点击显示 收起
Departments of Epidemiology and International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
International Centre for Diarrhoeal Disease Research, Bangladesh, Dhaka, Bangladesh
The burden of enterotoxigenic Bacteroides fragilis (ETBF)related diarrhea was determined in a birth cohort of 252 children in rural Bangladesh. Isolation rates of ETBF in stool and risk factors for acquisition of ETBF and disease were established. Of 382 B. fragilispositive specimens, 14.4% of the strains found in them produced enterotoxin, as determined by a tissue-culture assay. The overall isolation rate of ETBF was 2.3% (40/1750) from diarrheal specimens and 0.3% (15/5679) from nondiarrheal specimens collected throughout the 2 years of the study (P < .001). ETBF was isolated from 20.3% (40/197) of the B. fragilispositive diarrheal specimens and from 8.1% (15/185) of the B. fragilispositive nondiarrheal specimens (P < .001) and was significantly associated with acute diarrheal disease in children 1 year of age (P = .0001). The diarrheal illness was mild in nature. In conditional multivariate analyses that examined environmental and host risk factors, the presence of livestock in the household area was linked to the acquisition of ETBF (chickens, P < .05; cows, P = .06). ETBF was found to be a small but significant contributor to diarrheal disease in this rural community. Improved management of livestock may be useful for the prevention of ETBF infection.
Acute diarrheal disease remains a problem for millions of children worldwide. In developing countries, it is 1 of the 2 main causes of death and contributes greatly to the severity of malnutrition [1]. In the preceding 2 decades, significant advances have been made in the identification of enteropathogens and in the understanding of pathogenic mechanisms. Yet, some acute diarrheal illnesses still have unknown etiologies. Identifying new enteropathogens and establishing the pathogenicity of recently recognized ones remain important goals for understanding childhood diarrhea.
Bacteroides fragilis (BF) bacteria are gram-negative, anaerobic rods that are normal inhabitants of the large bowel of mammals. A subgroup of this species has been found to produce an enterotoxin. Enterotoxigenic BF (ETBF) secretes an extracellular 20-kDa metalloprotease toxin (termed BF toxin ) that causes swelling of human colonic epithelial cells without cell injury and alters the distribution of actin filaments in intestinal epithelial cells. BFT acts by causing a "leaky epithelium" and stimulating active chloride secretion by cells, which may contribute to the pathogenesis of diarrhea in humans [2].
ETBF strains were first described in 1984, when Myers et al. [3] found that they produced a secretory response in ligated ileal loops of calves and lambs. Shortly thereafter, this novel enteropathogen was proven to be associated with diarrheal disease in a variety of young domestic animals [46].
In 1987, isolation of ETBF from human adults and children with acute and chronic diarrhea was described [7]. The first controlled study of ETBF disease in Apache children with and without diarrhea who were attending an Arizona clinic found a strong association between acute diarrhea and ETBF in children >1 year of age [8]. To assess the significance of ETBF infections in the developing world, a subsequent controlled study [9] of children <5 years of age who were admitted to the hospital of the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) in Dhaka was performed. Again, ETBF was significantly associated with diarrhea in children >1 year of age.
To our knowledge, to date, no community surveillance studies of ETBF infections have been reported. The present study was conducted to describe the prevalence of ETBF organisms and the epidemiological and clinical characteristics of ETBF infections in children <2 years of age in rural Bangladesh.
SUBJECTS, MATERIALS, AND METHODS
Study population.
A household surveillance study was carried out in Mirzapur, Bangladesh, from 1993 to 1996. The field team was composed of 12 trained female community health workers (CHWs) and a physician. From July to August 1993, CHWs conducted a door-to-door census in 10 villages, identified pregnant and married women of childbearing age, and invited them to participate in the study. From August 1993 to October 1994, 288 newborns were enrolled into the study.
