Laboratory investigation of diarrhea in travelers to Mexico: evaluation of methods for detecting enterotoxigenic Echerichia coli.

A laboratory investigation was conducted on cultures collected from travelers before, during, and after a trip to Mexico to characterize the etiology of traveler's diarrhea. Four laboratory methods for detecting enterotoxigenicity of Escherichia coli were evaluated: the infant mouse assay, the Chinese hamster ovary (CHO) cell assay, the Y1 adrenal cell assay, and the rabbit ileal loop. Although a number of common enteric pathogens were identified as a cause of traveler's diarrhea, including six serotypes of Salmonella, two serotypes of Shigella, Vibrio parahaemolyticus, Giardia lamblia, and Entamoeba histolytica, enterotoxigenic Escherichia coli was most commonly isolated. Strains were identified that produced only heat-labile enterotoxin (LT), only heat-stable enterotoxin (ST), or both LT and ST. The infant mouse assay yielded results falling into two distinct groups, providing a clear separation of positive and negative cultures. The CHO assay also formed two groups, with positive cultures producing 11% or more of the elongated cells. There was good agreement between the CHO and the Y1 adrenal cell assays for detection of LT. The adrenal cell system for detection of LT was more suitable than the CHO assay for processing large numbers of specimens because of the miniculture modification of this method utilized in this study. The infant mouse method was a simple and reliable method for detecting ST.     

Hemagglutination by Escherichia coli in septicemia and urinary tract infections.

The agglutination of erythrocytes from various animal species by Escherichia coli was studied. The 405 strains of E. coli were isolated from urine in patients with urinary tract infections, from blood in septicemic patients, or from feces in persons without intestinal or urinary disorders. In urinary tract infections, d-mannose-resistant agglutination (MRHA) of human erythrocytes was the most common finding (23% of the strains). The highest frequency of mannose-sensitive hemagglutination (MSHA) attributed to type I (common type) pili occurred with guinea pig erythrocytes (11.5%). Of the 78 E. coli strains isolated from blood cultures, 11 (14%) produced MRHA of human erythrocytes and only one gave MSHA. In the stool cultures, only 1 of 170 E. coli strains was MSHA reacting, whereas 28 strains (16.5%) showed MRHA of human erythrocytes. No MRHA strain reacted with antiserum against colonization factor antigen (CFA)/I of pilus nature in enterotoxigenic human E. coli strains (O78:H12). MRHA of bovine erythrocytes, reputedly typical of enterotoxigenic E. coli of serogroups O6 and O8, was shown by only two strains, neither of which agglutinated with CFA/II antiserum. The most common hemagglutinating pattern of E. coli from urine and blood thus was MRHA for human erythrocytes. This agglutination may have been caused by pili or other surface properties of one or more serotypes. These may represent a new class of colonization-promoting antigens (adhesins).     

Detection of Escherichia coli enterotoxins in stools.

We determined whether enterotoxigenic Escherichia coli diarrhea could be diagnosed by direct examination of stools for heat-labile (LT) and heat-stable (ST) enterotoxins. The Y-1 adrenal cell and an enzyme-linked immunosorbent assay (ELISA) detected LT in 85 and 93%, respectively, of stool specimens obtained from adults with acute diarrhea from whom an LT- and ST-producing organism had been isolated. Furthermore, the ELISA assay detected LT in 8 of 35 stool specimens from which no LT-producing E. coli had been isolated. The infant mouse assay was utilized to detect ST in these stool specimens and was found to be an insensitive method, showing positive results in only 36% of the specimens from which an ST-producing organism was isolated. Further studies are warranted to determine the diagnostic value of direct detection of LT in stools, especially by the ELISA method.     

Development of PCR for screening of enteroaggregative Escherichia coli.

In this study, we determined the sequence of the EcoRI-PstI fragment of the plasmid pCVD432, also termed the enteroaggregative Escherichia coli (EAggEC) probe. A primer pair complementary to this probe was designed for PCR amplification of a 630-bp region. Comparison of the analysis of the EAggEC probe sequence with those in database libraries revealed no significant similarity to any known bacterial gene. Pure cultures of E. coli cells, as well as mixed cultures from stool specimens, were investigated with the PCR assay, the EAggEC probe test, and the adherence test. Of 50 E. coli strains which demonstrated aggregative adherence to HEp-2 cells, 43 (86%) were positive with the EAggEC PCR. All 43 of these strains reacted with the EAggEC probe. Six EAggEC strains gave negative results by both molecular techniques. In contrast, only 4 of 418 (0.96%) strains representing other categories of diarrheagenic E. coli demonstrated a positive PCR result. The PCR was also successful in screening for the presence of EAggEC in enriched cultures grown from stool specimens. Compared with cell culture assays and colony hybridization, our findings revealed that the PCR assay was more rapid, simple, and highly sensitive and can therefore be recommended as a screening method for EAggEC in the clinical laboratory.     

