This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Qadri, F. Articles by Sack, R. B. Search for Related Content PubMed PubMed Citation Articles by Qadri, F. Articles by Sack, R. B.

 Previous Article  |  Next Article 

Clinical Microbiology Reviews, July 2005, p. 465-483, Vol. 18, No. 3
0893-8512/05/$08.00+0     doi:10.1128/CMR.18.3.465-483.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Enterotoxigenic Escherichia coli in Developing Countries: Epidemiology, Microbiology, Clinical Features, Treatment, and Prevention

International Centre for Diarrhoeal Disease Research, Bangladesh, and Centre for Health and Population Research, Mohakhali, Dhaka 1212, Bangladesh,¹ Department of Medical Microbiology and Immunology, Göteborg University, 40530 Göteborg, Sweden ,² Department of International Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland³

SUMMARY

INTRODUCTION

HISTORICAL ASPECTS

From Discovery to Present Understanding of Public Health Importance

BIOLOGY

LT and ST Enterotoxins

Colonization Factors

ETEC Serotypes

ETEC Strains in Animals

EPIDEMIOLOGY

Age-Related Infections in Children and Adults

Relation to Presence of LT, ST, and Colonization Factors

Single Versus Mixed Infections

Seasonality of ETEC

Comparison of ETEC Diarrhea and Cholera in Children and Adults

Presence of ETEC in Food and Water in the Environment

ETEC Infections and Malnutrition

Infections in International Travelers

CLINICAL FEATURES

Disease Severity

Mortality from ETEC Diarrhea

DIAGNOSIS

Laboratory Assays

TREATMENT AND MANAGEMENT

Rehydration

Antimicrobials

Multidrug Resistance Patterns

Nutritional and Micronutrient Therapy

PREVENTION

Vaccine Development

Purified CFs and Enterotoxoids

Inactivated Whole-Cell Vaccines

Live Oral ETEC Vaccines

CONCLUSIONS

ACKNOWLEDGMENTS

REFERENCES

Top Next References

INTRODUCTION

Top Previous Next References

 
Acute infectious diarrhea is the second most common cause of death in children living in developing countries, surpassed only by acute respiratory diseases accounting for approximately 20% of all childhood deaths (96). The major etiologic agents that account for the estimated 1.5 million deaths per year are enterotoxigenic Escherichia coli (ETEC), rotavirus, Vibrio cholerae, and Shigella spp. (88, 96); all are known to be endemic in essentially all developing countries. Whereas V. cholerae, Shigella spp., and rotavirus can be readily detected by standard assays, ETEC is more difficult to recognize and therefore is often not appreciated as being a major cause of either infantile diarrhea or of cholera-like disease in all age groups. Since ETEC is a major cause of traveler's diarrhea in persons who travel to these areas, the organism is regularly imported to the developed world (18, 58, 75).

Among the six recognized diarrheagenic categories of Escherichia coli (118), ETEC is the most common, particularly in the developing world (214). Specific virulence factors such as enterotoxins and colonization factors differentiate ETEC from other categories of diarrheagenic E. coli. ETEC belongs to a heterogeneous family of lactose-fermenting E. coli, belonging to a wide variety of O antigenic types, which produce enterotoxins, which may be heat labile and/or heat stable, and colonization factors which allow the organisms to readily colonize the small intestine and thus cause diarrhea (118, 155, 211).

This review summarizes data on the recognition and importance of ETEC diarrhea in developing countries, emphasizing on its prevalence, toxin types, colonization factors, and morbidity in different population groups at risk. We have reviewed information on ETEC since its discovery almost 50 years ago (43) and used clinical and laboratory data from hospital and community-based studies around the world, in both urban and rural settings, to present a comprehensive picture of ETEC-mediated diarrheal disease with regard to epidemiology, diagnosis, treatment, and prevention through the use of vaccines. We hope that this review may increase the knowledge and awareness of the importance of ETEC infections, particularly in the developing world.

HISTORICAL ASPECTS

Top Previous Next References

 

The history of enterotoxigenic E. coli begins in 1956 in Calcutta (43). De and his colleagues injected live strains of E. coli, isolated from children and adults with a cholera-like illness, into isolated ileal loops of rabbits and found that large amounts of fluid accumulated in the loops, similar to that seen with Vibrio cholerae. However, they did not test the filtrates of these cultures to determine whether they produced an enterotoxin. These findings were not followed up until 1968, when Sack reported studies, also in Calcutta, of adults and children with a cholera-like illness, who had almost pure growth of E. coli in both stool and the small intestine (154). These E. coli isolates were found to produce a strong cholera-like secretory response in rabbit ileal loops, both as live cultures and as culture filtrates (74). The patients were also found to have antitoxin responses to the heat-labile enterotoxin produced by these organisms (163). At about the same time, similar studies were being done with animals that also demonstrated strains of E. coli to be responsible for diarrheal disease in several animal species: pigs, calves, and rabbits (80, 177, 178). Studies of these animal enterotoxigenic E. coli paralleled and sometimes preceded those done with human strains; these organisms were also found to produce enterotoxins and specific colonization factors.