A household baseline survey that collected socioeconomic and environmental information on families of enrolled newborns was conducted. Two hundred fifty-two children were followed for diarrheal morbidity from birth until 2 years of age. CHWs visited the participants' homes twice weekly to record occurrences of diarrhea and respiratory tract infections and clinical features of illnesses. Stool specimens or rectal swabs were collected from children when they had diarrhea, as well as on a monthly basis. Diarrhea was defined as 3 loose stools, or any number of stools containing blood, within a 24-h period. Diarrheal episodes were separated by at least 3 diarrhea-free days.
Informed consent was obtained from a parent or guardian for inclusion of each child in the study. During the study, children were provided with health care by a pediatrician in the home or at Kumudini Hospital, located 49 km from each study village. If a child had diarrheal illness, oral rehydration therapy was provided, and the physician administered antibiotics if indicated. The protocol was approved by the Ethical Review Committee of ICDDR,B, and the analysis plan was approved by the Committee on Human Research of the Johns Hopkins Bloomberg School of Public Health.
Microbiological assessment.
Stool specimens and rectal swabs were collected in the morning, kept in an ice cooler, and transported to the laboratory for processing that afternoon. The specimens were tested for recognized bacterial, viral, and parasitic enteropathogenssuch as Escherichia coli, Campylobacter jejuni, Shigella species, Salmonella species, Vibrio cholerae, Cryptosporidium species, Entamoeba histolytica, rotavirus, and astrovirususing methods that are routine at ICDDR,B [10].
Isolation of BF and assays for toxin production.
All stool specimens were tested for the presence of BF. Although earlier investigations of ETBF used PINN medium, which contains polymyxin B, irgasan, nalidixic acid, and novobiocin [8, 9, 11], it has been reported to be less than satisfactory for growing BF [9]. Thus, alternative media were used. Specimens collected between 1 October 1993 and 31 March 1995 were cultured for BF by streaking them onto Bacteroides bile esculin (BBE) agar for colony isolation. This medium allows the growth of several species of Bacteroides and has been shown to have the selectivity necessary for the isolation of BF [12]. Because information suggested that another medium was superior to BBE for growing BF, from 1 April 1995 to the end of the study (30 September 1996), colonies were grown on Reuben BF media (RBF), which was specifically formulated for the primary isolation of BF from stool specimens. The novel medium consisted of 3.7% Difco brain-heart infusion base, 1% yeast extract, 1% casamino acids, vitamin K, hemin additives, 20% ox bile, 4 g/mL gentamicin, 1 g/mL pefloxacin, and 100 g/mL kanamycin. When enriched with 5% lysed sheep blood, this medium provided good growth of isolates while inhibiting normal flora in stool [13]. Colonies were subcultured on blood agar, and BF colonies were identified by their characteristic mottled appearance. In all instances, plates were incubated anaerobically for 48 h at 37°C. Specimens were stored in chopped meat medium.
When diarrheal specimens tested positive for BF, colonies were tested for enterotoxin production. Additionally, 1 of every 5 of the monthly stool specimens was tested for ETBF. Laboratory workers were not aware of the source of the BF cultures. The cultures were coded only with numbers, handled in the order in which they were received, and decoded later. A tissue-culture assay using the cloned human colonic-epithelial-cell line HT29/C1 was used to detect ETBF in BF-positive specimens. HT29/C1 cells that are treated with the supernatants of ETBF strains develop morphological changes, including loss of cell-to-cell attachments, rounding, and swelling within as little as 1 h after treatment. Morphological changes progress over at least the first 24 h after treatment with supernatants of ETBF strains. This assay was proven to be 89% sensitive and 100% specific in detecting ETBF strains as defined by the previously used lambligated-intestinal-loop assay [14].
Statistical analyses.
To assess the pathogenicity of ETBF, children with ETBF-positive diarrheal specimens were compared with children with ETBF-positive nondiarrheal specimens. A similar comparison was done for children with BF-positive specimens. 2 and Fisher exact tests were used to compare differences between groups.
