Available online at www.sciencedirect.

com

Nutrition Research 29 (2009) 462 – 469

www.nrjournal.com

Comparison of yogurt, soybean, casein, and amino acid–based diets in

children with persistent diarrhea
Ângela P. de Mattos⁎, Tereza C.M. Ribeiro, Patrícia S.A. Mendes, Sandra S. Valois,
Carlos M.C. Mendes, Hugo C. Ribeiro Jr
Fima Lifshitz Research Center, Pediatric Hospital, Federal University of Bahia, 40.110-170 Bahia, Brazil
Received 4 December 2008; revised 20 June 2009; accepted 22 June 2009

Abstract

Although previous studies have shown successful treatment of persistent diarrhea (PD) with the
use of yogurt-based diets, some recent ones speculate the need of special formulas for the nutritional
management of PD complicated cases. In the present study, we tested the hypothesis that the
consumption of 3 lactose-free diets, with different degrees of complexity, is associated with lower
stool output and shorter duration of diarrhea when compared with the use of a yogurt-based one on
the nutritional management of PD. A total of 154 male infants, aged between 1 and 30 months, with
PD and with or without dehydration, were randomly assigned to 1 of 4 treatment groups. Throughout
the study, the patients were placed in a metabolic unit; their body weights and intakes of oral
rehydration solution, water, and formula diets, in addition to outputs of stool, urine, and vomit, were
measured and recorded at 24-hour intervals. Four different diets were used in this study: diet 1,
yogurt-based formula; diet 2, soy-based formula; diet 3, hydrolyzed protein-based formula; and diet
4, amino acid–based formula. Throughout the study, only these formula diets were fed to the
children. The data showed that children fed the yogurt-based diet (diet 1) or the amino acid–based
diet (diet 4) had a significant reduction in stool output and in the duration of diarrhea. The use of an
inexpensive and worldwide-available yogurt-based diet is recommended as the first choice for the
nutritional management of mild to moderate PD. For the few complicated PD cases, when available,
a more complex amino acid–based diet should be reserved for the nutritional management of these
unresponsive and severe presentations. Soy-based or casein-based diets do not offer any specific
advantage or benefits and do not seem to have a place in the management of PD.
© 2009 Elsevier Inc. All rights reserved.
Keywords: Persistent diarrhea; Children; Nutritional management; Malabsorption; Diet
Abbreviations: ANOVA, analysis of variance; ORS, oral rehydration solution; PD, persistent diarrhea; WHO, World Health
Organization.

1. Introduction episodes of diarrhea and 4.6 million diarrhea-related deaths

among children younger than 5 years in developing countries
More than 2 decades ago, the World Health Organization
[1]. Since then, much has been learned about the
(WHO) estimated an annual occurrence of 800 million
epidemiology, transmission, pathogenesis, and management
⁎ Corresponding author. Rua Padre Feijó, no. 29, Centro Pediátrico
of diarrheal diseases in childhood. Nevertheless, although
Professor Hosannah de Oliveira, CEP:40.110-170, Salvador, Bahia, Brazil.
there has been a reduction in the mortality rates (mainly in
Tel.: +55 71 33312027; fax: +55 71 33312027. children younger than 1 year), the most recent review
E-mail address: mattosangela@gmail.com (Â.P. de Mattos). conducted by the WHO has shown that in developing
0271-5317/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.nutres.2009.06.005
 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
countries, diarrheal diseases are still responsible for 21% of
the deaths of children younger than 5 years, corresponding to
2.5 million annual deaths [2].
This decline in diarrheal mortality is usually credited in
part to the improved case management of acute diarrhea,
including the widespread use of oral rehydration solution
(ORS) and the adoption of improved feeding practices [3].
However, dysentery and persistent diarrhea (PD) may cause
an increasing proportion of diarrheal deaths [2,4]. Persistent
diarrhea is defined as diarrheal episodes of presumed
infectious etiology, which begin acutely and last for at
least 14 days. This type of presentation of clinical symptoms
has been acknowledged to account for 3% to 20% of all
episodes of diarrhea in children younger than 5 years [5].
Although PD accounts for a small proportion of diarrheal
episodes, the great concern of health workers and public
health program planners is the high number of deaths
associated with PD [5-11]. In a few of the large prospective
studies, the death rate for PD is as much as 49%, 62%, or
73% [7,8,11]. A number of risk factors, such as age,
immune status, early weaning, specific pathogens, and
mainly malnutrition, have been associated with PD. Several
studies have shown the strong and bidirectional interaction
between malnutrition and PD. It seems that these 2
conditions coexist in the most situations, and some authors
even propose to call this combination the “malnutrition-
diarrhea syndrome” [12-16].
Although it is generally accepted that nutritional manage-
ment is by far the most important component of PD
treatment, no systematic approach to nutritional therapy
has been established so far. Thus far, the studies that have
emphasized nutritional management for PD have used a wide
variety of diets, which are based on traditional foods such as
milk (with or without lactose), soy protein, chicken protein,
hydrolyzed protein, and amino acids [15]. A multicenter
study coordinated by the WHO and conducted in India,
Bangladesh, Pakistan, Vietnam, Mexico, and Peru used
regional foods associated with yogurt or a low-lactose milk-
based formula. The success rate ranged from 65% to 89% in
different countries [16]. However, the only consensus
reached in literature until now is that, as an initial approach,
PD patients should have lower or nil amounts of lactose in
their diets [17-20]. Nevertheless, studies have shown that a
substantial proportion of patients who need to be hospita-
lized show nutritional impairment and continue to have
diarrhea despite the use of lactose-free diets [16].
