Scientific Affairs Advisory Committee Meeting Westin Diplomat Resort & Spa 3555 South Ocean Drive Hollywood, FL October 3, 2006

1:00 p.m. Opening Remarks Review of Minutes from March 15- 16, 2006 Meeting Discussion of the Agenda Dane Bernard, Committee Chair

Lm Intervention Updates FSIS – NTO Approval for Benzoate/Propionate/Sorbate FDA – FSIS Approval for Bacteriophage Octi-Gon update Tracie Sheehan, Sara Lee Corp. Tim Freier, Cargill Inc. Skip Seward, AMI Peter Bodnaruk, Ecolab, Inc.

Retail Deli Task Force on Lm Randy Huffman

FSIS Response to AMI Comments on Listeria Risk Assessment Update on Petition for Regulatory Limit for Lm – FSIS and FDA Petitions Randy Huffman

NACMCF Report on Campylobacter Methods Wafa Birbari, Sara Lee Corp. Tim Freier

Update on NAS Dioxin Review Randy Huffman

3:00 p.m. BREAK

IARC Update and Discussion on Nitrite Alternative Research Andy Milkowski

Preparation for WCRF 2007 - Discuss Concept of a Meat and Health Roundtable Jim Hodges, AMI Randy Huffman

CAST Report on Use of GMO Feed for Livestock Randy Huffman Dane Bernard

Antibiotic Resistance and NPB On-farm GMP Program Steve Larsen, National Pork Board (invited)

MDR Salmonella Update – FSIS View of Impact of Human Salmonella Serotypes in Production Facility – “Out of Control” Scenario Tim Freier

Mycobacterium avium subspecies paratuberculosis – NACMCF Update Tim Freier

Status of AMIF Collaboration with CDC on O157 Publication Randy Huffman Dane Bernard

AMIF Research Update Randy Huffman

USDA -MARC Research Update O157/STEC and Other Food Safety Research Mohammad Koohmaraie, ARS-MARC

Transfer of SAC Chairmanship Dane Bernard Tracie Sheehan

Discuss Future SAC Meeting Date and Location Discuss Future Joint Meetings with FMI Food Protection Committee Tracie Sheehan

5:00 p.m. Adjourn Minutes from March 15-16, 2006 Meeting

MEMORANDUM FOR AMI SCIENTIFIC AFFAIRS ADVISORY COMMITTEE

FROM: RANDY HUFFMAN, DANE BERNARD, JOHN BUTTS

SUBJECT: MINUTES – MARCH 15 - 16, 2006 MEETING, WASHINGTON DC

COMMITTEE MEMBERS IN ATTENDANCE:

Wafa Birbari, Sara Lee Corp. Kathleen O’Donnell, Wegmans Michael Bradley, Premium Standard Farms Kevin Ladwig, Johnsonville Sausage, Inc. John Butts, Land O’Frost Ann Marie McNamara, Silliker Inc. Jackie Caplinger, Kunzler, Inc. Phil Minerich, Hormel Foods Corp. George Cocoma, Professional Resource Org. Alan Oser, Hatfield Quality Meats Warren Dorsa, John Morrell and Co. Dennis Stifler, Coleman Natural Meats Bill Eisel, Fresh Mark Inc. Bill Sveum, Kraft / Oscar Mayer Margaret Hardin, Boars Head Provisions Steve Quickert, Kraft / Oscar Mayer Brian McFarlane, Tyson Fresh Meats Patricia Wester, SGS

GUESTS (in order of attendance during the meeting): Dr. David Goldman, Associate Administrator OPHS-FSIS Phil Derfler, Assistant Administrator, OFO-FSIS Dr. John Sofos, Colorado State University Dr. Martin Wiedmann, Cornell University FMI Staff and FMI Food Protection Committee members (see Appendix III for list of FMI attendees) Scott Brennan, Ecolab

AMI STAFF: Randy Huffman Skip Seward

Day 1

The AMI Scientific Affairs Advisory Committee (SAC) met on March 15 - 16, 2006, at the AMI Office in Washington, DC. The materials from the meeting have been posted as a PDF file on www.meatami.com with “members only” security access. The file can be located under the “committee resources” section, under the Scientific Affairs section, key documents.

