exceeding 30 g/kg of body weight for seven had diarrhea or receiving antibiotics in the 2 weeks
days after commencement of hospital treatment. preceding the present illness were excluded. None
had kwashiorkor, marasmic kwashiorkor, or any
skin lesions associated with nutritional deficiencies.
Received for publication Jan 31, 1985; accepted Aug 1, 1985. The day of admission to hospital was designated
Reprint requests to (M.D.B.) Institute of Child Health, Red
day 0, and initial treatment was the same in all
Cross War Memorial Children’s Hospital, Rondebosch, 7700,
Republic of South Africa. cases. Routine investigations included microscopic
PEDIATRICS (ISSN 0031 4005). Copyright © 1986 by the examination of stool and urine specimens for par-
American Academy of Pediatrics. asites and stool and urine cultures. Blood specimens

PEDIATRICS Vol. 77 No. 4 April 1986 477

Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018
were cultured and measured for serum electrolytes the drug (expressed as g, m, or c). The eight possible
and acid-base status. An inadequate circulating treatment combinations were: (1) gentamicin plus
blood volume, detected clinically by poor peripheral metronidazole plus cholestyramine (GMC), (2) gen-
capillary perfusion, was corrected by rapid infusion tamicin plus metronidazole (GMc), (3) gentamicin
of a plasma volume expander (Haemaccel). There- plus cholestyramine (GmC), (4) metronidazole plus
after, the patients were managed in the same way cholestyramine (gMC), (5) gentamicin (Gmc), (6)
as those without this sign. metronidazole (gMc), (7) cholestyramine (gmC), (8)
Severe metabolic acidosis (pH < 7.20) was par- no drugs (gmc).
tially corrected with intravenous sodium bicarbon- Gentamicin was given in a dosage of 50 mg/kg/d
ate administration. All fluid administered intrave- (maximum 360 mg/d) in six divided doses for three
nously to those without signs of an inadequate days. The dose of cholestyramine was 1 g every six
circulating blood volume was half-strength Dar- hours and that of metronidazole was 100 mg every
row’s solution in 5% dextrose in water. The calcu- eight hours for five days. Care was taken to give
lated volume of fluid for rehydration was based on the cholestyramine at least one hour before the
the clinical assessment of dehydration (ie, 50 mL/ gentamicin or metronidazole to avoid possible bind-
kg if 5% dehydrated and 100 mL/kg if 10% dehy- ing of the antibacterial agents.
drated). This, together with a maintenance volume The nutritional status of each infant was assessed
of 120 mL/kg/d, was given at a constant rate during by expressing the fully rehydrated weight as a per-
24 hours. Oral half-strength Darrow’s solution with centage of expected weight for age. The groups were
dextrose was given at three-hourly intervals start- compared with respect to age and nutritional status
ing three to 12 hours after admission, and the and in the pretreatment period (days 5 to 7) with
intravenous fluid rate was decreased as necessary. respect to daily stool weight per kilogram of body
The infants were frequently reassessed and fluid weight by analysis of variance. The effects of the
requirements recalculated according to their prog- drugs and the interactions between them on the
ress. About 24 hours after admission, the oral half- severity of the diarrhea were assessed by determin-
strength Darrow’s solution was replaced with cow’s ing the decline in stool weight from pretreatment
milk formula feedings. If the diarrhea persisted and levels. The daily stool weight per kilogram of body
intravenous fluids were required to maintain hydra- weight on days 8, 9, 10, and 1 1 was subtracted from
tion on day 4, the feedings were changed to a soya- the mean of the pretreatment stool weight of that
based formula (Isomil) given in a volume of 120 individual. These results were analyzed in the
mL/kg/d equally divided into three-hourly feed- standard way for a 2 x 2 x 2 factorial design.4
ings. The effect of the drugs on apparent nitrogen and
On day 5, those infants still requiring intravenous fat absorption was assessed on a three-day balance
fluid administration to maintain hydration were (days 9 to 11 inclusive) using multiple linear regres-
selected for further study. From day 5 to 11, they sion.5 The dependent variable was the percentage
were nursed on balance beds, and stool and urine of nitrogen or the percentage of fat absorption.
were collected separately. Stool weight and appar- Stool weight has a marked effect on each of these
ent fat and nitrogen absorption were measured variables,2 therefore, it was included as a continu-
daily. Those who still had severe diarrhea (stool ous independent variable. The factors gentamicin,
weight greater than 30 g/kg of body weight per day) metronidazole, and cholestyramine were included
on day 7 were selected for inclusion into the study. as categorical variables (1 when the drug was given
They were randomly allocated to receive one of and 0 when it was not).
eight treatment combinations from days 8 to 1 1. If
the initial bacteriologic screen showed a patient to RESULTS
have a urinary tract infection, a recognized bacte-
na! stool pathogen or parasite, or a bacteremia, he Details of the groups are shown in Table 1. There
was excluded from the study. The next infant to be were no differences between the groups with regard
studied was allocated to the treatment group from to age or percentage of weight for age (P > .05).
which the infant had been excluded, and the trial There were no differences in stool weights between
continued until 40 infants had completed the study. the groups in the pretreatment period (days 5 to 7);
The study was analyzed as a 2 x 2 x 2 factorial nor were there significant changes within each
with five replications (ie, five infants in each group between those days. All infants had severe
group).4 The three factors were gentamicin (G), diarrhea with a stool weight of 105.7 ± 8.5 g/kg/d
metronidazole (M), and cholestyramine (C). The (mean ± SEM).
two levels of each factor were the presence of the The mean decrease in stool weight from pretreat-
drug (capital letters G, M, or C) or the absence of ment levels on days 8, 9, 10, and 11 in the presence

