EPIDEMIOLOGIC AND CLINICAL CHARACTERISTICS OF ACUTE DIARRHEA
WITH EMPHASIS ON ENTAMOEBA HISTOLYTICA INFECTIONS IN PRESCHOOL
CHILDREN IN AN URBAN SLUM OF DHAKA, BANGLADESH
RASHIDUL HAQUE, DINESH MONDAL, BETH D. KIRKPATRICK, SELIM AKTHER, BARRY M. FARR,
R. BRADLEY SACK, AND WILLIAM A. PETRI, JR.
Centre for Health and Population Research, International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), Dhaka,
Bangladesh; Department of Medicine, Unit of Infectious Disease, University of Vermont College of Medicine, Burlington, Vermont;
Departments of Medicine, Microbiology and Pathology, University of Virginia, Charlottesville, Virginia; Department of International
Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland
Bangladeshi children (147 boys and 142 girls) 2−5 years old. The use of improved diagnostic tests for amebiasis enabled
for the first time analysis of the contribution of Entamoeba histolytica to total diarrheal illness in this community setting.
The average incidence rate of diarrhea was 1.8/child-year, and the average number of diarrheal days was 3.7 days/child-
year over an average observation period of 2.8 years/child. Seventy-five percent of the diarrheal episodes were ? 2 days
in duration. Persistent diarrhea was relatively uncommon (0.2% of the children) and chronic diarrhea was observed in
only one episode. Compared with malnourished and/or stunted children, better-nourished children experienced signifi-
cantly fewer diarrheal episodes. The diarrheal incidence rate for children with blood group A was significantly less that
that of the children with blood groups O and AB. The most frequent bacterial enteropathogens isolated from diarrheal
stool specimens were enterotoxigenic Escherichia coli (9%) and Aeromonas species (9%), followed by Plesimonas
shigelloides (4%) and Shigella flexneri (3.8%). Rotavirus was the most common viral agent isolated from diarrheal stool
samples (5%). Giardia lamblia, Cryptosporidium parvum, and E. histolytica were identified in 11%, 8.4%, and 8%,
respectively, of the diarrheal stool specimens. Dysentery was observed in 7.7% of all diarrheal episodes. The most
common pathogens isolated from dysenteric stool were S. flexneri (11.6%), Aeromonas sp. (10%), E. histolytica (8.7%),
Campylobacter jejunii (5.8%), P. shigelloides (4.3%), and A. caviae (4.3%). The overall incidence rate of E. histolytica-
associated diarrhea was 0.08/child-year. Visible blood and hemoccult test-detected blood loss was found in 7% and 25%,
respectively, of cases of E. histolytica-associated diarrhea. Children who had recovered from a diarrheal episode with E.
histolytica, but not E. dispar, had half the chance of developing subsequent E. histolytica-associated diarrhea, consistent
with the development of species-specific acquired immunity. In conclusion, the use of modern diagnostic tests demon-
strated that E. histolytica contributed to overall morbidity from diarrheal illness. Understanding the etiology, frequency,
and consequences of acute diarrhea in children from a developing country should aid in the design of interventions to
improve child health.
The epidemiology, clinical features, nutritional status, and causative agents of diarrhea were studied in 289
The rate of mortality from diarrheal diseases in the world
has decreased, mainly because of better therapy and inter-
ventions that promote sanitary conditions and that educate
inhabitants to encourage them to take part in primary health
care activities.1,2However, acute diarrheal diseases continue
to be one of the major causes of morbidity and mortality in
the developing world such as Bangladesh, where one in 10
children die before their fifth birthday.1,3An epidemiologic
study of an infectious disease in a community is an initial step
toward the introduction of the proper interventions for con-
trolling the disease because the features and the patterns of
isolation of etiologic agents of the disease vary from place to
place depending on the local meteorology, geography, and
Human infection with Entamoeba histolytica, known, as
amebiasis, is prevalent worldwide and is common in children
of the developing world. Entamoeba histolytica is associated
with diarrhea/dysentery in endemic countries.6Recently E.
histolytica has been reclassified into pathogenic (E. histolytica
sensu stricto) and nonpathogenic (E. dispar) species.7The
classic stool ova and parasite examination, whereby E. his-
tolytica is identified by its appearance in trichrome- or iron
hematoxylin-stained stool specimens, is insensitive and can-
not differentiate E. histolytica from the nonpathogenic but
identical-appearing parasite E. dispar. A few surveys in which
the two species are distinguished have been carried out in
endemic countries, and it has been found that the prevalence
of E. histolytica infection varies greatly from place to
Reliable epidemiologic data are essential to estimate the
burden of disease due to E. histolytica and to formulate policy
to control amebiasis. Amebiasis is endemic in Bangladesh.
Case series of patients at the International Centre for Diar-
rhoeal Disease Research, Bangladesh (ICDDR,B) Hospital
in Dhaka show peak incidences of infection among children
between two and 14 years of age and in adults > 40 years old.6
The field use of a rapid and specific antigen detection test for
E. histolytica infection is providing new insights into amebia-
sis. The prevalence of asymptomatic infection with E. his-
tolytica was unknown and until recently it was believed that
most asymptomatic infections were due to the non-pathogen
E. dispar. Currently, we are conducting a field study on ame-
biasis to understand the natural history of E. histolytica in-
fection in an urban slum of Dhaka. As a part of this study, we
are also conducting active diarrhea surveillance in a cohort of
2−5-year-old children at the time of enrollment. Here, we
describe the epidemiology, clinical features, nutritional status,
and causative agents of diarrhea, including E. histolytica, in
these children with the expectation that this information may
help in the ultimate control of diarrheal diseases.