Moreover, the tissue damage that occurs in these patients
is certainly not due to the lack of disaccharides activity alone,
which hence requires a discussion of more complex
approaches. Many studies have been carried out to compare
different types of diets for use in nutritional management, but
their results have been contradictory [21-24]. In the absence
of a definitive recommendation, a more judicious analysis of
a larger sample of patients becomes imperative. In addition,
the patients ought to be treated according to preestablished
algorithms, using systematic protocols that help to identify
463
the risk factors for unexpected clinical evaluations related to
inappropriate “initial diets.” It is possible that some groups of
patients, with ill-defined profiles, need a more specialized
nutritional management based on a more complex formula as
the initial nutritional approach. The objective of this study
was to test the hypothesis that the consumption of 3 full
lactose-free formulas—soy-protein, hydrolyzed casein, or
amino acid mixture—is associated with lower stool output
and shorter duration of diarrhea when compared with the use
of a yogurt-based formula on the nutritional management of
children with PD.
2. Methods and materials
Children with PD (defined as lasting for at least 14 days,
with the occurrence of 3 or more liquid stools during the 24
hours that preceded admission), who either appeared to be
dehydrated or not according to the WHO criteria, and who
fulfilled the following inclusion criteria were enrolled: male
sex, age between 1 and 30 months, weight to length index
greater than 50% of the National Center for Health Statistics
median, and totally or partially weaned (at least 50% of milk
diets based on artificial milk). The exclusion criteria
considered were absence of consent of the parent or legal
guardian, patients with systemic infections requiring anti-
biotics or with any other chronic disease, and patients with
edema. The main outcome variables considered were total
stool output in grams per kilogram, duration of diarrhea in
hours, and weight gain in grams from randomization to the
cessation of diarrhea. Secondary outcome variables con-
sidered were volume of intake and failure rate (defined as the
need to change the initial randomized dietary plan).
Stool output and the duration of diarrheal episodes in
children with acute and PD who were treated with the WHO/
United Nations Children's Fund ORS were highly variable.
Therefore, the sample size needed to detect a 20% reduction
in mean total stool output, and duration of diarrhea was
calculated, assuming a 5% level of significance and 80%
power. The calculation suggested that 40 patients per
treatment group (totaling 160 patients) would be needed.
The study was conducted in a 16-bed metabolic unit at
the Fima Lifshitz Research Center of the Federal University
of Bahia, where approximately 500 patients fulfilling the
eligibility criteria were admitted annually. The Research
Ethics Committee of the University Hospital, Federal
University of Bahia, approved the study, and all parents
or legal guardians gave written informed consent to
participate. Baseline examination was conducted to deter-
mine the eligibility of the patients for the study. Initial data
collected included symptoms before admission, previous
treatments attempted, and changes in feeding practices
before admission.
Children with some dehydration were administered the
standard salt ORS, at a dose of 100 mL/kg, during a period of
6 hours. Hydration status was evaluated every 3 hours, and
464 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
children with severe dehydration were administered intra-
venous fluids, at 100 mL/kg, for 2 to 3 hours or until
rehydration was achieved. In both cases, maintenance-phase
therapy was initiated at the end of the rehydration phase,
which was continued until the diarrheal episodes ceased
(passage of 2 semiformed/well-formed stools or no stools at
all for 24 hours). For proper separation of urine from stools,
only male infants were included. The infants were placed in
metabolic beds throughout the study, and stool losses (by
weight) were replaced with an equivalent volume of ORS.
The frequency and volume of vomiting were determined
using preweighed pads. Total water, ORS, and diet intakes
were recorded throughout the duration of diarrhea. Breast-
milk intake was determined by weighing the child before and
immediately after each breast-feeding. Nude body weight
was systematically measured at admission, daily, and at
discharge. Body length was measured on a standard length
board, both at admission and at discharge. Blood samples
were routinely taken at random intervals for the determina-
tion of Na, K, Cl, and red blood cells, in addition to the
possible presence of the human immunodeficiency virus.
Stool samples were dispatched for bacterial, viral, and
parasitological analysis, and other laboratory investigations
were carried out when clinically indicated.
2.1. Dietary plans
Four different diets were used in this study (Table 1). All
of them were nutritionally complete, with caloric equiva-
lence and similar nutrient proportions: diet 1 was yogurt
based; diet 2 was isolated soy-protein based; diet 3 was
hydrolyzed casein based; diet 4 was amino acid based. These
4 diets were randomly assigned to the patients studied. The
randomization code was prepared by a member of the study
center not involved in the trial according to 4 randomization
charts based on permuted blocks of different lengths.
Throughout the study, these diets were offered to the patients
as the only source of nutrients other than breast milk.
