# Performance Standards for the Production of Processed Meat and Poultry Products

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/fr%3A01-4420

## Record

- **Collection:** Federal Register
- **Document type:** Proposed Rule
- **Published:** February 27, 2001
- **Citation:** 66 FR 12590

## Text

DEPARTMENT OF AGRICULTURE
Food Safety and Inspection Service
9 CFR Parts 301, 303, 317, 318, 319, 320, 325, 331, 381, 417, and 430
[Docket No. 97-013P]
RIN No. 0583-AC46
Performance Standards for the Production of Processed Meat and Poultry Products

AGENCY:

Food Safety and Inspection Service, Agriculture.

ACTION:

Proposed rule.

SUMMARY:

The Food Safety and Inspection Service (FSIS) is proposing to amend the Federal meat and poultry inspection regulations by establishing food safety performance standards for all ready-to-eat (RTE) and all partially heat-treated meat and poultry products. The proposed performance standards set forth levels of pathogen reduction and limits on pathogen growth that official meat and poultry establishments must achieve in order to produce unadulterated products, but allow the use of customized, plant-specific processing procedures. The proposed RTE performance standards apply to all RTE meat and poultry products, which can be categorized as follows: Dried products (e.g., beef or poultry jerky); salt-cured products (e.g. country ham); fermented products (e.g., salami and Lebanon bologna); cooked and otherwise processed products (e.g., beef and chicken burritos, corned beef, pastrami, poultry rolls, and turkey franks); and thermally-processed, commercially sterile products (e.g., canned spaghetti with meat balls and canned corned beef hash).

Although FSIS routinely samples and tests some RTE products for the presence of pathogens prior to distribution, there are no specific regulatory pathogen reduction requirements for most of these products. The proposed performance standards will help ensure the safety of these products; give establishments the incentive and flexibility to adopt innovative, science-based food safety processing procedures and controls; and provide objective, measurable standards that can be verified by Agency oversight.

FSIS also is proposing environmental testing requirements intended to reduce the incidence of
Listeria monocytogenes
in RTE meat and poultry products. Specifically, FSIS is proposing to require establishments that produce RTE meat and poultry products to test food contact surfaces for
Listeria spp.
to verify that they are controlling the presence of
L. monocytogenes
within their processing environments. Establishments that have developed and implemented HACCP controls for
L. monocytogenes
would be exempt from these testing requirements.

Finally, FSIS is proposing to eliminate its regulations that require that both RTE and not-ready-to eat pork and products containing pork be treated to destroy
trichina
(
Trichinella spiralis
). These requirements are inconsistent with HACCP and some will be unnecessary if FSIS makes final the proposed performance standards for RTE meat and poultry products.

DATES:

Comments must be received on or before May 29, 2001.

ADDRESSES:

Submit one original and two copies of written comments to FSIS Docket #97-013P, U.S. Department of Agriculture, Food Safety and Inspection Service, Room 102, Cotton Annex, 300 12 St., SW., Washington, DC 20250-3700. All comments submitted in response to this notice will be available for public inspection in the Docket Clerk's Office between 8:30 a.m. and 4:30 p.m., Monday through Friday.

FOR FURTHER INFORMATION CONTACT:

Daniel L. Engeljohn, Ph.D., Director, Regulation Development and Analysis Division, Office of Policy, Program Development, and Evaluation, Food Safety and Inspection Service, U.S. Department of Agriculture (202) 720-5627.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Background

II. RTE Meat and Poultry Products

III. Performance Standards and HACCP

IV. The Proposed Performance Standards

A. Lethality

Compliance with the Proposed Lethality Performance Standards

Derivation of the Proposed Lethality Performance Standards

Selection of the Reference Organisms

Dried Products

Salt-cured Products

Fermented Products

Cooked and otherwise Processed Whole or Comminuted Meat Products

Meat Patties

Cooked and otherwise Processed Whole or Comminuted Poultry Products

B. Stabilization

V.
Listeria monocytogenes

A. Proposed Requirements for Controlling
L. monocytogenes

B. Shelf-life and Labeling

VI. Thermally Processed, Commercially Sterile Products

A. Lethality

B. Commercial Sterility

C. Training

VII. Elimination of
Trichina
Treatment Requirements

VIII. Other Proposed Revisions to the Regulations

IX. Scientific Information and Data Needs

X. Summary of the Proposed Rule

XI. Compliance with Executive Order 12866

XII. Compliance with the Regulatory Flexibility Act

XIII. Executive Order 12988

XIV. Risk Analysis

XV. Additional Public Notification

XVI. Paperwork Requirements

XVII. References

XVIII. Proposed Regulations

Appendix 1

I. Background

Under the Federal Meat Inspection Act (FMIA; 21 U.S.C. 601
et seq.
) and the Poultry Products Inspection Act (PPIA; 21 U.S.C. 451
et seq.
), FSIS issues regulations governing the production of meat and poultry products prepared for distribution in commerce. The regulations, along with FSIS inspection programs, are designed to ensure that meat and poultry products are safe, wholesome, unadulterated, and properly marked, labeled, and packaged. In this document, FSIS is proposing to establish new pathogen reduction regulations for ready-to-eat (RTE) and partially heat-treated meat and poultry products. This proposed action is compelled by recent outbreaks of foodborne illness related to the consumption of adulterated RTE meat and poultry products, as well as the need to provide objective, measurable pathogen reduction standards that can be met by official establishments and compliance with which can be determined through Agency inspection.

II. RTE Meat and Poultry Products

RTE meat and poultry products are products that have been processed so that they may be safely consumed without further preparation by the consumer, i.e., without cooking or application of some other lethality treatment to destroy pathogens. Although many of these products, such as frozen pizzas or country hams, customarily are cooked or otherwise reprocessed by the consumer, they would be safe to eat, if unpalatable, without this further preparation.

RTE meat and poultry products can be either non-shelf-stable or shelf-stable. Non-shelf-stable, RTE products must be refrigerated until consumption to prevent the growth of both pathogenic and spoilage organisms. Shelf-stable products remain ready-to-eat under ordinary temperature and humidity conditions and, if the package integrity is maintained during holding, shipping,

storage, display at retail, and in the home, throughout the manufacturer's shelf-life determination. Throughout the shelf-life, shelf-stable products are safe to eat when unrefrigerated (at temperatures over 50 °F or 10 °C) without additional preparation. Thermally processed, commercially sterile meat and poultry products are packaged in hermetically sealed containers (usually cans) and also remain shelf-stable under unrefrigerated conditions (over 50 °F or 10 °C).

For the purposes of this proposal, FSIS has divided ready-to-eat meat and poultry products into five categories, based on the type of processing they receive: dried products; salt-cured products; fermented products; cooked or otherwise processed whole and comminuted products; and thermally-processed, commercially sterile products. Many of these products can be either shelf-stable or non-shelf-stable.

Examples of RTE Products

Dried Products
Basturma, Pastirma, Basturmi.

Beef Sticks.

Carne Seca.

Dried Beef.

Dry Duck Breast.

Meat/Poultry Jerky.

Salt-Cured Products
Cappicola.

Coppa.

Country Ham.

Dry Cured Duck.

Parma Ham.

Prosciutto, Prosciutti.

Fermented Products
Alessandri (Dry Sausage).

Apenino (Dry Sausage).

Arles or D'Arles (Dry Sausage).

Blockwurst (Semi-Dry Sausage).

Cacciatore/Cacciatora (Dry Sausage).

Cervelat.

Cervelat, Soft.

Chorizo.

Lebanon Bologna.

Pepperoni.

Salami, Soft.

Salami: Genoa, Italian, German.

Summer Sausage.

Thuringer.

Thuringer, Soft.

Cooked or Otherwise Processed Whole or Comminuted Products

Meat

Berliner (Cooked, Smoked Sausage).

Bologna.

Bratwurst, Cooked.

Braunschweiger/Liver Sausage.

Breakfast Link Sausage or Patties.

Brown and Serve Sausage.

Burritos.

Cheese Smokies.

Cheesefurter.

Cheesewurst/Cheddarwurst.

Chili.

Chorizo.

Cooked Beef.

Cooked Ham.

Cooked Pork in BBQ Sauce.

Cotto Salami.

Entrees/Dinners.

Fleischkaese (Cured, Cooked Sausage).

Frankfurters.

Frozen Entrees/Dinners.

Gyros.

Meat Loaf.

Meat Salads.

Meat Soups, Frozen.

Nem-Chua (Cooked, Pickled Ham with Shredded Pork Skin).

Pasta with Meat Sauce.

Pastrami.

Pickled Pigs Feet in Vinegar.

Pickled Sausages/Meat in Vinegar.

Piroshki.

Pork Barbecue.

Pork Sausage Patties.

Ravioli.

Roast Beef.

Roast Pork.

Souse.

Stews.

White Hots.

Wieners.

Poultry (Includes Products Containing any Amount of Poultry).

Chicken Burritos.

Chicken BBQ.

Chicken Bologna.

Chicken Breast.

Chicken Franks.

Cooked Poultry.

Cooked Poultry Rolls.

Corn Chowder with Chicken.

Entrees/Dinners.

Poultry Loaf.

Poultry Patties.

Poultry Rolls.

Poultry Salads.

Poultry Soups, Frozen.

Turkey BBQ.

Turkey Franks.

Thermally-Processed, Commercially Sterile Products
Canned Spaghetti with Meat Balls.

Canned Corned Beef Hash.

Canned Ham.

Canned Chicken Salad.

Canned Soups with Meat or Poultry.

FSIS is proposing to require that the processing of each of these types of products achieve specific levels of pathogen reduction, as well as control over the growth of target pathogens so that they do not exceed specific levels. These levels are the performance standards. Establishments also would be required to maintain these levels of pathogen reduction and pathogen growth in their products, under normal handling conditions, until their products reach the consumer.

FSIS already has established pathogen reduction performance standards specific to certain types of not-shelf-stable, RTE meat and poultry products. On January 6, 1999, FSIS published a final rule in the
Federal Register
(FSIS Docket No. 95-033F; 64 FR 732) that established performance standards for RTE roast beef, corned beef, and cooked beef, all “fully-cooked” RTE poultry products, and partially-cooked meat patty and poultry products. Those standards are consistent with and, in fact, incorporated into the more comprehensive group of standards proposed in this document.

III. Performance Standards and HACCP

Under the regulations in 9 CFR 417, FSIS requires each official meat and poultry establishment to develop and implement a Hazard Analysis and Critical Control Point (HACCP) system, a science-based process control system for food safety that promotes systematic prevention of biological, chemical, and physical hazards. Establishments are responsible for developing and implementing HACCP plans that incorporate the controls necessary and appropriate to produce safe meat and poultry products. HACCP is a flexible system that enables establishments to tailor their control systems to the needs of their particular plants and processes. Performance standards can be usefully and seamlessly incorporated into HACCP systems.

When developing a HACCP plan, an establishment must conduct a hazard analysis to identify and list the physical, biological, or chemical food safety hazards reasonably likely to occur in the production process for a particular product and the preventive measures necessary to control those hazards. The establishment then must identify the critical control points (CCPs) in each of its processes. A CCP is a point, step, or procedure in a food process at which control can be applied to ensure that the occurrence of a food safety hazard is prevented, eliminated, or reduced to an acceptable level. Next, the establishment must establish critical limits for the preventive measures associated with each identified CCP. A critical limit is the maximum or minimum value to which a hazard must be controlled at a CCP to prevent, eliminate, or reduce to an acceptable level the occurrence of the identified food safety hazard. Critical limits are most often based on process parameters such as temperature, time, water activity, pH, or humidity. Significantly, critical limits must be designed to satisfy relevant FSIS regulations, including performance standards.

Therefore, performance standards are an integral part of the HACCP systems in official meat and poultry establishments. HACCP provides the framework for industry to set up science-based process controls. Performance standards tell establishments what those controls need to achieve for their HACCP plans to be effective and provide a necessary measure of accountability for achieving acceptable food safety. Performance standards and HACCP provide meat and poultry establishments with the incentive and flexibility to adopt innovative, science-based processing procedures and controls; ensure safety for consumers; and provide objective, measurable standards, compliance with which can be determined through Agency inspection.