These findings from Calcutta were soon confirmed by oral challenge of human volunteers (49, 100) and by corroboration of studies in Dhaka, Bangladesh (61, 113, 117, 149). ETEC were shown in these studies to be most frequently found in children; such findings have been subsequently corroborated in multiple studies in developing countries (23, 24). Thus, in most studies in the developing world, ETEC have been shown to be the most common bacterial enteric pathogen, accounting for approximately 20% of cases, as shown in Table 1, which summarizes findings from some of the more detailed studies done in several different countries.

Top Previous Next References

LT was found to be very similar physiologically, structurally, and antigenically to cholera toxin and to have a similar mode of action. The molecular mass (84 kDa) and the subunit structure of the two toxins were essentially identical, with an active (A) subunit surrounded by five identical binding (B) subunits (70, 83). Following colonization of the small intestine by ETEC and release of the LT, the LTB subunits bind irreversibly to GM1 ganglioside, and the A subunit activates adenylate cyclase, which results in increases in cyclic AMP, which stimulates chloride secretion in the crypt cells and inhibits neutral sodium chloride in the villus tips. When these actions exceed the absorptive capacity of the bowel, purging of watery diarrhea results (70).

ST is a nonantigenic low-molecular-weight peptide, consisting of 18 to 19 amino acids. There are two variants, STp and STh, named from their initial discovery from pigs and humans, respectively, and which have identical mechanisms of action. Released in the small intestine, ST binds reversibly to guanylate cyclase, resulting in increased levels of cyclic GMP (138). ST has also been implicated in the control of cell proliferation via elevation of intracellular calcium levels (174). As with LT, chloride secretion by the crypt cells is then increased and inhibition of neutral sodium chloride absorption occurs, leading to outpouring of diarrheal stool. The relative proportions of LT, ST, and LT/ST toxin-producing ETEC seems to vary from one geographic area to another in patients with ETEC diarrhea or asymptomatic carriers (Tables 1, 2, and 3). The rate of isolation from asymptomatic children has varied between 0% and 20% in numerous studies carried out with children in different settings but has in most instances been lower than the rates in children with diarrhea (7, 8, 19, 38, 79, 88, 143). On average ETEC is seen at least two to three times more frequently in symptomatic than asymptomatic children (Table 2).

A nomenclature for the CFs designating them as coli surface antigen (CS) was introduced in the mid-1990s (66). A list showing the old and new classifications of the CFs can be seen in Table 4. All except CFA/I have the CS designation in the present designation. Some of the better-characterized CFs can be subdivided into different families, i.e., the colonization factor I-like group (including CFA/I, CS1, CS2, CS4, CS14, and CS17) (66) and the coli surface 5-like group (with CS5, CS7, CS18, and CS20) (204) and those that are unique (CS3, CS6, and CS10 to CS12). Within each of these families there are cross-reactive epitopes that have been considered as candidates for vaccine development (147).

Based on an extensive database analysis of ETEC from a number of different countries all over the world, Wolf reported that in the ETEC antigen the largest variety was the O antigen (211). Thus, 78 different O groups were detected in the 954 ETEC isolates included in the study (hereafter called the ETEC database). In addition, there were several rough strains that lacked side chains, thus being nontypeable with regard to O antigen, or strains that had unknown serogroups. The most common O groups in this retrospective study were O6, O78, O8, O128, and O153; these accounted for over half of the ETEC strains.

In a more recent study in Egypt, a large variation of O groups was also recorded, with 47 O groups being represented among the 100 ETEC strains isolated; however, an entirely different O group pattern was recorded than in the database, the most common O groups in the Egyptian study being O159 and O43 (128).

Considerably fewer H serogroups than O serogroups are associated with ETEC. Thus, a total of 34 H groups were identified among the 730 ETEC strains included in the ETEC database (211). Five H types accounted for over half of those strains and they were widespread. Similarly, five different H types accounted for almost half of 100 ETEC strains isolated prospectively from Egyptian children, although different H types predominated from those reported for the ETEC database (128).