John Butts, former Chair of SAC, filled in for Chairman Dane Bernard who was unable to attend. John opened the meeting with a welcome and an overview of the agenda, which was accepted.

FSIS Update The first item on the agenda was a FSIS science policy update from Dr. David Goldman, and Phil Derfler, of FSIS. Highlights of this discussion are provided below.

Baseline data collection update: Microbiological (including E. coli O157) baseline for Raw Ground Beef Components: ƒ Started in mid November 2005 ƒ The goal is 250 samples/month or 3000 samples/year ƒ No positive results for E. coli O157 to date ƒ Additional general microbiological data is being collected in companion samples and to date have not been reported by the 3rd party lab

Future Raw Ground Beef Components baselines are being contemplated at the direction of the National Advisory Committee for Microbiological Criteria for Foods (NACMCF). These include the following potential components: AMR, LTRB, Imported beef, weasand / head / cheek meat. No timeline has been set by the agency at this point with respect to these baselines.

Young chicken baseline: ƒ 90 day “shakedown” period will commence in mid-April ƒ Will be a 12 – 36 month sampling period ƒ Will follow similar format to O157 baseline data collection ƒ Prevalence of Salmonella, including serotype, as well as quantitation of general microbiological data

Risk Assessments currently in progress at FSIS ƒ Salmonella in Raw products o Raw Beef o Raw poultry ƒ Further enhancement of Lm verification model is a high priority o Data gaps – validation of interventions is a key data gap recognized by the agency ƒ Harvard BSE reassessment is a key priority o A public meeting on the updated BSE risk assessment will be held within 2 months ƒ E. coli O157 risk assessment – will incorporate plant characteristics to stratify producers of ground beef o Past history o Production volume ƒ NAFS Listeria deli study, conducted by University of Tennessee, Auburn University, Michigan State University and University of California, Davis. o Comparative risk assessment of Listeria in retail vs. plant o Results to be presented during summer 2006 meetings, including the Conference on Food Protection. ƒ Building risk assessment for cross contamination beyond the plant floor to consumer practices

Note: AMIF requested a meeting with FSIS risk assessment division (Janell Kause) to discuss the AMIF submission of new data and FSIS risk assessment critique by Exponent Inc. This meeting will occur in Early May with key FSIS staff.

Research priorities ƒ While FSIS does not have authority to conduct research, they do for time to time establish research priorities based upon the needs of the agency ƒ December, 2004 last time FSIS provided priorities that were shared with sister agency, ARS ƒ FSIS developing improved relationship with CSREES ƒ In early March 2006, FSIS met with ARS for annual research sharing meeting. Below are some key highlights from that meeting: o Heavy focus on Salmonella at this meeting o Attribution of sources of Salmonella to human illness contribution was a key issue o Methods for Salmonella enumeration discussed o New methods for characterization of pathogens were discussed, i.e. genomics o Food defense – detection methods, vulnerability assessments ƒ Residues and HACCP – also discussed with ARS o ARS to work on live animal screening methods o Veal calf issue has generated this need – medium priority issue, but important o Agency is concerned about future violations related to residues ƒ FSIS has discussed the need for a consistent cycle of communicating research needs.