IV. The Proposed Performance Standards

A. Lethality

For each category of RTE product, FSIS is proposing at least one lethality performance standard. The term “lethality” refers to a required reduction in the number of specific pathogenic organisms. Further, FSIS is proposing lethality performance standards that reflect the destruction of “reference” organisms, i.e., microorganisms whose elimination or reduction most often indicates the elimination or necessary reduction of other pathogens of concern.

In this proposed rule, for all RTE products except thermally-processed, commercially sterile products, the lethality performance standards are

expressed as probabilities of remaining numbers of the reference pathogen in 100 grams of finished product after a successful lethality treatment is, or treatments are, applied to hypothetical “worst case” raw product. The lethality performance standards also are expressed as the number of decimal reductions of the reference pathogen required to achieve those probabilities in hypothetical worst case products. These decimal reductions are expressed as “x-log
10
”, meaning that the expected relative reduction of the reference organism would be a factor of 10
x
. FSIS has tentatively concluded that effecting these specific reductions ensure even a worst case product would present no health risk to consumers.

For all RTE meat and poultry products, other than thermally processed, commercially sterile products, FSIS is proposing to require that processing achieve one of the following probabilities that that no more than small numbers of
Salmonella
would remain in any 100 gram sample of a finished product made from worst case product:

>0 surviving
>1 surviving
>2 surviving
>3 surviving
>4 surviving

39.4
9.06
1.45
0.177
0.0174

Although an establishment's processing would be required to achieve these probabilities that there will be few, if any, remaining pathogens in finished product, any detectable levels of viable
Salmonella
in RTE product would render that product adulterated.

Alternatively, official establishments may employ processes validated to achieve specific levels of reduction of
Salmonella
organisms throughout their finished, RTE meat and poultry products: 6.5-log
10
throughout finished, RTE meat products and 7-log
10
throughout finished, RTE products containing any amount of poultry. The probabilities in Table 1 are derived from statistical models of hypothetical worst case meat and poultry products that have been successfully processed to achieve 6.5-log
10
and 7-log
10
reductions in
Salmonella
, respectively. A hypothetical, worst case raw meat product would contain 6.2-log
10
of
Salmonella
per hundred grams; a hypothetical, worst case raw poultry product would contain 6.7-log
10
of
Salmonella
per hundred grams. See the section entitled “Derivation of the Proposed Lethality Performance Standards” for further discussion.

The Agency has selected
Salmonella
as the reference organism for most RTE meat and poultry products because: (1) It is prevalent in raw poultry, beef, and pork; (2) it causes a high incidence of foodborne illness; and (3) foodborne illness associated with
Salmonella
is severe. See the section entitled “Selection of the Reference Organisms” for additional discussion of how FSIS determined the lethality performance standards and the target pathogen for each type of RTE meat and poultry product.

Because destruction of reference organisms may not always result in the elimination or necessary reduction of other pathogens of concern, FSIS also is proposing to clarify in the regulations that establishments must also reduce other pathogens and their toxins or toxic metabolites to the levels necessary to prevent product adulteration. It is the responsibility of the establishment to ensure that the final product is safe. If FSIS were to find certain viable pathogens in a RTE product at levels considered dangerous, even in product otherwise free of the reference pathogen, it would consider that product to be adulterated.

FSIS is not proposing any specific lethality performance standards in addition to those that target the reference pathogen,
Salmonella
, except for fermented RTE products that contain beef. Within its hazard analysis, each establishment will be responsible for determining which other pathogens might survive processing and then implementing the appropriate control measures. FSIS requests comment on whether it should enumerate, in its regulations, lethality performance standards for other pathogens and toxins that can pose hazards to specific products or within specific processing contexts.

FSIS is proposing an additional lethality performance standard for all fermented RTE products that include any amount of beef, except thermally-processed, commercially sterile products. The Agency is proposing to require that establishments that produce these products implement processes that result in the following probabilities that, at worst, only minute amounts of
E. coli
O157:H7 organisms would remain in any 100 gram sample of a finished product made from worst case product:

Table 2.—Probability (%) of E. Coli O157:H7 Surviving in 100 Grams of Finished Product Made From Worst Case Product

>0 surviving
>1 surviving

22.2
2.67

Although an establishment's processing would be required to achieve these probabilities of remaining pathogens in finished product, any detectable levels of viable
E. coli
O157:H7 in RTE product would render that product adulterated.

FSIS also is proposing that, alternatively, establishments may employ processes validated to achieve a 5.0-log
10
reduction of
E. coli
O157:H7 throughout fermented products containing beef. The probabilities in Table 2 are derived from statistical models applied to hypothetical worst case beef products that have been processed to achieve a 5-log
10
relative reduction in
E. coli
O157:H7. A hypothetical, worst case raw product that contained any amount of beef would contain 4.4-log
10
of
E. coli
O157:H7 per hundred grams. See the section entitled “Derivation of the Proposed Lethality Performance Standards” for further discussion.

The Agency is proposing this lethality performance standard in addition to the
Salmonella
standard for fermented products that contain beef for several reasons. In 1994, there was an outbreak of foodborne illness linked to
E. coli
O157:H7 in fermented beef sausages. Also, these products may not be fully cooked before fermentation and fermentation creates an acidic environment in which
E. coli
O157:H7 can survive.

Also, the FSIS Office of Public Health and Science (OPHS) recently sponsored a study entitled “Risk Assessment of the Public Health Impact of
Escherichia coli
O157:H7 in Ground Beef” (Ref. 1, available for viewing by the public in the FSIS Docket Room). The draft risk assessment shows that levels of
E. coli
O157:H7 in cattle represents a risk to consumers of ground beef and that unless there is a significant intervention on the farm or during processing, the risk is likely to remain. This draft risk assessment is discussed further under the sections entitled “Derivation of the

Proposed Lethality Performance Standards” and “Fermented Products.”

Cattle and sheep may carry
E. coli
O157:H7 in the intestinal tract at the time of slaughter. However, among commercially-prepared meat products, only those that contain beef have been implicated in a number of foodborne illnesses associated with this pathogen. Therefore, in regard to meat and poultry products, the Agency is proposing this standard only for fermented products that contain beef.

FSIS is not proposing this performance standard for fermented poultry products that do not contain beef.
E. coli
O157:H7 has been found to colonize the ceca of chickens and has been isolated from retail poultry in the United States (Ref. 2, available for viewing by the public in the FSIS Docket Room). However, FSIS has never found the pathogen in raw or ready-to-cook samples of poultry obtained from processing establishments. FSIS requests comment as to whether it should also apply this standard to RTE fermented poultry products that do not contain beef, as well as to RTE fermented meat products that do not contain beef.

FSIS is proposing performance standards for thermally-processed, commercially sterile meat and poultry products that are similar to these lethality standards but derived somewhat differently. See the section “Thermally-Processed Commercially Sterile Products” for a complete discussion.

Compliance With the Lethality Performance Standards

To meet the proposed lethality performance standards, establishments would need to employ processes validated either to achieve the proposed decimal reductions of pathogens throughout a finished product or that result in one of the stated probabilities that only small numbers of reference organisms would remain viable in a worst case finished product. To develop criteria for evaluating the effectiveness of processes that achieve one of the proposed probabilities, it will be necessary for the processor to define, using associated statistical criteria, the expected characteristics of the treated product after processing, assuming certain product conditions before processing. For example, an establishment would need to specify that the probability of there being more than
x
surviving organisms in the finished product is no more than
p
, given that the worst case pre-processed product contained at least
y
organisms.

By codifying acceptable probabilities of remaining reference organisms in finished product, FSIS would be allowing establishments to employ processes that achieve varying levels of lethality, therefore providing processing flexibility while ensuring product safety. By also proposing specific lethality performance standards in the regulations, FSIS provides clear performance standards to establishments that may not have the resources to derive an alternative lethality or the ability to demonstrate that their process achieves a specific probability that no more than a certain number of reference organisms might exist in the finished product.

As explained above, FSIS has tentatively determined that processes that achieve the proposed lethality performance standards will process hypothetical, worst case raw product into finished, RTE product that poses no health risk to the consumer and is thus safe. In reaching this tentative conclusion, the Agency made conservative assumptions concerning the actual lethality achieved throughout the product. The Agency acknowledges that it might be possible for producers to demonstrate scientifically that these lethality assumptions or the Agency's defined worst case would not be applicable for their particular processing situation. An establishment could then design a process with lethality values that are different from those provided in this rule, but that would still yield a product that meets the final conditions equivalent to those achieved by the specific levels of pathogen reduction contained in the lethality performance standards.

An establishment developing an alternative lethality treatment or treatments and assuming an initial product condition other than the worst case would need to include in its HACCP plan scientific data and statistical validation that would justify the assumed initial conditions and verify that these would remain constant over time. For example, an establishment may be able to demonstrate that the number of
Salmonella
is not uniformly distributed throughout a particular type of product. The establishment also might demonstrate that because of husbandry and slaughter practices, the worst case product processed within an establishment differs from the worst case scenarios developed for this rule. Demonstrations of initial product conditions solely by statistical means would likely be insufficient to ensure that processes that employ alternative lethalities will result in product that meets the performance standards.

Generally, an establishment will need to demonstrate in its HACCP plan how its lethality treatment results in a finished product equivalent to that provided by compliance with the probabilities set out in this proposal. The establishment will need to demonstrate the relationships between the lethality treatments and the specific characteristics of a product, such as physical and chemical properties. This demonstration could involve the use of heat transfer equations and should account for all variables that would affect lethality (e.g., size of product, humidity, density, thermal conductivity, specific heat, shape, product composition, and strain of organism).

Finally, establishments employing alternative lethalities will need to demonstrate, within their HACCP plans, that they have validated their processes as being effective in ensuring product safety. Section 417.4(a)(1) of the HACCP regulations sets forth the “initial validation” requirements for establishments under HACCP:

Upon completion of the hazard analysis and development of the HACCP plan, the establishment shall conduct activities designed to determine that the HACCP plan is functioning as intended. During this HACCP plan validation period, the establishment shall repeatedly test the adequacy of the CCPs, critical limits, monitoring and record keeping procedures, and corrective actions set forth in the HACCP plan. Validation also encompasses reviews of the records themselves, routinely generated by the HACCP system, in the context of other validation activities.

FSIS explains the derivation of the proposed lethality performance standards in the following section. A technical paper (Ref. 3, available for viewing by the public in the FSIS Docket Room and on the Internet.
1

) explaining the derivation of the lethality performance standards also is available. Establishments are encouraged to use this paper when developing alternative lethalities. In the paper, FSIS explains the methodology used to calculate the probability of remaining
Salmonella
organisms in treated product.

1
http://www.fsis.usda.gov /OPPDE/rdad/FRPubs/95- 033F_tech%20paper.pdf

Notably, with any final action, FSIS will provide compliance guides that give explicit processing instructions and time/temperature combinations proven to achieve the proposed decimal reductions of pathogens. Small and other establishments that do not have the technical resources to demonstrate that they are meeting the proposed

performance standards may use these compliance guides to develop their HACCP systems. FSIS has published compliance guides for meeting the lethality and stabilization performance standards already set forth in its January 6, 1999, final rule, has posted these documents to the FSIS web page (
http://www.fsis.usda.gov
), and has made the documents available free of charge via the Constituent Update (see section XIV Additional Public Notification) and the FSIS docket room. FSIS expects to make additional draft guidance documents available after publication of this proposed rule and as information becomes available in order to provide establishments with guidance for safely manufacturing RTE meat and poultry products. These draft guidance materials will be clearly identified as guidance materials and not as regulatory requirements. These guides would be applicable to the processing of many of the RTE meat and poultry products governed by these proposed regulations. FSIS plans to update these guides soon in accordance with ongoing Agricultural Research Service studies. Where possible, FSIS will base its compliance guides on existing industry practices and requests comment and information regarding processing that has been shown to meet the proposed performance standards.