There are clearly preferred combinations of serotypes, CFs, and toxin profiles in ETEC. For instance, certain H groups are strongly associated with an O serogroup, such as O8:H9, O78:H12, and O25:H42, and some O and H serogroups are associated with one or more CFs (112, 214). In a study of ETEC isolated from children in Argentina, it was shown that most CFs were expressed by strains exhibiting a limited array of serotypes, while the ETEC strains that lacked detectable CFs belonged to many different serotypes (207). However, the significance of these different combinations regarding enhanced virulence (211) or for vaccine development is still unclear.

Serotyping appears to give an indication of the variety of strains that are present in a particular ETEC type in a certain geographical area. A close genetic relationship has been found within ETEC strains belonging to a certain serotype, which is different from that noted in other serotypes, and the pattern of genetic relatedness did not change over a period of 15 years (125). The loss of CFs and toxin phenotypes did not affect the genetic relatedness of these strains and their clonal relationship, which suggests that serotype analysis can be coupled to genetic typing for studying the clustering of strains for epidemiological and pathogenetic studies of ETEC. Altogether, the great variation in O and H serogroups in ETEC makes serotyping less suitable for identification of these bacteria and makes O and H antigens less attractive as putative candidate antigens in an ETEC vaccine.

Like human ETEC, animal strains also have distinct binding proteins (adhesins and fimbriae), which allow the organisms to attach and colonize the small intestinal mucosa. Indeed, ETEC CFs display a remarkable species specificity, and colonization factors are clearly different from those of human and animal ETEC. Their genetic control may be in plasmids or in the chromosome. The most common of these have been designated K88, K99, and 987P, but there are at least another eight or more, which have other designations. The animal colonization factors are now being identified by F numbers, such as F4, F5, and F6 instead of K88, K99, and 987P (65). Because of the specificity of these adhesins, animal ETEC strains normally do not infect humans. This is in contrast to other diarrheagenic E. coli, such as those that produce Shiga-like toxins, e.g., O157:H7, which are found in animals, mainly cows, and produce severe disease in humans (92)

EPIDEMIOLOGY

Top Previous Next References

 

The susceptibility of infants and young children has also been observed in other settings which have poor public health and hygiene conditions. The characteristics of the toxin types and CFs present on ETEC strains isolated from young children vary among countries where ETEC is endemic (139, 183, 215). A comparison is shown for Bangladesh and other developing countries (Tables 1 to 5). Studies to better understand the natural infection pattern of ETEC are being conducted with cohorts of infants to discern the infection and reinfection pattern as well as the age group most at risk for infection. In studies of infants in West Africa, Egypt, and Bangladesh, the rate of ETEC infections in community-based studies increased from about 3 to 6 months of age, similar to the surveillance data of hospitalized patients in a diarrheal hospital in Bangladesh (139, 183, 215). The age at which a primary ETEC infection can be documented depends to some extent on the phenotype of ETEC that is infecting the child. In a study in Guinea Bissau, it was reported that in the youngest age group, 3 months, ETEC strains producing STh and LT were most common, whereas at 6 to 7 months ETEC strains producing STp, STpLT, and SThLT predominated (183).

The incidence of ETEC infections in developing countries decreases after 5 years of age with a decrease of infections between the ages of 5 to 15 years (Table 6). The incidence increases again in those over 15 years of age and about 25% of ETEC illness is seen in adults (113, 132). Limited epidemiological information is available for adults, and those available are mostly from India and Bangladesh. It was in these settings that ETEC was first described extensively and was shown to be a cause of adult diarrhea resembling cholera in the severity of infection (113, 149, 154). It thus became obvious that adults with severe dehydrating cholera-like illness attributable to ETEC infections are not uncommon (132, 214). In hospitalized patients, adults often present with more severe forms of ETEC diarrhea than children and infants (Table 7). Interestingly, further analyses have shown that the elderly are also susceptible to ETEC infections requiring hospitalization (62). ETEC was found to be the second most frequently isolated (13%) bacterial pathogen after V. cholerae O1 (20%). In this age group (>65 years), patients also presented with more severe dehydration than children.

Although over 22 CFs have been detected on ETEC (Table 4), only six to eight are more frequently isolated from diarrheal stools (Table 5). Of these, CFA/I and CS1 to CS6 are the predominant types (66). These CFs are mostly present on ETEC producing ST or both LT and ST. It is believed that immunity to strains that express the nonimmunogenic ST is derived from the anti-CF response to the protein adhesins. Thus, in the development of vaccines, these CFs as well as LT are being included to give a broad-spectrum protection (189, 192).

The relationship between the presence of colonization factors and the disease-producing capability in ETEC diarrhea has been analyzed in many different epidemiological settings. In community-based studies the risk of diarrhea increased when a CF was present on the infecting strain (1). In Bangladesh the presence or absence of CFs on ETEC could not be associated with the severity of diarrhea in hospitalized patients (Table 8) (132). Studies in Mexico suggest that there is a reduced risk of diarrhea in infants if there was reinfection with ETEC producing the same compared with different CFs (38).

