JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1995, p. 222–224
Copyright ? 1995, American Society for Microbiology
Vol. 33, No. 1
Evaluation of an Immunoassay for Direct Detection of
Escherichia coli O157 in Stool Specimens
BRENDA L. DYLLA,1EMILY A. VETTER,1JOHN G. HUGHES,1AND FRANK R. COCKERILL III1,2*
Division of Clinical Microbiology, Department of Laboratory Medicine and Pathology,1and Division of
Infectious Diseases, Department of Internal Medicine,2Mayo Clinic, Rochester, Minnesota 55905
Received 12 August 1994/Returned for modification 12 September 1994/Accepted 11 October 1994
An enzyme-linked immunosorbent assay (ELISA) produced by LMD Laboratories, Inc., Carlsbad, Calif., was
compared with culture for the detection of Escherichia coli O157. Nine of 185 stool specimens evaluated had
positive results by the LMD E. coli O157 ELISA and grew E. coli O157 on culture; 174 had negative LMD E.
coli O157 ELISA results and did not grow E. coli O157 on culture. Of 174 specimens negative by LMD E. coli
O157 ELISA, 117 specimens grew other enteric pathogens: Campylobacter spp. (46 isolates), Salmonella spp. (43
isolates), Yersinia spp. (20 isolates), and Shigella spp. (8 isolates). There were two indeterminant results by the
LMD E. coli O157 ELISA. One stool specimen did not grow other enteric pathogens on culture, and one grew
a Campylobacter sp. on culture. Both had negative LMD E. coli O157 ELISA results upon repeat testing. The
LMD E. coli O157 ELISA is an accurate, easy-to-read screening method for the detection of E. coli O157 in fecal
Escherichia coli O157:H7 causes hemorrhagic colitis and has
been associated with hemolytic-uremic syndrome (4, 5). The
rapid diagnosis of E. coli O157:H7 directly from stool samples
is preferred to avoid unnecessary diagnostic procedures (3, 11)
and inappropriate antimicrobial therapy (10, 13) and to iden-
tify common sources linked to transmission (1, 2).
Conventional testing methods for E. coli O157:H7 involve
the isolation of the organism on culture from stool samples
followed by biochemical and immunologic confirmation. This
method usually requires 72 h for complete identification.
Rapid, less labor-intensive methods are desirable. In this
context, we evaluated the utility of a rapid enzyme-linked
immunosorbent assay (ELISA) compared with that of our
standard culture method for detecting E. coli O157 directly
from stool samples.
(This study was presented in part at the 94th General
Meeting of the American Society for Microbiology, Las Vegas,
Nev., 23 to 27 May 1994.)
Materials and methods. Stool specimens (185) collected
from 161 patients between 5 August 1992 and 14 April 1994
were referred to the Mayo Clinic Clinical Microbiology Labo-
ratory for enteric pathogen culture and were included in the
present study. All patients were evaluated in either the outpa-
tient or inpatient facility associated with the Mayo Clinic in
Rochester, Minn. Seventy percent of the patients were Min-
nesota residents; the remaining patients were from 11 other
states and two other countries. The stool specimens were
transported to the laboratory in sterile containers and cultured
immediately. An aliquot of stool specimen was set aside for
ELISA testing the same day (30 specimens) or frozen at ?20?C
for batch testing (155 specimens).
Stool specimens were inoculated onto 5% sheep blood agar
(BBL, Cockeysville, Md.), eosin-methylene blue agar (BBL),
Hektoen enteric agar, cefsulodin-irgasan-novobiocin agar,
Campylobacter agar, and sorbitol-MacConkey agar and into
selenite F broth. The above-mentioned media were incubated
at 35?C in ambient air, with the exception of the Campylobacter
agar, which was incubated at 42?C in an atmosphere of 5% O2,
10% CO2, and 85% N2, and the cefsulodin-irgasan-novobiocin
agar, which was incubated at 30?C in ambient air. Colonies that
were sorbitol negative on sorbitol-MacConkey agar were sub-
cultured to 5% sheep blood agar, incubated overnight, and
tested with spot-indole reagent (p-dimethylaminocinnamalde-
hyde). Indole-positive isolates were tested for the enzyme
?-glucuronidase with the substrate 4-methylumbelliferone-?-D-
glucuronide. Agglutination with E. coli O157 antisera (Difco
Laboratories, Detroit, Mich.) was performed on 4-methylum-
belliferone-?-D-glucuronide-negative isolates. Isolates that
were agglutinated by E. coli O157 antisera were tested for the
presence of the H7 flagellar antigen by inoculating the organ-
ism into motility agar and into motility agar containing H7
antisera (Difco). E. coli cells possessing the H7 antigen are
immobilized in the H7 antibody-containing motility agar but
remain motile in the non-H7 antibody-containing motility agar.