478 TREATMENT OF SEVERE PERSISTENT DIARRHEA

Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018
TABLE 1. Characteristics of Infants in Treatment Groups
Characteristic Drug Regimen*

GMC GMc GmC gMC Gmc gMc gmC gmc

Age (mo)
Mean 4.33 3.98 3.50 4.64 5.12 3.62 4.62 5.00
Range 1.50-7.40 1.90-5.40 1.80-8.40 2.30-7.60 1.50-9.50 2.30-5.50 1.60-9.60 3.10-9.00
SEM ±0.98 ±0.66 ±1.24 ±1.46 ±1.68 ±0.59 ±1.05 ±1.05
% expected wt
for age
Mean 79.8 73.9 82.5 87.1 78.5 79.4 79.9 81.1
Range 64.2-93.9 64.4-88.6 65.1-100.8 59.0-101.7 54.4-89.4 63.0-94.0 70.4-90.8 40.2-113.9
SEM ±5.6 ±4.3 ±5.9 ±7.4 ±6.6 ±5.3 ±3.3 ±12.1
* G, gentamicin; M, metronidazole; C, cholestyramine. Administration of a particular drug is indicated by the capital
letters G, M, or C and the absence of a drug by the small letters g, m, or c.

TABLE 2. Decrease in Stool Weights in Infants on Different Drug Regimens*

Day G g M m C c
8 47.2 ± 11.9 28.3 ± 8.9 45.3 ± 10.3 30.2 ± 10.8 54.0 ± 11.2 21.5 ± 8.7
9 87.6 ± 12.1 40.0 ± 9.3 64.4 ± 11.5 63.1 ± 12.6 76.9 ± 13.2 50.7 ± 9.9
10 98.4 ± 12.5 38.7 ± 15.3 63.3 ± 16.4 73.8 ± 14.6 86.1 ± 15.0 51.1 ± 15.0
11 98.8 ± 13.8 49.1 ± 15.3 65.7 ± 16.7 82.7 ± 14.2 91.5 ± 15.9 56.5 ± 14.3
* Results are expressed as the mean decrease in stool weights (g/kg/d) ± SEM from
pretreatment (days 5to 7) levels in the presence (capital letter) and absence (lower case
letter) of each drug. G, gentamicin; M, metronidazole; C, cholestyramine.