In volunteer challenge studies, protection was observed to ETEC with the same CF as that present on the vaccine strain (97). Some CFs are seen more often in infants than in adults, suggesting that natural immunity to infection may develop. Thus, studies in Bangladesh have shown that almost all ETEC expressing CS7 and CS17 were isolated from children less than 3 years of age (132). LT-producing ETEC strains expressing CS7 were also most pathogenic in a birth cohort in West Africa whereas CF-negative strains were not, suggesting that the presence of a CF, even in the absence of ST, enhances the virulence of ETEC (183).

View larger version (12K): [in this window] [in a new window]   FIG. 1. Estimated numbers of enterotoxigenic E. coli and V. cholerae O1 isolated from diarrheal stools of children under 5 years of age from surveillance carried out from 1996 to 2002 at the International Centre for Diarrheal Disease Research hospital in Dhaka, Bangladesh. The figure indicates monthly isolation of ETEC () and Vibrio cholerae O1 ().

View larger version (33K): [in this window] [in a new window]   FIG. 2. Monthly incidence of ETEC diarrhea in Abu Homas, Egypt, showing that ST-producing ETEC was more common in warmer months, while LT-producing ETEC was present at similar levels throughout the year. Reproduced from reference 139 with permission.

A large proportion of patients with ETEC infection have short stays in the hospital and only about 5% of patients need to be hospitalized for longer periods of time. The length of stay at the hospital is similar for patients infected with any of the three toxin phenotypes of ETEC. Few patients go on to chronic illness (>14 days). Patients with ETEC diarrhea and cholera have similar clinical characteristics and differ mainly in the rates of severe dehydration (Table 7).

Transmission of ETEC by processed food products outside of the developing world is less commonly seen but well documented. In 1977, Sack et al. found that of 240 isolates of E. coli from food of animal origin in the United States, 8% were found to contain ETEC which produced either or both LT and ST (164). None of these food products were associated with diarrheal outbreaks. In studies carried out in the 1970s in Sweden, however, outbreaks of diarrhea due to food-borne ETEC were reported (41). Similar findings were reported from Brazil in 1980 (144); 1.5% of 1,200 E. coli strains from processed hamburger or sausage were found to be ETEC. ETEC transmission on cruise ships has now been reported on several occasions (40). These findings suggested that since ETEC is not uncommonly found in meat and cheese products, these organisms have the potential for producing diarrheal outbreaks in different parts of the world.

Contaminated weaning food is also a likely cause of ETEC diarrhea in infants (139, 146). Contaminated food and water sources both contribute to seasonal outbreaks which affect tourists. Thus, ETEC is a cause of traveler's diarrhea more often in the warm than in the cool season. In a study in Bolivia it has been shown that ETEC could be isolated from a sewage-contaminated river (121). Furthermore, contaminated food and water were found to be the source of ETEC infections in Peru (23).

Surface water sources in Bangladesh, in both rural and urban areas, are highly contaminated with ETEC. Thus, recently in a study in Bangladesh, ETEC strains were obtained from clinical samples as well from ponds, rivers, and lakes around the clinical field site. In this study it was found that 32% of water samples obtained from the surface water sources were contaminated with ETEC and that the toxin and CF phenotypes of strains isolated from the clinical and environmental samples were comparable (14). Furthermore, pulsed-field gel electrophoretic analysis of the ETEC strains showed that those present in the environment were similar to the clinical isolates, supporting that, as seen for V. cholerae, surface waters may be a major source for the survival and spread of ETEC.

Studies in communities where personal hygiene, education, and general living conditions are poor have shown that infection can spread within family groups. In one study on ETEC infections in Bangladesh, the bacteria were spread to 11% of contacts in a 10-day study period (25); transmission was dependent on socioeconomic status and living conditions. Contaminated food and water and the mothers themselves, who are food handlers, seem to be the reservoirs for such infections (54). It is not surprising therefore that the possession of a sanitary latrine significantly decreased the risk of ETEC diarrhea in children in Egypt (1). On the White Mountain Apache reservation in Whiteriver, Arizona, where ETEC was found to be an important cause of diarrhea in children, these organisms were also found in river water, sites of large gatherings of Apaches on festive occasions (162). Although ETEC has been detected as a cause of diarrhea in Apache children in Arizona (162), where water and sanitation were suboptimal, subsequent studies in the developed world where water supplies and sanitation are optimal show very low frequencies of ETEC in children with diarrhea (156).