A confirmatory conventional biochemical set was also per-
The LMD Laboratories, Inc., E. coli O157 antigen detection
kit provides plastic microwell test strips coated with anti-E. coli
O157 polyclonal antibodies. The ELISA was performed ac-
cording to the manufacturer’s instructions. A slurry of each
stool specimen was prepared by adding a minimal amount of
diluted wash buffer, vortexing the mixture, and allowing par-
ticulate material to settle out. Two drops (100 ?l) of stool
supernatant and a positive and a negative control sample were
added to the microwells, and the plates were incubated at
room temperature (15 to 25?C) for 20 min. The plates were
manually washed three times, after which 2 drops of peroxi-
dase-labeled anti-E. coli O157 antibodies were added. The
plates were incubated for 10 min at room temperature and
were then washed manually three times. The wells were rinsed
with deionized water. One drop each of substrate A (tetra-
methylbenzidine) and substrate B (peroxide) was added, and
the plates were incubated at room temperature for 5 min. Two
drops of stop solution (1 M phosphoric acid) were added, and
the plates were read spectrophotometrically at 450 nm. The
* Corresponding author. Mailing address: Division of Clinical Mi-
crobiology, Mayo Clinic, Hilton 4, 200 First St. SW, Rochester, MN
55905. Phone: (507) 284-2901. Fax: (507) 284-4272. Electronic mail
absorbance values provided by the manufacturer were used for
interpretation of patient specimens and controls.
Results. A total of 9 of 185 stool specimens were positive by
conventional culture for E. coli O157:H7; two of these speci-
mens were obtained from the same patient. All nine specimens
were also positive by the LMD E. coli O157 ELISA, yielding a
sensitivity of the LMD ELISA of 100%. A total of 176 of 185
specimens were negative by conventional culture for E. coli
O157:H7; two of these specimens were positive by the LMD E.
coli O157 ELISA, while the remaining 174 specimens were
negative by the LMD ELISA.
Stool specimens that were culture positive for other enteric
pathogens with negative results by LMD E. coli O157 ELISA
included Campylobacter spp. (45 isolates), Salmonella spp. (43
isolates), Yersinia spp. (20 isolates), and Shigella spp. (8
isolates). One exception was a specimen culture positive for
Campylobacter spp. This specimen contained particulate mate-
rial which resulted in a positive absorbance reading. Upon
repeat testing, the specimen was negative.
Fifty-nine of 60 specimens with negative cultures for enteric
pathogens had negative LMD E. coli O157 ELISA results. One
was positive by the E. coli O157 ELISA and contained
particulate material as described above. The specimen was
negative upon repeat testing. If one counted the two stool
specimens with particulate matter and initially positive results
by LMD E. coli O157 ELISA as positives, the overall specificity
for the LMD E. coli O157 ELISA was 98.9%.
Discussion. E. coli O157:H7 is the third most common cause
of community-acquired bacterial diarrhea diagnosed at the
Mayo-Rochester Medical Center. During the period from 1
January 1991 through 31 December 1993, the following rela-
tive frequencies of bacterial pathogens isolated from 14,699
stool specimens at the Mayo-Rochester Medical Center were
noted: for Campylobacter jejuni, 40.9%; for Salmonella spp.,
34.5%; for E. coli O157:H7, 10.6%; for Yersinia enterocolitica,
9.7%; for Shigella spp., 3.9%; and for Vibrio spp., 0.4%
(unpublished data). Rapid methods for the detection of E. coli
O157:H7 diarrhea are preferred to limit diagnostic evaluations
including invasive procedures for hemorrhagic diarrhea in the
clinical setting, decrease the time and expense associated with
conventional culture techniques, speed public health evalua-
tion of possible epidemics, and provide suitable isolation
procedures for infected patients in the hospital setting (1–3, 6,
10, 11, 13).
Our laboratory standard culture method used to identify E.
coli O157:H7 from stool specimens usually requires 72 h.