and absence of each drug is shown in Table 2. On TABLE 3. Effect of Stool Weight and Drug Regimen
day 8, the effect of cholestyramine treatment was on Apparent Nitrogen and Fat Absorption*
significant (F,,32 = 6.08; P < .05). The decrease in Y % Nitrogen % Fat
Absorption Absorption
stool weight was 32.5 g/kg greater in those infants
receiving cholestyramine treatment compared with A 84.19 90.31
B1 (stool wt) -.43 (±.06) -.42 (±.06)
those not receiving the drug. The effects of genta-
B, (gentamicin) -3.61 (±3.94) -6.65 (±4.04)
micin treatment (18.9 g/kg, F,,32 = 2.05; P > .05) B, (metronidazole) 1.70 (±3.71) -1.48 (±3.80)
and metronidazole treatment (15.5 g/kg, F1,32 = B4 (cholestyramine) -2.44 (±3.73) -13.51 (±3.82)
1,31; P > .05) were not significant. The effect of R’ .60 .61
cholestyramine treatment on day 8 was largely due * The partial regression coefficients for the multiple lin-
to a significant interaction between cholestyramine ear regression Y = A + B1X1 + B,X2 + B,X3 + B4X4,
and gentamicin (F,,32 = 5.75; P < .05). Stool weight where Y = % nitrogen absorption or % fat absorption,
declined 79.2 g/kg/d in those who received both x1 = stool weight (g/kg/d) ± SEM, and X2, X3, and X4
= categorical variables (±SEM) for gentamicin, metro-
drugs (GMC and GmC) compared with 15.2 g/kg/d
nidazole, and cholestyramine, respectively. When the
in those receiving gentamicin but not cholestyra- drug was given, the value of the categorical variable was
mine (Gmc and GMc), 28.8 g/kg/d in those given 1. The value 0 was allocated to those patients who did
cholestyramine but not gentamicin (gMC and not receive the drug.
gmC), and 27.9 g/kg/d in those given neither (gMc
and gmc).
On day 9 the effect of gentamicin treatment was
significant. The decrease in stool weight was 47.6 kg higher than those who did not receive it (GmC,
g/kg greater in those receiving gentamicin than Gmc, gmC, gmc).
those not (F,,32 = 9.95; P < .01). The effects of Between days 9 and 11, nitrogen and fat absorp-
cholestyramine treatment (26.2 g/kg; F1,32 = 3.01; tion improved as the stool weight decreased. In
P > .05) and metronidazole treatment (1.3 g/kg) Table 3, the partial regression coefficients for the
were not significant, and there were no significant multiple linear regression of percentage nitrogen
interactions. The results on days 10 and 1 1 were and fat absorption are shown. Except for their
similar to those on day 9. The effect of metronida- effect on stool weight, none of the drugs directly
zole treatment was not significant, but in those affected nitrogen absorption, but the administra-
patients receiving metronidazole (GMC, GMc, gMc tion of cholestyramine was associated with a reduc-
and gMC) on day 11, the stool weight was 17.0 g/ tion in fat absorption.

