AOAC RI ERP Final Action Recommendations eBook

Collaborative Study Approved Protocol Expert Review Panel Use Only

instructions for performing the study (Appendix 10.1) and required materials prior to the start of the 1 study. Training on the BAX System will be provided as necessary by a DuPont Qualicon representative.

2 3 4

3.0 5 6 A collaborative study will be conducted in accordance with the AOAC INTERNATIONAL Methods 7 Committee Guidelines for Validation of Microbiological Methods for Food and Environmental Surfaces 8 (4). The BAX method will be compared to the USDA-FSIS MLG method for the detection of Salmonella in 9 hot dogs. The BAX and USDA-FSIS MLG methods share a common enrichment for hot dogs, and thus 10 this matrix will be tested as a paired study. The BAX method will be compared to the FDA BAM method 11 for the detection of Salmonella in orange juice. A proprietary enrichment, the BAX® MP broth, will be 12 used to enrich orange juice for the BAX method, so this matrix will be tested as an unpaired study. 13 14 Each matrix will be screened for the presence of naturally contaminating Salmonella and total aerobic 15 plate count ( note : for the FDA BAM method, if the level of background microflora is > 10 4 cfu/g, then the 16 tetrathionate (TT) broth must be incubated at 43 ± 0.2°C in a circulating, thermostatically-controlled, 17 water bath, and if the level is ≤ 10 4 cfu/g, then TT should be incubated at 35 ± 2.0°C). Naturally 18 contaminated matrix is preferred if available. If naturally contaminated matrix is not available, an 19 artificially contaminated matrix will be prepared. A different serovar of Salmonella will be used for each 20 matrix (see Table 2). Each matrix will be divided into 3 samples. One sample serves as the 21 uncontaminated level, one sample is contaminated at a level that will produce a reference method POD 22 (POD R ) or candidate method POD (POD C ) in the range of 0.25 – 0.75 (approximately 0.2 – 2 cfu/test 23 portion), and finally one sample is contaminated at a level to assure a POD C of nearly 1.0 (approximately 24 5 cfu/test portion). For each matrix, twelve replicate test portions per sample will be tested by the 25 candidate method and the reference method. 26 27 On the day that the analysis of the test samples is initiated, a 5-tube 3-level Most Probably Number 28 (MPN) estimation of contamination levels will be conducted using the appropriate reference method. 29 The MPN analysis scheme may also make use of the reference method replicates (see Appendix X-A of 30 the AOAC Guidelines for details). The Least Cost Formulations, Ltd. (LCF) MPN Calculator-Version 1.6 (5) 31 will be used to determine the MPN values and 95% confidence intervals. The MPN is reported as 32 MPN/test portion with 95% confidence intervals. Study Design

33 34 35

Table 1. Food Preparation

Test Portion Codes Random number 100-199

Inoculation Organisms

Inoculation Level per Organism

Reference Method USDA FSIS- MLG Ch. 4.05

Matrix

Replicates

0 cfu/325 g

12 12 12

S . Worthington (DD4043)*

Hot dogs

0.2–2 cfu/325 g 2–5 cfu/325 g

5