Derivation of the Proposed Lethality Performance Standards

Salmonella

To derive the proposed lethality performance standards for
Salmonella
, FSIS first determined the levels of
Salmonella
in a hypothetical worst case raw product of a fixed weight. The hypothetical “worst cases” for
Salmonella
and
E. coli
O157:H7 were derived using data from FSIS's Nationwide Microbiological Baseline Data Collection Program surveys (Ref. 4, available for viewing by the public in the FSIS Docket Room). The baseline surveys conducted by FSIS were designed to provide estimates of the national prevalence and levels of selected bacteria of public health concern.
Salmonella
was one of the pathogens specifically addressed in all of the baseline surveys for the various classes of products. The baseline surveys were conducted over a specified period of time ranging from a half year to a full year. The baseline surveys were used to establish the pathogen reduction performance standards for
Salmonella
that were included as a component of the Pathogen Reduction-HACCP final rule of July 25, 1996 (61 FR 38806). The performance standards for
Salmonella
that were established as part of the Pathogen Reduction-HACCP final rule differ from the proposed lethality performance standards for
Salmonella
included as part of this proposed rulemaking.

The
Salmonella
performance standards for the Pathogen Reduction-HACCP final rule are designed as follows: they are applicable to establishments that produce raw products; FSIS collects and tests samples from raw product; the results of the raw product samples are reported to the establishment by FSIS after a specified number of samples are collected over time; and a positive result for
Salmonella
in raw product generally does not result in an adulteration determination. In contrast to this design, the
Salmonella
lethality performance standards of this proposed rule are designed as follows: they are applicable to establishments that produce ready-to-eat products (not raw product); the establishment may sample and test samples of RTE product as part of its verification activity associated with the production of RTE product and any testing by FSIS is conducted as part of the Agency's verification activity; and a positive result for
Salmonella
in RTE product does result in an adulteration determination. The premise and use of the lethality performance standards for
Salmonella
in this proposed rule are unchanged from those previously contained in the recent final rule for RTE products (64 FR 732, January 6, 1999). Consequently, the baseline surveys were used in the design of two separate performance standards: one performance standard identifies the prevalence of
Salmonella
in raw product over a specified period of time; the other performance standard (addressed as part of this proposed rule) identifies the expected reduction in the level of
Salmonella
in RTE product in a specified lot of product. Since these two performance standards apply to different types of establishments (i.e., the former applies to establishments producing raw product; the latter applies to establishments producing RTE product), they are not duplicative standards nor do they directly relate to each other. The only commonality between these two performance standards for
Salmonella
is that they are both derived from the same baseline surveys. The level of
E. coli
O157:H7 in raw products also was assessed in the same baseline studies as were used to determine the level of
Salmonella
in raw products.

Using the national baseline survey information to establish the levels of selected bacteria of public health concern (e.g.,
Salmonella
), the Agency then determined levels of lethality that would limit the probability of any remaining
Salmonella
or
E. coli
O157:H7 in finished product produced from worst case raw product. FSIS made conservative but reasonable assumptions concerning measurement error and distributions of organisms throughout the product. These assumptions are fully discussed in the technical paper (Ref. 3, available in the FSIS Docket Room and at the FSIS web page http://www.fsis.usda.gov). However, the assumptions are generally based on the following which are further discussed below: the number of organisms recovered from frozen samples; the sensitivity of the detection methodology; the confidence level of measurement variability; and the serving size. Thus, worst case levels in product are not expected to actually occur, provided products are handled appropriately before lethality treatments. The derived worst case levels are hypothetical constructs meant to represent upper limits of possibilities for raw product produced under appropriate, normal manufacturing conditions. These conditions include maintaining the raw product at or below temperatures known to prevent growth of
Salmonella
and most other pathogenic organisms (e.g., at or below 40 degrees Fahrenheit). In addition, they include processing the raw product into RTE product quickly before the raw product's surface temperature becomes elevated for sufficient amounts of time to allow
Salmonella
and most other pathogenic organisms to multiple exponentially. FSIS believes that under these conditions, processes that satisfy the performance standards established as a result of this rulemaking will be safe.

The Agency used the most probable number (MPN) method for measuring levels of
Salmonella
in the FSIS surveys of meat and poultry products. The MPN measurements were made on frozen samples. The calculations used to determine the number of organisms for the worst case product take into account non-recovery of organisms in frozen samples.

For
Salmonella
, the Agency assumed a 30 percent recovery of organisms from frozen samples (Ref. 3, available for viewing by the public in the FSIS Docket Room). The expected recovery is a function of how quick and long the sample was frozen. Based on FSIS experience with samples, the approximate detection limit for recovery of
Salmonella
is 0.5 cells per gram in 25-gram frozen samples. This means that there is a high probability that a 25-

gram sample with 13 organisms would be found positive. For the purposes of this regulation, the Agency assigned a 99% probability that a 25-gram sample with 13
Salmonella
cells would test positive. Even if one organism were recovered, the sample result would be positive, so that the probability of a positive sample result can be expressed as 1
τ
13
, where
τ
is the theoretical probability of a single injured or uninjured
Salmonella
organism not being recovered. With this assumption, for frozen samples,
τ
is approximately 70%, that is, there is a 70% probability that a single organism would not be recovered. Thus, there is a 30% recovery of
Salmonella
cells.

To account for measurement variability, the Agency calculated the 97.5 percent upper confidence limit associated with the measured MPN value (Ref. 3, available for viewing by the public in the FSIS Docket Room). FSIS did not use the average level of
Salmonella
reported for the various classes of product. Rather, in order to determine the highest estimate for the level of
Salmonella
in raw products, FSIS took the raw data, not the calculated average, and computed a number at the 97.5 percent upper confidence level. Using this upper limit, the Agency then computed the upper limit for 143 grams of raw product. The Agency used 143 grams of raw product as the basis for its calculations because after cooking, assuming a 70 percent yield, 143 grams would result in approximately 100 grams (3.5 ounces) of cooked product.

The Agency used the high MPN value for ground chicken (the highest MPN value measured for poultry products) from the FSIS national baseline surveys
2

to determine the proposed lethality for
Salmonella
for all RTE products containing poultry, other than thermally processed, commercially sterile products. For ground chicken, the upper 97.5 percent confidence limit for the highest measured MPN value of 2300 MPN per gram for
Salmonella
, assuming a 30 percent recovery, is approximately 37,500 cells/gram, which, when multiplied by 143 grams totals approximately 6.7-log
10
cells. Therefore, the level of
Salmonella
organisms in a hypothetical worst case raw product would be greater than 6.5-log
10
but just less than 7.0-log
10
. Consequently, to provide a margin of safety and to use either a whole or half integer lethality, FSIS is proposing to require a reduction in viable
Salmonella
of 7.0-log
10
, which is 0.3-log
10
above the worst case level, throughout RTE products that contain poultry, other than thermally processed, commercially sterile products. The consequence of this choice is that, for a hypothetical “worst case” product, the probability of surviving
Salmonella
organisms is 39.4%, assuming that the distribution of the number of survivors is binomial with number parameter equal to the number of organisms in the worst case and the probability parameter equal to 1/10
x
where x is the required decrease in viability.

2
While the numbers of samples in the FSIS national surveys are rather large, the largest MPN value, as an estimate of large densities of pathogenic organisms, from a statistical perspective, may have substantial statistical variation. Thus, to reduce differences in required lethality reductions caused by statistical variation, data sets of different species were combined if warranted by consideration of the prevalences and possibly the geometric means of the levels of the organisms for these species. The high value of combined data sets was used for determining the hypothetical worst case for these species. The criteria used for combining data sets of different species are easier stated as the converse of criteria for when data sets would not be combined and thus the lethality requirements for these species would be different. The criteria for determining when lethality requirements for two species, A and B, are different are: For a given type of product, the lethality requirement for species A is larger than that of species B if (1) the high MPN value for species A is larger than that of species B, and (2) the prevalence for species A is larger than that of species B, or the prevalences are approximately equal and the geometric mean for species A is larger than that of species B. Otherwise, the lethality requirements would be the same and the high value of the combined data set would be used for both species A and B. For the products and pathogen considered in this proposed regulation, the criteria for combining data depend upon the prevalences and the high values.

Alternatively, an establishment may use a processing procedure validated to achieve the probabilities in Table 1 above that no more than specific amounts of
Salmonella
would remain in any 100 gram sample of a finished, hypothetical worst case product. As stated above, these probabilities would result in hypothetical worst case poultry products that had been successfully processed to achieve a 7-log
10
reduction in
Salmonella
.

To determine the proposed lethality for RTE meat products that do not contain poultry, other than thermally processed, commercially sterile products, the Agency used the high MPN value for whole beef (the highest MPN value measured for all meat products): 240 MPN/cm
2
. To translate this value to a level per gram, FSIS assumed that, for a worst case level, a cut of meat is 0.8 cm and that the specific density of beef is approximately 1.1 grams/cm
3
(slightly lower than average) (Ref. 3, available for viewing by the public in the FSIS Docket Room). These factors are for practical purposes equal to 1, so that the MPN/cm
2
values are assumed to estimate the level per gram of product. Thus, for the worst case derivation, the starting value is 240 MPN/g.

The 97.5 percent upper confidence limit, assuming a 30 percent recovery, is 4100 cells/g. Because samples for the whole product surveys consisted of pooled tissue from 3 different carcass sections, and the prevalence was low (less than 3 percent), the Agency assumed that the high value used for determining the worst case product is 3 times that of the measured MPN value. Thus, the 97.5 percent upper confidence limit is multiplied by 3 and then multiplied by 143 grams. The resulting number of organisms for the worst case product is approximately 6.2-log
10
. Therefore, to provide the same margin of safety as provided for with poultry products, the proposed required lethality is obtained by adding 0.3 log
10
to the worst case level of 6.2 log
10
. Thus, FSIS is proposing to require either a relative reduction in viable
Salmonella
of 6.5-log
10
throughout finished, RTE meat products, or alternatively, one of the probabilities listed above in Table 1. FSIS has not specified the probability of worst case product actually occurring since the worst case was a hypothetical construct based, in part, on a high confidence level of the maximum observed level of microorganisms in a statistically designed national baseline. In addition, FSIS made additional assumptions that FSIS believes to be conservative. All the assumptions regarding the derivation of the worst case are contained in the technical paper (Ref. 3, available in the FSIS Docket Room and at the FSIS web page,
http://www.fsis.usda.gov).
FSIS requests comments regarding these assumptions.

E. coli O157:H7

After a 1994 outbreak of illnesses caused by
E. coli
O157:H7, FSIS recommended that producers of fermented RTE products that contain any amount of beef validate their processes to achieve a 5.0 log
10
lethality of
E. coli
O157:H7 (see additional discussion under
Fermented Products
). This recommended lethality was based on a report submitted to FSIS (The Task Force on Technical Issues Arising from the National Advisory Committee for Microbiological Criteria for Foods (NACMCF)). The 5-log
10
relative reduction was derived by adding 1 log
10
as a safety margin to an assumed worst case of 4.0 log
10
that was recommended by the NACMCF. If this lethality were applied in a product containing 10
4
cells per gram, then it would be expected that a single cell would remain. However, the conclusion that a

single
E. coli
O157:H7 cell per 10 grams (or a possible 10 cells per 100 grams) remaining in the product adequately prevents foodborne disease is in question. Some researchers now believe that low numbers of
E. coli
O157:H7 cells ingested are sufficient to cause foodborne disease (Ref. 5, available for viewing by the public in the FSIS Docket Room).

Presented in chapter 5 of the OPHS risk assessment are results of a derivation of the possible number of
E. coli
O157:H7 cells in combo bins of 2000 pounds or approximately 10
5.96
grams (Ref. 1, available for viewing by the public in the FSIS Docket Room). The highest number associated with a non-zero probability is 10
7
cells for which an upper bound probability of occurrence is 0.002% (1/50,000). As discussed above in the derivation of the proposed lethality requirements, the Agency considers the number of cells in 143 grams of raw product, accounting for a possible 70% yield when the product is processed. A bin with 10
7
cells implies that the expected number of cells in 143 grams of raw product would be about 3.2 log
10
cells per 100 grams. The assumptions used in deriving this number assume that the
E. coli
O157:H7 cells present are uniformly distributed throughout the bin, so that the 3.2 log
10
represent an average or expected number of cells per 143 grams of product. It is clearly possible that there would be in some 143-gram portion more than 3.2 log
10

E. coli
O157:H7 cells. Thus, a worst case level should be larger than 3.2 log
10

E. coli
O157:H7 cells.

To derive worst case levels for
E. coli
O157:H7 for the purpose of determining a performance standard, the Agency applied the algorithm, described above for
Salmonella,
using information presented in OPHS risk assessment. This risk assessment presented results of MPN analyses from the Agency's microbiological baseline surveys of bovine carcasses (Ref. 1, available for viewing by the public in the FSIS Docket Room). In total, out of about 4,000 samples, 4 samples were found positive. For each positive a matching sample was analyzed using the MPN procedure. Of the 4 analyzes, 2 were found positive. The highest reported MPN value was 0.93 cells/cm
2
, which, as described above, is assumed to represent level per gram value, or 0.93 MPN/gram. A 97.5 percent upper confidence limit for this value is 3.7 cells/cm
2
. FSIS did not use the average level of
E. coli
O157:H7 reported for the various classes of product. Rather, in order to determine the highest estimate for the level of
E. coli
O157:H7 in raw products, FSIS took the raw data, not the calculated average, and computed a number at the 97.5 percent upper confidence level.