Cyclic Adenosine Monophosphate and Alteration of Chinese Hamster Ovary Cell Morphology: a Rapid, Sensitive In Vitro Assay for the Enterotoxins of Vibrio cholerae and Escherichia coli

The major limitation to our understanding of the clinical importance of enterotoxigenic Escherichia coli in diarrheal illness has been the lack of a simple rapid assay for the enterotoxin produced by certain E. coli. On the basis of the activation of adenylate cyclase by heat-labile enterotoxin of E. coli (LT) and by cholera toxin (CT) in intestinal and other tissues, cultured Chinese hamster ovary (CHO) cells with known morphological responses to dibutyryl cyclic adenosine 5'-monophosphate (AMP) were exposed to these enterotoxins. Crude culture filtrates of LT-producing E. coli and CT stimulated cyclic AMP accumulation and cell elongation in CHO cells. The similarity of time course, concentration dependence, and potentiation by phosphodiesterase inhibitors suggested cyclic AMP mediation of the morphological change. Heat inactivated CT and LT in this system. Choleragenoid inhibited CT; antiserum against CT inhibited both enterotoxin effects. In contrast to culture filtrates of 16 strains of E. coli known to produce LT, culture filtrates from 13 E. coli that do not produce LT did not alter CHO cell morphology. The morphological change is a simple, specific assay for these enterotoxins and detect 3 x 10(-17) mol of CT or a 1:250 dilution of crude culture filtrate of LT-producing E. coli 334.     

Passive immune hemolysis for detection of heat-labile enterotoxin produced by Escherichia coli isolated from different sources.

Fifty-one strains of Escherichia coli isolated from humans, swine, food, and water and identified as enterotoxinogenic by the Y-1 adrenal cell assay, were examined for heat-labile enterotoxin (LT) production by the passive immune hemolysis test. Cholera antitoxin, anti-choleragenoid and anti-LT were used as antisera. Cholera antitoxin was much more potent than anti-choleragenoid and LT antiserum in the detection of LT-positive strains. All strains isolated from pigs and sausage were negative in tests made with LT antiserum. A few strains isolated from humans, food, and water also gave negative results. These data showed that the passive immune hemolysis test is not as efficient as the Y-1 adrenal cell assay in the detection of enterotoxinogenic E. coli strains.     

Escherichia coli K antigen in relation to serum-induced lysis and phagocytosis.

The presence of capsular polysaccharides (K antigens) and their relation to phagocytosis and sensitivity to the lytic action of serum of 26 strains of E. coli isolated from stools of healthy volunteers and from blood cultures were studied. Four of 12 strains isolated from stool cultures and 12 (86%) of the 14 strains isolated from blood cultures possessed K antigen. Three of the 12 strains isolated from stool cultures and seven of the 14 isolated from blood cultures were resistant to uptake by polymorphonuclear leucocytes; these resistant strains contained large amounts of K antigen. By contrast 10 strains, three with low amounts of K antigen and seven without detectable amounts of K antigen, were readily phagocytosed. Thus it appears that K antigen renders E. coli resistant to phagocytosis. Only four (15%) of the 26 strains were sensitive to serum lysis and there was no correlation between the presence of K antigen and the resistance to serum lysis.     