Because diet 1 was not an industrialized formula, it was
adjusted by adding an equal part of mineral water to the
natural yogurt, and the final solution was determined to
contain 5% carbohydrate (sucrose) and 2% fat (soy oil). In
Table 1
Composition of the 4 diets/100 mL
Diet 01 Diet 02 Diet 03
(yogurt) (soy) (hydrolyzed) acid–based)
kJ 318 276 285
Protein (g) 1.9 (15) 1.8 (11) 1.9 (11)
Carbohydrate (g) 7.4 (39) 6.7 (40) 6.9 (40)
Lactose (g) 2.4 — —
Lipids (g) 3.9 (46) 3.6 (49) 3.8 (49)
Zinc (mg)/ 1.8/625 0.6/254 0.5/260
Vitamin A (UI )
Osmolality (mOsm/L) 353 200 330
Values are in n (%), unless otherwise stated.
addition, to match the recommended dietary allowances of
the diets, patients who were randomized to diet 1 received a
micronutrient supplement (Centrum). Diets were initiated at
randomization (0 time), which was considered to be (a) the
time of admission for the patients with no dehydration and
(b) the end of the rehydration phase, for the dehydrated ones.
2.2. Statistical analyses
Data were analyzed statistically using the SPSS software,
version 10.0 for Windows (SPSS, Chicago, Ill), and
nonparametric tests were applied using the Stata Corp
software, 2001 (Stata Corp, College Station, Tex). Descrip-
tive statistics were obtained for all the variables in each
group. The Kolmogorov-Smirnov test was applied to
determine adherence to theoretical normal distribution [25].
Data regarding patient admission criteria were as follows:
age; duration of diarrhea before admission; duration of
exclusive breast-feeding; weaning age; serum levels of Na,
K, and Cl; hematocrit; and hemoglobin. When values were
found to have normal distribution, they were compared
among the 4 groups by analysis of variance (ANOVA).
When a significant difference was detected among groups by
ANOVA, the Bonferroni posttest was applied. In addition,
the nonparametric Kruskal-Wallis test, followed by a multi-
ple comparison posttest, was applied to analyze primary
variables such as duration of diarrhea (in hours), total stool
output (in grams per kilograms), and weight gain (in grams).
Finally, the χ2 test was applied to compare the proportion of
the qualitative variables [26]. For survival analysis, the
duration of diarrhea was considered the “survival time” and
the cessation of diarrhea was regarded as “failure event.” The
Kaplan-Meier test (Pearson logrank test) was used to
compare the survival curves [27,28]. The level of signifi-
cance was set at 5% for all analyses.
3. Results
A total of 155 male patients were randomized, and only 1
of them was not included in the analysis because on the
fourth day of hospitalization, he developed signs of sepsis
and worsening of diarrhea, thus requiring specialized clinical
management. Of the 154 patients studied, 40 were
randomized to diet 1 (yogurt-based formula), 38 to diet 2
(soy-based formula), 38 to diet 3 (hydrolyzed casein–based
formula), and 38 to diet 4 (amino acid–based formula). At
Diet 04 (amino admission, the characteristics of all groups were similar at
5% level of significance (Table 2). Most of the children,
297 77.9% (120/154), were on their habitual diet, consuming
1.9 (11) normal lactose levels during diarrhea and before admission;
8.1 (45)
14.3% (22/154) were started on a milk-free diet; 7.1% (11/
—
3.5 (44) 154) used a soy-based formula, and only 1 patient was
0.7/264 started on a chicken-based diet. According to this variable,
patient distribution was similar in all 4 groups (P = .68).
360 Only 27.3% of the patients (42/154) were dehydrated at
admission. Of these, only 4.8% (2/42) had severe dehydra-
 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469 465

Table 2
Baseline characteristics of subjects in groups a
Baseline characteristics at enrollment Diet 01 Diet 02 Diet 03 Diet 04 P
(yogurt; n = 40) (soy; n = 38) (hydrolyzed; n = 38) (amino acid–based; n = 38)
Mean age (mo) 11.3 ± 8.4 10.0 ± 7.2 12.3 ± 6.9 10.7 ± 7.4 .60 b
Mean weight (kg) 8.0 ± 2.6 7.7 ± 2.7 8.1 ± 2.4 8.2 ± 2.5 .89 b
Mean z score W/H −1.1 ± 1.0 −1.0 ± 1.1 −1.3 ± 1.3 −1.0 ± 1.1 .42 b
Mean length of diarrhea upon enrollment (d) 21.5 ± 8.9 22.6 ± 11.6 24.1 ± 15.6 22.4 ± 14.9 .83 b
Mean number of loose or watery stools 7.4 ± 3.5 8.0 ± 3.3 7.0 ± 3.3 7.9 ± 3.4 .62 b
Mean age of weaning (mo) c 2.7 ± 3.2 5.6 ± 5.9 4.3 ± 5.7 3.9 ± 3.7 .15 b
Mean duration of exclusive breast-feeding (mo) 1.9 ± 2.2 1.9 ± 2.1 1.8 ± 1.8 2.5 ± 2.6 .49 b
Percent of patients with bloody stools 20.0 26.3 18.4 21.4 .85 d
Percent of patients that used antibiotics 52.5 57.9 39.5 47.4 .41 d
Percent of patients with vomiting 52.5 68.4 68.4 63.2 .42 d
Percent of patients with fever 65.0 71.1 65.8 57.9 .69 d
Percent of patients who had previous episodes of PD 5.0 7.9 18.4 10.5 .26 d
Percent of households with basic sanitary conditions 12.5 10.5 7.9 5.3 .69 d
Percent of patients in use of full lactose-containing diets 67.5 78.9 84.2 81.6 .68 d
a
Data are presented as mean ± SD or percents.
b
There was no statically difference between the groups using ANOVA (P b .05).
c
n = 115 (39 children still breast-feeding).
d
There was no statically difference between the groups when using Q-Square test for the evaluation of the proportions.