To determine if household and biological factors were associated with acquisition of ETBF, a nested matched case-control study was conducted. Case patients were defined as all ETBF-positive children, and control subjects were defined as ETBF-negative children matched to case patients by age (within 3 months) and season of specimen collection. A second analysis to determine risk factors for diarrhea-associated ETBF infection was performed. ETBF-positive children with diarrheal specimens were matched by age and season of specimen collection to ETBF-negative children with nondiarrheal specimens. Independent variables in the regression models included factors recorded during household surveys (i.e., water sources, presence of livestock in the household area, sex, birth weight, and duration of exclusive breast-feeding). Indicators of nutritional status (normal versus low z scores of weight-for-age, height-for-age, and weight-for-height) were also studied in this analysis. Significant variables in the univariate analyses (at an level of 0.05) and potential confounders were included in conditional multivariate logistic regression models. Statistical analyses were performed using Stata (version 7.0; StataCorp).
RESULTS
A total of 7429 specimens (1750 diarrheal and 5679 nondiarrheal) were collected from the 252 children included in the present study. The overall isolation rate of BF, which included ETBF and non-ETBF strains, was 14.6% (1083/7429 specimens). From October 1993 to March 1995, a total of 3740 specimens were tested for BF using BBE medium, and from April 1995 to September 1996, a total of 3689 specimens were tested for BF using RBF medium. The overall distributions of diarrheal and nondiarrheal specimens were similar for the 2 culturing periods. From October 1993 through March 1995, there were 786 (21.0%) diarrheal specimens and 2954 (79.0%) nondiarrheal specimens. From April 1995 through September 1996, there were 964 (26.1%) diarrheal specimens and 2725 (73.9%) nondiarrheal specimens. The isolation rate for BF when BBE medium was used was 11.1% (417/3740), which was significantly less than the 18.1% (666/3689) isolation rate when RBF medium was used (P < .001).
BF was not isolated more frequently in children with diarrhea; actually, the reverse was true. BF was isolated in 221 (12.6%) of 1750 diarrheal specimens and in 862 (15.2%) of 5679 nondiarrheal specimens (P = .008).
A total of 382 BF-positive specimens were tested for enterotoxin production. These were composed of 89.1% of the BF isolates from the diarrheal specimens (some cultures were lost in storage) and a random sample of 21.5% of the BF isolates from the nondiarrheal specimens. The overall isolation rate of ETBF was 2.3% (40/1750) from diarrheal specimens and 0.3% (15/5679) from nondiarrheal specimens (P < .001).
Table 1 shows the isolation rates of ETBF in the 382 BF-positive specimens. A total of 14.4% (55/382) of the BF isolates were enterotoxin producers. ETBF was identified in 40 (20.3%) of 197 BF-positive diarrheal specimens and in 15 (8.1%) of 185 BF-positive nondiarrheal specimens (P < .001). ETBF was isolated more frequently from the stool specimens of children 1 year of age than from the stool specimens of children in other age groups. The difference between the isolation rates of ETBF in BF-positive children attained statistical significance in the oldest age group (1823 months). The numbers of ETBF-positive specimens in the younger age groups were low. The data, analyzed according to whether the specimens were from children <1 year of age or 1 year of age, showed that isolation rates of ETBF in diarrheal specimens versus those in nondiarrheal specimens for those <1 year of age were not statistically significantly different (11.9% vs. 6.4%; odds ratio [OR], 2.00; 95% confidence interval [CI], 0.509.53), but were statistically significantly different for children 1 year of age (24.6% vs. 9.0%; OR, 2.73; 95% CI, 1.517.62).
Both non-ETBF strains and ETBF were isolated in diarrheal and nondiarrheal specimens throughout the year. In Bangladesh, the dry, hot spring occurs from March to May; the wet, hot summer occurs from June to October; and the winter occurs from November to February. BF was significantly more prevalent in nondiarrheal specimens than in diarrheal specimens during all seasons (data not shown). The seasonal distribution of isolation of ETBF (figure 1) was also examined. When the isolation rates of ETBF in BF-positive nondiarrheal specimens were compared with those in diarrheal specimens, the diarrheal specimens had higher isolation rates of ETBF during each season, but they were statistically significantly higher during the summer (30.8% in diarrheal specimens vs. 8.6% in nondiarrheal specimens) (P = .001). In the nondiarrheal specimens, the isolation rates of ETBF did not change significantly from season to season. In the diarrheal specimens only, the peak in the isolation rate of ETBF (30.8%) in the summer differed significantly (P = .03) from the isolation rates of ETBF in diarrheal specimens in the spring and winter (15.1% and 15.3%, respectively).
