Comparison of commercially available kits with standard methods for the detection of coliforms and Escherichia coli in foods.

Three commercially available kits that were supplemented with substrates for enzyme reactions were evaluated to determine their abilities to detect coliforms and fecal coliforms in foods. Japanese and U.S. Food and Drug Administration standard methods, as well as two agar plate methods, were compared with the three commercial kits. A total of 50 food samples from various retailers were examined. The levels of detection of coliforms were high with the commercial kits (78 to 98%) compared with the levels of detection with the standard methods (80 to 83%) and the agar plate methods (56 to 83%). Among the kits tested, the Colilert kit had highest level of recovery of coliforms (98%), and the level of recovery of Escherichia coli as determined by beta-glucuronidase activity with the Colilert kit (83%) was comparable to the level of recovery obtained by the U.S. Food and Drug Administration method (87%). Isolation of E. coli on the basis of the beta-glucuronidase enzyme reaction was found to be good. Levine's eosine methylene blue agar, which has been widely used in various laboratories to isolate E. coli was compared with 4-methylumbelliferyl-beta-D-glucuronide (MUG)-supplemented agar for isolation of E. coli. Only 47% of the E. coli was detected when eosine methylene blue agar was used; however, when violet red bile (VRB)-MUG agar was used, the E. coli detection rate was twice as high. Of the 200 E. coli strains isolated, only 2 were found to be MUG negative, and the gene responsible for beta-glucuronidase activity (uidA gene) was detected by the PCR method in these 2 strains. Of the 90 false-positive strains isolated that exhibited various E. coli characteristic features, only 2 non-E.coli strains hydrolyzed MUG and produced fluorescent substrate in VRB-MUG agar. However, the PCR did not amplify uidA gene products in these VRB-MUG fluorescence-positive strains.