tion, and 95.2% (40/42) had some dehydration. Proportional
patient distribution in the presence or absence of dehydration
was similar among the 4 groups studied (P = .16), and the
average length of diarrheal condition was 40.8 hours. While
analyzing this variable, a statistically significant difference
was observed between the 2 groups that used the soy-based
and hydrolyzed casein–based formula and the 2 groups that
used the yogurt-based and amino acid–based formula. The
children receiving diets 1 and 4 showed a shorter duration of
diarrhea (Fig. 1), and the children receiving the yogurt-based
and amino acid–based formulas had a smaller stool output,
compared with the 2 groups receiving the soy-based and
hydrolyzed casein–based formula. No significant difference
was observed between groups 1 and 4 (Table 3).
As expected, ORS intake showed the same trend as stool
output, being significantly smaller in groups 1 and 4 than in
groups 2 and 3. Only 8.4% (13/154) of the patients needed
unscheduled intravenous rehydration therapy. No significant
Fig. 1. Survival analyses: Kalan-Meier, by random diet.
difference was observed between groups with reference to
this variable (P = .14). Only 24.6% (38/154) of the children
consumed human milk and that too in very small amounts,
with no significant difference in consumption among the
various groups (P = .31).
Statistical analysis of the intake of the 4 diets showed
significant differences among groups. The Bonferroni
posttest showed that the consumption of diets 3 and 4 was
smaller than the consumption of diet 1 (P = .03). There was
no significant difference among the 4 groups in regard to the
requirement of tube feeding that was needed in 25.3% (39/
154) of the patients (P = .14).
Most of the 154 patients studied (145, 94.2%) fulfilled the
diarrhea cessation and discharge criteria of the study after the
use of the randomized diet. The need to change the
randomized diet occurred in only 9 patients (5.8%), 5 of
whom had been randomized to diet 2 (13.2% failure) and 4,
to diet 3 (10.5% failure). No patient randomized to diets 1
and 4 needed to change his or her diet. Furthermore, of the 5
patients for whom diet 2 failed, 3 had a positive outcome
when switched to diet 3, and 2 had a positive outcome only
after a second change to diet 4. All the 4 patients who failed
with diet 3 recovered after being switched to diet 4.
Regarding the variable of weight gain, there was no
difference among groups. Regardless of the diet, all the 4
groups showed a positive average weight gain. In fact, at
discharge, the nutritional status had improved in all 4 groups.
According to the z score, the average weight/height index
was −1.11 at admission, whereas it was −0.9 at discharge
(Fig. 2). Of the total number of patients studied, 73.4% (113/
154) returned 1 week later for reevaluation. Of these, 93.8%
(106/113) had no diarrhea, and 81.4% (92/113) showed a
mean weight gain of 256 g. The statistical analyses of these 2
variables, that is, presence or absence of diarrhea (P = .81)
and weight gain after discharge (P = .75), did not suggest any
466 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
Table 3
Total stool output of the subject (g/kg)
Diet 01 (yogurt; n = 40) Diet 02 (soy; n = 38)
Mean ± SD 105.5 ± 288.7 279.1 ± 469.26
Minimum 0.00 0.00
P 25 0.00 5.1
Median 9.0 44.8
P 75 75.4 394.6
Maximum 1428.9 1851.2
There was difference between the groups (P = .01) identified by the Kruskal-Wallis test, assuming a P b .05. The posttest (Stata; Caci, 1999) for the Kruskal-
Wallis showed that the difference was between the group that consumed the diet 1 and the groups that consumed the diets 2 (P = .006) and 3 (P = .0008). P 25
indicates 25th percentile; P75, 75th percentile.
difference among the 4 groups. In addition, the incidence of
associated infections in the population was 38.9% (60/154),
with 22.7% (35/154) of them being bacterial infections; 9.1%
(14/154), viral respiratory infections; and 7.1% (11/154), oral
candidiasis. All the study groups were similar in terms of the
occurrence of associated infections (P = .56). The most
prevalent bacterial infection was dysentery, affecting 11.7%
(18/154) of the patients, followed by sepsis (identified in
5.2% of the cases or 8/154) and pneumonia (occurring in
2.6% or 4 of the 154 children studied). There were also 2
cases of sinusitis, 1 case of otitis, 1 case of urinary tract
infection, and 1 case of impetigo. Finally, the groups were
also similar in terms of the type of infection (P = .44).