Of the 55 specimens from which ETBF was isolated, 14 specimens were from children who already had tested positive for ETBF; therefore, 41 children (16.3% of all study children) had at least 1 ETBF-positive specimen. Repeat isolations of ETBF from diarrheal specimens came from separate diarrheal episodes, which indicates that these were sequential infections rather than persistent ETBF-related symptoms. The children who tested positive for ETBF more than once during the study had 24 ETBF episodes each. The repeat isolations of ETBF were separated by diarrheal episodes that were caused by other pathogens.
The data were further examined to determine whether specimens harboring ETBF had other enteropathogens. Fifteen of 55 ETBF isolates were not accompanied by diarrheal symptoms. There were a total of 40 ETBF-positive diarrheal episodes in 33 children. Of these 40 episodes, 22 (55%) were accompanied by other bacterial and/or parasitic agents (data not shown).
The clinical data were analyzed after exclusion of the ETBF-positive diarrheal episodes in which multiple enteropathogens were identified, which left for analysis 18 separate diarrheal episodes in 15 children. As is shown in table 2, there was a striking difference in ETBF-positive case patients by sex. With the exception of 4 diarrheal episodes that were accompanied by blood in the stool and 3 episodes that were persistent (lasting 14 days), the illnesses were generally found to be mild and secretory in nature.
Case-control studies.
A nested case-control study of 40 ETBF-positive children and 78 matched ETBF-negative control subjects was performed to assess the association of risk factors with the acquisition of ETBF. For 1 ETBF-positive child, a matched control subject was not identified. The household and host characteristics, as well as distributions of other enteropathogens, were similar in case patients and control subjects.
In the analysis to determine risk factors for diarrhea-associated ETBF infections, 32 of 33 children who had diarrhea-associated ETBF-positive episodes were matched by age and season of specimen collection to 55 ETBF-negative control subjects who had nondiarrheal specimens. One ETBF-positive child was not included in this analysis because a matched control subject was not identified. There were no significant differences between case patients and control subjects with regard to household characteristics, nutritional indicators, and other enteropathogens.
Univariate conditional logistic regressions for both case-control analyses are shown in table 3. No potential risk factors showed a significant association with ETBF-positive diarrheal specimens, but the presence of chickens or cows in the household area was statistically significantly associated with the acquisition of ETBF. As is shown in table 4, only the adjusted OR for the presence of chickens in the household area remained significantly associated with the acquisition of ETBF. The presence of cows in the household area had an association that approached statistical significance (P = .06).
DISCUSSION
Although numerous enteropathogens are known to cause diarrheal infections, the etiologies of a substantial number of cases of diarrhea remain unrecognized, and this suggests that additional, unidentified pathogens may be responsible for disease. Our results confirm that the isolation of ETBF is low but is significantly associated with acute diarrheal disease in children 1 year of age in rural Bangladesh. The disease is relatively mild and is readily treated with oral rehydration therapy. The present studywhich is, to our knowledge, the first longitudinal study to be conducted at the household level on rural children of a developing countryfound that 16.3% of all children were infected with ETBF during the first 2 years of life. To our knowledge, this study is also the second one [15] to indicate a seasonal pattern of ETBF infection and may be the first to demonstrate that ETBF infection is associated with environmental and household factors.
An interesting finding was that RBF medium was more sensitive in identifying BF organisms than was the traditional BBE medium. The isolation rate when RBF medium was used was about twice the rate seen when BBE medium was used (18.1% vs. 11.1%) (P < .001). In the future, studies of BF should benefit from the use of RBF medium.
