Annex T

Testing for Escherichia coli (E. coli) in Slaughter Establishments
(ref. 9 CFR Part 310.25 and Part 381, Subpart K)

T.1 Introduction

The "Pathogen Reduction (PR) and HACCP Systems (HACCP); Final Rule" requires that all slaughter establishments collect samples from the type of livestock or poultry that it slaughters in largest number and test these samples for generic Escherichia coli (E. coli - Biotype I). The purpose of this requirement is to have slaughter establishments use microbiological testing as a process control tool for measuring the effectiveness of their dressing procedures.

In order to maintain USA export access, these requirements must be implemented in an equivalent manner in Canadian establishments. Sampling for E. coli shall be done in a manner consistent with the requirements set out in this section.

T.1.1. Types Of Product To Be Sampled Under These Requirements

Carcasses from the following species are covered by this requirement:

• Cattle (All types including steer, heifer, cow, bull, veal (hide-off and hide-on));
• Swine
• Sheep
• Goats
• Horses, mules and other equines
• Chickens (All types including fryer, roaster, heavy broiler, fowl, rock Cornish hen, etc.);
• Turkeys (All types)
• Ducks
• Geese
• Guineas
• Squabs
• Ratites

Affected establishments which slaughter any of the above species must take samples for the species from the above list which it slaughters in the greatest amount. Where an establishment only slaughters species not found in the above list (e.g. quails, bison, rabbits, etc.) no sampling is required.

T.2 Technical Requirements

T.2.1 Written Sampling Protocol And Sampling Procedures; Written Records

The operator will develop and have a written protocol for E. coli sampling done at the establishment. The protocol shall be made available to the Veterinarian in Charge (VIC) upon request. The written protocol shall contain procedures covering all the establishment's areas of responsibility. The procedures shall be detailed to the extent necessary to enable verification. This would include the following information:

• Sample selection: e.g., who carries out the sampling, when and where sampling will be done, how random sampling of carcasses is achieved;
• Pre-sample preparation: e.g., check list of tasks to be performed prior to sample collection, materials needed for sample collection, verification of the suitability of materials before testing (e.g.: sponges checked to ensure they do not have bactericidal properties);
• Sampling procedure to be used;
• Shipping procedures (if applicable): e.g., who packages, where is packaging done, where are samples kept pending shipment, who ships samples; where are samples shipped(laboratory)/how are samples shipped (shipping agent);
• Tamper-proofing and protection of samples: e.g., how samples will be handled/packaged and shipped to ensure sample integrity (temperature maintenance and tamper proofing);
• Analytical method used: e.g., minimum sensitivity, official AOAC method used (or method published by other scientific body;
• Compilation and analysis of results // record keeping activities: e.g., who receives test results, who reviews and compiles data, where will lab reports/worksheets be kept, how long are records kept, how is CFIA inspector's access to results provided;
• Process Verification Criteria: e.g., "m" and "M" values, method used to set Process Verification Criteria when the sampling method does not have pre-set "m" and "M" values;
• Corrective action procedures: e.g., for inconclusive results, for non-compliance with procedures as found during the operator's internal verification activities, for E. coli results exceeding action levels; for E. coli sampling program results indicating possible loss of process control;

The VIC shall review the written program to verify that all the requirements set out in this Annex are met. Reference may be made to an establishment's general sample collection procedures provided that the references are to the procedure specific sections and that details specific to E. coli sampling procedures are provided. Procedures may also be written into the establishment's HACCP plans.

T.2.2 Sampling Frequency:

Sampling frequency for E. coli testing is determined by production volume. With the exception of Very Low Production Volume establishments, establishments shall collect samples at the rate set out in the following Table or at a minimum of at least once a week.

Table T.2.2

E. coli Testing Frequencies

Cattle, sheep, goats, horses, mules and other equines: 1 test per 300 carcasses.

Swine: 1 test per 1,000 carcasses.

Chickens: 1 test per 22,000 carcasses

Turkeys, ducks, geese and guineas, squabs and ratites: 1 test per 3,000 carcasses

Notes:

1. These testing frequencies do not apply to Very Low Volume Establishments. See Section T.2.2.1
2. A slaughter establishment with a HACCP System (recognized as per FSEP) may substitute an alternate E. coli sampling frequency in its HACCP plan for the species, provided that the establishment has data to show that the alternate frequency is suitable for verifying the effectiveness of the HACCP plan (CCP's and controls). The alternate frequency and scientific validation/justification would be assessed during the recognition process.

T.2.2.1 Sampling Rates for Very Low Volume Establishments:

Some establishments may be classified as Very Low Volume Establishments. The maximum yearly slaughter volumes for very low volume establishments are described in the following table.

Table T.2.2.1

Maximum Yearly Livestock Slaughter Volumes for Very Low

Volume Establishments

Very Low Volume Establishments shall conduct one series of sampling annually beginning the first full week that the establishment operates after June 1¹. Sampling shall continue at the rate of one sample taken each week that the establishment operates until:

• where the establishment uses a sampling procedure with pre-set Process Verification Criteria (see section T.2.8):
  • the establishment has taken a series of 13 tests that demonstrates that their process in under control (meet m and M established criteria as per the section T.2.8);
• where the establishment uses a sampling procedure for which pre-set Process Verification Criteria do not exist:
  • June 1 of the following year until 13 samples have been collected which ever comes first that demonstrates that the process is under control (see section T.2.8 );

It will not be required to take any further samples during the year in Very Low Volume Establishments unless:

• changes are made in establishment facilities, equipment, personnel or procedures;
• the establishment or CFIA determines that the changes may affect the adequacy of existing process control measures;

The VIC shall assess any changes made to determine if there is need for further sampling. Records of the assessment shall be conserved (record under E. coli inspection task). If it is determined that the change may required further sampling, the VIC shall contact the Director, Programs Network (DPN) to discuss the matter. If the DPN concurs that further sampling is required, the establishment will be so advised in writing.

T.2.3 Pre-sampling Preparation Procedures

Sample collection will be carried out by the individual designated in the establishment's written sampling protocol. Sampling supplies, such as sterile gloves, sterile sampling solutions, hand soap, sanitizing solution, etc., as well as specific materials needed for sampling different carcass types (i.e., specimen sponges in bags and template for sampling cattle or swine carcasses), will need to be assembled prior to beginning sample collection.

When sponges are used, a verification of the sponge's suitability must have been done. Sponges must not possess antimicrobial properties which might reduce bacterial counts. Supplier certification (e.g., letters of guarantee) shall be obtained for each new lot or in-house verification of each new lot of sponges shall be performed.

Such verification typically should consist of immersing a sponge for a number of hours within a seeded bath of transport medium (e.g. Butterfield's Phosphate Diluent) to simulate shipping conditions. The bath containing indicator organisms in a known quantity is cultured at the end of the period to verify if the concentration of bacteria has decreased significantly, which would indicate that substances within the sponge are killing bacteria.

For cattle, swine and equine carcass sampling, a template will be needed to mark off the area to sample. The template can be made of metal or aluminum foil, brown paper, flexible plastic, etc. Some disposable templates may come sterilized and individually prepackaged. To make a reusable template for cattle and swine, cut out a 10 centimetres (cm) x 10 cm square from a sheet larger than the area to be sampled (See Figure 1). For turkeys, ratites, sheep and goats, the template shall be 5 cm x 10 cm.

If a reusable template is used, it will need to be sanitized with an approved sanitizing solution [e.g., hypochlorite (bleach) solution or alcohol] for at least 2 to 3 minutes. However, the template needs to be totally dry before placing it on the carcass. Aluminum foil or paper templates can be used once and discarded. The foil for the template should be stored in a manner to prevent contamination. Since the area enclosed by the template will be sampled, take care not to touch this area with anything other than the sampling sponge. Using dirty or contaminated material may lead to erroneous results. If an autoclave is available, paper or aluminum foil templates can be wrapped in autoclavable paper and sterilized.