The samples used for determining
E. coli
O157:H7 levels in the FSIS surveys were frozen. In the OPHS risk assessment, information concerning the recovery rate is given. It is stated in the report that nine 25-gram samples of ground beef were inoculated with 0.7
E. coli
O157:H7 organisms per gram, and that eight of these samples subsequently were detected as positive. In determining the possible recovery for
E. coli
O157:H7 cells in a sample that is subsequently frozen, FSIS assumes that the actual number of cells in a specified 25-gram sample is a random variable, n, following a Poisson distribution, f(n, λ) = e
−
λ
λ/n!, with expected value λ = 17.5. If
τ
is the probability of not recovering a given single cell, then the probability of detecting the presence of
E. coli
O157:H7 in a 25 gram sample, is,
π
=
Σ
(1−
τ
n
)f(n, λ) = 1−e
−
λ
(1−
τ
)
. Thus, the probability of recovering a given cell is 1−
τ
= −ln(1−
π
)/λ. From nine samples, eight were detected positive, so that a 97.5% lower confidence bound for
π
is 0.6635. Using this value for
π
, the derived value for 1−
τ
is 0.062, representing the recovery. For the worst case level, the 97.5 percent upper bound, 3.7 cells/cm
2
, is divided by 0.062 to derive 59.45 cells/cm
2
.

As described above for deriving the worst case levels for
Salmonella
in beef, the measured levels are multiplied by 3, to account for the fact that samples from the bovine FSIS baseline surveys consisted of a composite from 3 sections of the carcass, and for a worst case derivation, FSIS assumes that all the cells existed in one of the three sections. Thus, for the worst case level, the 59.45 cells/cm
2
is multiplied by 3, and then multiplied by 143 grams to derive an approximate 4.4 log
10
cells for the worst case level.

The above derivation indicates that the “worst case” level of 4.4 log
10
cells per 143 grams is greater than the highest expected level of 3.2 log
10
cells per 143 grams derived in the OPHS risk assessment. Consequently FSIS will use the 4.4 log
10
as the “worst case” level.

To provide the same margin of safety as provided for with
Salmonella
in poultry and red meat products, the lethality is obtained by adding 0.3 log
10
to the worst case level of 4.4 log
10
. However, foodborne illness associated with
E. coli
O157:H7 might be more severe than that associated with
Salmonella,
as testified to by the severity of many reported cases in children and senior citizens. Also, as stated above, some researchers believe that low numbers of ingested
E. coli
O157:H7 cells are sufficient to cause foodborne illness. Furthermore, there is only a small amount of data from the Agency's microbiological baseline survey: four samples, of which only two were positive. This number of samples does not provide a high degree of confidence in the magnitude of the higher levels that might exist. Consequently, FSIS is requiring that processors of fermented products containing beef achieve a higher probability of no surviving cells of
E. coli
O157:H7 in treated worst case products than that required for
Salmonella.
Specifically, FSIS is proposing a 5-log
10
lethality, which can be obtained by adding 0.6 log
10
to the “worst case” level (instead of 0.3 log
10
added for
Salmonella
). The probability of no surviving
E. coli
O157:H7 cells given a “worst case” level of cells is about 78% (instead of 61% for
Salmonella
).

FSIS also examined measured levels of
E. coli
O157:H7 found in suspect lots of hamburger identified in foodborne disease outbreaks (Refs. 6 and 7, available for viewing by the public in the FSIS Docket Room). Direct count determinations were as follows: 50, 100, 5100, and 6200 colony forming units (CFU) per gram. Because of the possibility that the high
E. coli
O157:H7 levels represent product that has been abused and thus are not representative of product produced in an establishment and used in RTE product, FSIS could not, with complete justification, use these values for determining a required lethality. However, these results do suggest that a lethality of at least 5-log
10
is needed to help ensure an
E. coli
O157:H7 free RTE product.

The derivation for the proposed lethality of
E. coli
O157:H7, in using only a slightly higher probability of no surviving cells compared to that used for deriving the proposed lethalities for
Salmonella,
assumes only a slightly greater public health concern for
E. coli
O157:H7. However, foodborne illness associated with
E. coli
O157:H7 might be significantly more likely than that associated with
Salmonella.
As mentioned above, some researchers now believe that low numbers of
E. coli
O157:H7 cells ingested are sufficient to cause foodborne disease. This belief also is reflected in the recent OPHS draft risk assessment regarding
E. coli
O157:H7 in ground beef (Ref. 1, available for viewing by the public in the FSIS Docket Room). The dose response model used in this report allows the possibility of a 1% probability of illness when a random selected consumer ingests a

single cell; and when, ingesting 10 cells, the probability of illness could be as high as 10%.

Consequently, FSIS may need to require that processors of fermented products containing beef achieve a higher probability of no surviving cells of
E. coli
O157:H7 in treated worst case products. For example, if the proposed lethality were 5.5 log
10
, the probability of no surviving
E. coli
O157:H7 cells in the hypothetical worst case would be 92.4% instead of 77.8%; if the proposed lethality were 6.0, then the probability of no surviving
E. coli
O157:H7 cells would be 97.5%.

Since the number of sample results from which the worst case was derived is small, there is not a high degree of confidence in the magnitude of the higher levels of
E. coli
O157:H7 that might exist. Further information may require FSIS to adjust the worst case level and thus the required lethality, accordingly. It is important to note, however, that a fermentation process offers an extra degree of safety compared to a heat process, given the same lethality. Unlike ordinary cooked RTE products, the physio-chemical environment within fermented products is hostile to the survival of pathogens. Thus, within an ordinary cooked RTE product, sublethally injured bacteria may be able to resuscitate and then multiply when the temperature rises. Within fermented sausages, most of which are shelf-stable, resuscitation is not possible. FSIS specifically requests comment on the proposed performance standard for the pathogen
E. coli
O157:H7 in fermented products containing beef.

FSIS has not specified the probability of worst case product actually occurring since the worst case was a hypothetical construct based, in part, on a high confidence level of the maximum observed level of microorganisms in a statistically designed national baseline. In addition, FSIS made additional assumptions that FSIS believes to be conservative. All the assumptions regarding the derivation of the worst case are contained in the technical paper (Ref. 3, available in the FSIS Docket Room and at the FSIS web page,
http://www.fsis.usda.gov
). FSIS requests comments regarding these assumptions.

Selection of the Reference Organisms

An explanation of how the Agency established the proposed reference organisms for each category of RTE product, other than thermally processed, commercially sterile products, follows.

Dried Products

The pathogens associated with dried (but not fermented) RTE meat and poultry products are
Salmonella, Listeria monocytogenes, Staphylococcus aureus, E. coli
O157:H7 and
Trichinella spiralis. T. spiralis
is only associated with pork and game products. There are a limited number of studies on the reduction of pathogens during the processing of dried meat and poultry products.

J. A. Harrison and M. A. Harrison surface-inoculated one-third of a beef jerky strip (15 × 1.5 × 1.5 cm.) with 0.1 ml of a 10
8
CFU/ml cell suspension each of
L. monocytogenes, Salmonella typhimurium,
and
E. coli
O157:H7 (Ref. 8, available for viewing by the public in the FSIS Docket Room). Results show that higher log reductions of the three pathogens were obtained when beef jerky was preheated to 160 °F and when curing agents were added. In general,
L. monocytogenes
was more resistant to the treatments. However, after 10 hours of drying at 140 °F, the populations decreased to undetectable levels, resulting in a 5.5 to 6.0 log reduction of the three pathogens. After storage at 25 °C for 8 weeks, none of the pathogens were detected. Subsequent challenge studies on inoculated ground beef jerky, with or without curing agents, heated or unheated, showed that
Salmonella spp.
was in general more resistant than
L. monocytogenes
to the integrated process. However, after 6 hours of drying at 140 °F,
L. monocytogenes
and
Salmonella
had about the same population reduction in preheated samples (Refs. 9 and 10, available for viewing by the public in the FSIS Docket Room).

These studies show that the time and temperature of drying and other variables, such as the use of beef strips or formed ground beef jerky, the addition of curing agents, and preheating before drying, will affect the reduction of pathogens. Lethality of pathogens in dried products is achieved by dehydration to a water activity (a
w
) level that inhibits their growth. Preheating or precooking and the addition of curing agents facilitate and add to the lethality factor.

In 1995, a salmonellosis outbreak was associated with commercially produced beef jerky linked to three
Salmonella
serotypes (Ref. 11, available for viewing by the public in the FSIS Docket Room). The CDC Morbidity and Mortality Weekly Report (MMWR) report stated that the New Mexico Department of Health investigated five outbreaks of salmonellosis associated with locally produced beef jerky from 1966 to 1988 and one outbreak of staphylococcal food poisoning associated with beef jerky in 1982. Also according to the MMWR, four other states reported foodborne disease outbreaks associated with the consumption of locally produced or homemade jerky from beef, bear, or cougar meat. The outbreaks were caused by
T. spiralis
and by nitrite poisoning.

The MMWR set out the recommendations of CDC for the prevention of bacterial growth in jerky production. CDC recommended rapid drying at high temperatures (i.e., initial drying temperature >155 °F (68.3 °C) for 4 hours, then >140 °F (60 °C) for an additional 4 hours), and decreased water activity (
i.e.,
a
w
= 0.86).

E. coli
O157:H7 was implicated in one case in homemade venison jerky (Ref. 12, available for viewing by the public in the FSIS Docket Room).
L. monocytogenes
has not been reported to be associated with any foodborne illness attributable to the consumption of commercial jerky products. So, based on the epidemiological data and research studies on jerky, it does not appear that
E. coli
O157:H7 or
Listeria
represent serious hazards in commercially produced jerky. Consequently, FSIS chose
Salmonella
as the proposed reference organism for dried products.

If a process used to produce dried products achieves the proposed reduction in the number of
Salmonella
organisms, the number of
T. spiralis, E. coli
O157:H7, and
S. aureus
should also be reduced to safe levels because these organisms are generally less heat resistant than
Salmonella. L. monocytogenes
is a problem more often because of inadequate sanitation than inadequate processing. Under HACCP and Sanitation SOP requirements, establishments must ensure that their processing controls hazards in addition to
Salmonella,
such as
L. monocytogenes,
if they are reasonably likely to occur.

Salt-Cured Products

The microbiological stability (the lethality during processing) of salt-cured meats, such as salt-cured hams, is dependent on their low water activity, the presence of nitrite, and smoke applied between the salting and drying processes (Ref. 13, available for viewing by the public in the FSIS Docket Room). Lethality of pathogens in the salt-cured products is attained by low temperature salting and drying. Both of these processes reduce the water activity to levels that inhibit the growth of pathogens. The addition of nitrates or nitrites and smoke enhance the inhibitive effect of the process.

There were two salmonellosis outbreaks linked to salt-cured hams: Serrano variety cured ham in Spain and prosciutto ham in Italy (Refs. 14 and 15, available for viewing by the public in the FSIS Docket Room). Low levels of salt and relatively high water levels in some parts of the Serrano variety cured ham were judged to be the most probable cause of
Salmonella
growth and consequent illness. Aside from
Salmonella,
other pathogens of concern in salt-cured products are
S. aureus, L. monocytogenes,
and
T. spiralis.
The Agency is proposing to select
Salmonella
as the reference organism because outbreaks in salt-cured products have been associated with
Salmonella.
As with dried products, if the process used to produce salt-cured products achieves the proposed 6.5−log
10
or 7.0-log
10
reduction in
Salmonella
organisms, the number of these other pathogens should also be reduced to safe levels. In addition, establishments would have to ensure that processing also controls hazards other than
Salmonella,
including other pathogens, that are reasonably likely to occur.