A new heat-labile cytolethal distending toxin (CLDT) produced by Escherichia coli isolates from clinical material

A new heat-labile Escherichia coli toxin cytolethal to Vero, HeLa, HEp-2 and CHO cells and negative in Y-1 cells has been demonstrated in culture filtrates of 43 E. coli strains associated with diarrheal disease. This new toxin was termed a cytolethal distending toxin (CLDT) to reflect the progressive cell distention and cytotoxicity evidenced in all sensitive tissue cells. CLDT was distinct from the classic heat-labile (LT) and heat-stable enterotoxins, Verotoxins and hemolysins and was produced by some strains of the following E. coli serogroups (02, 07, 08, 018, 022, 039, 044, 055, 083, 086, 091, 0113, 0119, 0128 and 0167). Moderate cyclic AMP accumulation (75-fold) was observed in CHO cells exposed to E. coli CLDT for 24 h. In contrast to E. coli LT, cyclic AMP levels were optimal at 24 h and were observed to decrease over the following 72 h period in CHO cells exposed to E. coli CLDT. In addition to heat-lability at 70 degrees C for 15 min, E. coli CLDT demonstrated a molecular weight over 30,000, was nondialyzable and trypsin-sensitive. E. coli CLDT was negative in adult rabbit ligated ileal loop and suckling mouse assays and provoked only an erythematous response in rabbit skin. All CLDT-positive E. coli strains identified to date were non-hemolytic and non-invasive. Verotoxin and heat-labile enterotoxin were found in combination with CLDT in a few E. coli strains.     

Increased Escherichia coli enterotoxin detection after concentrating culture supernatants: possible new enterotoxin detectable in dogs but not in infant mice.

The heat-stable enterotoxin (ST) of Escherichia coli can be detected by infant mouse or dog intestinal loop tests. These tests differ in that the dog assay uses concentrated culture supernatants and is based on measurements of net intestinal absorption, whereas the mouse test uses unconcentrated supernatants and depends on gross fluid accumulation. To compare the relative sensitivities of these assays, culture supernatants of randomly selected E. coli isolates from 34 Bangalee diarrhea patients were tested for ST in dog loops and infant mice. Supernatants were also tested for heat-labile enterotoxin (LT) in dog loops, Y-1 adrenal cells, and Chinese hamster ovary cells. E. coli supernatants that produced positive responses for both ST and LT in the dog loop assay (ST+/LT+) also produced positive responses when tested for ST in infant mice and for LT in cell lines. Supernatants of strains negative for ST and LT in dog loop (ST-/LT) were also negative in other assays. Of 10 strains positive for just ST in the dog loop test (ST+/LT-), only 5 were ST positive in the standard infant mouse test. Supernatants of the other five strains (dog loop positive, mouse test negative) were then concentrated 100-fold and retested in mice. Three of these five gave consistently positive results after concentration, and two were only intermittently positive. Concentrated supernatants of negative control strains (ST-/LT-) were all negative in mice. The dog assay detects more strains producing ST than the infant mouse test. The infant mouse test, which detects only gross fluid accumulation, failed to detect approximately half of the 10 strains which produced ST alone (ST+/LT-; P = 0.025). Concentrating supernatants for the mouse assay increases sensitivity for detection of ST, but certain E. coli strains produce a variety of ST to which infant mice do not respond.     

Sensitivity of the different tests of enterotoxigenicity of Escherichia coli and their correlation

Enterotoxigenicity of E. coli isolates was tested in 136 cases of acute gastroenteritis. Heat labile toxin (LT) produced in-vitro was tested in rabbit ileal loop (RIL); vero cell line and Biktn plate. The results of live cultures were evaluated in RIL. The overall data of these four models were not statistically different. Elaboration of LT in these four models ranged from 14-21.4%. Out of the 20 LT producing strains 14 (70%) also revealed ST. Of the 6 positive reactors on vero cell line, appeared to produce vero toxin (VT) only. Out of 29 LT positive E. coli, 1 (3.45%) and 2 (6.89%) strains respectively revealed colonising factor antigen (CFA) I and II. The high incidence of ETEC showing both LT and ST has been highlighted in the age group 0-4 years, and its impact on nutritional status is discussed.     

Evaluation of non-radioactive trivalent DNA probe (LT, ST1a, ST1b) for detecting enterotoxigenic Escherichia coli.