Sterile sampling solutions, Butterfield's phosphate diluent (BPD), can be stored at room temperature. However, at least on the day prior to sample collection, check solutions for cloudiness. DO NOT use solutions that are cloudy, turbid or contain particulate matter. Place the number of containers of sampling solution (BPD) that will be needed for the next day's sampling in the refrigerator. Ensure that shipping containers, coolant packs and shipping documents are prepared as required.

Sterile Buffered Peptone Water (BPW) can be substituted for BPD. It must be noted however that there may be a slight rise in bacteria counts following this change (¼ log increase). If Process Verification Criteria have been published by USDA for the sampling method used, marginal ("m") and unacceptable ("M") limits are not allowed to be changed.

T.2.4 Procedure for the Random Selection of Carcasses:

Samples are to be taken randomly at the required frequency (See section T.2.2, "Sampling Frequency"). For example:

• For cattle (1 test required/per every 300 head slaughtered): a plant which slaughters 150 head of cattle an hour will select 1 sample at random for every 2 hour interval of production.
• For turkeys (1 test required/per every 3,000 turkeys slaughtered): a plant which slaughters 1,000 turkeys an hour will select 1 sample at random for every 3 hour interval of production.

Cattle half-carcasses and swine, sheep, goat, equine and ratites carcasses should be selected from those in the cooler 12 or more hours after slaughter². Poultry carcasses should be selected at random after chilling³.

The operator's written procedure must clearly explain how the random selection of samples is achieved. Every carcass (for cattle, every half-carcass, e.g. "leading" and "trailing" side) should have an equal chance of being selected from all eligible half-carcasses/carcasses. If multiple lines exist, randomly select the line for sample collection for that interval. Repeat the random selection process for the next sampling interval. Each line should have an equal chance of being selected at each sampling interval.

Use of random numbers tables or similar random number generating systems (e.g. computer-generated, calculator generated, drawing cards, etc.) is required. To ensure that all carcasses have an equal chance of being selected, the procedure must ensure:

1. that the TIME when sample is to be selected is determined in a random manner (alternatively, random selection of the carcass sequence number achieves this - see Note ²);
2. that, where there is more than one possible site for selecting the sample, the SITE is determined in a random manner. For example:
   • Selection of beef, hog, sheep, goat and equine carcasses when selection is done within coolers (12 or more hours after slaughter) or on transfer chains, grading chains or other rails: the selection of the cooler and site within the cooler where the carcass is to be selected for the sampling period (i.e. day's production) is determined in a random manner; AND the selection of the half-carcass or side of carcass to sample is randomly determined;
   • Selection of poultry carcass when the selection is done after chilling, at the end of the drip line or last readily accessible point prior to packaging (e.g., carcasses which are subjected to immersion freezing) - see note ³: that the selection of the chiller/drip line/packaging line where the sample is to be selected for the sampling period (i.e. day's production) is determined in a random manner.
3. where the sample selection procedure involves selecting a CARCASS/HALF-CARCASS at a randomly determined sampling time/sampling site, the following procedure is used to eliminate bias and ensure random selection.

   After having identified the carcass or half-carcass selected by the random procedure from the predetermined point, count back five (5) carcasses/half-carcasses and select the next carcass/half-carcass for sampling. Each carcass/half-carcass must have an equal chance of being selected. The reason for counting back five carcasses/half-carcasses is to avoid any possible bias during selection.

Additionally, for poultry carcasses only: because only WHOLE (untrimmed) carcasses are to be sampled, if the fifth carcass is not a WHOLE (untrimmed) carcass, count back an additional five carcasses for sample selection. Repeat if necessary.

If more than one shift is operating at the plant, sample(s) can be taken on any shift, provided the number of samples corresponds to the total day's production and that TIME, SITE and CARCASS requirements have been met.

T.2.5 Procedures to Ensure Aseptic Techniques/Sampling

Micro-organisms from the environment, hands, clothing, sample containers, sampling devices, etc., may lead to erroneous analytical results. Stringent requirements for microbiological analysis are necessary: use of aseptic sampling techniques and clean, sanitized equipment and supplies are of utmost importance.

There should be an area designated for preparing sampling supplies, etc. A stainless steel, wheeled cart or table would be useful during sampling. A small tote or caddy could be moved to the location of sampling and could be used for carrying supplies, supporting sample bags when adding sterile solutions to sample bags, etc.

Sterile gloves should be used for collecting samples. The only items which may contact the external surface of the glove are the exposed sample being collected and/or the sterile sample utensil (specimen sponge). Keep in mind that the outside surfaces of the sample container are not sterile. Do not handle the inside surface of the sterile sample containers. Do not touch anything else. The following procedure for putting on sterile gloves can be followed when collecting samples:

1.  Peel open the package of sterile gloves from the top without contaminating (touching, breathing on, contacting, etc.) the exterior of the gloves.
2. Remove the first glove by holding it from the wrist-side opening inner surface. Avoid any contact with the outer surface of the glove. Insert the washed and sanitized hand into the glove, taking care not to puncture the glove.
3. Remove the second glove by holding it by the cuff (outer surface). Avoid contaminating the outside of the first glove when pulling on the second glove. Take care not to touch the face, skin, clothes and other non-sterile surfaces.
4. If at any time you are concerned that a glove may be contaminated, discard it and begin again with Step (a) above.

T.2.5.1 Preparation for Sample Collection:

Prior to collecting samples, review appropriate sampling steps, random selection procedures, and other information that will aid in sample collection.

On the day prior to sample collection, after checking for cloudiness/ turbidity, place the number of BPD containers that will be needed for the next day's sampling in the refrigerator/cooler. If samples are to be shipped to an off-site facility, pre-chill shipping container and refrigerator packs.

On the day of sampling, gather all sample collection bags, sterile gloves, sanitizer, hand soap, sterile solutions for sampling, and specific materials listed under the Materials section (i) of the sample collection section for the type of carcass to be sampled. Ensure that all sampling supplies are on hand and readily available before beginning sample collection.

Label the sample bags before starting the sampling procedure. Use permanent ink. If you are using paper labels, it is important that the label be applied to the bag at normal room temperature; it will not stick if applied in the cooler.

Outer clothing (frocks, gloves, head gear, etc.) worn in other areas of the plant should be removed before entering the sampling area or preparing to collect samples. Replace outer clothing removed earlier with clean garments (i.e., laboratory coat) that have not been directly exposed to areas of the plant outside of the sampling area.

Sanitize the sample work area surfaces by wiping with a clean disposable paper towel dipped in a freshly prepared 500 ppm (parts per million) sodium hypochlorite solution (0.05% sodium hypochlorite) or other approved sanitizer which provides an equivalent available chlorine concentration. The sample work area surfaces must be free of standing liquid before sample supplies and/or product containers are placed on them.

Before sampling, thoroughly wash and scrub hands to the mid-forearm. Use antibacterial hand soap. If available, this should include a sanitizer at 50 ppm equivalence available chlorine. Dry the hands using disposable paper towels.

T.2.5.2 Specific Sample Collection Procedures:

The establishment's sample collection procedures must be clearly described in writing. The following procedures are those referred to in the USDA Pathogen Reduction and HACCP Systems Final Rule. Establishments which do sampling of E. coli on beef and hogs must use either the excision or carcass sponging technique described in this section. Establishments which do sampling of E. coli on ratites, sheep, goats and equine must use the carcass sponging technique described in this section. Establishments testing for E. coli in chickens, ducks and guineas shall use the carcass wash technique described in this section. Establishments testing for E. coli in turkeys and geese can either use the carcass wash technique or the carcass sponging technique described in this section.

Other sample collection procedures (e.g. Excision of samples instead of swabbing) may be acceptable provided they have been assessed by CFIA and found satisfactory. Establishments wishing to use alternative sampling methods should forward their proposal to the MPD.

Sampling/testing methods used for E. coli testing must achieve a sensitivity of a least 5 cfu/cm².

For methods where USDA has published Process Verification Criteria (see Section T.2.8), the establishment must adhere to the method and use the pre-set Process Verification Criteria. Where USDA Process Verification Criteria have not been published, the establishment should adhere to the procedure and must develop establishment specific Process Verification Criteria (see Section T.2.8.2).