MATERIALS AND METHODS
Study area and population. The study was conducted be-
tween January 1999 and July 2002 in Mirpur, an urban slum in
Dhaka as described elsewhere.12Two hundred eighty-nine
Am. J. Trop. Med. Hyg., 69(4), 2003, pp. 398–405
Copyright © 2003 by The American Society of Tropical Medicine and Hygiene
children (147 boys and 142 girls) 2−5 years old were enrolled.
Half of these children had IgG antibodies against the E. his-
tolytica galactose and N-acetyl-D-glactosamine (Gal/
GalNAc) lectin in their blood at the time of enrollment. The
amebic Gal/GalNAc lectin mediates adherence to and con-
tact-dependent cytolysis of human colonic epithelium. Serum
IgG antibodies to the lectin have been detected in patients
with amebiasis.8,9All enrolled children and their family mem-
bers received free primary health care services, including
medications, from the project office in Mirpur. Informed con-
sent was obtained from the parents or guardians and the hu-
man experimentation guidelines of the U.S. Department of
Health and Human Services, the University of Virginia, the
Johns Hopkins University Bloomberg School of Public
Health, and the Center for Health and Population Research,
ICDDR,B in Dhaka, Bangladesh were followed in conducting
The parents and children were visited and interviewed ev-
ery other day by health care workers for details about any
diarrheal episodes of the child, as well as other related ques-
tions. Children with diarrhea were also detected through their
parents contacting project personnel at the field clinic. When
diarrheal disease was detected, the child was examined and
treated with oral rehydration and/or antibiotics as appropri-
ate and a stool sample was collected for detailed investigation
of enteropathogens including E. histolytica. Stool specimens
were also collected monthly from every child for detection of
E. histolytica infection by antigen capture and culture.
Anthropometry. Anthropometric measurements were
taken by trained research assistants at the time of enrollment
and then every four months. Each child was weighed in light
clothes with an electronic weighing scale. The standing
heights of children were measured to the nearest 0.1 cm using
a locally constructed height stick. Nutritional status was as-
sessed by comparing the weight and height of the study chil-
dren with those of a National Center for Health Statistics
(Hyattsville, MD) reference population of the same age and
sex with the Epi-Info 6 version 6.04 (Centers for Disease
Control and Prevention, Atlanta, GA) computer program.13
Clinical definitions. Diarrhea was defined as having three
or more unformed stools in a 24 hour period. A diarrheal
episode was defined as being separated from another episode
by at least three diarrhea-free days. A new episode for asymp-
tomatic E. histolytica infection was defined as being separated
from another episode by at least two negative results in
monthly non-diarrheal stools for E. histolytica antigens.12The
presence of blood in the stool was defined by the gross ap-
pearance of blood in the stools, as indicated by the mother or
noted by study personnel. Diarrhea was further classified as
dysenteric or non-dysenteric. Dysenteric diarrhea was de-
fined by gross blood in the stools and/or microscopic stool
examination showing red blood cells ? 1/high-power field.
Diarrheal episodes were defined as acute (< 14 days), persis-
tent (? 14 days but < 30 days) and chronic (? 30 days) ac-
cording to duration. Episodes of E. histolytica-associated di-
arrhea or dysentery were defined as earlier in this report, but
were accompanied by the isolation of E. histolytica from di-
arrheal or dysenteric stools taken at the time of illness.
Asymptomatic E. histolytica infection was defined as having a
positive stool antigen detection test result for E. histolytica in
the absence of diarrheal illness. Fever was defined according
to the mother’s assessment. The project physician assessed
the degree of dehydration according to World Health Orga-
nization criteria.14The severity of each diarrheal disease epi-
sode was calculated by using a numeric scoring system.15
Stool sampling. Stool specimens were collected within 24
hours after the reporting of a diarrheal episode. Samples were
transported to the ICDDR,B laboratory within six hours after
collection. From January 1999 to December 2000, stool speci-
mens were transported to the laboratory without transport
media. From January 2001 onwards, Cary-Blair and buffer
glycerol saline BGS media were used for transportation of the
Stool microbiology. Stool specimens were cultured within
the same day of collection using standard methods.15The
stool samples were plated on MacConkey agar, Salmonella-
Shigella agar (SSA), taurocholate-tellerite-gelatin agar
(TTGA) and Campy-Brucella agar plate. The specimens were
also enriched in selenite F broth and bile peptone broth and
were subcultured onto SSA from the former and onto TTGA
from the latter. MacConkey agar and SSA were used for
isolation of Escherichia coli, Salmonella sp., and Shigella sp.,
TTGA was used for isolation of Vibrio cholerae, and Campy-
Brucella agar plate was used for isolation of Campylobacter
jejuni. All microbiologic media or their ingredients were ei-
ther Difco Products (Becton Dickinson Microbiology Sys-
tems, Sparks, MD) or Baltimore Biological Laboratories
products (Becton Dickinson and Company, Cockeysville,
Three lactose-positive colonies were picked from the Mac-
Conkey agar plates for identification of E. coli. Different cat-
egories of diarrheagenic E. coli: enterotoxigenic E. coli
(ETEC), enteropathogenic E. coli (EPEC), enteroaggrega-
tive E. coli, enterohemorrhagic E. coli, and enteroinvasive E.
coli were detected by a hybridization technique with specific
DNA probes.16The E. coli strains positive for hybridization
with EPEC adherent factor and/or the E. coli attachment and
effacement gene probe were identified as EPEC.17,18
Fecal specimens were also placed in phosphate-buffered
saline and kept in the freezer for later identification of ro-
tavirus, enteric adenovirus, and astrovirus by using commer-
cially available enzyme-linked imunosorbent assays
(IDEIA™Adenovirus, IDEIA™ Rotavirus, and IDEIA™
Astrovirus; Dako, Ely, United Kingdom).