Fermented Products

In late 1994, 23 cases of illness caused by the pathogen
E. coli
O157:H7 were reported in Washington State and northern California (Ref. 16, available for viewing by the public in the FSIS Docket Room). Epidemiological investigations by State and local health agencies associated the outbreak with the consumption of dry cured salami products. In October 1995, the Pennsylvania State Department of Health linked 26 cases of salmonellosis to the consumption of contaminated Lebanon bologna (Ref. 17, available for viewing by the public in the FSIS Docket Room).

After the 1994 outbreak of illnesses caused by
E. coli
O157:H7, FSIS met regularly with scientists from the Agricultural Research Service, representatives of the meat and poultry industry and members of the NACMCF to develop a policy for ensuring the safety of shelf-stable, RTE fermented sausages. This group developed several processing options that would ensure a 5-log
10
relative reduction of
E. coli
O157:H7 in fermented sausages. In addition, FSIS approved a processing option developed by the Blue Ribbon Task Force on
E. coli
O157:H7 of the National Cattleman's Beef Association.

As explained previously, the 5-log
10
reduction of
E. coli
O157:H7 in dry and semidry fermented sausages was originally based on the notion of adding a 1-log
10
safety margin over an assumed worst case of 10
4
CFU/gram in raw product. FSIS offered 4 options to either achieve the recommended 5-log
10
relative reduction of
E. coli
O157:H7 or control for its presence in finished product: (1) Apply the cooking treatment in either 9 CFR 318.17 or 318.23, (2) apply a validated integrated heat treatment of equal lethality, (3) test product using ICMSF lot acceptance criteria, or (4) apply a validated 5-log
10
relative reduction or process that results in less than 1
E. coli
O157:H7 per 100 gram of finished product. The Blue Ribbon Task Force of the National Cattlemen's Beef Association specifically addressed Option 2—a validated 5-log
10
inactivation treatment. The Task Force focused on the processing parameters of heat and acid sensitivity of the organism. The processes and the resultant level of reduction of
E. coli
O157:H7 were summarized in a table and flow chart. In addition, the report recommended a fifth option, combination of sampling of raw ingredients and a 2-log10 lethality treatment, and described the remaining 3 options.

On August 21, 1995, FSIS wrote to establishments producing fermented sausages and strongly encouraged that they implement one of the validated processing options contained in the document to ensure the processing used achieves at least a 5-log
10
relative reduction of
E. coli
O157:H7. While most establishments have implemented one of the processing options, not all have.

As discussed previously, in support of rulemaking, OPHS has sponsored a risk assessment of
E. coli
O157:H7 in ground beef (Ref. 1, available for viewing by the public in the FSIS Docket Room). The draft risk assessment presents data on the prevalence of
E. coli
O157:H7 among breeding herds and feedlots of cattle, and
E. coli
O157:H7 levels on carcass samples. This information shows that levels of
E. coli
O157:H7 in cattle represents a risk to consumers of ground beef, and that, unless there is a significant intervention on the farm or during processing, the risk is likely to remain.

In addition, because of the incidence of foodborne illness linked to
E. coli
O157:H7 in fermented sausages and because these products ordinarily are not fully cooked before being fermented (which creates a situation that may allow the survival of
E. coli
O157:H7), the Agency is proposing to include
E. coli
O157:H7, in addition to
Salmonella,
as a reference organism for fermented RTE meat and poultry products that contain beef.

Under this proposal, processing of fermented products that contain beef would be required to meet lethality performance standards for both
Salmonella
in § 430.2(a) and for
E. coli
O157:H7 in § 430.2(b). As discussed under the “Lethality” heading above, for fermented RTE meat and poultry products that contain beef, the Agency is proposing that processing achieve either specific probabilities of remaining organisms in 100 grams of finished product, or a 5.0-log
10
relative reduction of
E. coli
O157:H7 throughout the product, which would achieve those probabilities in a hypothetical, worst case raw product. FSIS is not proposing this performance standard for fermented meat and poultry products that do not contain beef.

The Agency tests fermented sausage products for
Salmonella, L. monocytogenes , E. coli
O157:H7, and staphylococcal enterotoxin. Isolation or detection of any of these pathogens and enterotoxin results in product recall and destruction of product. With the exception of
L. monocytogenes,
these pathogens and staphylococcal enterotoxin have been linked to foodborne illness associated with fermented sausage products. With regard to
S. aureus,
the production of a heat stable enterotoxin (staphylococcal enterotoxin) after it has achieved a density of at least 10
5
CFU/g rather than the bacterium itself is responsible for foodborne illness. Growth of
S. aureus
is inhibited by the competitive growth of lactic acid bacteria, such as lactobacilli and pediococci, which are often used in fermented sausage products (Refs. 18 and 19, available for viewing by the public in the FSIS Docket Room).

A suboptimally active fermentation culture or an initial large number of
S. aureus,
as has occurred when contaminated starter culture is used, may result in the growth of
S. aureus
and the production of enterotoxin. However, since 1980, the industry has implemented fermentation controls, and no repeat of the previous type outbreaks has occurred. Therefore, FSIS is not proposing
S. aureus
as a reference organism for these products.

L. monocytogenes
is the most frequently isolated pathogen of those included in the FSIS monitoring program for fermented sausages. Despite its prevalence in fermented sausage products, no foodborne illnesses have been linked to
L. monocytogenes
in fermented sausages. Thus, the Agency is not proposing that
L. monocytogenes
be a reference organism for fermented sausages; however, if the Agency were to find
L. monocytogenes
in the finished

product, the product would be adulterated and subject to recall.

In a Lebanon bologna process, a 3- to 4-log
10
reduction of
Salmonella dublin
and a reduction of
Salmonella typhimurium
to undetectable levels was observed by the end of fermentation if starter culture was used (Ref. 20, available for viewing by the public in the FSIS Docket Room). Similarly, Bacus noted that contamination of fermented meat products with
Salmonella
most likely results from an inadequate lactic acid production or a highly contaminated raw product (Ref. 21, available for viewing by the public in the FSIS Docket Room).

Various studies have shown that fermentation and drying resulted in about a 2-log
10
reduction of
E. coli
O157:H7 (Refs. 22 through 24, available for viewing by the public in the FSIS Docket Room). In one study, Glass, et al., reported that
E. coli
O157:H7 decreased by about 2-log
10
CFU/g after fermentation, drying, and storage at 4 °C for 6 weeks following the end of an 18-21 day drying cycle for a fermented sausage formulation. In another, however, Faith et al., observed a 5- to 6-log
10
reduction of
E. coli
O157:H7 in pepperoni sticks following fermentation, drying, and 2 weeks of storage at an ambient (unrefrigerated) temperature of 21 °C.

In one of the few studies that compared the combined effect of fermentation and drying on both
Salmonella
and
E. coli
O157:H7, Ellajosyula,
et al.,
observed that the reduction of
Salmonella
and
E. coli
O157:H7 in Lebanon bologna was less than 2- log
10
after fermentation to pH 4.7 (Ref. 22, available for viewing by the public in the FSIS Docket Room). In this study,
Salmonella
was equally or significantly (P<0.01) less resistant than
E. coli
O157:H7 to various combinations of pH levels achieved after fermentation and subsequent heating at 110 °F to 120 °F. Fermentation to pH 5.2 or 4.7 followed by heating at 110 °F to 120 °F for specified times (e.g., 110 °F for 20 hours or 120 °F for 3 hours) resulted in a greater than 7- log
10
reduction of both
Salmonella
and
E. coli
O157:H7. This study shows that a final heating step may be necessary to achieve the proposed reduction of both
Salmonella
and
E. coli
O157:H7 in fermented sausage products.

Salmonella
and
E. coli
O157:H7 have been the cause of foodborne illnesses linked to fermented sausage products. Although, as noted above,
Salmonella
may be less resistant than
E. coli
O157:H7 to the processes for the different fermented meat products, it has not been demonstrated that processes resulting in a 5.0-log
10
reduction of
E. coli
O157:H7 will result in a 6.5- log
10
or 7.0- log
10
reduction of
Salmonella
in meat and poultry products, respectively. Conversely, processes resulting in a 6.5- or 7.0- log
10
reduction of
Salmonella
have not been shown to produce a 5.0-log
10
reduction of
E. coli
O157:H7. Therefore, a process for fermented RTE products that contain beef must be validated for both pathogens.

Cooked and Otherwise Processed Whole or Comminuted Meat Products

As stated above, FSIS already has made final lethality performance standards for certain RTE meat products, including RTE cooked beef, corned beef, and roast beef. In this document, FSIS is proposing to extend these performance standards to all other cooked and otherwise processed (e.g., cured) meat products. Under this proposal, establishments would be required to employ processing validated to achieve specific probabilities (Table 1) that only small numbers of
Salmonella
organisms could remain in finished cooked or otherwise processed, whole and comminuted, RTE meat products. Alternatively, an establishment could use a process validated to achieve a 6.5-log
10
reduction of
Salmonella
throughout a finished RTE meat product.

As with cooked beef, corned beef, and roast beef, the primary pathogenic microorganism of concern in these other RTE meat products has been
Salmonella.
FSIS tentatively finds that the destruction of
Salmonella
in these products will result in the destruction of most other pathogens. FSIS is not proposing to require that any particular means be used to meet the lethality standard. Cooking, for example, would not need to be the sole means by which lethality would be achieved. Other applicable treatments, such as curing or other controls, could be used in combination with cooking to achieve the required lethality.

Meat Patties

In the proposal preceding the January 1999 final rule that established performance standards for certain RTE meat and poultry products, FSIS identified
Salmonella
as the target pathogenic microorganism in fully-cooked, uncured meat patties and proposed a 5-log
10
reduction in
Salmonella
as the lethality performance standard. FSIS made a tentative finding that a 5-log
10
reduction in
Salmonella
in cooked, uncured meat patties would effectively eliminate most other bacterial pathogens of concern. Notably, compliance with the time/temperature requirements already contained in the regulations effectively achieved a 5-log
10
reduction in
Salmonella.

However, FSIS did not make final the lethality performance standards proposed for RTE comminuted meat patty products. In the course of developing the final regulation, FSIS determined that a higher lethality was likely necessary to produce RTE, uncured meat patties that would pose no health risk to consumers. The Agency could find no conclusive information demonstrating that the distributions of bacteria on ground and whole product produced under normal manufacturing conditions would present comparatively higher or lower risks to consumers. Furthermore, most, if not all, RTE meat and poultry products will be manufactured from the same supply of raw product examined in the FSIS national baseline surveys. So, using performance standards that would render any hypothetical, worst case raw product safe should be applicable to all categories of RTE meat and poultry products.

Consequently, FSIS is proposing to require that establishments achieve a 6.5-log
10
reduction of
Salmonella
in all RTE meat products, including RTE meat patties. FSIS believes that many establishments are achieving this higher lethality already, either through a cooking step or a combination of treatments. Furthermore, new and innovative processing technologies, including irradiation of raw product, should allow establishments to achieve this lethality without significantly altering the quality of their products through overcooking.

Cooked and Otherwise Processed Whole or Comminuted Poultry Products

Again, FSIS recently made final lethality performance standards for all fully cooked, RTE poultry products, such as poultry rolls. In this document, FSIS is proposing to extend these performance standards to all other cooked and otherwise processed (e.g., cured) RTE poultry products. Under this proposal, establishments would be required to employ processing validated to achieve specific probabilities that only small numbers of
Salmonella
organisms could remain in finished cooked or otherwise processed, whole and comminuted, RTE products that contain any amount of poultry. Alternatively, an establishment could use a process validated to achieve a 7-log
10
reduction of
Salmonella
throughout a finished product.