T.2.5.2.1 Carcass Sample Collection Procedures - Cattle (including veal), sheep, goats, horses, mules and other equines.

T.2.5.2.1.1 Surface sponging method

(i) Materials

1. Sterile specimen sponge in sterile wire twist top-type bag or equivalent
2. 25 ml sterile Butterfield's phosphate diluent (BPD)
3. Sterile plastic self-sealing-type or stomacher bag
4. Template for 100 cm² sampling area ( except for sheep and goats where a template of 50 cm² is required)
5. Sterile gloves
6. Wheeled ladder, sampling platform, or step ladder
7. Sanitizing solution
8. Small tote or caddy for carrying supplies

(ii) Collection

Read sections T.2.3 "Pre-Sampling Preparation" and T.2.5.1 "Preparation for Sample Collection" before beginning the sampling procedure. Use the random selection procedures for selecting the half-carcass (ref. section T.2.4).

A sampling sponge (which usually comes dehydrated and prepackaged in a sterile bag) will be used to sample all three sites on the carcass (flank, brisket, and rump--see Figure T.2). It is important to swab the areas in the order of least to most contamination in order to avoid spreading any contamination. Therefore, swab the areas in the sequence indicated in this sampling protocol.

Nondestructive surface sampling will be conducted as follows:

1. Ensure that all bags have been pre-labeled and all supplies are on hand, including the sampling template. (An assistant may be helpful during the sampling process.)
2. IF a reusable template is used, immerse the sampling template in an approved sanitizing solution for at least 1-2 minutes. Just prior to swabbing the first sample site on the carcass (step 13), retrieve the sampling template from the sanitizing solution. Shake excess solution from the utensil, then protect the portion of the template that will contact the carcass from contamination.
3. Locate the flank, brisket, and rump sampling sites using illustrations and directions in Figure T.2 (cattle carcass sampling locations).
4. Position the wheeled ladder, sampling platform, or step ladder near the carcass so the rump sample area (Figure T.2) is within easy reach from the ladder.
5. While holding the sponge bag at the top corner by the wire closure, tear off the clear, perforated strip at the top of the bag.
6. Remove the cap from sterile BPD bottle, being careful not to touch the bottle opening.
7. Carefully pour about half the contents of the sterile BPD bottle (approximately 10 ml) into the sponge bag to moisten the sponge.
8. Close the top of the bag by pressing the wire closures together. Use hand pressure from the outside of the bag and carefully massage the sponge until it is FULLY HYDRATED (moistened).
9. With the bag still closed, carefully push the moistened sponge to the upper portion of the bag orienting one narrow end of the sponge up toward the opening of the bag. Do NOT open the bag or touch the sponge with your fingers. While holding the bag, gently squeeze any excess fluid from the sponge using hand pressure from the outside. The whole sponge should still be in the bag.
10. Open the bag containing the sponge, being careful not to touch the inner surface of the bag with your fingers. The wire closure at the top of the bag should keep the bag open. Set bag aside.
11. Put on a pair of sterile gloves.
12. Carefully remove the moistened sponge from the bag with the thumb and fingers (index and middle) of your sampling hand.
13. With the other hand, retrieve the template by the outer edge, taking care not to contaminate the inner edges of the sampling area of the template.
14. Locate the flank sampling area (Figure T.2). Place the template over this location.
15. Hold the template in place with one gloved hand (Remember, only the sponge should touch the sampling area. Take care not to contaminate this area with your hands)
16. With the other hand, wipe the sponge over the enclosed sampling area(10 cm x 10 cm) for a total of approximately 10 times in the vertical and 10 times in the horizontal directions. The pressure for swabbing would be as if you were removing dried blood from the carcass. However, the pressure should not be too hard as to crumble or destroy the sponge. (Note: The template may need to be "rolled" from side to side during swabbing since the surface of the carcass is not flat. This ensures that the 100 cm² area is enclosed while swabbing.)
17. Repeat steps 14-16 for the brisket area, using the SAME side or surface of the sponge used to swab the flank area.
18. After swabbing the brisket area, transfer the template to the same hand holding the sponge. Do not contaminate the sponge or inner edges of the sampling area of the template.
19. Climb the ladder or platform, holding onto the handrail with the hand used to hold the template. Once at a convenient and safe height for sampling the rump, transfer template back to "climbing" hand (hand used to hold onto the rail while climbing the ladder), taking care not to contaminate the inner edges of the template.
20. Repeat steps 14-16 for the rump area, using the "clean" surface or side (the side that was NOT previously used to swab the flank/brisket areas) of the sponge.
21. After swabbing the rump area, carefully place the sponge back in the sponge sample bag, taking care not to touch the sponge to the outside of the sample bag.
22. While holding the handrail, climb down from the ladder.
23. Add the additional BPD (about 15 ml) to the sample bag to bring the total volume to approximately 25 ml.
24. Expel excess air from the bag containing the sponge and fold down the top edge of the bag 3 or 4 times to close. Secure the bag by folding the attached wire tie back against the bag. Place closed sponge bag into second bag and close the second bag securely.
25. (a) If samples are to be analyzed at an ON-SITE LABORATORY, begin sample preparation (T.2.7 "ANALYTICAL METHODS")
    (b) If samples are to be analyzed at an OUTSIDE (OFF-SITE) LABORATORY, follow procedure in section T.2.6 "Sample Shipment".
26. Results must be analyzed using establishment specific statistical process control - the moving window and pre-set ("m") and ("M") criteria CANNOT BE USED (see section T.2.8.2.).

T.2.5.2.1.2 Excision method - Cattle (including veal)

This procedure is available upon request from the MPD. When this method is used, the Process Verification Criteria found in section T.2.8 must be utilized.

T.2.5.2.2 Carcass sample collection procedure - Swine

T.2.5.2.2.1 Carcass sponging method - Swine

(i) Materials

1. Sterile specimen sponge in sterile wire twist top type-bag or equivalent
2. 25 ml sterile Butterfield's phosphate diluent (BPD)
3. Sterile plastic self-sealing-type bag or stomacher-type bag
4. Template for a 100 cm² sampling area
5. Sterile gloves
6. Wheeled ladder, sampling platform, or step ladder
7. Sanitizing solution
8. Small tote or caddy for carrying supplies

(ii) Collection

Read sections T.2.3 "Pre-Sampling Preparation" and T.2.5.1 "Preparation for Sample Collection" before beginning the sampling procedure. Use the random selection procedures for selecting the half-carcass (ref. section T.2.4).

A sampling sponge (which usually comes dehydrated and prepackaged in a sterile bag) will be used to sample all three sites on the swine carcass (belly, ham, and jowl-see Figure T.3). It is important to swab the areas in the order of least to most contamination in order to avoid spreading any contamination. Therefore, swab the areas in the sequence indicated in this sampling protocol. Nondestructive surface sampling will be conducted as follows:

1. Ensure that all supplies are on hand. (An assistant may be helpful during the sampling process.)
2. If a reusable template is used, immerse the sampling template in a sanitizing solution for at least 1-2 minutes. Just prior to swabbing the first sample site on the swine carcass (step 12), retrieve the sampling template from the sanitizing solution. Shake excess solution from the utensil, then protect the portion of the template that will contact the carcass from contamination.
3. Locate the belly, ham, and jowl sampling sites using illustrations and directions in Figure T.3 (swine carcass sampling locations).
4. Position the wheeled ladder, sampling platform, or step ladder near the carcass so the ham sample area (Figure T.3) is within easy reach from the ladder.
5. Hold the sponge bag at the top corner by the wire closure, then tear off the clear perforated strip at the top of the bag. Open the bag.
6. Remove the cap from sterile BPD bottle, being careful not to touch the bottle opening. Do not contaminate the lid.
7. Carefully pour about half of the contents of the sterile BPD bottle (10 ml) into the sponge bag to moisten the sponge. Put the lid back on the BPD bottle.
8. Close the top of the bag by pressing the wire closures together. Use hand pressure from the outside of the bag and carefully massage the sponge until it is FULLY HYDRATED (moistened).
9. With the bag still closed, carefully push the moistened sponge to the upper portion of the bag orienting one narrow end of the sponge up toward the opening of the bag. Do NOT open the bag or touch the sponge with your fingers. While holding the bag, gently squeeze any excess fluid from the sponge using hand pressure from outside. The whole sponge should still be inside the bag.
10. Open the bag containing the sponge, being careful not to touch the inner surface of the bag with your fingers. The wire closure at the top of the bag should keep the bag open.
11. Put on a pair of sterile gloves.
12. Carefully remove the moistened sponge from the bag with the thumb and fingers (index and middle) of your sampling hand.
13. With the other hand, retrieve the template by the outer edge, taking care not to contaminate the inner edges of the sampling area of the template.
14. Locate the belly sampling area (Figure T.3). Place the template over this location.
15. Hold the template in place with one gloved hand. Remember, only the sponge should touch the sampling area. Take care not to contaminate this area with your hands.
16. With the other hand, wipe the sponge over the enclosed sampling area (10 cm x 10 cm) for a total of approximately 10 times in the vertical and 10 times in the horizontal directions. The pressure for swabbing would be as if you were removing dried blood from the carcass. However, the pressure should not be too hard as to crumble or destroy the sponge.
    Note: The template may need to be "rolled" from side to side during swabbing since the surface of the carcass is not flat. This ensures that the 100 cm² area is enclosed while swabbing.
17. After swabbing the belly area, transfer the template to the same hand that is holding the sponge. Do not contaminate the sponge or the inner edges of the sampling area of the template.
18. Climb the ladder or platform, holding onto the handrail with the hand used to hold the sampling template in place. Once at a convenient and safe height for sampling the ham, transfer template back to the "climbing" hand (hand used to hold onto the rail while climbing the ladder), taking care not to contaminate the sponge or the inner edges of the template.
19. Repeat steps 14-16 for the ham sampling area, using the SAME surface of the sponge used to swab the belly area.
20. After swabbing the ham area, carefully place the template back to the same hand that is holding the sponge. Do not contaminate the sponge or the inner edges of the sampling area of the template.
21. While holding the handrail, climb down from the ladder.
22. Transfer the template back to the "climbing" hand (hand used to hold onto the rail while descending the ladder), taking care not to contaminate the sponge or the inner edges of the template.
23. Repeat steps 14-16 for the jowl area, using the "clean" surface or side (the side that was not previously used to swab the belly/ham areas).
24. After swabbing the jowl area, carefully place the sponge back into the sponge bag. Do not touch the surface of the sponge to the outside of the sponge bag.
25. Add the additional BPD (about 15 ml) to the bag to bring the total volume to approximately 25 ml.
26. Press wire closures of the sponge bag together, expel excess air, then fold down the top edge of the bag 3 or 4 times. Secure the bag by folding the attached wire tie back against the bag. Place the closed sponge bag into the second bag and close the second bag securely.
27. (a) If samples are to be analyzed at an ON-SITE LABORATORY, begin sample preparation (T.2.7 "ANALYTICAL METHODS")
    (b) If samples are to be analyzed at an OUTSIDE (OFF-SITE) LABORATORY, follow procedure in section T.2.6 "Sample Shipment".
28. Results must be analyzed using establishment specific statistical process control - the moving window and pre-set ("m") and ("M") criteria CANNOT BE USED (see section T.2.8.2.).

T.2.5.2.2.2 Carcass Excision Method - Swine

This procedure is available upon request from the Meat Programs Division. When this method is used, the Process Verification Criteria found in section T.2.8 must be utilized.

T.2.5.2.3 Chicken, Ducks, Guineas and Squabs Carcass Rinse Sampling Procedure

Note: There is no approved excision or carcass swabbing method for chicken carcasses.

i) Materials

1. 2 Sterile 3500 millilitre (ml) stomacher-type or self-sealing type bags or equivalent. (The bag must be sterile and should be large enough to hold the carcass while rinsing.)
2. 400 ml sterile, Butterfield's phosphate diluent (BPD).
3. Plastic tie wraps or equivalent (if needed to secure the bag).
4. Sterile gloves.
5. Optional--(See alternate sampling--step 10)--Sterile leak-proof container.

(ii) Collection

Read sections T.2.3 "Pre-Sampling Preparation" and T.2.5.1 "Preparation for Sample Collection" before beginning the sampling procedure. Use the random selection procedures for selecting the carcass (ref. section T.2.4). Handle the selected carcass in the following manner:

1.  Ensure all sampling supplies are present and have been properly labeled. An assistant may be helpful during sampling.
2. Open a large stomacher-type bag without touching the sterile interior of the bag. (Rubbing the top edges of the bag between the thumb and forefinger will cause the opening to gap for easy opening.)
3. Put on sterile gloves.
4. With one hand, push up through the bottom of the sampling bag to form a "glove" over one hand with which to grab the bird, while using your other hand to pull the bag back over the hand that will grab the bird. This should be done aseptically without touching the exposed interior of the bag.
5. Using the hand with the bag reversed over it, pick up the bird by the legs (hocks) through the stomacher bag. (The bag functions as a 'glove' for grabbing the bird's legs.) Take care not to contaminate the exposed interior surface of the bag. Allow any excess fluid to drain before reversing the bag back over the bird. (Alternately, have an assistant hold open the bag. Using your gloved hand, pick up the bird by the legs, allow any fluid to drain, and place the bird in the sampling bag.)
6. Rest the bottom of the bag on a flat surface. While still holding the top of the bag slightly open, add the sterile BPD (400 ml) to the bag containing the carcass, pouring the solution over the carcass. (Alternately, with the aid of an assistant holding the bag open, add the sterile BPD (400 ml) to the bag containing the carcass, pouring the solution over the carcass.)
7. Expel most of the air from the bag, then close the top of the bag. While securely holding the bag, rinse the bird inside and out using a rocking motion for 30 shakes (approximately one minute). This is done by holding the bird through the bottom of the bag with one hand and the closed top of the bag with the other hand. Hold the bird securely and rock it in an arcing motion, alternating the weight of the bird from one hand to the other (motion like drawing an invisible rainbow or arch), assuring that all surfaces (interior and exterior of the carcass) are rinsed.
8. Rest the bag with the bird on a flat surface and, while still supporting the bird, open the bag.
9. With a gloved hand, remove the carcass from the bag. Since the carcass was rinsed with a sterile solution, it can be returned to the chill tank. Be sure not to touch the interior of the bag with your gloved hand.
10. Secure the top of the bag so that the rinse fluid will not spill out or become contaminated. (Alternately, at least 30 millilitres of rinse fluid can be poured into a sterile leak-proof container to be sent to the lab for analysis.)
11. Place the sample bag (or leak-proof container) into another bag and secure the opening of the outer bag.
12. (a) If samples are to be analyzed at an ON-SITE LABORATORY, begin sample preparation for the selected method of analysis.
    (b) If samples are to be analyzed at an OUTSIDE (OFF-SITE) LABORATORY, follow the procedure in section T.2.6 "Sample Shipment".
13. Results must be analyzed using the moving window and pre-set ("m") and ("M") (see section T.2.8.2.).