Detection methods for E. coli O157:H7 disease other than
culture have been reported, but because of special procedures
(e.g., fluorescence microscopy and cell monolayer cultures,
etc.), they are not easily adapted to most laboratories. Ritchie
and colleagues (12) reported the utility of a direct fecal assay
for the detection of verotoxins, or Shiga-like toxin I (SLT-I)
and Shiga-like toxin II (SLT-II), associated with E. coli
O157:H7 disease. This assay requires the observation of cyto-
toxicity of stool supernatants which are applied to a Vero cell
monolayer. Neutralization of cytotoxicity with monoclonal
antibodies directed against SLT-I and SLT-II is also required
to confirm cytotoxicity related to these toxins. This assay
requires a minimum of 48 h and is likely as labor-intensive as
culture. Theoretically, this method has the advantage of de-
tecting Shiga-like toxins from E. coli serotypes other than E.
coli O157:H7 which have been shown to produce the same
clinical syndrome and are not as easy to recover by conven-
tional culturing methods (12). However, the frequency with
which these E. coli serotypes cause hemorrhagic colitis or
hemolytic-uremic syndrome appears to be low (12). Detection
of Shiga-like toxins from Shigella isolates is also possible, as
Shigella spp. possess them (12). Such a nonspecific assay may
therefore preclude clinical usefulness, especially for Shigella
spp., for which isolation of the organism is preferred so that
susceptibility testing can be performed.
Paton and colleagues (9) described a method for detecting
the presence of nucleic acids which encode SLT-I and SLT-II
directly from stool samples. This method incorporates PCR
and DNA probe hybridization detection formats. Like both the
cytotoxin cell culture and conventional culture methods, this
method is not a rapid procedure and requires at least 24 h to
complete. Furthermore, Paton and colleagues observed a
significant number of positive results for patients who were
asymptomatic. In this study, direct comparisons of PCR results
with results from toxin cytotoxicity assays and E. coli O157:H7
culture isolation were not done and PCR sterilization proce-
dures to prevent contamination were not used.
In another study, Yamada and colleagues (14) used an
ELISA for the direct detection of SLT-I and SLT-II from
stools from patients who were infected during an E. coli
O157:H7 outbreak. Of 14 inpatients, 5 had fecal Shiga-like
toxins in stools obtained within 5 days of the onset of symp-
toms. Among these five patients, three had E. coli O157:H7
recovered from culture whereas the other two did not. Of the
14 inpatients, 4 additional patients had E. coli O157:H7
recovered from stools but did not have Shiga-like toxins
detected by ELISA. In summary, this Shiga-like toxin ELISA was
relatively easy to perform but was less sensitive than culture.
Park and colleagues (8) recently described a rapid (?2-h
turnaround time) direct immunofluorescent antibody staining
method to detect E. coli O157:H7 directly from stool speci-
mens. For 336 fecal samples evaluated, direct immunofluores-
cent antibody staining of the fecal smear detected all isolates of
E. coli O157 that were recovered by culture, including nonmo-
tile strains, strains possessing the H7 flagellar antigen, and one
with a flagellar antigen other than H7.
Like the assay evaluated by Park and colleagues, the LMD
E. coli O157 ELISA described herein appears to be an
acceptably sensitive, specific, and rapid (?1-h turnaround
time) method for directly screening stool samples for E. coli
O157. According to the manufacturer, there was no demon-
strated cross-reactivity with other E. coli organisms not pos-
sessing the O157 antigen, including the following serotypes:
O26:H11, O55, O88:H49, O91:21, O111:NM, and O163:NM.
Additionally, no cross-reactivity was demonstrated for the fol-
lowing gram-negative bacteria: Escherichia hermanii, Aeromonas
hydrophila, Brucella abortus, Citrobacter freundii, Enterobacter
cloacae, Hafnia alvei, Pseudomonas aeruginosa, Salmonella
typhimurium, Xanthomonas maltophilia, and Y. entero-colitica O9.
An advantage of the LMD E. coli O157 ELISA over the direct
assay used by Park and colleagues is that it does not require
fluorescent microscopy. For the LMD E. coli O157 ELISA,
additional prospective studies are required to corroborate our
results, as the total number of stools positive for E. coli
O157:H7 was low (9 of 185 specimens). Nevertheless, our
results suggest that this assay could be used to screen stool
samples prior to culture for E. coli O157 and that negative
LMD E. coli O157 ELISA results may not require confirma-
tory culture. The LMD E. coli O157 assay may detect nonmo-
tile (negative H serotypes) as well as H serotypes other than
H7 which may not produce Shiga-like toxins. Cultures per-
formed with O157-positive specimens would allow for H-
antigen confirmation. In this study, inoculation of 176 fecal
specimens onto sorbitol-MacConkey agar could have been
avoided, which could represent savings in materials and sub-
sequent evaluation. In some clinical settings, especially when
VOL. 33, 1995NOTES 223
recovery of the organism on culture is not required for
additional study (e.g., strain analysis for epidemiologic pur-
poses ), this assay could replace culture. As antimicrobic
therapy is not currently recommended for E. coli O157:H7
disease, isolation of the organism for susceptibility testing is
also not required (10, 13). The LMD E. coli O157 ELISA is a
rapid direct test which is easy to perform and can be readily
adapted to most laboratories. This assay may be particularly
useful for large-scale batch screening of routine stool speci-
mens submitted to reference laboratories or large-scale screen-
ing of stool specimens from exposed persons by public health
laboratories during outbreaks of food-borne diarrhea.