In the present study, BF of any type was isolated from 12.6% of diarrheal specimens and from 15.2% of nondiarrheal specimens. Thus, unlike the finding from an Oklahoma hospital study by San Joaquin et al. [15], but in accordance with the finding from a study of Apache children by Sack et al. [8], there was no significant association between the presence of any BF and diarrhea in Mirzapur. That BF was consistently more often isolated from nondiarrheal specimens supports the idea that, like E. coli, it is a normal inhabitant of the colon.
Approximately 14% of BF isolates tested were enterotoxin producers. Similarly, San Joaquin et al. [15] discovered that, in Oklahoma, 15% of BF in stool specimens (regardless of diarrheal status, age, and season) were enterotoxin producers. After enumeration of the organisms in municipal sewage in Montana, it was found that 9.3% of BF was ETBF [16]. ETBF may make up a considerable portion of the BF flora of the intestinal tract, and, like diarrheogenic E. coli, BF may have additional virulence factors that are important in the pathogenesis of diarrheal disease.
Recently, in epidemiological studies conducted throughout the world [8, 9, 15, 1721], ETBF has been explored as a human diarrheal pathogen in persons presenting at clinics or hospitals. Seven of these studies have focused on the association between ETBF and childhood diarrhea [8, 9, 15, 1720]. A 1997 study by Pantosti et al. [18] in Italy showed that the rate of ETBF carriage was high and was not significantly different between children with diarrhea (17%) and healthy control subjects (12%) 1 month9 years of age. A Bangladeshi hospitalbased study [9] found ETBF rates in children <5 years of age to be lower (6.1% of case patients and 1.2% of control subjects) than those in the Italian study. The rates at this Bangladeshi hospital are therefore similar to the childhood rates found in the Oklahoma study (4.4% of case patients and 3.1% of control subjects) and in the study of Apache children (12% of patients and 6% of control subjects). In all 3 studies, differences in isolation rates of ETBF in case patients and control subjects attained statistical significance only for children >1 year of age. In the present study, overall isolation rates of ETBF from BF-positive diarrheal and nondiarrheal specimens were high (20.3% from diarrheal specimens and 8.1% from nondiarrheal specimens). Similar to the isolation rates in the 3 previous studies mentioned, the isolation rates for the 2 groups in the present study were significantly different only for those specimens taken from children 1 year of age, in which the isolation rates of ETBF rose to 24.6% in diarrheal specimens but remained virtually constant in nondiarrheal specimens.
Our isolation rates of ETBF from diarrheal specimens and from nondiarrheal specimens (2.3% and 0.3%, respectively) were close to those seen for a number of other organisms detected in this cohort: Shigella flexneri (5.4% in diarrheal specimens and 0.02% in nondiarrheal specimens), Salmonella species (1.83% in diarrheal specimens and 0.2% in nondiarrheal specimens), V. cholerae O1 Ogawa (0.2% in diarrheal specimens and 0.0% in nondiarrheal specimens), and Cryptosporidium species (0.6% in diarrheal specimens and 0.2% in nondiarrheal specimens). The burden of ETBF in our study may have even been underestimated because of the medium initially used to detect BF, as was indicated above.
To our knowledge, this is the second study to report a seasonal pattern in ETBF infections. San Joaquin et al. [15] found no appreciable differences in isolation rates of ETBF from season to season in Oklahoma in 43 case patients and 18 control subjects. In contrast to those results, we found a peak in diarrhea-associated ETBF during the summer, compared with the rates in spring and winter. Isolation rates of ETBF in nondiarrheal specimens did not change seasonally. This seasonal difference suggests some as-yet-unspecified property of the diarrhea-associated ETBF that is lacking in those strains not associated with diarrhea.
