Validation of three rapid screening methods for detection of verotoxin-producing Escherichia coli in foods:...

An interlaboratory study was conducted for the validation of 3 methods for the detection of all verotoxin-producing Escherichia coli (VTEC) in foods. The methods were a multi-analyte 1-step lateral flow immunoassay (LFIA) for detection of E. coli O157 and verotoxin (VT); an enzyme-linked immunosorbent assay targeted against VT1, VT2, and VT2c (VT-ELISA); and a polymerase chain reaction (PCR) method for detection of VT genes (VT-PCR). Aliquots (25 g or 25 mL) of 4 food types (raw minced [ground] beef, unpasteurized milk, unpasteurized apple juice [cider], and salami) were individually inoculated with low numbers (<9 to 375 cells/25 g) of 6 test strains of E. coli (serogroups O26, O103, O111, O145, and O157) with differing VT-producing capabilities. Five replicates for each test strain and 5 uninoculated samples were prepared for each food type. Fourteen participating laboratories analyzed samples using the LFIA, 9 analyzed the samples by ELISA, and 9 by PCR. The LFIA for O157 and VT had a specificity (correct identification of negative samples) of 92 and 94%, respectively, and a sensitivity (correct identification of positive samples) of 94 and 55%, respectively. The VT-ELISA and VT-PCR had a specificity of 98 and 99%, respectively, and a sensitivity of 89 and 72%, respectively.