Stool parasitology. The TechLab (Blacksburg, VA) Enta-
moeba test (designed to detect but not differentiate E. his-
tolytica and E. dispar antigen in stool specimens) and the
TechLab E. histolytica II test (designed to detect specifically
E. histolytica antigen in stool specimens) were performed on
the stool specimens according to the manufacturer’s instruc-
tions.11Cryptosporidium was also detected by an antigen de-
tection test (Cryptosporidium TEST; Techlab) according to
the manufacturer’s instructions. Stools were examined for ova
and parasites by direct microscopy, also for the identification
of Giardia lamblia, Ascaris lubricoides, Trichuris trichiura,
Strongyloides stercoralis, Hymenolepsis nana, and Cyclospora
Stool lactoferrin was detected by a latex agglutination assay
(Techlab), and fecal occult blood was detected by a hemoccult
test (Beckman Coulter, Inc., Palo Alto, CA). Blood typing
was done for all 235 children who remained in the study at the
beginning of the third year of follow-up. ABO and Rh blood
typing was done by conventional techniques.19
EPIDEMIOLOGY AND ETIOLOGY OF DIARRHEA IN AN URBAN SLUM
Statistical methods. Basic demographic information, sur-
veillance data, and clinical and laboratory findings of each
diarrheal episode for which stool sample was collected were
stored in data files using Fox-Pro®(Microsoft, Redmond,
WA). Categorical data were compared by chi-square analysis
with Fisher’s exact test. The relationship between numeric
and categorical data was analyzed by analysis of variance or a
Kruskal-Wallis test. Spearman correlation was used for cor-
relation analysis between numeric variables. Logistic regres-
sion was used to calculate adjusted odd ratios. The percentage
protection against subsequent infection was calculated as (1 -
odds ratio [OR]) × 100%. The statistical package SPSS ver-
sion 10.01 (SPSS, Inc., Chicago, IL) was used for data analy-
Socioeconomic characteristics. A total of 289 children from
252 households were enrolled in the study; 221 were followed
for 36 months. The average follow-up period was 1,036 days
(2.8 years). Baseline characteristics of the study population
are shown in Table 1. The average family size was six. The
percent of mothers who never attended school was 64%. Sev-
enty-two percent of the families had very low annual incomes
(< 5,000 Taka [Tk]; 1 US$ ? 58 Tk). Complete information
regarding sanitation could be given by 240 households. The
majority of the study population provided themselves with
water from a municipal supply by using plastic pipes, 18% of
which were near to latrines and 27% to drains. Only 33% had
direct accessibility to water from a municipal supply. Water
storage in the home was found in 86% of the households. The
use of food container covers was found in 95% of the house-
holds. An attached bedroom with a kitchen was found in
56%. Sixty-four percent of all adults and 26% of the study
children used sanitary latrines; pit latrines were used by 2.5%
of the study population. Mud floors in the house were seen in
27%. The proportion of malnourished and stunted children
was more common in the younger age groups (Table 1).
Diarrheal morbidity. Two hundred eight-nine children con-
tributed 299,616 child-days of observation for this study. Of
these, 3,046 child-days from 254 children (88%) were associ-
ated with diarrhea. The total number of episodes was
1,447. The average incidence rate of diarrhea was 1.8/child-
year, and the average number of diarrheal days was 3.7 days/
child-year. Seventy-five percent of the diarrheal episodes
were ? 2 days in duration (Figure 1). Persistent diarrhea was
relatively uncommon (0.2% of the children) and chronic di-
arrhea was observed in only one episode. The youngest age
group of children had significantly more diarrheal episodes,
but not E. histolytica-associated diarrheal incidence (Table 2).
Overall diarrheal morbidity was more in male children com-
pare with female children (Table 3); however, this difference
was not observed for E. histolytica-associated diarrheal illness
Relationship between diarrheal diseases and baseline nu-
tritional status of the children. Compared with malnourished
children, better-nourished children experienced significantly
fewer diarrheal episodes (Table 4). Similarly, children not
stunted at baseline had significantly less diarrheal morbidity
(Table 4). Entamoeba histolytica-associated diarrheal inci-
dence was also significantly less in better-nourished and not-
stunted children compared with malnourished and stunted
children (Table 4). However, asymptomatic E. histolytica in-
fection and E. histolytica-associated dysenteric incidence
were not related to the base line nutritional status of the
children (Table 4).
Effect of breast-feeding on diarrheal incidence. A complete
history of breast-feeding was available for 285 children.
