Comparison of beta-glucuronidase and indole-based direct plating methods for enumerating Escherichia coli in artificially inoculated ground meats

Peptone tergitol glucuronide (PTG) agar containing 4-methylumbelliferyl-beta-D glucuronide (MUG) (for beta-glucuronidase activity), the Holbrook, Anderson, Baird-Parker (HABP) method (for detecting indole production), and the standard 3-tube most probable number (MPN) method were compared with plate count agar (PCA) for enumerating three strains of unstressed Escherichia coli artificially inoculated into ground beef and chicken at 1-6 X 10(6) cells/g. No significant difference (P0.05) was determined between PTG agar and PCA in the recovery of E. coli. The MPN method enumerated a significantly greater (P0.05) number of E. coli cells than PCA. Compared with PCA, the HABP method recovered a significantly lower (P0.05) number of E. coli cells from chicken, whereas no significant difference (P0.05) was obtained with ground beef. When combining all data from chicken and beef, the recovery of E. coli cells by the HABP method was also significantly lower (P0.05). Overall, based on the enumeration of E. coli on PCA, the HABP method, PTG agar, and MPN method recovered 57, 102, and 144%, respectively