Oral antibiotics were used by 17.5% (27/154) of the
patients and intravenous antibiotics by 8.4% (13/154), with
no significant difference between the groups (P = .61 and
P = .82). Regarding the etiological pattern of PD, an
enteropathogenic agent was identified in 48% (74/154) of the
cases and more than 1 pathogenic agent was detected in 16%
(12/74) of the cases. The prevalence of rotavirus was 13.3%
(19/143), with no significant difference in the rate of
positivity among the 4 groups (P = .89). In addition,
adenovirus was identified in 4.9% (7/143) of the patients,
with no significant difference in the rate of positivity among
the 4 groups (P = .13). The percentage of positive stool
cultures did not differ significantly among the groups (P =
.73). The isolation rate of enteropathogenic bacteria was
11.2% (16/143), and enteropathogenic Escherichia coli was
the most prevalent agent, being detected in 68.7% (11/16) of
Fig. 2. Nutritional status of the studied patients on admission and on
discharge according to the z-score of weight for length, by random diet.
Diet 03 (hydrolyzed; n = 38) Diet 04 (amino acid–based; n = 38)
177.7 ± 242.6 135.5 ± 261.9
0.00 0.00
22.0 6.5
108.1 35.4
233.1 119.8
1228.1 1346.9
all positive stool cultures. Finally, Salmonella sp and
enteroinvasive E coli were also isolated from the fecal
cultures of 3 (2%) and 2 (1.4%) patients, respectively.
Parasites were isolated from 29.7% (44/148) of the
patients, with no significant difference in positivity among
the 4 groups (P = .48). In fact, by far, the most predominant
parasite identified was Cryptosporidium, which was detected
in 52.3% of all positive samples (23/44), with no significant
difference in distribution among the 4 groups (P = .81).
Considering the 148 patients as a whole, the following
parasite prevalence was observed: 15.5% (23/148) of Cryp-
tosporidium parvum, 6.1% (9/148) of Giardia lamblia, 5.4%
(8/148) of Ascaris lumbricoides, 2% (3/148) of Blastocystis
hominis, 1.3% (2/148) of Trichocephalus trichiurus, 1.3%
(2/148) of Endolimax nana, and 0.7% (1/148) of Isospora
belli and Chilomastix. The samples were tested positive for
more than 1 agent in only 5 patients (3.4%). Parasite
distribution was similar for all the groups (P = .48). Finally,
the prevalence of human immunodeficiency virus–positive
patients was 4% (6/150), with no statistically significant
difference among the 4 groups (P = .88).
4. Discussion
The potential benefits of improved nutrition for the
clinical outcome of diarrheal diseases have been recognized
for a long time [17,21-24]. Persistent diarrhea imposes
severe disruption, to different extents, of the small-intestinal
microbiota and of the mucosal, a fact that necessitates proper
nutritional management of PD cases [29-35]. When design-
ing a proper nutritional management regime, the transitory
difficulties of the digestive-absorptive process should be
considered, in addition to the need to promote an ideal
nutrient incorporation with a consequent improvement of
nutritional status, which—in turn—will favor the reestab-
lishment of normal physiology of the small intestine.
Although many agree that a more efficient approach to
PD should emphasize on nutritional management, no
systematic protocol has been validated in international
literature thus far. The multiplicity of dietary options that
have been proposed in many studies is the reflection of this
lack of both uniformity and consensus about the most
efficient nutritional approach for treating children with PD
 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
[36,37]. Surely, the broad spectrum of manifestations related
to the duration of the episodes, degree and extent of mucosal
damage, and previous nutritional status leads to different
tolerance patterns of various nutrients. This condition adds
much complexity to the issue of deciding the ideal diet.
A rational selection of an efficient diet (preferentially
complete) requiring little or no manipulation will greatly
contribute to the standardization of a more successful
nutritional management of PD. In the present study,
considering all the patients tested, a 94.2% success rate was
obtained, showing that the 4 diets proved to be effective in the
nutritional management of PD. Nevertheless, it was possible
to identify different degrees of efficacy between them.
First, the patients who were randomized to the yogurt-
based diet (diet 1) and to the amino acid–based diet (diet 4)
showed a shorter duration of diarrhea and less stool output
when compared with the patients who were randomized to
the isolated soy-based diet (diet 2) and to the hydrolyzed
casein-based diet (diet 3). Second, in addition to the superior
impact on stool output and duration of diarrhea, another
aspect that seems to show the greater efficacy of diets 1 and 4
was the fact that no patient randomized to these diets needed
to change his original diet, indicating that the rate of success
of these diets was 100%. Diets 2 and 3 showed a 13.2% (5/
38) and 10.5% (4/38) failure rate, respectively.
Diet 4 was superior for the management of patients whose
previous diet had failed, with a success rate of 100%. The
amino acid–based diet promoted the improvement of the 4
patients who were previously randomized to diet 3 (and had
failed) and of the 2 patients who, having failed the
randomized diet 2, were switched to diet 3 and had failed
all over again. Diet 3 was successful in managing 3 of the 5
patients who failed diet 2. This consistency of findings
strongly suggests that although showing distinct organolep-
tic aspects and different degrees of complexity and protein
sources, diets 1 and 4 had characteristics that promoted the
better performance observed.