Forty-seven children had a history of breast-feeding for less
than the first 12 months of life and had a mean diarrheal
incidence rate of 2.32 episodes/child-year. Two hundred
thirty-eight children were breast-fed for a mean of 12 months
and had an average diarrheal incidence of 1.68/child-year
(P < 0.05) (Table 4). Although not statistically significant, the
incidence rates of E. histolytica-associated diarrhea, dysen-
tery, and asymptomatic E. histolytica infection were higher in
children who were breast-fed less than 12 months (Table 4).
Relationship between blood group and diarrheal diseases.
Overall diarrheal illness showed a significant relation with
blood type. Children with blood group A experienced signifi-
cantly fewer episodes of diarrhea compare with the children
with blood group O and blood group AB (Table 5). However,
such a relationship was not observed for E. histolytica-
associated diarrheal morbidity (Table 5).
Baseline characteristics of the study population in Mirpur, Dhaka, Bangladesh*
(WAZ <−2) (male:female)
(HAZ <−2) (male:female)
Household information (n ? 252)
Average family size (minimum, maximum)
Maternal schooling, no. (%)
Never attended school
Attended 1–5 years
Attended 6–10 years
Attended >10 years
Annual income in Tk, no. (%)
<5,000 (<$U.S. 68)
5,000–10,000 ($US 68–172)
>10,000 (>$US 172)
6 (2, 17)
* WAZ ? weight-for-age Z-score; HAZ ? height-for-age Z-score; Tk ? Taka.
HAQUE AND OTHERS
Season and diarrheal incidence. The incidence of diarrhea
showed a marked seasonality, with significantly higher rates
being recorded in the months of March−August than in the
months of September−February; P < 0.0001) (Figure 2). En-
tamoeba histolytica−associated diarrheal incidence did not
show such a relationship with season.
Enteropathogens isolated from diarrheal stools. Of 1,447
diarrheal episodes, diarrheal stool samples from 893 episodes
(62%) were available for analysis. All were examined for
routine bacterial enteropathogens (Salmonella sp., Shigella
sp., V. cholerae, Campylobacter sp., Plesiomonas sp., Aero-
monas sp.) and parasitic agents. Of these 893 samples, a ran-
domly selected subset of 210 were also examined for diarrhe-
agenic E. coli and selected viral enteropathogens. Analysis of
these 210 stool samples that received a complete bacteriolog-
ic, parasitologic, and virologic work-up resulted in an isola-
tion rate for enteropathogens of 59%. Mixed infection with
two or more pathogens was found in 19%. The prevalence of
enteropathogens isolated from all 893 diarrheal stools is
shown in Table 6.
The most frequent bacterial enteropathoges isolated from
diarrheal stool specimens were ETEC (9%) and Aeromonas
sp. (9%), followed by P. shigelloides (4%) and S. flexneri
(3.8%). Rotavirus was the most common viral agent isolated
from diarrheal stool samples (5%). Giardia lamblia, C. par-
vum, and E. histolytica were identified in 11%, 10%, and 8%,
respectively, of the diarrheal stool specimens. Co-infection
was most frequent in 63% (12 of 19) of the cases with ETEC
and 44% (11 of 25) of the cases with C. jejuni. Co-infection of
E. histolytica and other bacterial enteropathogens was found
in 32% (22 of 69). Re-infection (isolation of a given pathogen
from different diarrheal stools of different diarrheal episodes
of the same child) was comparatively more common with E.
histolytica (14 of 43, 32%), C. parvum (6 of 16, 37%), Aero-
monas sp. (11 of 65, 16%), and A. hydrophilia (3 of 17, 17%).
Dysenteric and non-dysenteric episodes were observed in
7.7% and 92.3% of the specimens, respectively. The most
common pathogens isolated from dysenteric stool were S.
flexneri (11.6%), Aeromonas sp. (10%), E. histolytica (8.7%),
C. jejuni (5.8%), P. shigelloides (4.3%), and A. caviae (4.3%).
Entamoeba histolytica−associated diarrhea incidence and
association with diarrhea and dysentery. The overall inci-
dence rate of E. histolytica-associated diarrhea was 0.08/child-
year for the sampling rate of 62%. Of the 651 E. histolytica
antigen-positive monthly stool specimens, 25 (3.84%) were
followed by diarrhea and 6 (0.92%) with dysentery. Only 44
(0.92%) and 13 (0.15%) of 8,842 E. histolytica antigen-
negative monthly stool samples were followed by diarrhea
and dysentery, respectively. The age-adjusted OR for the as-
sociation between diarrhea and E. histolytica was 4.7 (95%
confidence interval [CI] ? 2.9−7.6). The age-adjusted OR
between dysentery and E. histolytica was 2.9 (95% CI ?
1.2−7.1). Of 43 children who contributed 69 E. histolytica-
associated diarrhea specimens, only 14 had a subsequent E.
histolytica-associated diarrheal episode. The crude OR for
developing a subsequent episode of E. histolytica-associated
diarrhea was 0.48 (95% CI ? 0.26–0.91).
Clinical findings of the first E. histolytica-associated diar-
rhea/dysentery episodes. The clinical data and some labora-
tory findings of E. histolytica-associated diarrhea are shown in
Table 7. The mean weight-for-age Z-score was -2.01 and 47%
of the affected children were malnourished. Abdominal pain
and mild-to-moderate dehydration were the most frequent
findings. The average duration of diarrheal episodes was
three days. Visible blood was found in 7% of the cases,
whereas blood loss was detected by the hemoccult test in 25%
of the cases. A fecal lactoferrin test result was positive in 20%
of the cases.