Dioxin Reassessment ƒ NAS review of the EPA Dioxin risk assessment is set for end of April (recently postponed until June, 2006) ƒ NAS is putting the finishing touches on the peer review questions ƒ FSIS will not get much of a heads up on the content of the report, with an embargoed release of the report 1 day in advance to FSIS ƒ Interagency Working Group has developed a Q&A document ƒ HHS – USDA – EPA coordinated response is desired ƒ FSIS response is unknown at this point since they have not yet viewed the NAS report and will have limited time with the document ƒ The recommendation from NAS could be the risk assessment should be sent back to EPA, but is unlikely to go that extreme ƒ In terms of FSIS response, the agency tentatively plans to conduct a follow up survey of dioxin in livestock species – likely 2007 – 2008. This will be part of the agency response, i.e. they will continue to conduct periodic surveys on about a 5 year basis

NACMCF ƒ Meeting in Arlington, VA, mid March. ƒ The Subcommittee on Safe Cooking Temperatures for Poultry has been meeting and they will finalize the report. This will be presented to the full committee on Friday ƒ The NACMCF will receive new work charges during the plenary session. The following are on the list of potential charges: o MAP – mycobacterium paratuberculosis o Other potential charges, not final ƒ Determination for micro sampling methods – mapping studies for FSIS baseline studies ƒ Determination of appropriate microbial subtyping, i.e. genomics ƒ Non-O157 STEC’s (CDC rates show these going up) ƒ Change reporting for O157:H7 to include non-O157 STEC • This could really change the regulatory issue, but NACMCF will be given the charge Dr. Goldman and Mr. Derfler fielded questions from the group and departed the meeting.

Preparation for Joint FMI meeting on Lm at retail

The committee had discussions concerning preparation for the joint meeting with FMI Food Protection Committee. After working lunch, the group traveled to Hogan and Hartson for the joint meeting.

The joint meeting agenda is attached as an appendix to these minutes. Presentations were made by Bill Sveum, Kraft-Oscar Mayer, John Butts, Land O’Frost, and Tracie Sheehan, Sara Lee Foods on various in-plant aspects of Listeria control. Roger Hancock of Albertsons provided a retail perspective on Listeria control. Dr. Martin Weidman and Dr. John Sofos provided an overview of research efforts related to evaluating the Listeria risk in retail delis. (all PowerPoint presentations are available for download at http://www.meatami.com//Template.cfm?Section=Scientific_Advisory_Committee&Template=/Me mbersOnly.cfm&NavMenuID=615&ContentID=4504&DirectListComboInd=D The proposed AMIF funded retail deli study was presented by Dr. Ann Marie McNamara of Silliker Laboratories and significant discussion ensued. The outcome of these discussions was the suggestion that AMI and FMI should establish a small working group to determine what the needs of each group are and design a study to address those needs.

During the wine and cheese reception, Randy Huffman gave a presentation on the current issues involving the use of low-oxygen carbon monoxide MAP with fresh meat. This was followed by an extensive question and answer session.

Day 2

The meeting began with a wrap-up of day 1 joint meeting with FMI Food Protection Committee. The detailed agenda is attached as an appendix to this document. Generally, SAC members expressed that the meeting was a very worthwhile endeavor and good dialog was had between the two groups.

With respect to the Listeria at retail deli survey, the outcome of the discussion was that FMI felt that the proposal needed modification. As a result, a task force of 4 members, led by Warren Dorsa as co-chair was formed to work with a group of 4 FMI members. The joint task force will invite Dr. Weidman and Dr. Sofos to participate as science advisors on the task force to develop a valid and useful protocol for further understanding Listeria sources within the deli operation. Other volunteers in addition to Warren Dorsa were: Wafa Birbari, Sara Lee; Margaret Hardin, Boars Head Provision; Ann Marie McNamara, Silliker, Inc. The task force will convene via conference call to develop the protocol.

John Butts moved a resolution that the AMI SAC should continue active dialog with FMI Food Protection Committee and work to form an on-going partnership to address food safety or technical issues of mutual concern. This resolution was passed unanimously.

New Listeria Intervention Research – Scott Burnett of Ecolab presented new data on the effectiveness of a surface treatment based upon octanoic acid chemistry that has recently received FSIS letter of no objection as a post lethality treatment. Scott presented the recent data and answered questions for the group related to this technology.