Although great care in the design and use of appropriate
methodology was followed for the study, the presence of
specific characteristics that may have acted as the covari-
ables in each diet could not be avoided. The fact that the
hydrolyzed casein-based and amino acid–based diets have
potential digestive advantages over the other diets, theore-
tically, in situations associated with malabsorption, they
could even provide a better and shorter recovery. However,
in the present study, only the amino acid–based diet had this
effect, whereas the performance of the hydrolyzed casein-
based diet was not as good as the yogurt-based diet.
Despite the qualitative variations, all diets had similar
proportions and quantities of macronutrients. Thus, the
yogurt-based diet was carefully reformulated to contain
similar amounts of protein, fat, and carbohydrate. In addition,
to make this diet comparable to the others, micronutrients
were offered as a supplement as suggested in a previously
conducted WHO study .The same dosage and source
(Centrum) as described in the above-cited study were used
467
[16]. However, this intervention led to an imbalance of diet 1,
which had a greater concentration of zinc. In fact, a high zinc
concentration has been reported to have a positive effect on
the clinical course of PD patients [38-41]. More recently, the
WHO and United Nations Children's Fund even included
zinc supplements (10-20 mg/d for 10-14 days) in the new
guidelines for the treatment of diarrhea [42]. In the present
study, whereas the average daily zinc consumption was 19
mg for the diet 1 group, it was 6, 5.3, and 5.4 mg for the diets
2, 3 and 4 groups, respectively. Because of this increased
average, the possibility that the yogurt-based formula may
have had the highest zinc concentration as a coadjuvant
therapeutic agent (apart from the low-lactose content and the
presence of probiotic agents) must be admitted.
The present population was basically composed of
patients with mild to moderate PD because the variables
that classically characterize greater severity, such as
dehydration, concurrent bacterial infections, moderate to
severe malnutrition (z score, b−2), age less than 6 months,
and occurrence of previous episodes of PD were present only
in a small proportion of the patients studied, that is, 23.7%,
22.7%, 22.7%, 30.5%, and 10.4%, respectively.
Notwithstanding the apparent contradiction of the
suboptimal performance of a hydrolyzed casein-based diet
compared with a whole protein diet, it is herein speculated
that, similar to the process in acute diarrhea in which the
intestinal mucosal damage does not necessarily require the
use of hypoallergenic or more complex diets, perhaps in
PD also, the particular symptoms of severe patients should
not be extrapolated to the general population of PD
patients. Despite their different protein sources, all 4 diets
promoted the cessation of diarrhea and improved the
nutritional status of most patients studied. In conclusion, it
seems reasonable to assume that the absorption of
macromolecules and its well-known antigenic potential do
not play a major role in the pathophysiology of mild to
moderate PD. Probably, the differentiated performance of
diet 4 was more due to its extreme digestibility and
excellent nutrient proportion rather than to its hypoaller-
genic feature. Although a more elemental diet is recognized
to have a lower capacity of inducing mucosal trophism, it
seems that in the acute phase of PD, digestibility is more
important in terms of short-term recovery.
Regarding the source of protein in diet 1, whole-milk
protein (known worldwide as the most allergenic protein for
infants and children) did not seem to interfere with the
improvement and recovery of the patients. Possibly, in
addition to this nonallergenic PD pattern (which was clearly
shown in the present study), the functional benefits of
yogurt, such as probiotic effects and the presence of enzymes
that greatly contribute to intraluminal digestion, also led to
this differentiated performance of diet 1. When considering
the hazards of milk protein against the lower antigenicity of a
hydrolyzed protein, it seems that some peculiarities of the
yogurt composition played a more important role in patient
outcome in the present study. In addition, it also is very clear
468 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
that the benefits of yogurt use for the nutritional management
of PD cannot be explained only by its low-lactose levels
because its performance was superior to that of the other 2
lactose-free diets.
Two potential limitations of this study should be
considered. First, because it was a hospital-based study, the
applicability of these findings to community settings can
only be guessed. Second, the current data do not permit a
conclusion about the management of severe patients.
However, considering that delayed recovery after successive
failures using less complex diets is strongly associated with
the worsening of clinical and nutritional status, it is, hence,
reasonable to admit that when available, the amino acid–
based diet should be the first choice for the nutritional
management of severe PD cases. Furthermore, even not
offering any specific advantage and despite their respective
failure rates of 13.2% and 10.4%, the use of soy-based and
casein hydrolysate–based diets is acceptable but only if
yogurt-based or amino acid–based diets are not available. In
a recent publication on the evidence gaps in PD, the authors
affirm that although, in general, there is no reason to provide
specialized formula, the hospital-based management of PD-
complicated cases may require parenteral feeding and the use
of special formulas [43]. Also, the concerns about the
malabsorption of key nutrients in PD remain and are
important and a very actual issue [44].