Asymptomatic E. histolytica infection incidence. To deter-
mine the incidence rate of asymptomatic E. histolytica infec-
tion and its relationship with other variables, including age,
sex, and nutritional status of the study children, we tested
9,493 non-diarrheal monthly stool specimens for E. histolytica
antigen. Six hundred seventy specimens were positive. Nine-
teen samples were positive at baseline and the remaining 651
were positive during follow-up. One hundred ninety-nine
(69%) children were positive at least once by the E. histolytica
antigen detection test. The total number of new episodes was
359. The incidence rate of asymptomatic infection with E.
histolytica was 0.44 episodes per child-year (95% CI ? 0.39–
0.49). The incidence rate was not related to baseline nutri-
tional status, age category, and duration of breast-feeding of
Diarrheal morbidity by age of the children at Mirpur, Dhaka, Bangladesh
* P < 0.001 versus age group 24–36 months.
to their duration in days in an urban slum of Dhaka, Bangladesh.
Frequency distribution of diarrheal episodes according
EPIDEMIOLOGY AND ETIOLOGY OF DIARRHEA IN AN URBAN SLUM
the children (Tables 2 and 4). It was also not related to the sex
of the children or their blood group (Tables 3 and 5). How-
ever, it was significantly related to the baseline E. histolytica
anti-lectin IgG status of the children, as previously reported.11
The incidence rate was 0.51 episodes/child-year (95% CI ?
0.44–0.58) in children positive for IgG antibody to lectin and
0.34 episodes/child-year (95% CI ? 0.28–0.40) (P ? 0.001) in
children negative for IgG antibody to lectin. The average
duration of asymptomatic infection was 1.96 months.
Asymptomatic E. histolytica infection and season. During
the surveillance of this cohort, one or more children had a
positive antigen detection test result for E. histolytica during
all months (100%). This demonstrated that the reservoir of E.
histolytica was present within the pediatric population in Mir-
pur during all the calendar year, with two peaks of incidence
rate during April and October (Figure 2).
The most important outcome of the present study is the
description of the epidemiology, clinical features, nutritional
status, and causative agents of diarrhea, with special emphasis
on E. histolytica-associated diarrhea. This study of 289 Bang-
ladeshi children 2−5 years old at Mirpur, an urban slum of
Dhaka, used modern diagnostic techniques to allow for the
first time a more accurate picture of the contribution of ame-
biasis to the overall burden of diarrheal disease. It addition-
ally provided additional insight into the natural history of
amebiasis, including evidence of species-specific immunity to
Our study demonstrated that diarrheal disease is a major
health concern for the children of this community. Overall
diarrheal illness rates ranged from 1.37 to 2.57/child-year de-
pending on age groups. The rate was highest for those 2−3
years old and lowest for children 4−5 years old. Our results
are similar with those reported by Ferrecio and others,5but
different from the results of other studies in Bangladesh.20,21
These studies, which showed incidence rates of diarrhea rang-
ing from 1.94 to 4.41/child-year, were conducted in a rural
setting of Bangladesh in which 78% of the children were mal-
nourished.20,21Average diarrheal days ranged from 2.72 to 6
days/child-year depending on age and were less than those of
children in northeastern Brazil.22Male children had a higher
diarrheal incidence than female children. Similar results
have been found in other studies, although the explanation
for this difference is not known.23–25Entamoeba histolytica-
associated diarrhea did not show any relationship with the sex
of the children.
Children in this community were better nourished com-
pared with children of the same age from other studies con-
ducted in Bangladesh.19,20In the present study 39% and 32%
of the children, respectively, were malnourished and stunted,
which is still very high and proved to be major risk factors for
diarrheal morbidity. Most episodes of diarrhea were of short
duration; 87% of episodes were ? 3 days in duration. This
may explain the comparatively low collection rate of diarrheal
stool samples for etiologic diagnosis (62%). The percentage
of children having persistent diarrhea was only 0.4%, which
was considerably less than that observed in another field
study.21The low incidence of persistent diarrhea could be
explained by the average better nutritional status of these
The association of diarrheal illness with seasons is a well-
established fact.23In this study, we have found the most pro-
nounced seasonality with a large number of diarrheal epi-
sodes in the warm months of the year (March−August).
One of the interesting findings of our study was the asso-
ciation between ABO blood group types of the children with
overall diarrheal illness. The association between different
enteropathogens, especially ETEC and V. cholerae O1, and
Diarrheal morbidity by sex of the children at Mirpur, Dhaka, Bangladesh
* P < 0.001.
Relationship between diarrheal morbidity, nutritional status, and duration of breastfeeding at baseline of the study children*
Number of Entamoeba
Malnourished (WAZ <−2)
Better-nourished (WAZ ?−2)
Stunted (HAZ <−2)
Not-stunted (HAZ ?−2)
* WAZ ? weight-for-age Z-score; HAZ ? height-for-age Z-score.
† P < 0.01.
‡ P < 0.038.
HAQUE AND OTHERS
ABO blood groups has been well studied.26,27Our data sug-
gests that overall diarrheal illness is significantly higher in
children with blood group O and AB, compared with children
with blood group A and blood group B. This is the first report
of a relationship between blood group and overall diarrheal
incidence. In other studies, investigators have not looked for
an association of blood groups with overall diarrheal diseases.
Children with blood groups O and AB also were more prone
to E. histolytica-associated diarrhea, but this association was
not statistically significant.