T.2.5.2.4 Turkey, geese and ratites carcass sample collection procedures:

T.2.5.2.4.1 Carcass Rinse Sampling Procedure - Turkey and Geese

(i) Materials

1. 2 Sterile 3500 ml stomacher-type or self-sealing type-bags or equivalent. (The bag must be sterile and should be large enough to hold the carcass while rinsing, the bags USDA used for their sampling program measure approximately 18" x 24". Large turkeys should be placed in a plain, clear polypropylene autoclave bag , about 24" x 30" to 36").
2. 600 ml sterile, Butterfield's phosphate diluent (BPD)
3. Plastic tie wraps or thick rubber bands or equivalent, if needed to secure sample bag
4. Sterile gloves
5. Optional--sterile, leak-proof container (see step 12 Alternate procedure)

(ii) Collection

Read sections T.2.3 "Pre-Sampling Preparation" and T.2.5.1 "Preparation for Sample Collection" before beginning the sampling procedure. Use the random selection procedures for selecting the half-carcass (ref. section T.2.4). Handle the selected carcass in the following manner:

1.  Ensure that all supplies are on hand and readily available. An assistant will be needed to hold the bag for collecting the bird.
2. Have an assistant open the large sterile stomacher-type bag (designated for rinsing the carcass) and be ready to receive the turkey carcass. (Rubbing the top edges of the bag between the thumb and index finger will cause the opening to gap open). Alternately: If no assistant is available, place the closed large sampling bag into a bucket or pail (e.g., use the bag to "line" a bucket like a trash-can liner), then open the bag. The bucket will be used as a holder or stand to support the bag. Do not contaminate the inner surfaces of the sampling bag.
3. Put on sterile gloves.
4. Remove the selected turkey from the drip line by grasping it by the legs and allowing any fluid to drain from the cavity.
5. Place the turkey carcass, vent side up, into a sterile sampling bag. Only the carcass should come in contact with the inside of the bag.
6. Manipulate the loose neck skin on the carcass through the bag and position it over the neck bone area to act as a cushion and prevent puncturing of the bag. The assistant will need to support the carcass with one hand on the bottom of the bag.
7. While still supporting the bottom of the bag, have the assistant open the bag with the other hand. Alternately, rest the bottom of the bag on a pre-sanitized surface (i.e., a table), and while still supporting the carcass in the bag, open the bag with the other hand.
8. Add the sterile BPD (600 ml) to the bag containing the carcass, pouring the diluent over the carcass.
9. Take the bag from the assistant and expel excess air from the bag and close the top. While securely holding the bag, rinse the bird inside and out using a rocking motion for 30 shakes (approximately one minute). This is done by holding the carcass through the bag with one hand and the closed top of the bag with the other hand. Holding the bird securely with both hands, rock in an arcing motion alternating the weight of the bird from one hand to the other (motion like drawing an invisible rainbow or arch), assuring that all surfaces (interior and exterior of the carcass) are rinsed.
10. Hand the bag back to the assistant.
11. With a gloved hand, remove the carcass from the bag letting excess fluid drain back into the bag. Since the carcass was rinsed with a sterile solution, it can be returned to the chill tank. Be sure not to touch the interior of the bag with your gloved hand.
12. Expel excess air, taking care not to expel any rinse fluid. Secure the top of the bag so that the rinse fluid will not spill out or become contaminated. (Alternately, at least 30 millilitres of rinse fluid can be poured into a sterile, leak-proof container and sent to the lab for analysis.)
13. Place the sample bag (or container) into another bag and secure the opening of the outer bag.
14. (a) If samples are to be analyzed at an ON-SITE LABORATORY, begin sample preparation (T.2.7 "ANALYTICAL METHODS")
    (b) If samples are to be analyzed at an OUTSIDE (OFF-SITE) LABORATORY, follow procedure in section T.2.6 "Sample Shipment".
15. Results must be analyzed using establishment specific statistical process control - the moving window and pre-set ("m") and ("M") criteria CANNOT BE USED (see section T.2.8.2.).

T.2.5.2.4.2 Carcass Sponging Sampling Procedure - Turkeys, Geese and Ratites

(i) Materials

1. Sterile specimen sponge in sterile wire twist top type-bag or equivalent
2. 25 ml sterile Butterfield's phosphate diluent (BPD)
3. Sterile plastic self-sealing-type bag or stomacher-type bag
4. 5 x 10 cm sampling area template
5. 5. Sterile gloves
6. Sanitizing solution
7. Small tote or caddy for carrying supplies

(ii) Collection

Read sections T.2.3 "Pre-Sampling Preparation" and T.2.5.1 "Preparation for Sample Collection" before beginning the sampling procedure. Use the random selection procedures for selecting the carcass (ref. section T.2.4).

A sampling sponge (which usually comes dehydrated and prepackaged in a sterile bag) will be used to sample two sites on the turkey carcass (thigh and back - see figure T.4.) It is important to swab the areas in the order of least to most contamination in order to avoid spreading any contamination (i.e. back then thigh). Therefore, swab the areas in the sequence indicated in this sampling protocol. Nondestructive surface sampling will be conducted as follows:

1. Ensure that all supplies are on hand. (An assistant may be helpful during the sampling process.)
2. If a reusable template is used, immerse the sampling template in a sanitizing solution for at least 1-2 minutes. Just prior to swabbing the first sample site on the carcass (step 12), retrieve the sampling template from the sanitizing solution. Shake all excess solution from the utensil, then protect the portion of the template that will contact the carcass from contamination.
3. Place the carcass to be swabbed on a work surface which has been cleaned, disinfected and dried. To prevent the carcass from sliding, place paper towels or other suitable material between the carcass and the work surface. Place the carcass on its breast to allow access to both sampling sites. It is ok if the carcass leans to one side, provided the sampling sites are not contaminated.
4. Locate the back and thigh sampling sites using illustrations and directions in Figure T.4 (turkey carcass sampling locations).
5. Hold the sponge bag at the top corner by the wire closure, then tear off the clear perforated strip at the top of the bag. Open the bag.
6. Remove the cap from sterile BPD bottle, being careful not to touch the bottle opening. Do not contaminate the lid.
7. Carefully pour about half of the contents of the sterile BPD bottle (10 ml) into the sponge bag to moisten the sponge. Put the lid back on the BPD bottle.
8. Close the top of the bag by pressing the wire closures together. Use hand pressure from the outside of the bag and carefully massage the sponge until it is FULLY HYDRATED (moistened).
9. With the bag still closed, carefully push the moistened sponge to the upper portion of the bag orienting one narrow end of the sponge up toward the opening of the bag. Do NOT open the bag or touch the sponge with your fingers. While holding the bag, gently squeeze any excess fluid from the sponge using hand pressure from outside. The whole sponge should still be inside the bag.
10. Open the bag containing the sponge, being careful not to touch the inner surface of the bag with your fingers. The wire closure at the top of the bag should keep the bag open.
11. Put on a pair of sterile gloves.
12. Carefully remove the moistened sponge from the bag with the thumb and fingers (index and middle) of your sampling hand.
13. With the other hand, retrieve the template by the outer edge, taking care not to contaminate the inner edges of the sampling area of the template.
14. Locate the back swabbing area (Figure T.4). Place the template over this location. The template should straddle the backbone with the long axis of the template parallel to the spine.
15. Hold the template in place with one gloved hand. Remember, only the sponge should touch the sampling area. Take care not to contaminate this area with your hands.
16. With the other hand, wipe the sponge over the enclosed sampling area (5 cm x 10 cm) for a total of approximately 10 times in the vertical and 10 times in the horizontal directions. The pressure for swabbing would be as if you were removing dried blood from the carcass. However, the pressure should not be too hard as to crumble or destroy the sponge.
    Note: The template may need to be "rolled" from side to side during swabbing since the surface of the carcass is not flat. This ensures that the 50 cm² area is enclosed while swabbing.
17. After swabbing the back area, transfer the template to the same hand that is holding the sponge. Do not contaminate the sponge or the inner edges of the sampling area of the template.
18. Repeat steps 14-16 for the thigh sampling area, using the surface of the sponge WHICH WAS NOT USED to swab the back area. The template needs to be locates in the following manner: the long axis of the template is placed parallel to the thigh bone and the top of the template is just below the hip joint.
19. After swabbing the thigh area, carefully place the sponge back into the sponge bag. Do not touch the surface of the sponge to the outside of the sponge bag.
20. Add the additional BPD (about 15 ml) to the bag to bring the total volume to approximately 25 ml.
21. Press wire closures of the sponge bag together, expel excess air, then fold down the top edge of the bag 3 or 4 times. Secure the bag by folding the attached wire tie back against the bag. Place the closed sponge bag into the second bag and close the second bag securely.
22. (a) If samples are to be analyzed at an ON-SITE LABORATORY, begin sample preparation (T.2.7 "ANALYTICAL METHODS")
    (b) If samples are to be analyzed at an OUTSIDE (OFF-SITE) LABORATORY, follow procedure in section T.2.6 "Sample Shipment".
23. Results must be analyzed using establishment specific statistical process control - the moving window and pre-set ("m") and ("M") criteria CANNOT BE USED (see section T.2.8.2.).