AIMS: To evaluate a digoxigenin-labelled trivalent DNA probe (LT, ST1a, ST1b) for detecting enterotoxigenic Escherichia coli (ETEC), by comparison with a cell culture assay for detecting LT, individual DNA probes for LT, ST1a and ST1b, and an enzyme immunoassay for detecting ST1. METHODS: A 1268 base pair DNA fragment, containing parts of the genes for E coli heat labile enterotoxin (LT) and heat stable enterotoxins (ST1a and ST1b), was random prime labelled with digoxigenin-dUTP. The labelled DNA was used as a probe in colony hybridisation reactions to examine 180 E coli strains of which 92 had previously been shown by a cell culture assay to produce LT. Six LT negative ST1 positive E coli, 34 Verotoxin producing E coli (VTEC), and 84 organisms from other genera were also examined. All organisms other than VTEC were isolated from travellers returning from abroad with diarrhoea. All E coli strains were retested by cell culture for LT, and were tested by enzyme immunoassay (EIA) for ST1, and by the trivalent and individual DNA probes. RESULTS: All 81 isolates, that on retesting by cell culture were positive for LT, also hybridised with the trivalent and LT probes; 27 of these were also enzyme immunoassay (EIA) positive for ST1 of which 24 hybridised with the ST1b probe and three with the ST1a probe. Of 99 isolates, that on retesting by cell culture were negative for LT, all were negative by LT probe and only three were EIA positive for ST1; these three were positive by both trivalent and ST1b probes. Four isolates were positive by the trivalent probe but negative by cell culture and EIA; all four were positive by ST1b probe. Compared with the cell culture assay for LT, the probe had a sensitivity and specificity both of 100%; compared with the EIA for ST1, the probe had a sensitivity of 100% and specificity of 88%. CONCLUSIONS: The trivalent DNA probe is a sensitive, specific, and reliable method for detecting ETEC that should be considered for use by diagnostic microbiology laboratories.     

Detection of enterotoxigenic Escherichia coli colonization factor antigen I in stool specimens by an enzyme-linked immunosorbent assay.

An enzyme-linked immunosorbent assay (ELISA) was employed to detect and quantitate the fimbrial colonization factor antigen (CFA/I) of enterotoxigenic Escherichia coli in stool specimens obtained from adult cases of diarrhea in which CFA/I-positive E. coli was the known causative agent. The inhibition method, or blocking technique, was used. In this method, a standardized dilution of human anti-CFA/I serum was preincubated with dilutions of stool extract before transfer to CFA/I-coated microtiter plate wells, and then ELISA was performed with alkaline phosphatase-conjugated anti-human immunoglobulin. CFA/I purified from E. coli strain H-10407 (O78:H11) was used. Acute-phase diarrheal stool specimens were found to contain approximately 3.0 mg of antigen (mean value) per g stool, whereas control (CFA/I-negative) specimens contained insignificant amounts (less than 0.03 mg/g) of antigen. Also, CFA/I was detected in culture fluids of CFA/I positive enterotoxigenic E. coli belonging to a variety of serotypes and was undetectable in similar preparations from P-strains (spontaneous CFA/I-negative derivatives) of the same test cultures. Equivalent results were obtained in ELISA tests by using bacterial cells taken from isolated colonies grown on CFA agar. These results indicate that the ELISA technique will be useful for the diagnosis of diarrhea caused by CFA/I-positive enterotoxigenic E. coli.     

Immunological nonidentity of heat-labile enterotoxins from human and porcine enterotoxigenic Escherichia coli.

The Ouchterlony double-gel diffusion test demonstrated that heat-labile enterotoxin (LT) from human strains of enterotoxigenic Escherichia coli was antigenically related but not identical to that from porcine E. coli. All human strains examined produced immunologically identical LTs, and all porcine strains examined produced immunologically identical LTs. The Biken test (modified Elek test), developed for detection of E. coli LT (Honda et al., J. Clin. Microbiol. 13: 1-5, 1981), was useful for confirmation of these results.     

Improved GM1-enzyme-linked immunosorbent assay for detection of Escherichia coli heat-labile enterotoxin.