In the crude analysis of risk factors for acquisition of ETBF, the presence of cows and chickens in the household area showed statistically significant, positive associations with the acquisition of ETBF. After controlling for cows in the household and sex of the child, there was still a 5-fold increase in the odds of acquiring ETBF for children who had any chickens in the household area, compared with the odds for children who had no chickens. Although it has been determined that different domestic animals, including lambs and calves, can have ETBF-associated diarrhea [3, 11], to our knowledge there has not been a report of ETBF being isolated from the feces of chickens. It is possible that feces from chickens could yield ETBF, as has been seen in the case of Campylobacter [22].
The results of the present study suggest that both cows and chickens could be sources of human infection or vehicles for the transmission of ETBF. This finding points to areas of future investigation. Further surveillance studies could include collection of specimens from livestock with diarrhea, particularly when a case of ETBF is discovered in humans. ETBF-positive stool specimens from livestock and ill children within the same time frame would give strength to the epidemiological association we found. Because a large proportion of families in Mirzapur keep livestock in the household area (i.e., 50% of families keep >3 chickens), the discovery of a definitive animal-to-human transmission of ETBF infection would have implications for the prevention of this infection in this community and in comparable populations.
No significant associations between ETBF-associated diarrhea and environmental or biological risk factors could be demonstrated. The inhabitants of Mirzapur are a largely homogeneous population; few disparities in living conditions and behaviors exist. Because they have access to the same water sources and exposure to the same unsanitary conditions, the children of these families had high levels of illness. For example, in this cohort, the incidence of infant diarrhea was 4.25 episodes per child per year (P.P., K.Z.H., E.R., F.H., A.K.S., and R.B.S., unpublished data). Thus, because almost all children in this cohort are exposed to the same potential disease-causing conditions, it is difficult to determine which ones specifically are linked with diarrheal disease. Additionally, the use of ETBF-associated diarrhea as the outcome of a study could be problematic, given that there is a high prevalence of mixed infections in this population and that ETBF-positive diarrheal specimens were also often found to be positive for other enteropathogens. To be able to attribute diarrhea solely to ETBF would have required omitting children who had additional enteropathogens from the study, which would have greatly reduced the numbers studied and the statistical power of the analysis.
In summary, ETBF has been shown to be a small but important contributor to disease burden in young children living in rural Bangladesh. Approximately 16% of all children <2 years of age were infected with ETBF at some time. This community study had the advantage of detecting illnesses of a mild nature; most children with ETBF-associated diarrhea would probably not be seen at a treatment center. Approximately 22% of ETBF-positive diarrheal episodes were associated with bloody stool; similar findings have been reported elsewhere [7], and, in the present study, these episodes may have been the result of unrecognized coinfections. Prevention of ETBF infection through the improved management of livestock, as well as adequate treatment of the infection with oral rehydration therapy, would help to lessen the impact of diarrheal diseases, which are a persistent problem in this part of the world.
Acknowledgments
We thank B. P. Pati of Mirzapur Hospital, for providing medical services to the study population, and to the hospital staff, the field staff, and the people of Mirzapur, for their cooperation in the study. The researchers affiliated with the International Centre for Diarrhoeal Disease Research, Bangladesh (Dakha) acknowledge with gratitude the commitment of the Department for International Development, United Kingdom, and the Ford Foundation to the center's research efforts. We thank Cynthia Sears (Johns Hopkins School of Medicine, Baltimore, MD), for her valuable comments on the manuscript.
