Comparison of methods for detection and isolation of cold- and freeze-stressed Escherichia coli O157:H7 in raw ground beef

A comparison was made of the relative efficiencies of three enrichment media, RapidChek Escherichia coli O157:H7 enrichment broth (REB), R&F broth (RFB), and modified E. coli broth containing novobiocin (mEC+n), and four selective plating media for detection of cold- and freeze-stressed E. coli O157:H7 in raw ground beef. Ground beef (25 g) was inoculated with E. coli O157:H7 at < or =0.5 and < or =2 CFU/g, and samples were then enriched immediately or were stored at 4 degrees C for 72 h or at -20 degrees C for 2 weeks and then enriched. After 8 or 20 h of enrichment, the cultures were plated onto R&F E. coli O157: H7 chromogenic plating medium, cefixime-tellurite sorbitol MacConkey agar, CHROMagar O157, and Rainbow agar O157 and tested using the RapidChek E. coli O157 lateral flow immunoassay and a multiplex PCR assay targeting the E. coli O157: H7 eae, stx1, and stx2 genes. Recovery of E. coli O157:H7 on the four agar media was 4.0 to 7.9 log CFU/ml with the REB enrichment, 1.4 to 7.4 log CFU/ml with RFB, 1.7 to 6.7 log CFU/ml with mEC+n incubated at 42 degrees C, and 1.3 to 3.3 log CFU/ml from mEC+n incubated at 35 degrees C. The percentages of positive ground beef samples containing nonstressed, cold-stressed, and freeze-stressed E. coli O157:H7 as obtained by plating, the immunoassay, and the PCR assay were 97, 88, and 97%, respectively, with REB, 92, 81, and 78%, respectively, with RFB, 97, 58, and 53%, respectively, with mEC+n incubated at 42 degrees C, and 22, 31, and 25%, respectively, with mEC+n incubated at 35 degrees C. Logistic regression analyses of the data indicated significant main effects of treatment, type of medium, enrichment time, inoculum concentration, and detection method. In particular, a positive result was 1.1 times more likely to occur after 20 h of enrichment than after 8 h, 25 times more likely with RFB and REB than with mEC+n at 35 degrees C, 3.7 times more likely with an initial inoculum of < or = 2.0 CFU/g than with < or = 0.5 CFU/g, 2.5 to 3 times more likely using freeze-stressed or nonstressed bacteria than with cold-stressed bacteria, and 2.5 times more likely by plating than by the immunoassay or the PCR assay. REB had better overall performance for enrichment of cold- and freeze-stressed E. coli O157:H7 present in ground beef than did the other media examined.