Process Lethality spreadsheet upgrade Report from Dr. Brad Marks, Michigan State University. Dr. Brad Marks joined via conference call and provided an overview of the AMIF – funded research project to enhance the AMI process lethality spreadsheet. The 2 – year project is now complete. Dr. Marks is seeking input on how to make the model upgrades applicable and appropriate for use in industry. He has some reservations about the use of the spreadsheet as currently constructed, since frequently processors are using model reference values that may not reflect true conditions in the product and/or process. Currently, the model reference values are only comprised of ground beef and poultry meat data and only fat and moisture variables are known due to the limited data availability that met the requirements for use. He stressed concern over the uncertainty of many data sources that could be used from the literature, such as, lab variables, different lab methods, (e.g. selective media, TSA plates, irradiated products to eliminate background), different cocktails of Salmonella, methods for growing cultures, and incubation times. These are some examples of data uncertainty that he believes may lead to uncertainty in lethality prediction for certain product types. The modified model incorporates the concept of confidence intervals around the lethality estimate, and given the variability in the data sets included, Dr. Marks has concern over the width of these confidence bands.

Dr. Marks proposed that a collaborative study could address these data gaps, by using multiple labs using same methods, same samples, isolates, etc. This would help define the true uncertainty of the method, rather than the uncontrollable factors. This has been proposed in a NIFSI-USDA study that MSU has submitted to USDA.

Committee members provided comment and suggestions on the project. It was pointed out that in industry settings, it is not standard practice for process authorities to consider confidence intervals, and this would be a new concept. The question was raised about how FSIS would deal with this concept.

IARC monograph on ingested nitrites and nitrates—

Steve Quickert, Kraft Oscar Mayer, provided the SAC with an overview of the recent announcement by the International Agency for Research on Cancer (IARC) that they will be publishing a monograph concerning the carcinogenicity of ingested nitrites and nitrates. Since IARC is viewed as an authoritative body globally by many government agencies, the implications for a negative outcome of the monograph committee conclusions could have significant implications for the industry. AMI has nominated Dr. Jim Coughlin to represent AMI interests as an official observer during the meeting in Lyon France, June 14 – 21. Dr. Coughlin will play a key role in making sure that the expert panel considers basic facts about nitrite use in industry and that they consider all of the relevant factors when evaluating the current literature.

AMIF Research update Huffman reported on the recently approved research projects and stated that the Executive Committee of AMI has continued to endorse and fund the process. Titles of newly approved research projects are listed below:

• Refinement of Listeria monocytogenes Low Dose Data from Pregnant Guinea Pigs for Human Risk Assessment, University of Georgia.

• Minimum Nitrite Levels Required to Control Listeria monocytogenes on Ready-to-eat Meat and Poultry Products, University of Wisconsin.

• Risk Factors and Consequences Associated with Condensation in Fresh and Ready-to-Eat Processing Facilities, Texas Tech University.

• Determine the Likelihood that Salmonella Develops Heat Resistance During Thermal Processing of Commercial, Whole-Muscle, Ready-to-Eat Meat Products. Michigan State University.

• Determination of the Efficacy of Chlorine Dioxide as an Anti-Listerial Agent in RTE Brine Chilling Solutions, Penn State University.

• White paper of Human Illness Caused by E. coli O157:H7 from Food and Non-Food Sources, University of Wisconsin.

Each spring, the Research Advisory Committees (RAC) is asked to evaluate the existing priorities and the request for proposal and make necessary modifications. As part of this process, a targeted research priority area has been established to address very specific industry needs. A research priority setting conference call with the RAC will take place in April – May to develop the 2007 priorities.

The next meeting time was discussed and it was decided to not attempt to hold a meeting during the summer IAFP meeting this year. A date will be established and communicated to the committee via email.

