Enumeration of Escherichia coli in Foods by the Hydrophobic Grid-Membrane Filter Method (HGMF) Method

HPB Method MFHPB-26
March 2001

Si vous avez besoin d'aide pour accéder aux formats de rechange, tels que Portable Document Format (PDF), Microsoft Word et PowerPoint (PPT), visitez la section d'aide sur les formats de rechange.

---

1. APPLICATION

The method may be used for enumeration of E. coli biotype 1 in foods other than bean or alfalfa sprouts, to determine compliance with the requirements of Sections 4 and 7 of the Food and Drugs Act. Where an

Official Method for a food is specified, that method shall be followed. This revised method replaces MFHPB-26, dated April 1997.

2. DESCRIPTION

The method has been shown to produce satisfactory results with beef and ground beef, chicken and deboned chicken, fish, nuts, black pepper, cheese and green beans (9.2, 9.3, 9.4). The HGMF method was granted

Official Final Action Status by the Association of Official Analytical Chemists (9.2). Since this time it has been used for the enumeration of E. coli biotype 1 in most foods and food ingredients, generally without reports of problems. The exceptions are bean and alfalfa sprouts, in which large populations of indole- positive Klebsiella spp. preclude an accurate determination of E. coli.

3. PRINCIPLE

The HGMF analysis takes 24-30h and yields counts that are higher and more precise than the Most Probable Number method (MFHPB-19), and equivalent to or higher than the Direct Plating Method (MFHPB-27) (9.4).

A single dilution gives an accurate count over a wide range of contamination levels. Counting precision may be better than on conventional plates or membrane filters because the HGMF reduces the effect of individual visual acuity on the count (9.5). If a low count is expected, the detection limit can be lowered by filtering more of the suspension.

The enumeration of presumptive E. coli is based on incubation at a selective temperature and the demonstration of indole production through formation of pink to red grid cells (9.5). It detects those E. coli classed as late lactose fermenting or anaerogenic (approx. 10% of strains) which are missed by the MPN method, but does not detect indole-negative E. coli (3-5% of strains). The specificity of the test is high; overall confirmation rates of indole-positive colonies on membrane filters is 95% for E. coli biotype 1 (IMViC ++--), while a further 3.4% were confirmed as faecal coliforms on the basis of IMViC +++- or +--- reactions (9.1). This level of specificity was independently confirmed (9.6). Thus, confirmation of E. coli should not normally be necessary. Since the indole staining reaction kills bacterial cells, further examination of the sample would require replication of the HGMF prior to staining.