Two rapid methods for detection of Escherichia coli exceeding 10(4)/g action levels: precollaborative study

The current AOAC Method 966.24 for enumeration of Escherichia coli in foods uses a most probable number (MPN) procedure with extensive confirmation steps. Two new methods based on membrane filtration (MF) were compared to the MPN reference method for detection of high levels of E. coli in 5 food types, some of which represent categories for which the U.S. Food and Drug Administration (FDA) mandates additional testing if an action level of 10(4)/g E. coli is exceeded. Ground beef, which is not FDA regulated, was also tested. The 5 food types were all inoculated at 3 levels: 10(2)/g, > or = 10(4)/g, and > or = 10(5)/g E. coli. An MF protocol using either m-ColiBlue24 (CB) or lauryl sulfate tryptose plus BCIG (LST/BCIG) was an effective potential alternative to the reference method. Sensitivity and specificity for both CB and LST/BCIG were 98 and 100%, respectively. Agreement between MPN and both CB and LST/BCIG was 98%. The 2 proposed methods allow completion of both presumptive and confirmatory steps in 1-3 days, whereas the reference method requires as many as 11 days. Exclusivity testing with 50 non-E. coli strains indicated 100% were correctly ruled out by the proposed protocols. Inclusivity testing was used to determine whether typical results were obtained after incubation of E. coli cultures on CB or LST/BCIG for 24 h. Of 50 E. coli strains tested, 100% yielded typical results after incubation on CB, and 98% yielded typical results after incubation on LST/BCIG.