We thank Roberta Kondert for her efforts in preparing the manu-
script, the Mayo Clinic laboratory personnel for processing of speci-
mens, and the LMD Laboratories for providing the LMD E. coli O157
1. Centers for Disease Control and Prevention. 1993. Preliminary report:
foodborne outbreak of Escherichia coli O157:H7 infection from hamburgers.
Western United States, 1993. Morbid. Mortal. Weekly Rep. 42:85–86.
2. Centers for Disease Control and Prevention. 1993. Update: multistate
outbreak of Escherichia coli O157:H7 infections from hamburgers—Western
United States, 1992–1993. Morbid. Mortal. Weekly Rep. 42:259–261.
3. Griffin, P. M., S. M. Ostroff, R. V. Tauxe, et al. 1988. Illness associated with
E. coli O157:H7 infections. Ann. Intern. Med. 109:705–712.
4. Karmali, M. A. 1989. Infection by verocytotoxin-producing Escherichia coli.
Clin. Microbiol. Rev. 2:15–38.
5. Karmali, M. A., M. Petric, P. C. Fleming, G. S. Arbus, and H. Lior. 1985. The
association between idiopathic hemolytic uremic syndrome and infection by
verotoxin-producing Escherichia coli. J. Infect. Dis. 151:775–782.
6. Martin, D. L., K. L. MacDonald, K. E. White, J. T. Soler, and M. T.
Osterholm. 1990. The epidemiology and clinical aspects of the hemolytic
uremic syndrome in Minnesota. N. Engl. J. Med. 323:1161–1167.
7. O’Brien, A. D., A. R. Melton, C. K. Schmitt, M. L. McKee, M. L. Batts, and
D. E. Griffin. 1993. Profile of Escherichia coli O157:H7 pathogen responsible
for hamburger-borne outbreak of hemorrhagic colitis and hemolytic uremic
syndrome in Washington. J. Clin. Microbiol. 31:2799–2801.
8. Park, C. H., D. L. Hixon, W. L. Morrison, and C. B. Cook. 1994. Rapid
diagnosis of enterohemorrhagic Escherichia coli O157:H7 directly from fecal
specimens using immunofluorescence stain. Am. J. Clin. Pathol. 101:91–
9. Paton, A. W., J. C. Paton, P. N. Goldwater, and P. A. Manning. 1993. Direct
detection of Escherichia coli Shiga-like toxin genes in primary fecal cultures
by polymerase chain reaction. J. Clin. Microbiol. 31:3063–3067.
10. Proulx, F., J. P. Turgeon, G. Delaze, L. Lafleur, and L. Chicoine. 1992.
Randomized, controlled trial of antibiotic therapy for Escherichia coli
O157:H7 enteritis. J. Pediatr. 121:299–303.
11. Riley, L. W. 1985. Hemorrhagic colitis—a ‘‘new’’ disease. Clin. Microbiol.
12. Ritchie, M., S. Partington, J. Jessop, and M. T. Kelly. 1992. Comparison of
a direct fecal Shiga-like toxin assay and sorbitol-MacConkey agar culture for
laboratory diagnosis of enterohemorrhagic Escherichia coli infection. J. Clin.
13. Tarr, P. E., M. A. Neill, D. L. Christie, and D. E. Anderson. 1988. Escherichia
coli O157:H7 hemorrhagic colitis. N. Engl. J. Med. 318:1697.
14. Yamada, S., S. Matsushita, A. Kai, M. Sasaki, A. Tsuji, T. Kanemitsu, N.
Yamashita, E. Anzai, and Y. Kudoh. 1993. Detection of verocytotoxin from
stool and serological testing of patients with diarrhea caused by Escherichia
coli O157:H7. Microbiol. Immunol. 37:111–118.
224NOTESJ. CLIN. MICROBIOL.