T.2.6 Sample Shipment Procedures:

To obtain the most accurate results, samples should be analyzed as soon as possible after collection. However, if samples must be sent off-site, they need to be maintained at refrigeration temperatures until transport and shipped refrigerated to the laboratory performing the analysis. The sample must be analyzed no later than the day after collection. The following section gives information on shipping containers and transporting samples to off-site facilities.

T.2.6.1 Shipping Containers and Coolant Packs:

It is important that samples fit easily into the shipping containers so that the sample bags do not break. Correct use of refrigerant gel-ice packs and proper packing of the shipping container are necessary to ensure that samples arrive at the laboratory at an acceptable temperature. Some bacteria may be damaged by temperatures that are too cold, while temperatures that are too warm can allow bacteria to reproduce. Maintaining samples at improper temperatures may cause inaccurate sample results. Frozen samples or samples which are too warm (>10°C) are not considered valid and must not be analyzed.

The sample should be kept refrigerated, NOT FROZEN, in the shipping container prior to pickup by the courier service. The shipping container, itself, should not be used as a refrigerator. However, multiple samples (if needed) for one day may be stored in an open shipping container kept in the cooler or refrigerator.

The following packaging procedure can be used:

1. Prechill shipping container by placing the open shipping container in the refrigerator at least the day before sampling.
2. Place the appropriately-labeled, double-bagged sample(s) in the prechilled shipping container in an upright position to prevent spillage. Newspaper may be used for cushioning the sample and holding it in the upright position. If more than one sample is collected during the day, take steps to ensure that samples are maintained at refrigeration temperature. Refrigeration temperatures help limit multiplication of any microorganisms present which ensures the most accurate results.
3. Place a corrugated cardboard pad on top of samples. This corrugated cardboard pad prevents direct contact of frozen gel packs with the samples. Next place the frozen gel pack(s) on top of the corrugated pad. Use sufficient frozen coolant to keep the sample refrigerated during shipment to the designated laboratory. Insert foam plug and press it down to minimize shipper head space.
4. Ship samples (via overnight delivery or courier) to the assigned laboratory.

When shipping samples, the method used to protect against sample tampering shall be described in the written sampling protocol.

T.2.7 Analytical Methods

The laboratory shall not analyze samples which are frozen or too warm (e.g., > 10°C). Samples must be analyzed using one of the E. coli (Biotype I) quantitation methods found in the Official Methods of Analysis of the Association of Official Analytical Chemists (AOAC), International, 16th edition⁴, or by any method which is validated by a scientific body in collaborative trials against the three tube Most Probable Number (MPN) method and agreeing with the 95% upper and lower confidence limits of the appropriate MPN index.

T.2.7.1 Suggested Quantitation Schemes

If a generic 1 ml plating technique is used for E. coli quantitation for cattle or swine carcass sponging sample analysis, the plate count needs to be divided by 12 to equal the count per cm² of carcass surface area, according to the following calculations:

Total surface swabbed: 3 x 100 cm² = 300 cm²

Total quantity (volume) of fluid: 10 + 15 ml = 25 ml

Conversion factor calculation:

if 25 ml for 300 cm²

then 1 ml for x cm²

and x = 300 ÷ 25 = 12

Where a generic 1 ml plating technique is used for E. coli quantitation for turkey carcasses, the plate count needs to be divided by 4 to obtain the count per cm² (i.e.: 100 ÷ 25 = 4 ).

Similar conversions are not necessary for quantitation of E. coli/ml of poultry rinse fluid samples. To cover the marginal and unacceptable range for E. coli levels (described in section T.2.8.1), the undiluted sample extract, as well as 1:10, 1:100, 1:1,000 and 1:10,000 dilutions should be plated, preferably in duplicate. Higher or lower dilutions may need to be plated based on the specific product.

If a hydrophobic grid membrane filtration method were used, the only difference is that filtrations shall be performed on 1 ml of the undiluted sample extract, as well as of 1:10, 1:100, 1:1,000 and 1:10,000 dilutions.

Additional dilutions of the original extract may need to be used if a three tube Most Probable Number (MPN) protocol is used. The three highest dilutions that were positive for E. coli are to be used to calculate the MPN. For cattle and swine samples collected by swabbing technique, MPN values from the appropriate MPN Table represent the count per ml of original extract and therefore needs to be divided by 12 to obtain the count per cm² of carcass surface area (divide by 4 for turkey samples collected by swabbing.)

Laboratories shall report the exact E. coli count. Where values are quantified below 1 cfu/cm², the exact count should be reported (e.g. 0,01 cfu/cm²) - zero (0) shall not be reported as a test result. Methods used must have a minimum sensitivity of at least 5 cfu/cm² carcass surface area for cattle and swine.

T.2.8 Analysis of Results and Statistical Process Controls

The goal of generic E. coli (biotype I) testing is to assess the effectiveness of a process through microbiological data and statistical process control. The objective is to meet the specific Process Verification Criteria (PVC) developed for the specified combination of process sampled and sample collection method used at the establishment.

The USDA published PVC values for specific combinations of processes sampled and sample collection methods. These are found in table T.2.8. If the establishment is using one of the listed combinations, they must use the PVC provided in the table. If they are not using one of the listed combinations, they MUST develop their own specific PVC criteria using Statistical Process Control (SPC) techniques and cannot use the PVC criteria developed for a different combination.

Note:

• Establishments testing chickens are obliged to use the carcass rinse sampling method and the published PVC.
• Beef and hog slaughter establishments are not required to sample carcasses by excision and the corresponding PVC. The sponging methods may be used.
• There are no PVC published for sampling by sponging sampling method and turkeys, ducks, geese and guineas carcasses rinse sampling methods. Establishments using these methods will have to develop their own specific PVC criteria.

Table T.2.8

Process Verification Criteria for the assessment of generic E. coli (biotype I) results

Combination process sampled x sampling method		Process Verification Criteria (PVC) to use
Process sampled	Sampling method	Size of moving window ("n")	Max. # of marginal results ("c")	Acceptable results	Marginal limit ("m")	Unacceptable limit ("M")
Cattle, sheep, goats and equine slaughter	excision	13	3	Negativea	Positivea	10,000 cfu/cm2
	sponge	No published PVC - the operator must develop their own PVC using Statistical Process Control (SPC) techniques.
Swine slaughter	excision	13	3	10 cfu/cm2 or less	10 cfu/cm2	10,000 cfu/cm2
	sponge	No published PVC - the operator must develop their own PVC using SPC techniques.
Chicken, ducks, guineas and squabs slaughter	carcassb rinse	13	3	100 cfu/ml or less	100 cfu/ml	1,000 cfu/ml
	sponge	Not applicable - the operator must use the carcass rinse method.
Turkey and geese slaughter	carcass rinse	No published PVC - the operator must develop their own PVC using SPC techniques.
	sponge	No published PVC - the operator must develop their own PVC using SPC techniques.
Ratite slaughter	carcass rinse	Not applicable - the operator must use the sponge method.
	sponge	No published PVC - the operator must develop their own PVC using SPC techniques.

Notes:

a: The test used for analyzing samples must have a minimum sensitivity of 5 cfu/cm².

b: PVC applicable only to chicken

Statistical Process Control techniques available

The use of statistical process control (SPC) techniques is based on the following principles :

• every product is manufactured through a process
• every process is subject to variation
• every process can be understood, described and/or measured in mathematical or statistical terms
• the ongoing performance of the process can be assessed if regular measurements are made and improvements to the process can be made until an optimum level is reached.

According toSPC principles, a process is said to be "in control" when it is stable and performs in an acceptable manner i.e. test results range close to average and are within predicted limits. SPC techniques allow the detection of process variations which would not have been identified otherwise. Through early detection of unexpected variations, deficiencies can be readily detected and corrected.