Our study demonstrated that the etiologic agents of diar-
rhea in the Mirpur community were essentially identical to
those found in other studies,18,24except for E. histolytica and
Aeromonas, which were isolated more frequently with diar-
rheal stools. The status of Aeromonas sp. as a human enteric
pathogen is controversial.28,29A case-control study in this
community in the future may explain this controversy. The
isolation rate of Cryptosporidium from diarrheal stool was
higher (8.4%) when compared with the clinical study con-
ducted at the ICDDR,B Hospital in Dhaka which isolated
Cryptosporidium from only 3.5% of the diarrheal stools.30
This difference may be caused by the different methods used
for the diagnosis of Cryptosporidium infection. In our study,
we used an antigen detection test for diagnosis, whereas the
study at the ICDDR,B Hospital used microscopy. The
present study has shown that along with Aeromonas sp.,
ETEC, and Cryptosporidium parvum, E. histolytica is also a
major enteropathogen in the community children of Mirpur,
Dhaka. The isolation rate of E. histolytica was 8%, which was
higher than that found in another study.31The low isolation
rate of E. histolytica reported by Zaki and others31may be
due to the exclusive use of stool microscopy, which has a low
sensitivity and is no longer recommended. In Egypt, a recent
survey using antigen detection tests showed an even higher
isolation rate of E. histolytica in individuals presenting with
acute diarrhea to an outpatient clinic.32Additional studies in
other communities will be required to derive the overall con-
tribution of amebiasis to diarrheal disease.
Diarrheal morbidity by blood group of the children at Mirpur, Dhaka, Bangladesh
a,b,cIndicate p ? <0.05 between groups.
* P < 0.05 versus group A.
FIGURE 2.Seasonal variation of diarrheal diseases in preschool children in an urban slum of Dhaka, Bangladesh.
EPIDEMIOLOGY AND ETIOLOGY OF DIARRHEA IN AN URBAN SLUM
Our present study has demonstrated that the asymptomatic
carrier state of E. histolytica infection increased by five-fold
the risk of E. histolytica-associated diarrhea and by three-fold
the risk of E. histolytica-associated dysentery. Our results are
in contrast with the results of the case-control study of the
ICDDR, B, Hospital in Dhaka that failed to demonstrate an
association between E. histolytica and diarrhea.18The lack of
an association between E. histolytica and diarrhea, as ob-
served by Albert MJ and others,18might have been caused
by sampling bias because 84% (679 of 814) of their cases
were less than 24 months old or by their failure to use an
E. histolytica-specific diagnostic test. A previous study at the
ICDDR, B Hospital demonstrated that E. histolytica infection
was very low in children less than two years old.6
Of all clinical symptoms, abdominal pain and mild-to-
moderate dehydration were the most frequent findings in E.
histolytica-associated diarrheal episodes. These data support
the findings reported by Wanke and others.6Stool examina-
tion showed that E. histolytica-associated dysentery was un-
derestimated when dysentery was defined by the gross ap-
pearance of blood in stool, which was found in only 5% of the
cases of E. histolytica-associated diarrhea. Dysentery was
demonstrated in 12−25% of cases of E. histolytica-associated
diarrhea if stool samples were examined microscopically for
red blood cells or examined for occult blood. Test results for
fecal lactoferrin, another indicator of invasive infection, were
positive in 20% of cases of E. histolytica-associated diarrhea.
Thus, our study demonstrates that 12−25% of E. histolytica-
associated diarrhea illnesses are invasive.
Of 43 children with a first episode, only 14 had a subse-
quent episode of E. histolytica-associated diarrhea. The OR
was 0.48, indicating that children after a diarrheal episode
with E. histolytica had half the chance of developing subse-
quent E. histolytica-associated diarrhea. This also supports
our previous observations from this cohort that immunity ex-
ists after natural E. histolytica infection.12
Received July 10, 2003. Accepted for publication July 27, 2003.
Acknowledgments: We thank Dr. Lyerly of TechLab Inc. for provid-
ing parasite fecal antigen detection tests. The ICCRD,B acknowl-
edges the commitment of the National Institutes of Health (NIH)
(Bethesda, MD) and the University of Virginia to its effort. We thank
the parents and children of Mirpur for their participation, and the
field team, including the study supervisor Lutfar Rahaman; field as-
sistants Janata Rani Shaha, Salma Akther, Sahina Parveen, Nurjahan
Akther Baby, and Dulari Begum; the Data Management Officer
Mahbubur Rahman; and Dr. Hamidur Rahman from the laboratory
staffs for their contributions to this study.
Financial support: The study was conducted at the ICDDR, B Centre
for Health and Population Research with the support of a grant (AI-
43596) from the NIH. This work was also supported by NIH grant
AI-43596. Rashidul Haque is a Howard Hughes Medical Institute
International Research Scholar and William A. Petri, Jr. is a Bur-
roughs Wellcome Fund Scholar in Molecular Parasitology.
Disclosure: The authors wish to disclose that the University of Vir-
ginia has a license agreement with Techlab, Inc. for diagnostic tests
for amebiasis. Dr. Petri donates all of his royalties from this agree-
ment to the American Society of Tropical Medicine and Hygiene
Authors’ addresses: Rashidul Haque, Dinesh Mondal, and Selim Ak-
ther, Laboratory Sciences Division, Center for Health and Population
Resaerch, International Centre for Diarrhoeal Disease Research,
Bangladesh (ICDDR,B), GPO Box 128, Dhaka 1000, Bangladesh.