Previously described GM1 ganglioside enzyme-linked immunosorbent assays (GM1-ELISA) for the detection of Escherichia coli heat-labile enterotoxin (LT) showed sensitivity equal to the Y-1 adrenal cell assay when anti-LT (a reagent not commercially available) was used. However, when antitoxin to immunologically related (commercially available) cholera toxin was substituted, a marked loss in sensitivity occurred. We modified the GM1-ELISA that employed anti-cholera toxin to make it comparable in sensitivity to the Y-1 adrenal cell assay. When five media commonly used for LT production were compared, Mundell's Casamino Acids medium was shown to be significantly superior. Lincomycin (45 micrograms/ml) added to E. coli cultures significantly increased net optical densities in the GM1-ELISA, a direct measure of the amount of LT. Treatment of broth cultures or bacterial cell pellets with polymyxin B or extension of culture time to 48 h also significantly increased net optical density by allowing enhanced release of periplasmic LT. A major innovation involved the direct culture of E. coli strains in GM1-coated wells of microtiter plates followed by ELISA. This direct culture method GM1-ELISA (DCM-GM1-ELISA) saved not only assay time, but also materials and reagents. The net optical densities that result from this assay allow the test to be read visually without a spectrophotometer. Three independent observers read plates with E. coli tested by DCM-GM1-ELISA. Thirty-four of 35 adrenal cell-positive strains (97% sensitivity) and 30 of 30 LT-negative control E. coli strains (100% specificity) were identified by all three observers reading coded plates. The DCM-GM1-ELISA provides a simple, practical and efficient assay for LT for less sophisticated laboratories.     

Immunomagnetic separation and DNA hybridization for detection of enterotoxigenic Escherichia coli in a piglet model.

Enterotoxigenic Escherichia coli (ETEC) strains were detected by stool blot hybridization assays using different oligonucleotide probes for the colonization fimbrial antigen F4, heat-stable enterotoxin I (ST I), and heat-labile enterotoxin (LT I) genes. Forty-eight fecal samples and seven samples of intestinal content from ETEC-challenged newborn piglets were processed in two ways: (i) by direct inoculation of bacterial suspension onto nylon membranes overlaying blood agar and (ii) by immunomagnetic enrichment of F4+ ETEC using magnetic beads coated with F4 monoclonal antibodies before inoculation onto nylon membranes. In samples obtained from nondiarrheic piglets pre- and postchallenge, E. coli genes for F4, ST I, and LT I could be detected only after immunomagnetic enrichment. No difference between the two methods in detection of these E. coli genes was observed when stool blots from diarrheic piglets were examined. By using magnetic separation, it was easy to decrease background bacterial flora, intestinal cells, and fecal debris and thus produce purer specimens. The method evaluated in this animal model appeared simple and quick and increased the sensitivity of detection of ETEC strains 100-fold compared with the direct stool blot hybridization assays. Prior bacterial isolation and identification were not necessary.     

Evaluation of a ganglioside immunosorbent assay for detection of Escherichia coli heat-labile enterotoxin.

The GM1 ganglioside enzyme-linked immunosorbent assay (GM1-ELISA), an immunological method for detection of Escherichia coli heat-labile enterotoxin (LT), was quantitatively and qualitatively compared with the conventional adrenal cell test for the identification of LT-producing strains. A micromodification model of the assay was developed. Enterotoxin preparations from 120 E. coli isolates from individuals with diarrhea, which had been previously shown to be enterotoxigenic by the adrenal cell test, and from 44 control strains of E. coli were compared in parallel by the two methods. Quantitatively the covariation of the enterotoxin titers was highly significant (RS = 0.98, P less than 0.001), the GM1-ELISA being somewhat more sensitive than the adrenal cell test. The methodological error was less than 5% in both tests. Qualitatively the overall agreement for positive and negative reactions for the two methods was 89%. The GM1-ELISA is practical for routine use in the diagnosis of enterotoxigenic E. coli, especially in laboratories without facilities for cell culture.     

Partition of heat-labile-enterotoxin genes between human and animal Escherichia coli isolates.

Heat-labile toxin (LT) genes from human and animal Escherichia coli isolates from a restricted geographical region (Thailand) exhibited a segregated pattern of dissemination that was revealed by a restriction enzyme site polymorphism. The results suggest an exclusion of LT genes from or a low transfer rate for LT genes between E. coli strains infectious for humans and those infectious for animals in natural settings.     

Probing for enterotoxigenicity among the salmonellae: an evaluation of biological assays.