ARTICLES 479
Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018
DISCUSSION gentamicin treatment (groups GMC, GMc, GmC,
Gmc) did not have a significantly greater decrease
Prior to treatment, all of the infants in this study in stool output in the first 24 hours of treatment
had equally severe persistant diarrhea. The mean compared with those not receiving gentamicin. On
daily stool weight was more than 100 g/kg of body days 9 to 11, the administration of gentamicin was
weight, and a previous study has shown that at this accompanied by a significantly greater decrease in
magnitude of diarrhea nitrogen and energy require- stool weight in comparison to those not receiving
ments cannot be met with oral nutrient intake.2 the drug (Table 2). It appears that the effect of
Such infants are in negative nitrogen balance and cholestyramine on day 8 is largely due to an inter-
apparent fat absorption is of the order of 30% of action with gentamicin. During the first 24 hours
intake. Carbohydrate absorption is also impaired.6 of treatment, the decrease in stool weight of infants
Termination of the diarrhea is essential if nutrition receiving cholestyramine but not gentamicin
is to be maintained. (groups gMC and gmC) was similar to the effect in
To assess the effects of the drugs, a 2 x 2 X 2 those receiving gentamicin without cholestyramine
factorial method of analysis was chosen to over- (groups GMc and Gmc) or neither drug (groups
come the problem of having small numbers in each gMc and gmc). The effect on stool weight of infants
group. Use of this method allows the main effect of receiving both gentamicin and cholestyramine
a particular drug to be assessed by comparing all (GMC and GmC) compared with those receiving
patients receiving the preparation with all of those neither (gMc and gmc) is illustrated in the Figure.
not receiving it. In effect, this means that for each Metronidazole treatment appears to have little ef-
drug used in this study 20 patients receiving a fect on stool weight either alone or in combination
particular drug were compared with 20 not receiving with the other drugs.
it. The factors responsible for the persistence of
During the first 24 hours of treatment, infants severe diarrhea are not entirely clear. None of the
who received cholestyramine treatment (groups infants included in this study had a recognized
GMC, GmC, gMC, gmC) had a significantly greater bacterial pathogen in their stools. A previous study
decrease in stool weight than those who did not using the same criteria for patient selection dem-
receive the drug (Table 2). Thereafter, on days 9 to onstrated a grossly abnormal overgrowth of micro-
1 1, although the decline in stool weight was greater, organisms in the small bowel of infants with diar-
the difference was not significant. Infants receiving rhea persisting for seven days after admission.7

20

C
H
A
N
G
E
20

_40

S
I
0
0 6O
L

80

G
H
I -100

g
, _120

‘I
7 8 9 10
d

DAY OF HOSPITAL TREATMENT

Figure. Decrease in stool weight from pretreatment levels in infants receiving a combi-
nation of gentamicin and cholestyramine (-) and those receiving neither of these drugs
(- - -). Horizontal line at 0 represents the mean daily stool weight (g/kg of body weight!
d) of each group before starting treatment.

480 TREATMENT OF SEVERE PERSISTENT DIARRHEA

Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018

 Although the duodenal microflora of the infants in lestyramine together was a marked reduction in
this study was not investigated, there is no reason daily stool output and an improvement in nitrogen
to believe it should be any different from those in and fat absorption.
the previous study. We believe the bacterial over- It is a minority of infants with acute gastroenter-
growth plays a role in the pathogenesis of persistent itis who go on to have persistent diarrhea of the
diarrhea. The microorganisms may be directly in- severity of the infants in this study. In patients
volved by damaging the intestinal mucosa.8 Alter- such as these, maintenance of nutrition via the oral
natively, or in addition, they may be indirectly route is not possible because of the large stool
involved by the elaboration of toxins or by decon- nitrogen and energy losses. Termination of the
jugating and dehydroxylating the bile salts.8 The diarrhea and reduction of these losses is essential.
early beneficial effect noted with the use of choles- This study has shown the use of a combination of
tyramine treatment supports the indirect mecha- oral gentamicin and cholestyramine to be an effec-
nism as a cause of persistent diarrhea. The domi- tive and rapid means of achieving this. It is also
nant role of gentamicin noted after the first 24 safe and, except for a mild metabolic acidosis during
hours of treatment supports the view that elimi- the administration of cholestyramine, no compli-
nation of the bacterial overgrowth is most impor- cations have been encountered in 8 years of using
tant in controlling ongoing diarrhea. this treatment. Measurements of serum gentamicin
As has been noted in a previous study, the greater levels in a number of similar infants receiving oral
the stool weight the poorer is apparent nitrogen gentamicin in the dose prescribed have shown ab-
and fat absorption.2 Nitrogen and fat absorption sorption to be insignificant (unpublished data). It
decreased by about 0.4% with each gram per kilo- is recommended that oral gentamicin and choles-
gram of body weight increase in stool output (Table tyramine be given to infants with severe, persistent,
3). Gentamicin treatment had a major effect on dehydrating diarrhea who show no signs of im-
stool weight and indirectly on apparent nitrogen provement after seven days of hospital treatment.
and fat absorption. As the stool weight decreased,
apparent nitrogen and fat absorption improved. No
direct effect of gentamicin was noted. In contrast, ACKNOWLEDGMENTS
metronidazole treatment did not produce a decrease
This work was supported by the South African Medical
in stool weight; therefore, no indirect effect on
Research Council.
apparent nitrogen and fat absorption occurred. Nei- We thank the Mobil Research Associateship for tech-
ther was a direct effect noted. nical assistance.
By reducing stool weight (Table 2), cholestyra-
mine indirectly improved nitrogen absorption, but
no direct effect was demonstrated. The decrease in REFERENCES
stool weight also improved fat absorption, but the 1. Bowie MD, Mann MD, Hill ID: The bowel cocktail. Pedi-
drug directly impaired uptake of fats (Table 3). atrics 1981;67:920-921