The primary pathogenic microorganism of concern in these other

RTE poultry products has been
Salmonella.
FSIS tentatively finds that the destruction of
Salmonella
in these products will result in the destruction of most other pathogens. For example,
Campylobacter jejuni
was not selected as a reference organism in RTE poultry product, even though it is present at high levels in poultry, because it is generally recognized as being very sensitive to heat. As with the analogous meat products, FSIS is not proposing to require that any particular means be used to meet the lethality standard. For example, various treatments, such as curing or other controls, can be used in combination with cooking to achieve the required lethality.

B. Stabilization

In addition to lethality standards, FSIS is proposing that processing used to produce all RTE products, other than thermally processed, commercially sterile products, and processing used to produce partially heat-treated products, meet stabilization performance standards. The proposed stabilization standards require that establishments control their production processes to prevent the multiplication of spore-forming microorganisms. Stabilization is typically achieved through cooling a product after cooking. Specifically, the Agency is proposing to require that establishments producing these products ensure that there is no multiplication of toxigenic microorganisms, such as
Clostridium botulinum,
that potentially would create harmful toxins in the product, and that there is no more than a 1-log
10
multiplication of
Clostridium perfringens
within the product.

FSIS is proposing this performance standard because the means applied to products to bring about the lethality of certain microorganisms in RTE products, particularly heat treatment, can create a model environment for the multiplication of spore-forming bacteria. The processing for many RTE products includes a heat treatment. Spores of
C. botulinum, C. perfringens,
and other spore-forming bacteria can survive cooking and, in fact, can thrive in the warm product following cooking after competitive microorganisms, such as
Salmonella
or lactic acid bacteria, have been eliminated. Anaerobic, non-refrigerated conditions also facilitate multiplication and growth of these organisms.

Similarly, during processing, partially-heat treated meat and poultry products are partially cooked and then cooled, which creates a model environment for the growth of
C. perfringens, C. botulinum,
and other spore-forming, toxigenic bacteria. Cooking by the consumer, retailer, or other end-user may not eliminate these bacteria or the toxins that they create in these products. Therefore, it is important that bacterial growth be controlled in these products to the extent possible before they reach the end consumer.

The stabilization performance standards are identical to the standards made final in the January 1999 performance standard rulemaking, cited above, for RTE products and partially-cooked poultry and meat patties. The purpose for imposing the no (zero) multiplication of
C. botulinum
standard was to ensure that harmful toxins would not be created in the product during cooling. Toxins are created only when there is multiplication of
C. botulinum,
or other spore-forming, toxigenic bacteria. When spores germinate and reach the outgrowth stage, even slight temperature abuse to the product can result in cell multiplication and, if there are sufficient numbers of cells, subsequent toxin formation. Thus, logically, ensuring no growth of these bacteria would provide the greatest amount of safety. Microscopic examination of cells can be used to determine whether cells have germinated and reached outgrowth stage.

The Agency requests comments on whether the
C. botulinum
standard should be no (zero) multiplication as proposed. The Agency also requests any data to support a tolerance in place of the proposed
C. botulinum
standard. The primary purpose for the zero growth standard is to ensure that harmful toxins will not be created in cooked product during cooling. If there were cell multiplication during cooling and sufficient numbers of cells, there could be subsequent toxin formation. Thus, ensuring no growth
C. botulinum
provides for the safety of the product with the greatest amount of confidence.

It is possible that there can be a small amount of
C. botulinum
growth within the time of a 1-log
10
relative growth of
C. perfringens.
If the relative growth of
C. botulinum
were greater than zero, but less than some small amount, the affected product could possibly be considered safe for consumption, provided it is also assumed that the initial levels of
C. botulinum
were not high. This assumption would be a reasonable one, since generally the levels of
C. botulinum
in raw meat are low. However, in this situation, the consequence of the low-level
C. botulinum
assumption being incorrect and of the possible toxin production would be severe.

It is possible that compliance with the proposed zero growth standard for
C. botulinum
could impose a significant burden on industry. Because there may be growth of
C. botulinum
during a 1-log
10
relative growth of
C. perfringens,
compliance with the proposed zero growth standard for
C. botulinum
could effectively require establishments to meet a more restrictive standard than that for
C. perfringens.
Further, demonstrating “no multiplication” by experiments (microscopic examination of cells to determine whether cells have germinated and reached outgrowth stage) could be expensive. Also, to the Agency's knowledge, there are not extensive data on which to build mathematical models for predicting the time before cell germination or outgrowth and using data from growth curves to develop predictive models for cell population growth is not propitious for demonstrating no multiplication. Usually with predictive growth models, it is very difficult or impossible to show a no occurrence event (zero-growth) with high probability. Consequently, FSIS requests comment on this issue, and data to support a possible relative growth tolerance in place of the zero growth proposed
C. botulinum
standard.

The proposed stabilization performance standard provides that any more than 1-log
10
multiplication of
C. perfringens
will adulterate the product for the following reasons: Viable counts of 10
5
or greater of
C. perfringens
/gram in finished product have been listed by the CDC as one criteria for incriminating
C. perfringens
as the causative agent of foodborne illness (Ref. 25, available for viewing by the public in the FSIS Docket Room), although foods responsible for
C. perfringens
outbreaks usually contain at least 10
6
vegetative
C. perfringens
cells per gram (Refs. 26 and 27, available for viewing by the public in the FSIS Docket Room). In the FSIS microbiological product surveys, some samples were found to contain more than 10
4
, but less than 10
5
,
C. perfringens
/gram. It is a conservative assumption with respect to public health that the great majority of
C. perfringens
in the raw product are spores. Heating activates the spores that, during the cooling, become vegetative cells that can multiply to hazardous levels. Given that there can be more than 10
4

C. perfringens
(spores) per gram on raw product, it is possible that there could be as many as 10
4
vegetative
C. perfringens
/gram of these surviving, after cooking, in the product. Therefore, the Agency, using the aforementioned CDC criteria as an upper limit that should not be exceeded, has tentatively

determined that a limit of no more than 1 log
10
growth of
C. perfringens
is appropriate to ensure that there would be no more than 10
5

C. perfringens
per gram on the finished product after cooling.

An academic researcher recently suggested to the Agency that the stabilization performance standard for
C. perfringens
should apply only to the surface of intact, whole muscle, RTE products. This researcher stated that there is no data indicating that the interior of whole muscle products would ever contain
C. perfringens
. FSIS requests comment on this issue, as well as any relevant research data.

V. Listeria monocytogenes

L. monocytogenes
grows at low oxygen conditions and refrigeration temperatures, and survives for long periods of time in the environment, on foods, in processing plants, and in household refrigerators. Although frequently present in raw foods of both plant and animal origin, it also can be present in cooked foods due to post-processing contamination. Consumption of food contaminated with
L. monocytogenes
can cause listeriosis, an uncommon but potentially fatal disease in newborns, the elderly, and persons with weakened immune systems, such as those with chronic disease, HIV infection, or persons taking chemotherapy for cancer. Listeriosis also is a major concern in pregnant women. Even though symptoms may be relatively mild in the mother, the illness can be transmitted to the fetus, causing serious illness or fetal death.

Each year, according to the FDA-FSIS draft risk assessment on
L. monocytogenes
(Ref. 28, available for viewing by the public in the FSIS Docket Room), the bacteria cause an estimated 2,493 cases of listeriosis. Of these, 2,298 persons are hospitalized and 499 persons die. The case-fatality rate is high across the whole population—20 deaths per 100 cases of illness. Epidemiologic surveillance data indicates that the case-fatality rate varies by age, with a higher case-fatality rate among newborns (<1 year) and the elderly (>60 years). For a full discussion on case-fatality rate, refer to the “Baseline Number of Listeriosis Cases and Deaths and the Potential Benefits from the Proposed Rule” section in Appendix 1.

Since 1987, FSIS has conducted a microbiological testing program in which the Agency randomly samples, in-plant, RTE meat and poultry products produced in federally inspected establishments for
L. monocytogenes
, including cooked and fermented sausages, cooked corned beef, sliced ham and luncheon meats, beef jerky, cooked uncured poultry, and salads and spreads. FSIS treats RTE products in which
L. monocytogenes
is found as adulterated under the FMIA or the PPIA (21 U.S.C. 453(g) or 601(m)). This testing of approximately 7,000 RTE product samples per year for
L. monocytogenes
is an indicator of possible public health problems, but FSIS believes that more discriminating approaches are in need of development. (A comprehensive presentation on the FSIS testing program, entitled “FSIS Ready-to-Eat (RTE) Sampling in Transition,” is available from the FSIS Docket Room.)

During the late 1980's,
L. monocytogenes
emerged as a problem in deli meats and other processed food products. FSIS and the Food and Drug Administration (FDA) worked with processing plants to improve their procedures and emphasized the “zero” tolerance (no detectable level of viable pathogens permitted) for the pathogen in RTE products. Between 1989 and 1993, the rate of illness from
L. monocytogenes
declined 44 percent. This reduced incidence of foodborne listeriosis remained level until recently.

In the fall of 1998, state health departments and the CDC began investigating an increased number of reported cases of illness due to
L. monocytogenes
. CDC and state and local health departments identified the vehicle of transmission as hotdogs and possibly deli meats produced by one manufacturer under many brand names. On December 22, 1998, in response to reports of illness, the manufacturer voluntarily recalled specific production lots of these products that might be contaminated. Subsequently, CDC and FSIS investigators isolated the outbreak strain of
L. monocytogenes
from an opened and a previously unopened package of hotdogs manufactured by one plant. In addition, a different strain of the pathogen was isolated from unopened packages of deli meats produced at the same plant. CDC has since reported 101 illnesses, 15 adult deaths, and 6 stillbirths or miscarriages associated with this outbreak.

With this outbreak in mind, on May 7, 1999, the FDA, in consultation with FSIS, announced plans to conduct a risk assessment to determine the prevalence and extent of exposure of consumers to foodborne
L. monocytogenes
and to assess the resulting public health impact of such exposure (64 FR 24661). FDA and FSIS published this draft risk assessment for comment on January 19, 2001 (Ref. 28, available for viewing by the public in the FSIS Docket Room). Significantly, it identifies certain RTE meat and poultry products, among the food products assessed, as posing a relatively high health risk of listeriosis to consumers because of potential RTE product contamination by
L. monocytogenes
.

In this document, FSIS is proposing regulatory requirements and considering other options to address the relatively high risk ranking of these RTE meat and poultry products. Significantly, the draft risk assessment was designed to estimate the predicted relative risk of serious illness and death that may be associated with consumption of different types of ready-to-eat foods. The draft risk assessment document, unlike more complete risk assessments, did not attempt to account for the level or sources of contamination of ready-to-eat meat and poultry products in a farm-to-table approach such as during processing in Federally inspected facilities. Rather, the draft risk assessment accounted for the retail foodborne exposure to human listeriosis (i.e., after the ready-to-eat product is out of the control of the Federal establishment). The data included in the draft risk assessment were gleaned from both international and domestic sources, with FSIS providing a substantial amount of data from its various microbiological programs associated with Federally inspected meat and poultry. The draft risk assessment was designed to address data only associated with listeriosis, providing a distinction between foodborne illness associated with mild, flu-like symptoms (referred to as listerial gastroenteritis) and severe and life-threatening outcomes (i.e., listeriosis). For this reason, some Federally inspected meat and poultry products were not addressed in the draft risk assessment (e.g., canned meat and poultry and partially- and fully-cooked meat patties). Except for the canned products and the meat patties, FSIS believes that the risk assessment addresses the remaining meat and poultry products contained in this proposed rule (i.e., frankfurters, dry/semi-dry fermented sausages, deli meats, and pate

and meat spreads).

A. Proposed Requirements for Controlling
L. monocytogenes

In the risk assessment, FDA and FSIS note that although pasteurization or cooking by an establishment will kill
L. monocytogenes
, there is risk of recontamination of RTE foods during processing, after the lethality is applied (Ref. 28, (Interpretive Summary, p. 24; Exposure Assessment, p. 24), available for viewing by the public in the FSIS

Docket Room). Significantly, FDA, FSIS and other authors point out, that deli meats in particular are most likely to be recontaminated by
L. monocytogenes
after cooking, during processing such as slicing (Ref. 28 (p. 167); Ref. 32; Ref. 33; all available in the FSIS Docket Room), although no data were available to distinguish between the risks of slicing product in a retail environment rather than an official establishment.
3

3
Although pate and meat spreads also are identified in the draft FDA/FSIS risk assessment as having high predicted relative risk of causing listeriosis on a per serving basis, much of the reported foodborne outbreaks are associated with foreign populations. However, FSIS is aware of one foodborne outbreak in the U. S. involving pate produced in a federally inspected facility in 1999. In this outbreak, pate was prepared by cooking the product in open containers and then over-wrapped with film. Product was then distributed to multiple states and sold in gourmet shops. The pate was implicated as the food vehicle for
L. monocytogenes
.
L. monocytogenes
was cultured from an unopened package of pate at retail. FSIS was not able to determine whether
L. monocytogenes
was present in the unopened package as a consequence of underprocessing (i.e., inadequate lethality) or post-lethality contamination.