In a study in India, diarrheal illness including that caused by ETEC was found to be more severe in children with malnutrition (107). Micronutrient deficiency such as vitamin A and zinc is quite common in developing countries and generally increases the morbidity due to diarrheal illnesses (137, 140), although the effect on the morbidity of ETEC diarrhea has not yet specifically been studied. It has been estimated that in Bangladesh over 40% of children younger than 5 years of age may have zinc deficiency (131, 168). Supplementation with zinc increases the adaptive immune responses to cholera vaccination in children and adults (9, 91, 131) and in children with shigellosis (140). The effect of micronutrient deficiency on the morbidity and protective immune responses in ETEC diarrhea has not been specifically studied but is an area that needs attention. However, repeated diarrheal episodes including those induced by ETEC may be an important cause in predisposing the child to malnutrition (22, 106).

Other factors, such as breast feeding, may have the capacity to prevent ETEC diarrhea. Factors in milk such as specific secretory immunoglobulin A antibodies and receptor analogues (85) as well as innate and anti-inflammatory factors may all contribute to decrease the infection. Hyperimmune bovine colostrum containing high titers of ETEC CF antibodies has been shown to provide temporary protection against ETEC challenge (64) but is not suitable for public health application (198). Breast feeding reduces overall diarrhea and mortality (71, 205). A reduction in diarrheal episodes has been seen in infants who had been breastfed for the first 3 days of life, irrespective of other dietary practices, emphasizing the positive effects of colostrum (34). Studies in Bangladesh have shown that breast milk antibodies against cholera toxin and lipopolysaccharide do not protect children from colonization with V. cholerae but do protect against disease in those that are colonized (72).

Protection from cholera in breastfed infants of mothers immunized with killed cholera vaccine could not be correlated to antibacterial and antitoxic antibodies in breast milk, suggesting that the reduced transmission of pathogens from the mother to the infant had a protective effect (35). Since secretory immunoglobulin A antibodies to CFs and enterotoxin are present in breast milk samples from mothers in developing countries (39, 84, 187), it would be natural to assume that breastfed infants should be protected from ETEC diarrhea. However, epidemiological studies show that partial breast feeding does not result in a reduced risk of ETEC diarrhea. However, in data obtained in various studies, it appears that exclusive breast feeding practices have a positive effect of decreasing the severity and/or incidence of ETEC infections (34, 102, 136). This effect is short term and does not last long after infancy, and an overall protection is not seen in the crucial first 2 to 3 years of life (1, 34, 139).

The limited capacity of breast milk to protect against ETEC diarrhea in developing countries can also be attributed to other social and behavioral factors. These include the introduction of contaminated water and weaning food in the child's diet, leading to increased symptomatic as well as asymptomatic ETEC infections. In Mexico, the incidence of diarrhea increased even in the first 3 months of age if a barley drink was given to the infant (102). Since mixed feeding is started quite early in life in a majority of infants in developing countries, sometimes as soon as after birth, contaminated water may also be the cause of a multitude of infections (146). The importance of personal hygiene rather than breastfeeding appeared to be more protective against ETEC diarrhea in Egypt (1).

CLINICAL FEATURES

Top Previous Next References

 

The diarrhea produced by ETEC is of the secretory type: the disease begins with a sudden onset of watery stool (without blood or inflammatory cells) and often vomiting, which lead to dehydration from the loss of fluids and electrolytes (sodium, potassium, chloride, and bicarbonate) in the stool (25, 157). The loss of fluids progressively results in a dry mouth, rapid pulse, lethargy, decreased skin turgor, decreased blood pressure, muscle cramps, and eventually shock in the most severe forms. The degree of dehydration is categorized from mild to severe, and this clinical distinction is important in the provision of adequate therapy. The patients are afebrile. Usually the diarrhea lasts only 3 to 4 days and is self-limited; if hydration is maintained, the patients survive, and without any sequelae. With adequate treatment, the mortality should be very low (<1%).

The pathophysiology of the illness caused by ETEC is essentially the same as that caused by Vibrio cholerae (94) and the clinical picture is identical, especially in adults (Table 7). Studies with human volunteers have shown that the infecting dose is high for both diseases. For ETEC, the dose is around 10⁶ to 10¹⁰ CFU, with lower doses being less pathogenic (100). The need for a large infectious dose, the proliferation of the bacteria in the small bowel through colonization factors and the production of enterotoxins, and the watery, secretory type of diarrhea which produces clinical dehydration are comparable in both diseases. Both organisms produce an immunologic protective response, reflecting the observation that the attack rates are higher in children and decrease with age (24, 134).