Beth D. Kirkpatrick, Unit of Infectious Diseases, University of Ver-
mont/Fletcher Allen Health Care, Burgess 303, MCHV Campus,
Burlington, VT 05401. Barry M. Farr and William A. Petri, Jr., De-
partment of Medicine, University of Virginia Health System, Char-
Organisms identified in diarrheal stools of children in Mirpur,
Diarrheal stool %
(no. with agent/no. tested)
Enterotoxigenic Escherichia coli
Heat-stable toxin (+)
Heat-labile toxin (+)
Enteropathogenic E. coli
Enteroinvasive E. coli
Enteroaggregative E. coli
Enterohemorrhagic E. coli
Aeromonas species (non-typable)
Vibrio cholerae 01 El tor inaba
V. cholerae O139
Clinical and laboratory findings of Entamoeba histolytica–associated
diarrhea in 43 children*
Age in months, mean (SD)
Mean WAZ (SD)
Children with WAZ <−2, no. (%)
Fever present, no. (%)
Abdominal pain, no. (%)
Vomiting present, no. (%)
Average peak stool frequency/24 hours (SD)
Stool consistency (loose-liquid not watery), no. (%)
Other family member with diarrhea, no. (%)
Mild-moderate dehydration present, no. (%)
Average duration in days (SD)
Average diarrheal severity score (SD)
Visible blood in stool, no. (%)
Stool RBCs ?1/hpf, no. (%)
Fecal hemoccult positive, number
positive/number tested (%)
Fecal lactoferrin positive, number
positive/number tested (%)
* WAZ ? weight-for-age Z-score; RBCs ? red blood cells; hpf ? high-power field.
HAQUE AND OTHERS
lottesville, VA 22908. R. Bradley Sack, Department of International Download full-text
Health, Johns Hopkins University Bloomberg School of Public
Health, 615 North Wolfe Street, Baltimore, MD 21205-2179.
Reprint requests: Rashidul Haque, Laboratory Sciences Division, In-
ternational Centre for Diarrhoeal Disease Research, Bangladesh
(ICDDR, B), GPO Box 128, Dhaka1000, Bangladesh, Telephone:
880-2-881-751, extension 2411, Fax: 880-2-8812529 or 880-2-8823116,
1. Bern C, Martines J, de Zoysa I, Glass RI, 1992. The magnitude of
the global problem of diarrhoeal disease: a ten-year update.
Bull World Health Organ 70: 705–714.
2. World Health Organization, Division of Diarrhoeal and Acute
Respiratory Disease Control, 1995. Integrated management of
the sick child. Bull World Health Organ 73: 735−740.
3. Petri, WA Jr, Haque R., Lyerly D, Vines RR, 2000. Estimating
the impact of amebiasis on health. Parasitol Today 16:
4. Yamashiro T, Nakasone N, Higa N, Iwanaga M, Insisiengmay S,
Phounane T, Munnalath K, Sithivong N, Sisavath L, Phan-
thauamath B, Chomlasak K, Sisulath P, Vongsanith P, 1998.
Etiologic study of diarrheal patients in Vientiane, Lao People’s
Democratic Republic. J Clin Microbiol 36: 2195–2199.
5. Ferreccio C, Prado V, Ojeda A, Cayyazo M, Abrego P, Guers L,
Levine MM, 1991. Epidemiologic patterns of acue diarrhea
and endemic Shigella infections in children in a poor periurban
setting in Santiago, Chile. Am J Epidemiol 134: 614–627.
6. Wanke C, Butler T, Islam M, 1988. Epidemiologic and clinical
features of invasive amebiasis in Bangladesh: a case-control
comparison with other diarrheal diseases and postmortem
findings. Am J Trop Med Hyg 38: 335–341.
7. World Health Organization, 1997. World Health Organization
(WHO)/Pan American Health Organization/UNESCO report
of a consultation of experts on amebiasis. World Health Organ
Wkly Epidemiol Rep 72: 97−99.
8. Gathiram V, Jackson TFGH, 1987. A longitudinal study of
asymptomatic carriers of pathogenic zymodemes of Entam-
oeba histolytica. S Afr Med J 72: 669–672.
9. Haque R, Ali IKM, Petri WA Jr, 1999. Prevalence and immune
response of E. histolytica infection in preschool children in
Bangladesh. Am J Trop Med Hyg 60: 1031–1034.
10. Gathiram V, Jackson TFHG, 1985. Frequency distribution of En-
tamoeba histolytica in a rural South African population. Lancet
11. Haque R, Faruque ASG, Hahn P, Lyerly DM, Petri WA Jr, 1997.
Entamoeba histolytica and Entamoeba dispar infection in chil-
dren in Bangladesh. J Infect Dis 175: 734–736.
12. Haque R, Ali IKM, Sack RB, Farr BM, Ramakrishnan, Petri WA
Jr, 2001. Amebiasis and mucosal IgA antibody against the En-
tamoeba histolytica adherence lectin in Bangladeshi children. J
Infect Dis 183: 1787−1793.
13. Hamill PVV, Drizd TA, Johnson CL, Reed RB, Roche AF,
Moore WM, 1979. Physical growth: National Center for Health
Statistics percentiles. Am J Clin Nutr 32: 607–629.