The development of SPC programs involve an initial collection of data to determine baseline parameters characterizing the process of interest (also called baseline data; baseline level; process control level). These parameters describe what the typical or "normal" process looks like. Using a statistical method, the baseline data is used to develop Process Verification Criteria (PVC). Many different statistical approaches can be used for SPC (for example, the moving window approach, Shewhart Charts, trends analysis, standard deviations, time plots, CUSUM, control charts, etc.) Once the SPC approach has been selected and the PVC developed, regular monitoring is performed and results checked against the PVC to verify if the process is operating within "normal" limits. If not, this indicates that the process is not "in control" and it is time to find and correct the possible causes for the abnormal results.

This section only discusses the basic principles of the moving window approach which the USDA used to develop published PVC for assessing generic E. coli biotype I in slaughter processes. The reader should refer to one of the following specialized references for more information on other SPC methods:

• Ishikawa, K. 1986 - Guide to Quality control. Kraus International Publications, White Plains, NY
• Kane, V.E. 1989 - Defect Prevention. ASQ Quality Press, Milwaulkee WI
• Kume, H. 1985 - Statistical Mehtods for Quality Improvement. UNIPUB/Kraus International publications, White Plains, NY
• Surak, J.G. 1999 - SPC for the food processing Industry. Clemson University, Clemson, SC
• Surak, J.G. 1999 - Integrating HACCP and SPC. Clemson University, Clemson, SC.

Note: Many of these documents were developed for industrial applications other than food and examples used may not lend themselves to food production.

T.2.8.1.2 The moving window approach used by the USDA

The USDA used a statistical process control technique referred to as the "moving window approach" when they developed the PVC listed in table T.2.8. The following paragraphs describe in layman's terms how the "moving window approach" used by the USDA was designed and how it can be used to assess E. coli testing results.

Establishment of "M", "m", "n" and "c"

• Data is collected in order to obtain a baseline for the process.
• The data is ranked from the lowest result to the highest and converted into percentiles i.e. divided into 100 classes with equal size (containing an equal number of units/individuals), for example,

  the 1^st percentile = 1^st class of results

  the 50^th percentile = 50^th class of results (within that class, the median is the value for which 50% of the results are smaller or larger)

  the 100^th percentile = 100^th class of results

• The size of the moving window ("n" value") is established and the control limits are defined. The moving window approach used by the USDA sets the size of the moving window at "n"= 13 and uses two control limits, an Absolute Control Limit ("M" value) and a Lower Control Limit ("m" value).
• The Absolute control limit is the maximum value beyond which any single result means that the process is no longer "in control" and corrective action is to take place. The moving window approach used by the USDA sets the "M" value at the "M"=98^th percentile of results.
• The Lower control limit ("m" value) is an intermediate value somewhere between the average and "M". Test results which are between "m" and "M" are normal and the process is still "in control" when this happens. However, if the number of these test results exceeds a pre-determined number ("c" value), this trend indicates that the process is no longer "in control". Different combinations of "m" and "c" values may be used when evaluating a process. The selection of the combination to use is based on statistical factors (for e.g., confidence level sought, size of population sampled, prevalence of defect, etc.) and consultation with a statistician usually takes place when the sampling plan is designed. The moving window approach used by the USDA sets these values at "c"= 3 and "m"= 80^th percentile of results.

Use of the moving window after "m" and "M" values have been identified:

• The operator conducts testing as per requirements.
• Results are recorded as they become available either in a log book or on a control diagram.
• The results are analyzed on an ongoing basis, i.e. each time a new result is received and recorded.
• When analyzing the results, only the results which appear in the window (i.e., last "n" number of results, in this case the last 13 results) are used to assess the process.
• If the "M" value is exceeded at any time within the window (last 13 results), the process is said to be not "in control". Immediate corrective actions are required as soon as the "M" value is exceeded.
• If the "m" value is exceeded more than 3 times ("c" value) within the window (last 13 results), the process is said to be not "in control". Immediate corrective actions are required as soon as the "m" value is exceeded more than 3 times..
• Each time a new result is received, the window moves forward by 1 result and the results are analyzed again on the basis of the 13 most recent results. This is why the method is called the "moving window approach".
• Periodically, as the process improves, the 80^th and 98^th percentile can be recalculated and new "m" and "M" values used.

Example:

The following example corresponds to records of results for a chicken slaughter plant required to do two tests per day (using a carcass wash technique). As per table T.2.8 the "M" and "m" values for E. coli (biotype I) are set to 1000 and 100 cfu/ml respectively, c to 3 times and n to 13 results.

Test #	Date	Time collected	Test result (cfu/ml)	Is the result above the "unacceptable" limit? > ("M")	Is the result above the "marginal" limit? > ("m")	# of results exceeding the "marginal" limit in last 13 tests	Is the process "under control"? (# of results > "m") # 3 AND no result >"M"
1	10-07(Mon)	08:50	200	No	Yes	1	Yes
2	............	14:00	50	No	No	1	Yes
3	10-08 (Tue)	07:10	500	No	Yes	2	Yes
4	............	13:00	5	No	No	2	Yes
5	10-09(Wed)	10:00	80	No	No	2	Yes
6	............	12:20	50	No	No	2	Yes
7	10-10 (Thu)	09:20	800	No	Yes	3	Yes
8	............	13:30	50	No	No	3	Yes
9	10-11 (Fri)	10:50	80	No	No	3	Yes
10	............	14:50	50	No	No	3	Yes
11	10-14(Mon)	08:40	500	No	Yes	4	No
12	............	12:00	80	No	No	4	No
13	10-15 (Tue)	09:30	80	No	No	4	No
14	............	15:20	50	No	No	3	Yes
15	10-16(Wed)	07:30	80	No	No	3	Yes
16	............	11:40	5	No	No	2	Yes
17	10-17 (Thu)	10:20	80	No	No	2	Yes
18	...........	14.45	80	No	No	2	Yes
19	10-18 (Fri)	08:30	50	No	No	2	Yes
20	...........	16:00	80	No	No	2	Yes
21	10-19 (Sat)	09:10	50	No	No	1	Yes
22	..........	13:00	1500	Yes	Yes	2	No
23	10-21(Mon)	10:30	80	No	No	2	Yes

The following observations can be made upon studying this example:

1. As of October 14 (10-14) at 08:40, there are four results in the last 11 which exceed the "marginal" limit ("m"); this exceeds the maximum of tests above "m" permitted ("c" = 3) within the window ("n" = 13). The process can no longer be said to be "in control".
2. The limit of 3 results above "m" in 13 consecutive samples stays exceeded for the next two tests. The process is still considered not to be "in control", but since no new result above the "marginal" or "unacceptable" limits have occurred, the conclusion should not be considered as evidence of a new problem. The log or documentation of corrective action taken following the initial loss of control will indicate if action was taken to address the identified deviation.
3. On 10-15 at 15:20 the number of results above the "marginal" limit (>"m") in the last 13 tests goes down to 3 because the original result above "m" (10-07 at 08:50) is no longer in the window after it moves ahead by one. The process can again be said to be "in control".
4. The result for 10-19 at 13:00 exceeds the "unacceptable" limit (result > "M"). Even if the number of results above the marginal limit ("m") is less than the maximum allowed ("c" = 3), the process is still considered to no longer be "in control". Corrective action must be undertaken and logged by the operator because of this unacceptable result.
5. The result for 10-21 at 10:30 is below the "marginal" limit (<"m"). The total of marginal results is 2. The process is said to be "in control".

Figure T.5 shows the same results as the above example but the results are displayed in chart form. The numbers along the horizontal axis of the graph (x-axis), refers to the test number in the chart above. The information for each test result, such as the time and date the sample was collected could also be recorded on the chart.

T.2.8.2 Establishments with either a process and/or E. coli sampling method which does not correspond to published USDA Process Verification Criteria listed in table T.2.8

An establishment which does generic E. coli (biotype I) testing for a process and which is and/or uses a sampling method for which there is as yet no USDA published Process Verification Criteria (PVC) (see table T.2.8), the establishment cannot use the published PVC and must develop establishment-specific PVC using SPC techniques (see Section T.2.8).