In Bangladesh, the majority of cases of acute watery diarrhea, especially in children, are caused by three pathogens, rotavirus, V. cholerae, and ETEC (7, 8, 24). Hospital-based studies during the early 1980s have demonstrated that the purging rate is higher in cholera compared to the other two illnesses (114). A comparison of the clinical features of the disease in adults with ETEC and V. cholerae infections seeking care at the hospital in Bangladesh shows that ETEC disease differs significantly from V. cholerae infections in the severity of dehydration (Table 7), although both infections can result in severe dehydration. In comparison to children, adults with ETEC diarrhea seem to have more dehydrating illness, requiring longer hospitalization and more intravenous fluid management. This may be because of more delay in reaching a treatment facility. In children, rotavirus and ETEC diarrhea share similar clinical characteristics but differ from cholera in being less severe (Table 7).

It should be mentioned that the adult form of ETEC-related disease (of considerable severity) seems to be identified more in the Indian subcontinent. There are few (if any) reports of ETEC in adults, other than in traveler's diarrhea. This may be due to the lack of diagnosis in adults, again because of the lack of easily available laboratory techniques.

DIAGNOSIS

Top Previous Next References

 

Direct identification of the enterotoxins of ETEC has evolved over the past 35 years. Physiologic assays, the rabbit ileal loop model for LT (43), and the infant mouse assay (44) for ST were initially used as the gold standards before other simpler assays could be identified. Because LT was strongly immunogenic whereas ST was not, diagnostic assays developed along different lines. In 1974 the direct action of LT on two tissue culture cell lines, Y1 adrenal cells (46) and Chinese hamster ovarian cells (78), was found to be provide physiological responses that could be detected by morphological changes in tissue culture. These changes were specific for LT and could be neutralized by antitoxin. The two tissue culture assays were widely used for LT recognition until the development of the enzyme-linked immunosorbent assay technology in 1977 (217). Other assays such as staphylococcal coagglutination (32), passive latex agglutination (173), immunoprecipitation in agar, and the Biken test (86) were found to be specific but were not used widely for diagnostic purposes. Enzyme-linked immunosorbent assays became a widely used method for detecting LT, particularly using microtiter GM1 ganglioside methods (190, 196). Subsequently, combined GM1 enzyme-linked immunosorbent assays for ST and LT were developed (196, 197) and have been used in different epidemiological studies (1, 16, 126, 128, 132).

ST testing in infant mice continued to be used widely and could be enhanced by the use of culture pools, thereby minimizing the numbers of infant mice. In 1981 Gianella developed a radioimmunoassay for ST which compared favorably with the infant mouse assay (68).

In 1980, methods using molecular diagnostic techniques began. Moseley et al. (104) showed that the genes controlling the enterotoxins could be detected using ³²P-labeled DNA probes derived from plasmids for both LT and ST. This method was shown to be specific and sensitive and could detect as few as 1 to 100 CFU per gram of material (53, 82). Variations of this technology, including both polynucleotide and oligonucleotide probes with both radioactive and nonradioactive labeling, have been found to be useful in detecting ETEC both in clinical and environmental samples and is widely used (7, 53, 82).

In 1993, PCR was first used in ETEC diagnosis (123). It was found to be useful for diagnosis directly on fecal material as well as of isolated colonies (172). It was also adapted to a multiplex form so that the diagnosis of LT- and ST-producing organisms as well as other diarrheagenic E. coli can be made simultaneously (181, 200, 208, 209).

During recent years DNA probes, with either radioactive or nonradioactive detections or GM1 enzyme-linked immunosorbent assays using monoclonal antibodies against ST or LT have been the most widely used methods for detection of ETEC toxins (7, 133, 183, 188).

For detection of ETEC colonization factors a number of different methods have been used during the years. Initially the capacity of E. coli CFs to agglutinate certain species of erythrocytes in a mannose-resistant manner was used for demonstration of CFA/I and CS1, CS2, and CS3 (60). This nonprecise method was soon replaced by more specific slide agglutination and immunodiffusion tests initially using polyclonal sera and subsequently monoclonal antibodies against different CFs (5, 76, 110). Other methods that were used included nonspecific salting-out tests (16) and binding to tissue culture cell lines (42). These assays have now been replaced by different molecular methods, e.g., DNA probes and PCR methods against most of the known CFs or dot blot assays using several different anti-CF monoclonal antibodies (133, 139, 180, 182).

The method of choice varies from one laboratory to another and is dependent on the capability of the investigator and the level of development of the laboratory where the work is being carried out. The phenotypic methods can be set up relatively easily in different laboratories and are useful for prospective studies; most reagents are not available commercially but may be obtained from different laboratories. One point to bear in mind is that the virulence antigens are encoded by plasmid genes and can be easily lost or become silenced due to the loss of regulatory genes (182). The more recently developed DNA probe methods have the capacity to detect the structural genes for toxins and CFs and thus have the advantage of detecting ETEC from samples which have been stored for long periods of time and where phenotypic changes may have taken place. These procedures are more difficult to adapt to field sites in developing countries, where laboratory facilities may be inadequate for molecular microbiological methods. Furthermore, in some instances ETEC CFs can only be detected by molecular but not phenotypic methods, since they are not exposed on the bacterial surfaces due to mutation of genes required for surface expression (119).