This is not surprising because cholestyramine is a 2. Mann MD, Hill ID, Peat GM, et al: Protein and fat absorp-
tion in prolonged diarrhoea in infancy. Arch Dis Child
bile salt-binding resin. The net effect on fat absorp- 1982;57:268-273
tion is a balance between the improvement derived 3. Hill ID, Mann MD, Bowie MD: Successful management of
persistent diarrhoea in infants. S Afr Med J 1980;58:241-
from the reduction of stool weight and the direct
243
effect of cholestyramine’s impairment of fat ab- 4. Snedecor GW, Cochrane WG: Statistical Methods, ed 6.
sorption. By calculation, a decrease in stool weight Ames, Iowa, Iowa State University Press, 1967
5. Bottenberg RA, Ward JH: Applied Multiple Linear Regres-
exceeding 32 g/kg of body weight per day would
sian. Lackland AF Base, TX. 6570th Personnel Research
offset the reduction in fat absorption associated Laboratory, Aerospace Medical Division, report No. PRL-
with cholestyramine treatment. Between days 9 and TDR-63-6, 1963
6. Mann MD, Hill ID, Moore L, et al: Xylose absorption in
11, infants receiving cholestyramine had a mean
infants with severe prolonged diarrhoea. S Afr Med J
reduction in daily stool weight that was 32 g/kg of 180;58:598-599
body weight greater than those not receiving the 7. Hill ID, Mann MD, Moore L, et al: Duodenal microflora in
drug (Table 2). On average, cholestyramine treat- infants with acute and persistent diarrhoea. Arch Dis Child
1983;58:330-334
ment alone neither improved nor impaired fat ab- 8. Lifshitz F: The enteric flora in childhood disease-diar-
sorption. The overall effect of gentamicin and cho- rhoea. Am J Clin Nutr 1977;50:1511-1516

ARTICLES 481
Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018
 Use of Oral Gentamicin, Metronidazole, and Cholestyramine in the Treatment of
Severe Persistent Diarrhea in Infants
Ivor D. Hill, Michael D. Mann, Keith C. Househam and Malcolm D. Bowie
Pediatrics 1986;77;477

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/77/4/477
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its
entirety can be found online at:
https://shop.aap.org/licensing-permissions/
Reprints Information about ordering reprints can be found online:
http://classic.pediatrics.aappublications.org/content/reprints

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has
been published continuously since . Pediatrics is owned, published, and trademarked by the American
Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007. Copyright ©
1986 by the American Academy of Pediatrics. All rights reserved. Print ISSN: .

Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018

 Use of Oral Gentamicin, Metronidazole, and Cholestyramine in the Treatment of
Severe Persistent Diarrhea in Infants
Ivor D. Hill, Michael D. Mann, Keith C. Househam and Malcolm D. Bowie
Pediatrics 1986;77;477

The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://pediatrics.aappublications.org/content/77/4/477

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has
been published continuously since . Pediatrics is owned, published, and trademarked by the American
Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007. Copyright ©
1986 by the American Academy of Pediatrics. All rights reserved. Print ISSN: .

Downloaded from http://pediatrics.aappublications.org/ by guest on March 5, 2018