14. World Health Organization, 1990. A Manual for the Treatment of
Diarrhea For Use by Physicians and Other Senior Health
Workers. Geneva: World Health Organization, CDD/SER/
80.2 Rev. 2.
15. Ruuska T, Vesikari T, 1990. Rotavirus disease in Finnish chil-
dren: use of numerical scores for clinical severity of diarrheal
episodes. Scand J Infect Dis 22: 259–267.
16. Lennett EH, Balows A, Hausler WJ Jr, Shadmoy HJ, eds., 1985.
Manual of Clinical Microbiology. Fourth edition. Washington.
DC: American Society for Microbiology.
17. Levine MM, Prado V, Robins-Browne R, Lior H, Kaper JB,
Moseley SL, Gicquelais K, Nataro JP, Vial P, Tall B, 1988. Use
of DNA probes and Hep-2 cell adherence assay to detect di-
arrheagenic Escherichia coli. J Infect Dis 158: 224–228.
18. Albert JM, Faruque ASG, Faruque SM, Sack RB, Mahalanabis
D, 1999. Case-control study of enteropathogens associated
with childhood diarrhea in Dhaka, Bangladesh. J Clin Micro-
biol 37: 3458–3464.
19. Miller RV, ed., 1977. Technical Manual. Seventh edition. Wash-
ington, DC: American Association of Blood Banks.
20. Baqui AH, Black RE, Sack RB, Chowdhury HR, Yunus M,
Siddique AK, 1993. Malnutrition, cell-mediated immune defi-
ciency, and diarrhea: a community-based longitudinal study in
rural Bangladeshi children. Am J Epidemiol 17: 355–365.
21. Baqui AH, Black RE, Sack RB, Yunus MD, Siddique AK,
Chowdhury HR, 1992. Epidemiological and clinical character-
istics of acute and persistent diarrhoea in rural Bangladeshi
children. Acta Paediatr 381 (suppl): 15–21.
22. Guerrant RL, Kirchhoff, Shields DS, Nations MK, Leslie J, de
Sousa A, Araujo JG, Correia LL, Sauer KT, McClelland KE,
Trowbridge FL, Hughes JM, 1983. Prospective study of diar-
rheal illness in northeastern Brazil: patterns of diseases, nutri-
tional impact, etiologies, and risk factors. J Infect Dis 148:
23. Santosham M, Sack RB, Reid R, Black R, Croll J, Yolken R,
Aurelian L, Wolff M, Chan E, Garrett S, Froehlich J, 1995.
Diarrhoeal diseases in the White Mountain Apaches: epide-
miologic studies. J Diarrhoeal Dis Res 13: 18–28.
24. Baqui AH, Sack RB, Black RE, Haider K, Hossain A, Abdul
Alim ARM, Yunus M, Chowdhury HR, Siddique AK, 1992.
Enteropathogens associated with acute and persistent diarrhea
in Bangladeshi children < 5 years of age. J Infect Dis 166:
25. Huilan S, Zhen Lu G, Mathan MM, Mathew MM, Olarte J,
Espejo R, Maung K, Ghafoor MA, Khan MA, Sami Z, Sutton
RG, 1991. Etiology of acute diarrhea among children in devel-
oping countries: a multicentric study in five countries. Bull
World Health Organ 69: 549–555.
26. Frederik PL, van Loon, Clemens JD, Sack DA, Rao MR, Ahmed
F, Chowdhury S, Harris JR, Ali M, Chakraborty J, Khan MR,
Neogy PK, Svennerholm AM, Holmgren J, 1991. ABO blood
groups and the risk of diarrhea to enterotoxigenic Escherichia
coli. J Infect Dis 163: 1243−1246.
27. Glass RI, Holmgren J, Haley CE, Khan MR, Marisvennerholm
A, Stoli BJ, Hossain KMB, Black RE, Yunus M, Barua D,
1985. Predisposition for cholera of individuals with O blood
group. Am J Epidemiol 121: 791–796.
28. Janda JM, Abbott SL, 1998. Evolving concepts regarding the
genus Aeromonas: an expanding panorama of species, diseases
presentation, and unanswered questions. Clin Infect Dis 27:
29. Teka T, Faruque ASG, Hossain MI, Fuchs GJ, 1999. Aeromonas-
associated diarrhoea in Bangladeshi children: clinical and epi-
demiological characteristics. Ann Trop Paediatr 19: 15–20.
30. Bhattacharya MK, Teka T, Faruque ASG, Fuchs GJ, 1997. Cryp-
tosporidium infection in children in urban Bangladesh. J Trop
Pediatr 43: 282–285.
31. Zaki A.M, Herbert L, Pont Du, Alamy MAEA, Arafat RR,
Amin K, Awad MM, Bassiouni L, Imam IZ, Malih GSE, Mar-
safie AE, Mohieldin MS, Naguib T, Rakha MA, Sidaros M,
Wasef N, Wright CE, Wyatt RG, 1986. The detection of en-
teropathogens in acute diarrhea in a family cohort population
in rural Egypt. Am J Trop Med Hyg 35: 1013−1022.
32. Abd-Alla MD, Ravdin JI, 2002. Diagnosis of amoebic colitis by
antigen captures ELISA patients presenting with acute diar-
rhoea in Cairo, Egypt. Trop Med Int Health 7: 365–370.
EPIDEMIOLOGY AND ETIOLOGY OF DIARRHEA IN AN URBAN SLUM