Only samples negative for
V. cholerae were tested for ETEC, starting from July 20, 2004, when the patient numbers increased at the ICDDR,B hospital for ≈6 weeks, until the patient numbers decreased and the floods had receded. Information, including age, sex, fever, vomiting, dehydration status, and related clinical features, was also collected from patients. For ETEC surveillance, we used lactose-fermenting
E. coli colonies cultured on MacConkey agar plates that had been cultured from fresh stool specimens (
4). Six lactose-fermenting individual colonies of
E. coli were tested for the presence of LT, ST, and colonization factors. Detection of LT and ST was carried out with ganglioside GM1 enzyme-linked immunosorbent assays (
4). The colonies that tested positive for the toxins were also plated onto colonization factor antigen (CFA) agar plates with and without bile salts for testing colonization factors (
4). Trypticase soy agar containing 5% sheep blood (TSA) was used to test for the colonization factor CS21 (
5).
The strains were cultured at 37°C overnight; those grown on CFA agar without bile were tested for colonization factors CFA/1, CSI, CS2, CS3, CS4, and CS6, and those on CFA agar plus bile were tested for CS5, CS7, CS17, CS8, CS12, and CS14 (
4). Those strains grown on TSA were tested for CS21 only (
5). Of the patients included in this study, 67% had severe-to-moderate dehydration; of these, 51% were children <5 years of age, while 39% were >15 years of age. They were treated for diarrhea with oral (61%) or intravenous (39%) rehydration therapy and other medications as needed.
Of 350 stool specimens tested during the epidemic, 78 (22.2%) were positive for
V. cholerae O1 (22 Ogawa and 56 Inaba serotype), and 63 (18.0%) were positive for ETEC.
Shigella spp. (3.4%, n = 11) and
Salmonella spp. (1.7%, n = 5) were seen at lower rates. Children with ETEC diarrhea were negative for
V. cholerae O1 as well as
Shigella spp. and
Salmonella spp. We did not test
V. cholerae–positive samples for ETEC and therefore cannot rule out possible concomitant infection with ETEC in these 78 cholera patients (
4).
Isolation of ETEC and V. cholerae O1 remained high throughout the epidemic (), and during 1 week, comparable numbers of ETEC and V. cholerae were isolated from stools of patients. We compared demographic and clinical features of patients with ETEC and V. cholerae infections (). Most patients with ETEC diarrhea were <2 years of age (56%) or >15 years of age (36%) (median 1.5 years), whereas those with V. cholerae O1 infection were mostly >5 years of age (median age 15.5 years). Although more cholera patients had severe dehydration (60%), 22% of the patients with ETEC diarrhea also had severe dehydration (p<0.001). Intravenous rehydration was needed for both ETEC- and V. cholerae–infected patients, but it was more frequently used in the latter.
| Table 1Characteristics of the patients with enterotoxigenic Escherichia coli (ETEC) and Vibrio cholerae O1 infection during the diarrheal epidemic, July–August 2004, Bangladesh |
With regard to toxin profile, ETEC expressing ST alone was the most common (67%), followed by strains producing both ST and LT (19%) and LT alone (14%). Dominance of the ST-expressing ETEC has been documented earlier during seasonal outbreaks and epidemics in Bangladesh (
4) and in Egypt and the Middle East (
11,
12). Patients infected with the different toxin phenotypes of ETEC had dehydration status ranging from severe to none, although no significant association was seen between toxin phenotype and degree of dehydration.
A high proportion of the ETEC strains (78%) expressed 1 or more colonization factors (), a much higher frequency than that seen in other hospital or community-based studies (
10,
12). In earlier studies in Bangladesh, we found 56% of strains positive for these colonization factors (
4). In the present study, ≈92% of ST/LT-, 79% of ST-, and 56% of LT-expressing ETEC expressed 1 or more colonization factors. CFA/I was the most common phenotype, followed by the strains expressing CS4 + CS6 or CS5 + CS6, followed by others. Thus, most of the colonization factor types were those known to be present in clinical strains and those that have previously been isolated from hospitalized patients (
4,
5). These antigens have been given priority for designing vaccines to protect against a wide range of colonization factors (
10). In addition, 3 strains co-expressed CS21, a type IV pilus antigen (
4). Of these, 2 strains expressed CFA/I and CS21, and 1 was positive for CS1, CS3, and CS21.
| Table 2Colonization factor (CF) types of enterotoxigenic Escherichia coli (ETEC) isolated from patients during diarrhea epidemic, Bangladesh* |
We used 13 colonization factor–specific monoclonal antibodies in testing; however, >22 colonization factors have been described, not all of which could be tested in this study. In addition, although precautions were taken to rule out the loss of phenotypic properties of colonization factors, some may have been lost on culture. By using polymerase chain reaction or DNA hybridization procedures, more colonization factor–specific genes and those that have undergone phenotypic changes could have been detected (
13).
* Ashraful I. Khan,* Abu Syed G. Faruque,* Yasmin Ara Begum,* Fahima Chowdhury,* Gopinath B. Nair,* Mohammed A. Salam,* David A. Sack,* and Ann-Mari Svennerholm†
This article has been cited by other articles in PMC.