Sixty-eight Salmonella strains representing 39 serotypes were variously screened for enterotoxigenicity by using the Chinese hamster ovary (CHO), Y1 adrenal, and Vero cell tests, rabbit skin tests for delayed permeabiltity factor (DPF) and rapid permeability factor (RPF), the rabbit ileal loop test, and the infant mouse test. An iron-sufficient medium, YT-1, and a deferrated medium, DF, were compared. Of the culture supernatant fluids of strains grown in DF medium, 66% yielded positive reactions in the CHO cell test compared with only 10% with TY-1 medium. The corresponding performances with supernatant fluids of DF medium cultures in Y1 adrenal and Vero cell tests were 85 and 69% positive, respectively. The overall agreement between the Y1 adrenal or CHO cell test and the rabbit skin test for DPF, i.e., positive or negative in both tests, was about 70%. Positivity in DPF tests was a better predictor of positivity in either the Y1 adrenal or rabbit ileal loop test than vice versa. CHO cell, DPF, and rabbit ileal loop reactivities of unheated culture filtrates were each neutralized by anticholera antitoxin. Only four strains gave positive reactions in the infant mouse test, whereas up to 66% were positive for RPF in rabbit skin, based on positivity in Ty-1 or DR medium or both. DPF and RPF were produced by 35% of the strains. Of the 28 isolates from human stools, 82 and 92% and all of 11 strains tested were positive in the DPF, Y1 adrenal cell, and rabbit ileal loop tests, respectively. The corresponding data for 17 sewage isolates, representing 17 different serotypes rarely isolated from human stools in Sweden, were 63 and 69% and 8 of 8 tested. On the basis of this investigation, rabbit skin tests for both DPF and RPF provide the most reliable means of screening for enterotoxigenicity among salmonellae.     

Detection of enterotoxigenic Escherichia coli after polymerase chain reaction amplification with a thermostable DNA polymerase

The direct identification of enterotoxigenic Escherichia coli from clinical specimens was examined by using the polymerase chain reaction (PCR) for amplifying the heat-labile toxin (LT) gene. Two synthetic primers, each of which was 20 bases in length, were used with the thermostable DNA polymerase from Thermus aquaticus to amplify the LT gene. The amplified PCR products were detected by either gel electrophoresis or hybridization to a 24-base synthetic oligonucleotide probe conjugated to alkaline phosphatase. The PCR method detected LT-positive bacteria but did not react with E. coli producing the heat-stable toxin, enteroinvasive E. coli, Salmonella typhi, Salmonella typhimurium, or Shigella dysenteriae. By the PCR method, a single bacterium could be detected following 30 cycles of amplification. The T. aquaticus DNA polymerase was inhibited by more than 10(3) organisms in the amplification reaction mixture. A group of 40 clinical specimens consisting of 16 LT bioassay-positive and 24 LT bioassay-negative stool specimens were tested by PCR for the presence of toxigenic E. coli. The total DNA from 100 microliters of stool specimen was extracted and partially purified with a commercially available ion-exchange column. All 16 of the bioassay-positive stool specimens were positive by PCR. In addition, one stool specimen which was bioassay negative for LT but positive for LT in a previous hybridization assay with a different LT probe was also positive by PCR. This may indicate that the LT gene is present but either is not expressed or is expressed below detectable levels. Amplification of specific DNA sequences by PCR provides a highly sensitive and specific tool for the detection of pathogenic microorganisms directly from clinical specimens without the need for prior isolation. This technique may find wide application in the detection of other organisms in addition to enterotoxigenic E.coli.     

Comparison of rapid methods for detection of heat-labile (LT) and heat-stable (ST) enterotoxin in Escherichia coli.

Oxoid VET-RPLA, ST-EIA and Pharmacia Phadebact ETEC-LT enterotoxin tests were compared to find a simple but reliable method for detecting enterotoxigenic Escherichia coli (ETEC) in Hungary. In the Oxoid tests, all six reference LT- or ST-producing strains, except one ST-producer, gave positive results. Of 11 reference porcine enterotoxigenic strains, all four LT-producers gave positive reactions for LT but three of 10 ST-producers gave negative reactions for ST. Thirteen of 50 strains from culture collections of H. Steinrück (Germany) were LT+ and nine of 33 were ST+. When 31 isolates were tested simultaneously with the Oxoid and the Pharmacia LT tests, 12 strains were LT+ by the Oxoid LT test but by the Phadebact LT test only seven of these strains were LT+ and, of the remainder, three gave uncertain results and two gave negative results. Of 69 porcine strains, seven were LT+ and three ST+. Of 901 human strains isolated in Hungary, 10 were LT+ and one of 24 tested was ST+. In two cases, ETEC strains were isolated from contacts of travellers returning from Mongolia and Bangladesh. Results of comparative studies with reference strains corresponded well to those of the classical toxin detection tests. The Oxoid test was rapid, sensitive, specific and easy to perform and is recommended for use in screening ETEC isolates.