FSIS is proposing to require that all establishments that produce RTE meat and poultry products conduct environmental testing of food-contact surfaces for
Listeria spp
., after lethality treatment and before final product packaging, unless they have identified
L. monocytogenes
as a hazard reasonably likely to occur and so have incorporated into their HACCP systems one or more controls validated to eliminate it from their products. This testing will verify that an establishment's Sanitation Standard Operating Procedures (Sanitation SOPs) are preventing direct product contamination by
L. monocytogenes
after the lethality treatment, thus addressing the risk assessment assertion that RTE foods often are recontaminated by
L. monocytogenes
after lethality is applied.

After an establishment finds one of its food contact surfaces to be positive for
Listeria spp
., it must take corrective actions defined in its Sanitation SOP that must include product testing, as well as any other activities that it deems necessary to determine and demonstrate that the affected lot or lots of product are not adulterated with
L. monocytogenes
. The establishment must have in place procedures: to determine which lots of product might be affected; to hold, sample, and test that product; and to dispose of affected product appropriately.

Establishments that have identified
L. monocytogenes
as a hazard reasonably likely to occur in their HACCP plans and that have consequently established CCPs for
L. monocytogenes
would be exempt from this mandatory testing requirement. For example, establishments that produce thermally processed, commercially sterile, hermetically-sealed (canned) products should be relatively unaffected by this proposed requirement. Neither should many other establishments that produce meat and poultry products that receive lethality treatment in their final packaging, such as beef cooked in an impervious bag. In most cases, these and similar establishments would need only to modify their HACCP plans to reflect that
L. monocytogenes
is likely to occur at some point during their processing, but their existing CCPs for lethality would eliminate the pathogen.

FSIS believes that
L. monocytogenes
contamination is reasonably likely to occur in the production of all RTE meat and poultry products. On May 26, 1999, FSIS published in the
Federal Register
a Notice advising manufacturers of RTE meat and poultry products of the need to reassess their HACCP plans to ensure that the plans are, in fact, adequately addressing
L. monocytogenes
(64 FR 28351). If this reassessment revealed that
L. monocytogenes
was a hazard reasonably likely to occur in an establishment's production process, the Notice stated that the establishment must address the hazard in its HACCP plan.

FSIS acknowledges, however, that there may be certain processing environments in which
L. monocytogenes
is not a hazard reasonably likely to occur. In such environments, verification through testing that the establishment's Sanitation SOP is controlling
Listeria spp.
would be necessary, at a minimum.

Notably, Tompkin, et al., have recommended plant-wide environmental testing that

* * * should focus on a non-pathogenic indicator such as
Listeria spp.
or
Listeria
-like organisms * * * , because these organisms will be found more frequently in the environment than
L. monocytogenes
and because test results are available more quickly.

(Ref. 29, available for viewing by the public in the FSIS Docket Room)

FSIS agrees, although the Agency is proposing to require only the testing of food contact surfaces. Were an establishment to find
Listeria spp.
on a food contact surface, that finding would be indicative of a sanitation problem that could cause product adulteration, even though the contaminant on the surface may not be
L. monocytogenes
.

FSIS is proposing to require that establishments without HACCP controls for
L. monocytogenes
test food contact surfaces for
Listeria spp.
at one of the following frequencies, depending on establishment size:

• If the plant is large, at least four tests, per line, per month;

• If the plant is small, at least two tests, per line, per month;

• If the plant is very small, at least one test, per line, per month;

FSIS is proposing to employ the same Small Business Administration (SBA) size standards that it used to determine the implementation dates for its HACCP/Pathogen reduction final rule. Large establishments would be defined as all establishments with 500 or more employees. Small establishments would be defined as all establishments with 10 or more employees but fewer than 500. Very small establishments would be defined as all establishments with fewer than 10 employees or annual sales of less than $2.5 million.

These frequencies ensure a very minimal amount of testing and, because they are progressive, mitigate some of the economic impact on small businesses. FSIS has not been able to correlate risk of product contamination with production volume or establishment size. However, assuming that large establishments produce a greater volume of product than do small establishments, and that a large insanitary establishment would be more likely to contaminate more product and thus pose more risk to the public health, FSIS is proposing to require large plants to test more often. Because these frequencies are not based on research but represent what the Agency believes to be minimal levels, FSIS requests comment on these proposed testing frequencies, their efficacy in preventing product adulteration, and the costs to industry. FSIS also specifically solicits information the current state of knowledge about the relationship between
Listeria spp.
on food contact surfaces and
L. monocytogenes
on the product; the appropriate timing of the test (pre-start-up or post-start up), seasonality and other risk based considerations that might be important in creating effective testing protocols; and, the testing methodologies that are currently available and the current practice and use of the tests by industry or others Agencies. FSIS will use the information to develop testing frequencies and methodologies that protect the public health, while providing flexibility to establishments. FSIS plans to hold one or more technical conferences during the comment period for this proposed rule, at which these testing issues and other can be discussed. FSIS plans to provide for discussion of the latest testing

methodologies, including those used by other Federal Agencies and industry, as well as an ongoing ARS study on the testing of intact RTE product for
L. monocytogenes
.

FSIS is proposing to require that establishments take certain actions after food contact surfaces test positive for
Listeria spp.
After an establishment finds one of its food contact surfaces to be positive for
Listeria spp.
, it must take the corrective actions defined in its Sanitation SOP. According to § 416.15(a), Sanitation SOP corrective actions may include “procedures to ensure appropriate disposition of product(s) that may be contaminated, restore sanitary conditions, and prevent the recurrence of direct contamination or adulteration of product(s).”

The presence of
Listeria spp.
may be indicative of serious sanitation problems, especially if positive findings recur. Further,
Listeria spp.
positives on food contact surfaces indicate a potential for product adulteration by
L. monocytogenes.
Therefore, an establishment's corrective actions following a positive must include product testing and any other activities that it deems necessary to determine and demonstrate that the affected lot or lots of product are not adulterated with
L. monocytogenes.
The establishment must have in place procedures: to determine which lots of product might be affected; to hold, sample, and test that product; and to dispose of affected product appropriately. FSIS acknowledges that some establishments would have to modify their Sanitation SOP corrective actions to include these elements.

FSIS requests comments on the proposed testing provisions and any data that would support the approach proposed. FSIS requests comments concerning whether
Listeria
positive test results on different food contact surfaces should be treated differently (e.g., positives on food contact surfaces that have undergone listericidal treatment versus other food contact surfaces). FSIS also requests comments on whether it should establish more specific requirements regarding product sampling and testing following a finding of
Listeria spp.
on a food contact surface. And, FSIS request comment on whether it should allow establishments that find
Listeria spp.
on a food contact surface to determine if the positive sample is in fact
L. monocytogenes
before having to initiate product testing.

If a sampled lot is found to be positive for
L. monocytogenes,
and is already in commerce, it will be subject to recall. Further, if product is found to be positive for
L. monocytogenes,
the establishment likely will need to establish controls within its HACCP plan for
L. monocytogenes.
Also, reoccurring positives for non-pathogenic
Listeria spp.
may indicate that the establishment has a serious sanitation problem, even if
L. monocytogenes
is never found. FSIS enforcement action will vary depending on the establishment's efforts to correct its sanitation and processing problems and its disposition of affected product. FSIS acknowledges that establishments that develop one or more CCPs to control
L. monocytogenes
would not necessarily be testing for
Listeria spp.
to verify the efficacy of their Sanitation SOPs and requests comments on this issue.

The two provisions for
Listeria
control contained in this proposed rule (i.e., Sanitation SOPs and HACCP) require specific daily action regarding controls to ensure product is not adulterated. FSIS does not, at this time, consider control programs outside of Sanitation SOPs and HACCP to be sufficient for controlling hazards associated with post-lethality contamination with
Listeria
in the manufacturing of RTE meat and poultry products microbiological results and documentation of corrective and preventive actions generally are not provided to FSIS. FSIS has received a petition from a group of industry organizations regarding the issue of prerequisite programs. FSIS will address this issue separately from this proposed rule. In addition, FSIS will be further addressing this issue as part of its response to an Office of Inspector General report on HACCP implementation (Ref. 35, available in the FSIS Docket Room and at the FSIS web page,
http://www.fsis.usda.gov
).

With any final action FSIS will publish guidance to establishments regarding testing frequencies and methodologies and appropriate corrective actions following food-contact surface positives. FSIS also will publish guidance regarding available listericidal interventions establishments can implement as CCPs. FSIS expects to make draft guidance documents available after publication of this proposed rule and as information becomes available in order to provide establishments with appropriate guidance regarding sampling and testing to verify sanitation procedures. FSIS will consider comments on this draft guidance in developing any final regulations. These draft guidance materials will be clearly identified as guidance materials and not as regulatory requirements. FSIS expects to post these guidance materials to the FSIS web page
(http://www.fsis.usda.gov)
and will make the documents available free of charge via the Constituent Update (see section XIV Additional Public Notification) and the FSIS Docket Room.

Eventually, FDA and FSIS may allow establishments to treat RTE products with ionizing radiation. If applied within a HACCP system, irradiation could eliminate
L. monocytogenes
from a RTE product. FSIS also is aware that industry is developing edible, antimicrobial coatings that could be applied to RTE meat and poultry after cooking or other lethality treatments. However, FDA has not yet approved any of these coatings for meat and poultry. FSIS also will make available its directives to inspection personnel that will explain how to verify whether an establishment has implemented a testing regime sufficient to verify the efficacy of Sanitation SOPs in preventing direct product contamination by
L. monocytogenes
prior to the effective date of any final regulation.

Finally, FSIS notes that on January 13, 2000, it received a petition from the Center for Science in the Public Interest (CSPI) requesting that FSIS require all establishments that produce RTE meat and poultry products to conduct environmental testing for
Listeria spp.
and product testing for
L. monocytogenes.
FSIS will respond to this petition completely along with other public comments submitted in response to this proposal. CSPI also requested that FSIS require RTE products produced by establishments without CCPs for
L. monocytogenes
to bear warning labels. FSIS discusses this request in the following section and also will respond more completely in any final action that stems from this proposal.

B. Shelf-Life and Labeling

In the petition discussed above, CSPI also requested that FSIS require establishments that have not incorporated microbial testing for
L. monocytogenes
into their HACCP plans to label their products so as to alert “consumers that the products may be contaminated and should not be eaten by at-risk consumers without reheating.” FSIS will respond to this petition fully in any final action stemming from this proposed rule.

FSIS considered, but did not propose in this document, the option of requiring that the labeling of certain RTE meat and poultry products state the product's shelf-life, and that shelf-life be based on product safety (“use-by” date labeling). If after processing, a RTE product that could support growth of
L. monocytogenes
were to be

recontaminated by even a single cell of the pathogen, that cell could multiply during storage at refrigeration temperatures to levels that could pose a risk of illness to vulnerable individuals (e.g., pregnant women, the elderly, or the immunocompromised). “Use-by” date labeling may provide further reductions in risk of listeriosis if the labeling increases the likelihood that high-risk RTE products would be consumed before very low levels of
L. monocytogenes,
undetectable at the establishment, could grow to dangerous levels.

