To identify the reservoirs of shiga toxin-producing Escherichia coli O157, sensitive detection and
isolation methods are necessary. The sensitivity of traditional culture methods can be improved significantly
by the inclusion of an immunoconcentration step, resulting in less false-negative results. In this study,
enrichment procedure and immunomagnetic separation (IMS) were compared for use in conjunction with a
multiplex polymerase chain reaction (M-PCR) method for detection of genes stx1, stx2, eaeA and hlyA. A
total number of 975 faecal samples were collected from 26 dairy farms in Shiraz area, Shiraz, southern Iran.
The samples were cultured at 37°C for 18–24 hrs in modified tryptic soy broth (m-TSB). Each of five
enriched samples were pooled and examined in two ways—direct PCR and IMS. The detection limit of the
M-PCR protocol for seeded E. coli O157:H7:ATCC:43895 in m-TSB without stool was 1.23 × 102 CFU/ml,
whereas it was 1.23 × 106 CFU/ml with enriched faecal sample. In direct PCR of enriched samples, no
positive sample was detected. However, in IMS of enriched samples one specimen was positive. The
prevalence of E. coli O157:H7 in faeces of cows in examined farms was 0.51% and the herd prevalence was
3.86%. Isolation of this serotype from faecal samples indicates that cattle are reservoirs of this pathogen and
potentially a source of human infection. This finding is of considerable public health importance