References
1. Rice AL, Sacco L, Hyder A, Black RE. Malnutrition as an underlying cause of childhood deaths associated with infectious diseases in developing countries. Bull World Health Organ 2000; 78:120721. First citation in article
2. Sears CL. The toxins of Bacteroides fragilis. Toxicon 2001; 39:173746. First citation in article
3. Myers LL, Firehammer BD, Shoop DS, Border MM. Bacteroides fragilis: a possible cause of acute diarrheal disease in newborn lambs. Infect Immun 1984; 44:2414. First citation in article
4. Myers LL, Shoop DS, Byars TD. Diarrhea associated with enterotoxigenic Bacteroides fragilis in foals. Am J Vet Res 1987; 48:15657. First citation in article
5. Myers LL, Shoop DS. Association of enterotoxigenic Bacteroides fragilis with diarrheal disease in young pigs. Am J Vet Res 1987; 48:7745. First citation in article
6. Myers LL, Shoop DS, Collins JE, Bradbury WC. Diarrheal disease caused by enterotoxigenic Bacteroides fragilis in infant rabbits. J Clin Microbiol 1989; 27:202530. First citation in article
7. Myers LL, Shoop DS, Stackhouse LL, et al. Isolation of enterotoxigenic Bacteroides fragilis from humans with diarrhea. J Clin Microbiol 1987; 25:23303. First citation in article
8. Sack RB, Myers LL, Almeido-Hill J, et al. Enterotoxigenic Bacteroides fragilis: epidemiologic studies of its role as a human diarrhoeal pathogen. J Diarrhoeal Dis Res 1992; 10:49. First citation in article
9. Sack RB, Albert MJ, Alam K, Neogi PKB, Akbar MS. Isolation of enterotoxigenic Bacteroides fragilis from Bangladeshi children with diarrhea: a controlled study. J Clin Microbiol 1994; 32:9603. First citation in article
10. Albert MJ, Faruque ASG, Faruque SM, Sack RB, Mahalanabis D. Case-control study of enteropathogens associated with childhood diarrhea in Dhaka, Bangladesh. J Clin Microbiol 1999; 37:345864. First citation in article
11. Border M, Firehammer BD, Shoop DS, Myers LL. Isolation of Bacteroides fragilis from the feces of diarrheic calves and lambs. J Clin Microbiol 1985; 21:4723. First citation in article
12. Livingston SJ, Kominos SD, Yee RB. New medium for selection and presumptive identification of the Bacteroides fragilis group. J Clin Microbiol 1978; 7:44853. First citation in article
13. Reuben J. Characterization of enterotoxigenic Bacteroides fragilis and studies in its enterotoxin production [PhD thesis]. Baltimore, MD: Johns Hopkins School of Public Health, 1995. First citation in article
14. Weikel CS, Grieco FD, Reuben J, Myers LL, Sack RB. Human colonic epithelial cells, HT29/C1, treated with crude Bacteroides fragilis enterotoxin dramatically alter their morphology. Infect Immun 1992; 60:3217. First citation in article
15. San Joaquin VH, Griffis JC, Lee C, Sears CL. Association of Bacteroides fragilis with childhood diarrhea. Scand J Infect Dis 1995; 27:2115. First citation in article
16. Shoop DS, Myers LL, LeFever JB. Enumeration of enterotoxigenic Bacteroides fragilis in municipal sewage. Appl Environ Microbiol 1990; 56:22434. First citation in article
17. Niyogi SK, Dutta P, Mitra U, Pal DK. Association of enterotoxigenic Bacteroides fragilis with childhood diarrhoea. Indian J Med Res 1997; 105:1679. First citation in article
18. Pantosti A, Menozzi MG, Frate A, Sanfilippo L, D'Ambrosio F, Malpeli M. Detection of enterotoxigenic Bacteroides fragilis and its toxin in stool samples from adults and children in Italy. Clin Infect Dis 1997; 24:126. First citation in article
19. Kato N, Liu C, Kato H, et al. Prevalence of enterotoxigenic Bacteroides fragilis in children with diarrhea in Japan. J Clin Microbiol 1999; 37:8013. First citation in article
20. Krzyzanowsky F, Avila-Campos MJ. Detection of non-enterotoxigenic and enterotoxigenic Bacteroides fragilis in stool samples from children in So Paulo, Brazil. Rev Inst Med Trop So Paulo 2003; 45:2257. First citation in article
21. Zhang G, Svenungsson B, Karnell A, Weintraub A. Prevalence of enterotoxigenic Bacteroides fragilis in adult patients with diarrhea and healthy controls. Clin Infect Dis 1999; 29:5904. First citation in article
22. Rao M, Naficy AB, Savarino SJ, et al. Pathogenicity and convalescent excretion of Campylobacter in rural Egyptian children. Am J Epidemiol 2001; 154:16673. First citation in article