FSIS is not proposing to require “use-by” dates on the labels of any RTE products at this time because further information regarding the potential effects of use-by date labeling is needed. For instance, there is sparse information on current consumer understanding of use-by date labeling, the likelihood that consumer practices will change, and on the effect of changes in consumer behavior on listeriosis cases. Similarly, FSIS currently does not possess all the data necessary to assess the reduction in risk that will occur from this change. Also, FSIS does not have information concerning how use-by date labeling would affect the production and shipment patterns of labeled ready-to-eat meat and poultry products and the structure of the industry. FSIS requests comments on all of these issues and on the feasibility of requiring “use-by” date labeling on RTE meat and poultry products. Significantly, FDA and FSIS will present “use-by” date labeling issues to NACMCF for their review. FSIS has conducted a more thorough analysis of use-by date labeling in Appendix 1, Compliance with Executive Order 12866, under the “Alternatives” section.

Related to “use-by” date labeling is the issue of consumer preparation of hotdogs and similar RTE foods. In the draft risk assessment, FSIS and FDA state that “the factor that has the greatest effect on the predicted health impact of frankfurters is the extent of post-retail reheating by the consumer” (Ref. 28 (p. 161); Ref. 33; Ref 34; all available in the FSIS Docket Room). Obviously, testing for
L. monocytogenes
in the establishment will not directly affect consumer preparation of frankfurters or other RTE foods. However, if in-plant testing verifies that establishments are effectively preventing the contamination of frankfurters and other RTE products by
L. monocytogenes,
consumer preparation or handling of these RTE products will no longer be so inappropriately crucial to ensuring their safety. Furthermore, once FSIS is more confident that establishments are adequately addressing the safety of their RTE products, especially for frankfurters and deli meats, throughout the shelf-life of their products, FSIS will consider modifying its consumer message to vulnerable populations and remove the current recommendation for these populations to either not consume these RTE products or to fully re-cook these products before consuming them.

Finally, as discussed below, FSIS is proposing that the labeling of RTE products state that the product requires refrigeration after opening, as applicable. Current regulations require that labels of perishable products include such instructions, but the Agency is proposing to expand the required label instructions to include RTE shelf-stable products that require refrigeration after opening. FSIS also considered proposing to change the “keep refrigerated” and the “refrigerate after opening” statements (see proposed in §§ 317.2(k) and 381.125(a)) to reflect the guidance developed by FDA on February 24, 1997 (62 FR 8248). In the guidance, these statements were modified to read “Important Must Be Kept Refrigerated to Maintain Safety” or “Important Must Be Refrigerated After Opening To Maintain Safety.” FDA provided this guidance in response to the recommendations from the NACMCF, the National Food Processors Association, the Association of Food and Drug Officials, and the CDC regarding the labeling of foods that need refrigeration. FDA stated in this policy document that “[t]his guidance, which represents FDA's policy on adequate safe handling instructions for food, should reduce the likelihood of temperature abuse of certain foods by consumers, and it is intended to reduce the potential for foodborne illness and death.” FSIS is not proposing to require these provisions because further information regarding the potential effects of this labeling is needed. FSIS requests comment on the statements and their appropriateness for RTE meat and poultry products which are not shelf stable.

VI. Thermally-Processed, Commercially Sterile Products

Thermally-processed, commercially sterile meat and poultry products generally have a water activity above 0.85 and have received a thermal process either before or after being packed in a hermetically sealed container. They are typically canned, although other types of packaging can be used. The thermal process renders the product shelf-stable and commercially sterile, that is, free of microorganisms capable of growing in the product in nonrefrigerated conditions (temperatures over 50 °F or 10 °C), under which the product will be held during distribution and storage, until consumed.

Sections 318.300 to 318.311 and 381.300 to 381.311 of the regulations prescribe the exact means by which official establishments must produce thermally processed, commercially sterile meat and poultry products. These regulations include detailed requirements regarding containers and container closures, equipment specifications and operations, measurements and instrument calibration, recordkeeping and record review, corrective actions in the case of processing deviations, finished product inspection, personnel training, and product recalls. They also require that official establishments implement process schedules validated to render treated meat and poultry commercially sterile and shelf-stable. These process schedules must be developed or validated by processing authorities, persons or organizations with expert knowledge of thermal processing requirements for foods packaged in hermetically sealed containers.

Processors that produce thermally processed, commercially sterile meat and poultry products also must meet all other regulations applicable to meat and poultry establishments, such as sanitation and HACCP requirements. Significantly, however, under § 417.2(b)(3), FSIS exempts producers of thermally processed, commercially sterile products from addressing in their HACCP plans “food safety hazards associated with microbiological contamination.” FSIS granted this exemption in response to comment on the proposal to require HACCP systems:

FSIS agrees that the microbial hazards associated with canned meat and poultry products are eliminated by complying with the regulations in 9 CFR Secs. 318.300-311 and 381.300-311. These regulations are based on HACCP concepts and provide for the analysis of thermal processing systems and controls to exclude microbial hazards. Accordingly, the final rule provides that HACCP plans for thermally processed/commercially sterile products do not have to address the food safety hazards associated with microbiological contamination if the product is produced in accordance with the canning regulations. However, because the current regulations exclusively address microbial hazards, processors of canned meat, meat food and poultry products must develop and implement HACCP plans to address chemical and physical hazards that are reasonably likely to occur.

(61 FR 38824)

The regulations governing the processing of thermally processed, commercially sterile meat and poultry products are, in a sense, a prescribed HACCP system that official establishments must implement along with controls to address other hazards not addressed in those regulations. Maintaining this prescriptive regulatory approach to a single category of meat and poultry products, however, is inconsistent with FSIS's other regulatory initiatives intended to grant industry maximum flexibility to innovate in processing, while clarifying industry's responsibility and accountability for the safety of meat and poultry products. Therefore, FSIS is proposing to replace the prescriptive regulations governing thermally processed, commercially sterile products with performance standards. FSIS is also proposing to remove §§ 320.2(b)(6) and 381.175(b)(3) because they refer to recordkeeping requirements in the canning regulations that FSIS is proposing to eliminate. FSIS has discussed this proposed action in previous documents, including the final rule that established the HACCP requirements:

The current canning regulations contain numerous prescriptive features, including extensive FSIS involvement in the decision making process, that are inconsistent with the philosophy underlying HACCP. In the advance notice of proposed rulemaking “FSIS Agenda for Change: Regulatory Review” (60 FR 67469; December 29, 1995), FSIS stated its intention to convert the canning regulations to performance standards, which are more consistent with HACCP.

(61 FR 38824)

FSIS is proposing lethality performance standards to ensure the elimination or control of the pathogen
C. botulinum
in thermally processed, commercially sterile meat and poultry products. FSIS also is proposing a revised requirement ensuring the commercial sterility of these products. This requirement is consistent with the existing shelf-stability/commercial sterility definitions in § 318.300(u) and 381.300(u) and the FDA regulations for commercial sterility of canned products contained in 21 CFR 113.3(e).

A. Lethality

FSIS is proposing different lethality performance standards, depending on whether the product is a low-acid product or a product in which pathogen growth is controlled by acidification or factors other than the thermal process. A low-acid, thermally processed, commercially sterile product is a canned or other hermetically sealed product in which any component has a pH value above 4.6 and a water activity above 0.85. Such products include canned poultry and canned uncured meat products, such as beef stew and chili con carne, and certain canned cured meats, such as vienna sausages and corned beef. An acidified thermally processed, commercially sterile product is a canned product that has been formulated or treated so that every component of the finished product has a pH of 4.6 or lower, usually within 24 hours after the completion of the thermal process, but sometimes longer. Such products include spaghetti sauce with meat and meat with tomato sauce. In addition, there are some canned, hermetically sealed products in which pathogen growth is controlled by factors other than the thermal process, such as a heat treatment in combination with salt or nitrite (e.g., canned luncheon meat).

FSIS is proposing to require that an establishment's process for producing a low-acid canned product result in a probability of 10
-9
or less that there are spores of
C. botulinum
in a container of the product that are capable of growing, assuming an initial load of ≤ 1000 spores per container. Alternatively, the establishment may achieve a 12-log
10
reduction of
C. botulinum.
A process carried out for a certain number of minutes at a given temperature that reduces
C. botulinum
by a factor of 12 decimal units, often referred to in the canning industry as a “botulinum cook,” is one that meets a 12-log
10
standard, also known as a 12-D standard. A 12-D process has been demonstrated to be sufficient to destroy
C. botulinum
in a low-acid canned product. Under this proposal, the level of safety that a process other than a 12-D process would have to achieve would be a probability of 10
-9
or less of any
C. botulinum
spores in a container of the product that are capable of growing, assuming an initial load of ≤1000 organisms.

The 12-D concept arose from studies on the thermal resistance of
C. botulinum
conducted in the early 1920's by scientists of the National Canners Association (predecessor of the National Food Processors Association). These scientists inoculated a phosphate buffer with spores of the most heat-resistant strain of the organism then known. They determined, by extrapolating from the exponential survival curve for the organism, the temperature and duration of the heat process necessary to reduce the population from 6 × 10
11
spore/unit to less than one spore/unit. Subsequent studies on products inoculated with
C. botulinum
and other organisms essentially confirmed the results of these studies.

These products undergo a botulinal cook to achieve an acceptable safety level. It should be noted that the intensity of the process is not related to the actual number of
C. botulinum
organisms that may be in the product. That number is usually very low in a meat product (less than a spore per kilogram). So the 12-D process provides a tremendous safety margin to consumers.

The level of safety achieved by a 12-D process in low-acid canned products is understood by thermal processing experts to be a 10
-9
probability of any live botulinum organisms (Refs. 30-31, available for viewing by the public in the FSIS Docket Room). That means that the odds are one in a billion that a can is contaminated with the organism. This result is arrived at by assuming that a process that reduces botulinum spores by 10
-12
—a 12-D process—is applied to a test pack of product inoculated with 10
3
spores per unit. The probability that any containers that are subjected to the process harbor spores capable of growing is 10
-9
. Thus, FSIS is proposing to require that establishments producing low-acid products achieve a probability of 10
-9
or less that there are spores of
C. botulinum
in a container that are capable of growing or a 12-log
10
reduction of
C. botulinum.

FSIS is proposing to require that the processing of acidified low-acid products and of some cured products and other canned products in which pathogen growth is controlled by factors other than the thermal process, prevent multiplication of
C. botulinum.
For these products, processing (formulation and environment) must prevent growth rather than achieve any specific decimal reduction of
C. botulinum.
Therefore, there can only be one level of performance for acidified low-acid products and other thermally processed, commercially sterile products in which pathogen growth is controlled by factors other than the thermal process—prevention of
C. botulinum
multiplication. However, the prevention of multiplication can be achieved by a variety of methods.

Acidified low-acid meat and poultry products are generally acidified by ingredients, such as tomato sauce, or by additives, such as glucono-delta-lactone, which increase the acidity (i.e., lower the pH) of the products. The acidity of these products (pH at or below 4.6) is sufficient to prevent the germination of
C. botulinum
and other bacterial spores. The heat processing of these products does not include a botulinum cook or retort but is achieved at pasteurizing

temperatures below 100 °C. (212 °F.) and is sufficient to kill or inactivate molds, yeasts, and vegetative bacterial cells. This processing is important because, if canned acidified foods are contaminated by yeast or mold, the pH of the foods could be raised above 4.6, thus providing an environment for possible
C. botulinum
growth. These products—spaghetti sauce, for example—can be heat-treated before being placed in a container (i.e., hot-filled) rather than retorted and still achieve commercial sterility.

Other thermally processed, commercially sterile products can be rendered commercially sterile by a heat treatment in combination with other factors. For example, the shelf-stability of canned luncheon meat is a combined effect of heat treatment, the presence of nitrite and salt, and a low pre-processing level of
C. botulinum.
A 10-percent salt concentration or about 2 tenths of a percent of nitrite in the product formulation is usually considered sufficient to inhibit growth of the organism. The shelf-stability of dried meat-filled pasta results from a heat treatment and a water activity of less than 0.92 in the product. (Water activity is a measure of free moisture, or water available for microbial growth, in a food; the lower the number, the less moisture.)
C. botulinum
and other spore-forming organisms cannot grow at water-activity levels below 0.93. The heat treatment of these products destroys the vegetative cells of both pathogenic and nonpathogenic organisms, and the outgrowth of spores is prevented by the other inhibiting factors.

B. Commercial Sterility

FSIS also is proposing a specific requirement that all thermally processed, commercially sterile products, in fact, be commercially sterile and hermetically sealed. This requirement is consistent with the existing shel

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A01-4420. Public record. Not legal advice.