To conclude, based on our findings, the yogurt-based diet
should be granted priority in the nutritional management of
mild to moderate PD. Nevertheless, in the absence of a
satisfactory response to a yogurt-based diet, an amino acid–
based diet, if available, should be considered the second
choice in the nutritional management of those patients. When
these results are compared with the results of the classical
WHO multicenter trial on nutritional PD management
(WHO, 1996), in which 11% to 35% of the patients did
not recover during the study and a longer duration of
hospitalization was needed, it can be surmised that a more
systematic and aggressive nutritional approach is necessary
because more complex diet options can lead to better success
rates in the severe cases. Finally, the establishment of
objective intervention criteria that can assure an earlier
feeding of an appropriate diet and the standardization of the
therapeutic changes of diet that might be necessary seem to
have a definite influence on the recovery course of these
patients. The establishment of these criteria will greatly
contribute to the adoption of a universal approach in PD
management to achieve more successful recovery rates and
to guide the rational utilization of the resources of public
health programs.
Acknowledgment
This study was supported by SHS International, which
also provided the amino acid–based formula. The authors
thank Danone, for providing the natural yogurt; Support, for
providing the soy-based formula; and Mead-Johnson, for
providing the hydrolyzed protein formula.
References
[1] World Health Organization. The magnitude of the global problem of
acute diarrhoeal disease: a review of active surveillance data. Bull
World Health Org 1982;60:605-13.
[2] World Health Organization. The global burden of diarrhoeal disease, as
estimated from studies published between 1992 and 2000. Bull World
Health Org 2003;81:197-204.
[3] Gracey M. Diarrheal disease in perspective. In: Gracey M, Walker-
Smith JA, editors. Nestlé Nutrition Workshop Series. Philadelphia:
Lippincott-Raven Publishers; 1997. p. 1-11.
[4] Victora CG, Huttly SR, Fuchs SC, Barros FC, Garenne M, Leroy O,
et al. International differences in clinical patterns of diarrhoeal deaths: a
comparison of children from Brazil, Senegal, Bangladesh and India.
J Diarrhoeal Dis Res 1993;11(1):25-9.
[5] World Health Organization. Persistent diarrhoea in children in
developing countries: memorandum from a WHO meeting. Bull
World Health Org 1988;66:709-17.
[6] Bhan MK, Bhandari N, Sazawal S, Clemens J, Raj P, Levine MM, et al.
Descriptive epidemiology of persistent diarrhea among young children
in rural northern India. Bull World Health Org 1989;67:281-8.
[7] Fauveau V, Henry FJ, Briend A, Yunus M, Chakraborty J. Persistent
diarrhoea as a cause of mortality in rural Bangladesh. Acta Paediatr
Suppl 1992;381:12-4.
[8] Victora CG, Huttly SR, Fuchs SC, Nobre LC, Barros FC. Deaths due to
dysentery, acute and persistent diarrhoea among Brazilian infants. Acta
Paediatr Suppl 1992;81(s383):7-11.
[9] Mbori-Ngacha DA, Otieno JA, Njeru EK, Onyango FE. Prevalence of
persistent diarrhoea in children aged 3-36 months at the Kennyata
National Hospital, Nairobi, Kenya. East Afr Med J 1995;72:711-4.
[10] Andrade JAB, Gomes TAT, Fagundes-Neto U. Letalidade em lactentes
com diarréia persistente: fatores de risco associados ao óbito. Arq
Gastroenterol 1998;35:62-8.
[11] Lins MGM, Silva GAP. Doença diarréica em crianças hospitalizadas:
importância da diarréia persistente. J Pediatr (Rio J) 2000;76:37-43.
[12] Bhandari N, Bhan MK, Sazawal S, Clemens JD, Bhatnagar D,
Khoshoo V. Association of antecedent malnutrition with persistent
diarrhea: a case control study. Br Med J 1989;298:1284-97.
[13] Schorling JB, McAuliffe JF, De Souza MA, Guerrant RL. Malnutrition
is associated with increased diarrhoea incidence and duration among
children in an urban Brazilian slum. Int J Epidemiol 1990;19:728-35.
[14] Baqui AH, Black RE, Sack RB, Chowdhury HR, Yunus M, Siddique
AK. Cell-mediated immune deficiency and malnutrition are indepen-
dent risk factors for persistent diarrhea in Bangladeshi children. Am J
Clin Nutr 1993;58:543-8.
[15] Bhutta ZA, Hendricks KM. Nutritional management of persistent
diarrhea in childhood: a perspective from developing world. J Pediatr
Gastroenterol Nutr 1996;22:17-37.
[16] World Health Organization. Evaluation of an algorithm for the
treatment of persistent diarrhoea: a multicentre study. Bull World
Health Org 1996;74:1479-89.
[17] MacLean WC, Lopez RG, Massa E, Graham GG. Nutritional
management of chronic diarrhea and malnutrition: primary reliance
on oral feeding. J Pediatr 1980;97(2):316-23.
[18] Bhan MK, Arora NK, Khoshoo V. Chronic diarrhea in infants and
children. Indian J Pediatr 1985;52:483-95.
[19] Penny ME, Brown KH. Lactose feeding during persistent diarrhoea.
Acta Paediatr Suppl 1992;381:133-8.
[20] Bhatnagar S, Bhan MK, Singh KD, Shrivastav R. Prognostic factors in
hospitalized children with persistent diarrhea: implications for diet
therapy. J Pediatr Gastroenterol Nutr 1996;23:151-8.
[21] Molla AM, Molla A, Khatun N, Khatun M. Feeding in diarrhea during
the acute stage and after recovery: experience in developing countries.