Unfortunately, in spite of all these available techniques, there are still no simple, readily available methods that can be used to identify these organisms in minimally equipped laboratories. For that reason many laboratories conducting studies on diarrhea in developing countries do not include ETEC in their routine diagnostic capabilities, and special research or referral laboratories are necessary to identify these bacteria.

TREATMENT AND MANAGEMENT

Top Previous Next References

 
The treatment of diarrheal disease due to ETEC is the same as that for cholera or any other acute secretory diarrheal disease. The correction and maintenance of hydration is always most important. Antimicrobials are useful only when the diagnosis or suspicion of ETEC-related diarrhea or cholera is made. Provision of adequate nutrition is critical in children in the developing world, where all diarrheal diseases are frequent. The guidelines for therapy of all diarrheas have been widely disseminated by the World Health Organization (216).

Antimicrobials, however, are of definite benefit in the treatment of diarrhea of travelers, a diarrheal syndrome in which the clinical symptom is well recognized and ETEC is known to be the most frequent pathogen (90). It should be noted, however, that antimicrobials used for traveler's diarrhea will treat not only ETEC but also most of the other known causes (enteroaggregative E. coli, Shigella, and Campylobacter) of the diarrhea.

The antimicrobial treatment of traveler's diarrhea has changed over the years because of increasing antimicrobial resistance (58). When ETEC were first recognized, the bacteria were usually highly sensitive to all antimicrobials, including tetracyclines and trimethoprim-sulfamethoxazole (159). However, with time, antibiotic resistance emerged, necessitating the use of newer antimicrobials for treatment of traveler's diarrhea. Antimicrobials that have been used in effective treatment include doxycycline, trimethoprim-sulfamethoxazole, erythromycin, norfloxacin, ciprofloxacin, ofloxacin, azithromycin, and rifamycin. A summary of these studies over the years is given in several references (58, 159). The general history of the evolving antibiotic resistance patterns in ETEC is given in Table 9.

In areas where ETEC is endemic, antimicrobial treatment is usually not given because the diagnosis cannot be easily made microbiologically and there are no controlled studies to provide recommendations.

Although no sensitivities were reported in ETEC strains first isolated in Calcutta in 1968, when ETEC strains were used in volunteer studies (49), they were sensitive to ampicillin, which was used for treatment. ETEC strains described for the first time in Apache children in Arizona in 1971 (162) showed a completely uniform sensitivity pattern.

Because of the high sensitivity of ETEC to doxycycline, and because it has a long half-life and high levels in stool, this drug was first chosen to study antibiotic prophylaxis among travelers to developing countries (75, 111, 151). The first studies of doxycycline prophylaxis were done in Peace Corps volunteers in Kenya (150) and Morocco (160), who showed high degrees of protection (85%). In the Kenyan study (150) all ETEC strains were sensitive to tetracycline, and only a few were resistant to streptomycin and sulfonamide in the Moroccan study (160). An interesting finding in these two traveler's diarrhea studies (150, 160) and the study in Apaches (162) was that nontoxigenic E. coli strains showed more antimicrobial resistance than ETEC. This pattern was also seen in a study of large numbers of ETEC isolated before 1978 (45), suggesting that there may be some protective effect of harboring enterotoxin plasmids; it was also shown that ST-producing strains were more likely to be resistant to antimicrobials than either LT or LT/ST strains.

In 1973, Gyles (81) found that a single conjugative plasmid carried genes for both antibiotic resistance and enterotoxin production, the result of recombination of an R factor with an enterotoxin-carrying plasmid. A few years later, Echeverria (56) found that antibiotic resistance and the ability to produce enterotoxin were frequently transferred together and suggested that the widespread use of antibiotics could result in an increase of enterotoxigenic strains. Plasmids coding for both antibiotic resistance and ST could be transferred in vitro to E. coli K-12 (56) and in vivo in suckling mice, suggesting that antibiotic selective pressure could result in a wider distribution of ETEC (105). This hypothesis, however, has never been conclusively verified.

A marked increase in resistance in ETEC began to be reported in 1980, when it was found that during a cruise ship outbreak the epidemic strain O25:NM was resistant to tetracycline and sulfathiazole (104) and in a hospital outbreak, all isolates of the epidemic strain were also resistant to tetracycline (89).