There is no requirement to use the moving window approach that was used by USDA to develop the published PVC. The moving window approach as described in section T.2.8.1 may be used, provided that the establishment use its own historical data to develop a baseline and establish PVC. Percentile values can be easily determined through computer spreadsheets, use of calculators with statistical functions and even manually. The establishment must make available to the CFIA upon request the raw data used to calculate their "m"=80th percentile and "M"=98th percentile values. Because sponging methods are less sensitive than excision, the establishment's "m" and "M" values should not be higher than those published for the same process with the excision method.

In many cases it may not be appropriate for the establishment to use the moving window approach with m=80th percentile M=98th percentile. An example of this is when an establishment samples hog carcasses using a swabbing collection method and obtains results consisting mostly of values equal or close to zero. If the establishment used the same moving window approach as USDA, their "m" value could be near the detection level for the test and the "M" value very low (e.g., 10 cfu/cm²).

Unless there are published "m" and "M" values for the combined process x sampling technique, the operator can choose to use a different SPC(for e.g., trend analysis approach, 3 standard deviations, etc.). In such cases, the operator needs to make available to the CFIA information regarding how they developed their PVC. This information should include details information regarding the SPC method which was used (with appropriate references where possible), the baseline data used to define the PVC, etc. Provided the approach used is reasonable, the PVC developed will be acceptable.

T.2.9 Corrective Actions when the Process is not "in control"

It is expected that the establishment will occasionally find that their process is not "in control". This situation does not correspond to an inspection finding of deficiency or a failure to meet generic E. coli (biotype I) testing requirements. It does however mean that something in the establishment's process has caused a variance of test results from normal. The establishment must look for the cause of the variance and implement any required corrective measures. These actions (operator's investigation and measures taken) must be documented in writing. If the measures are effective, subsequent test results should indicate that the process is again "in control"

Situations where test results indicate repeatedly that the process is not "in control" are not usual. This situation indicates that the operator is not successfully identifying and correcting the cause(s) of the variance and is not successfully bringing the process back "in control". However, because PVC are not performance standards, this situation in itself does not warrant the same type of CFIA enforcement response as would repeated failure to meet the Salmonella Performance Standards.

The CFIA I/VIC conducts an investigation at the establishment to verify that the generic E. coli (biotype I) testing is being done according to requirements and that test results are being assessed properly. They also review the corrective measures taken by the operator each time that the process was found to be not "in control". The I/VIC tries to determine what the establishment is doing when the PVC values are exceeded and if this is acceptable. For example:

• If the establishment is consistently exceeding the PVC values but does nothing, this is unacceptable.
• If the establishment is taking action, but the action is always the same and the problem keeps repeating itself, this is unacceptable.
• If the establishment is making a deliberate effort to find the cause behind the higher than normal results, for e.g., is trying different things, but these measures haven't worked, this is conditionally acceptable, i.e., the establishment must continue to work towards a solution which corrects the situation in a definitive manner.

Part 5 of the Basic Compliance Checklist should be completed by the I/VIC at this time. If there is a "No" answer to any question on the checklist, the I/VICis to contact the Program Network Director (PND). In cases where testing requirements are not being met, the procedure outlined in section Q.3 applies.

In cases where the testing requirements are being met but there is a chronic inability to meet PVC, the procedure outlined in section Q.3. DOES NOT apply, but the PND may consider other measures (for e.g., Salmonella testing by the operator, HACCP review of the process by CFIA.)

T.2.10 Record Keeping

Results of each test result must by recorded in terms of colony forming units per square centimetre (cfu/cm²) for sponging or excision methods, and in terms of colony forming units per ml rinse fluid (cfu/ml) for whole bird carcass rinse methods. Results must be reported to the minimum sensitivity of the test used. Zero shall not be reported as a test result. A process control table or chart can be used to record the results and facilitate evaluation.

Results should be recorded as they are collected and need to be logged in the order of sample collection. Records should also include information useful for determining appropriate corrective actions when problems occur. The information needed for each sample must include date and time of sample collection, and, if more than one slaughter line exists, the slaughter line from which the sample was collected.

E. coli testing records are to be maintained at the establishment for twelve months and must be made available to Inspection staff on request.

T.2.11 CFIA Verification Activities

CFIA inspectors will verify that establishments meet E. coli testing requirements through The BCC and Multi-Commodity Activity Program (MCAP) verification.

If the VIC discovers that the establishment is failing to perform required testing and/or to record test results, the provisions outlined in Annex Q, section Q.3 apply.

In principle the FSIS requirements for E. coli testing are that the operator must develop testing procedures, conduct routine testing of carcasses in slaughter operations and record results as provided in this annex. The operator is also responsible to develop and implement an action plan at the satisfaction of the Veterinarian in Charge (VIC) upon receiving results exceeding the established process verification criteria.

If the test results indicate a possible loss of control over the process, the VIC verifies that the establishment has taken or is taking the required actions (see T.2.9). The VIC also ensures that dressing procedures remain in compliance with program requirements and, if required, take further action as appropriate (ex: increased verification at boning room carcass re-inspection station or shipping re-inspection station).

The Process Verification Criteria are not standards, enforcement activities can not be initiated because of failure to meet Process Verification Criteria only. Whenever the VIC finds that there is repeated failure to meet the Process Verification Criteria and that the actions taken by the operator are not effective, they shall contact the Program Network Director to advise the situation and discuss further action. At this point one option would be to conduct an in-dept inspection of the establishment.

T.2.12 Figures And Illustrations

Figure T.1 SWABBING TEMPLATE FOR CATTLE AND SWINE - EXAMPLE

Figure T.2 SAMPLING LOCATIONS FOR SWABBING- CATTLE, SHEEP, GOATS AND EQUINE

Figure T.3 SAMPLING LOCATIONS FOR SWABBING - SWINE

Figure T.4 SAMPLING LOCATIONS FOR SWABBING - TURKEYS AND RATITES

Figure T.5 PROCESS CONTROL CHART - EXAMPLE

* These drawings and descriptions of sampling sites are available from CFIA inspectors.

---

¹ Where an establishment's normal operating week is less than 5 days/week, testing shall begin following the first week that the establishment operates during all of it's scheduled operating days. June 1 was set as the starting date for sampling because the highest number of food borne diseases occur during the 3 month period from June to September.

² The selection of carcasses (or half-carcasses) can be carried out before the carcass enters the cooler. The selected carcass can be tagged and placed in the coolers with other carcasses and directed to the sample taking site after 12 hours of chilling has been completed (where conventional chilling methods are used). Carcasses should not be placed in a special area at the beginning of the chilling process as this may misrepresent actual carcass chilling conditions.

³ Where an establishment has been authorized to slaughter, dress or chill types of livestock or poultry by using non-traditional methods (ex: hot boning of swine and poultry, chilling of split turkey carcasses) and is not able to collect samples in the cooler after 12 hours or is unable to sample a whole bird carcass at the end of the drip line, sample collection may be done after the final wash. For poultry, where it is otherwise impractical or unsafe to select the carcass after the chiller, the carcass may be selected at the end of the evisceration line; (ref. Federal Register, Vol. 42, No. 220, Nov. 14, 1997)

⁴ AOAC has approved the following methods for generic E. coli quantitation in foods:

• 3-tube MPN method - AOAC 17.2.01 - 17.2.02;
• Modified 3-tube MPN method - AOAC 17.3.07 - Substrate Supporting Disc Method (ColiComplete®).
  • Modified 3-tube MPN method - AOAC 17.4.01 - Fluorogenic Assay for Glucuronidase, Lauryl sulfate tryptose broth with added 4-methylumbelliferyl-β-D-glucuronide (MUG) is used in a 3-tube MPN method.
  • Plating Method - AOAC 17.3.04 - Dry Rehydratable Film (Petrifilm E. coli Count Plate) Method (Note: Use in lieu of plating ONLY - Not for use as a carcass direct contact medium: see AOAC method for details.)
• Filtration /Plating Method - AOAC 17.3.09 - Hydrophobic Grid Membrane Filter/MUG (ISO-GRID) Method

---

Next page: Annex U | Previous page: Annex Q