 Â.P. de Mattos et al. / Nutrition Research 29 (2009) 462–469
In: Lifschitz CH, Nichols B, editors. Malnutrition in chronic diet-
associated infantile diarrhea. California: Academic Press; 1990.
p. 293-303.
[22] Bhutta ZA, Molla AM, Issani Z, Badruddin S, Hendricks K, Snyder
JD. Nutrient absorption and weight gain in persistent diarrhea:
comparison of a traditional rice-lentil-milk diet with soy formula.
J Pediatr Gastroenterol Nutr 1994;18:45-52.
[23] Nurko S, Garcia-Aranda JB, Fishbein E, Perez-Zuniga MI. Successful
use of a chicken-based diet for the treatment of severely malnourished
children with persistent diarrhea: a prospective, randomized study.
J Pediatr 1997;131:405-12.
[24] Bhutta ZA, Molla AM, Issani Z, et al. Dietary management of
persistent diarrhea and malnutrition: comparison of a traditional rice-
lentil-based diet with soy formula. Pediatrics 1991;88:1010-8.
[25] Fleiss JL, Levin B, Paik MC, Fleiss J. Statistical methods for rates and
proportions. 3rd ed. Malden, MA: Wiley-Interscience; 2003.
[26] Rosner B. Fundamentals of biostatistics. 3rd ed. Boston, MA: PWS
Kent pub; 1990.
[27] Cleves M, Gould W, Gutierrez R, Marchenko Y. An introduction to
survival analysis using Stata. 2nd ed. College Station, TX: Stata Press;
2008.
[28] Lee ET, Wang JW. Statistical methods for survival data analysis.
3rd ed. Malden, MA: Wiley-Interscience; 2003.
[29] Challacombe DN, Richardson JM, Rowe B, Anderson CH. Bacterial
microflora of the upper gastrointestinal tract in infants with protracted
diarrhea. Arch Dis Child 1974;49:70-7.
[30] Barnes GL, Townby RW. Duodenal mucosal damage in 31 infants with
gastroenteritis. Arch Dis Child 1973;48:343-9.
[31] Mann MD, Hill ID, Peat GM, et al. Protein and fat absorption in
prolonged diarrhea in infancy. Arch Dis Child 1982;57:268-73.
[32] Lebenthal E, Lee PC. Glucoamylase and disaccharidase activities in
normal subjects and in patients with mucosal injury of the small
intestine. J Pediatr 1980;97:389-93.
[33] Khoshoo V, Bhatnagar S, Bhan MK. Monosaccharide intolerance
complicating protracted diarrhea in infants. J Pediatr Gastroenterol
Nutr 1989;9:131-2.
469
[34] Schneider RE, Viteri FE. Luminal events of lipid absorption
in protein-caloric malnourished children; relationship with nutri-
tional recovery and diarrhea. I. Capacity of the duodenal content to
achieve micellar solubilization of lipids. Am J Clin Nutr 1974;
27:777-87.
[35] Schneider RE, Viteri FE. Luminal events of lipid absorption in protein-
caloric malnourished children; relationship with nutritional recovery
and diarrhea. II. Alterations in bile acid content of duodenal aspirates.
Am J Clin Nutr 1974;27:788-96.
[36] Iyngkaran N, Davis K, Robinson MJ. Cow's milk protein-sensitivity
enteropathy. An important factor prolonging diarrhea in acute infective
enteritis in early infancy. Arch Dis Child 1978;53:150-3.
[37] Jirapinyo P, Young C, Srimaruta N, Rossi TM, Cardano M, Lebenthal
E. High-fat semi-elemental diet in the treatment of protracted diarrhea
of infancy. Pediatrics 1990;86(6):902-8.
[38] Castillo-Duran C, Vial P, Uauy R. Trace mineral balance during acute
diarrhea in infants. J Pediatr 1988;113:452-7.
[39] Roy SK, Tomkins AM. Impact of experimental zinc deficiency on
growth, morbidity and ultrastructural development of intestinal tissue.
Bangladesh J Nutr 1989;2:1-7.
[40] Roy SK, Behrens RH, Haider R, Akramuzzaman SM, Mahalanabis
D, Wahed MA. Impact of zinc supplementation on intestinal
permeability in Bangladeshi children with acute diarrhea and
persistent diarrhea syndrome. J Pediatr Gastroenterol Nutr 1992;
15:289-96.
[41] Sazawal S, Black RE, Bhan MK, Bhandari N, Sinha A, Jalla S. Zinc
supplementation in young children with acute diarrhea in India. N Engl
J Med 1995;333:839-44.
[42] WHO, UNICEF, MOST. Diarrhoea treatment guidelines: including
new recommendations for the use of ORS and zinc supplementation;
2005.
[43] Bhutta ZA, Nelson EA, Lee WS, Tarr PI, Zablah R, Phua KB, et al.
Recent advances and evidence gaps in persistent diarrhea. Pediatr
Gastroenterol Nutr 2008;47:260-5.
[44] Pawlowski SW, Warren CA, Guerrant R. Diagnosis and treatment of
acute or persistent diarrhea. Gastroenterology 2009;136:1874-86.