During a study of traveler's diarrhea in Mexico, in 1989 to 1990, 49% of 74 ETEC strains were resistant to doxycycline, 9% to trimethoprim-sulfamethoxazole, 35% to ampicillin, but none to norfloxacin or aztreonam (47) and in studies of outbreaks of ETEC diarrhea aboard three different cruise ships during 1997 to 1998, tetracycline resistance as high as 84% (27/32) was reported while 30% were resistant to more than three antimicrobials (40). This was a marked change from previous outbreaks before 1990 when no ETEC were resistant to more than three antimicrobials.

More recently, studies from Bangladesh and India have also shown multiple antimicrobial resistance of ETEC isolates. A comparison of the resistance pattern in strains isolated recently with those obtained 30 years back highlights the increase of resistance to commonly used drugs (48). Studies of ETEC strains isolated between 1999 and 2001 show intermediate to complete resistance to multiple drugs and combined resistance to four to six drugs (including erythromycin, ampicillin, cotrimoxazole, tetracycline, streptomycin, and doxycycline); however, not a single strain was found to be resistant to ciprofloxacin. In studies in India, multidrug resistance including resistance to nalidixic acid and to fluoroquinolones is increasing (31). In Bangladesh, ETEC strains are still sensitive to drugs which are generally used for the treatment of invasive diarrhea, but there needs to be more awareness of changing drug sensitivity patterns of ETEC when erythromycin is used for treatment of acute watery diarrhea in children.

Nutritional therapy for all childhood diarrheas, including those due to ETEC, is an integral part of diarrhea treatment. Episodes of diarrhea due to any cause, including ETEC, result in decreased nutritional status and thus inhibit growth in children (106). Attention to providing food, particularly breast milk, early in the course of therapy is essential. Additional food during and following the diarrheal episode will help in catch-up growth (3).

PREVENTION

Top Previous Next References

 

Based on the great impact of ETEC infections on morbidity and mortality, and probably also on nutritional status (106), particularly of children in areas where they are endemic, an effective ETEC vaccine is highly desirable. Such a vaccine is feasible since epidemiologic evidence and results from experimental challenge studies with human volunteers have demonstrated that specific immunity against homologous strains follows ETEC infection. Furthermore, multiple infections with antigenically diverse ETEC strains seem to lead to broad-spectrum protection against ETEC diarrhea (38). Experimental studies with animals and indirect evidence from clinical trials (191) suggest that protective immunity against ETEC is mediated by secretory immunoglobulin A antibodies directed against the CFs, other surface antigens, and LT; ST, which is a small peptide, does not elicit neutralizing antibodies following natural infection.

To provide broad-spectrum protection, an ETEC vaccine should probably contain fimbrial antigens representative of the most prevalent ETEC pathogens. The great diversity of ETEC serotypes, with regard to both O and H antigens, makes such antigens less attractive as vaccine components. Since CFA/I and CS1 to CS6 are the most common human ETEC fimbriae, they are key candidate immunogens in an ETEC vaccine. Other fimbrial CFs may also be considered, based on their relative importance in certain geographic areas (see Table 5). Since a majority of ETEC strains that produce both LT and ST or ST only produce CFs, it has been postulated that a multivalent ETEC vaccine containing CFA/I and CS1 to CS6 may provide protection against approximately 50 to 80% of ETEC strains in most geographic areas (189). If an LT toxoid such as the nontoxic B subunit LTB or a mutant LT is included, a multivalent toxoid-CF vaccine might provide relatively broad protection against 80 to 90% of ETEC strains worldwide. Inclusion of, e.g., CS7, CS12, CS14, and CS17 might expand the potential spectrum of coverage to up to 90% of all ETEC strains (189). A number of different strategies have been taken to deliver fimbrial and toxin antigens of ETEC to the human immune system to elicit protective immune responses and functional immunological memory.

An alternative administration route that has been considered is to give an ETEC vaccine by the transcutaneous route. Such administration of E. coli CS6 together with LT has induced immune responses against CS6 in about half of the volunteers and anti-LT responses in all of them (77). Work is in progress to evaluate E. coli LT as a candidate vaccine after transcutaneous immunization (73).

In an initial pilot study, the rCTB-CF ETEC vaccine was shown to confer 82% protective efficacy (P < 0.05) against ETEC disease in European travelers going to 20 different countries in Africa, Asia, and Latin America (213). However, the number of cases fulfilling the inclusion criteria was low. In a large placebo-controlled trial in nearly 700 American travelers going to Mexico and Guatemala, the rCTB-CF ETEC vaccine was shown to be effective (protective efficacy, 77%; P = 0.039) against nonmild ETEC d