With no further business, the meeting was adjourned.

cc: J. Patrick Boyle Jim Hodges Appendix I

Scientific Affairs Advisory Committee Meeting Washington, DC March 15 - 16, 2006

Day 1, March 15, AMI Office 1150 Connecticut Ave, NW

9:00 a.m. Opening Remarks, Discussion of the Agenda, Review of Minutes from September meeting John Butts, Land O’Frost (filling in as chair for Dane Bernard)

9:00 – 10:30 FSIS science policy update: Recent Salmonella Initiative Upcoming Listeria directives Dr. David Goldman, FSIS-OPHS Phil Derfler, FSIS-OPPED 10:30 – 12:00 Prepare for joint AMI-FMI meeting: Lm at Retail CO packaging of fresh meat

Working lunch to be provided at AMI office. Committee will travel together to the law offices of Hogan and Hartson, 555, 13th street, NW

Joint AMI – FMI meeting and wine and cheese Reception

See separate agenda below.

Day 2, March 16, AMI Office, 1150 Connecticut Ave, NW

8:00 a.m. Wrap up and Next Steps from Joint FMI meeting

9:00 – 9:30 New Listeria intervention research Dr. Peter Bodnaruk, Ecolab (Scott Burnett substituted for Peter) 9:30 – 9:45 Break

9:45 – 10:30 Process Lethality Spreadsheet upgrade – report from Michigan State University project Dr. Brad Marks (via phone)

10:30 – 11:00 AMIF research update: currently initiated projects Priorities for 2007 Randy Huffman 11:00 –12:00 Update from the Nitrite task force: IARC monograph

Adjourn Appendix II

FMI Food Protection Committee -- AMI Scientific Affairs Committee Listeria Control in RTE Meat and Poultry Wednesday, March 15, 2006

New Location: Columbia Square Building Law Offices of Hogan & Hartson Fulbright Room

555 Thirteenth Street, NW Washington, DC 20004 Phone: +1.202.637.5600

From the Crowne Plaza Hamilton Hotel: Go South on 14th Street; turn left on F Street; turn right on 13th. Enter the Columbia Square Building next to the Warner Theater. (About a 5 minute walk)

Once you enter the Columbia Square Building:

You want to go to the Hogan & Hartson Fulbright Room - From the elevators, go the 13th floor, West Tower. Name tags will be ready and waiting for your arrival at the reception area. You will be escorted from the reception area to the Fulbright Room by FMI staff. Coats and luggage can be kept in a coat closet near the Fulbright Room. The Hogan & Hartson offices are approximately 2 blocks from FMI’s Office and about a 5 minute walk from the Crowne Plaza Hamilton Hotel. .

Wednesday, March 15, 2006 Please arrive at 1:15 p.m.

1:30 – 5:30 p.m. FMI-AMI joint meeting - Hogan & Hartson, Fulbright Room

• See Agenda below

5:30 p.m. – 7:30 p.m. FMI-AMI Reception - Hogan & Hartson, Fulbright Reception Center

AGENDA

Introductions, review of objectives for meeting, and Jill Hollingsworth, FMI Approval of agenda Randy Huffman, AMIF Committee Chairs

Overview of AMI efforts regarding Listeria Control Randy Huffman

FSIS Interim Final Rule for Listeria Control Bill Sveum Kraft/Oscar Mayer

Lessons Learned in RTE Meat Processing: John Butts Sanitation, Sampling, Data, Corrective Actions Land O’Frost

Interventions to Control Lm in Deli Meats: Tim Freier Inhibitors, Cook-in-bag, other technology Cargill Meat Solutions

Retail Perspective on Listeria Risks and Controls Roger Hancock Albertsons, Inc.

Break

Review of Findings from Draughon et al., USDA funded Listeria at Retail Deli study (Tentative)

Preliminary Findings from NYSA&M/Cornell Retail Deli Survey of Listeria Martin Weidman Cornell University

Plans for newly funded USDA, Multi-Institution John Sofos, Colorado State Listeria Control Study Martin Weidman, Cornell

Proposed AMIF-funded Listeria study of Listeria Ann Marie McNamara at Retail Deli Silliker, Inc.

Discussion All

Appendix III

FMI meeting attendees (indicated as of 3/08/06)

Prince Gale The Kroger Company Brown Pat Pathmark Stores Inc. Oswald Steve Wakefern Food Corporation Ambrosio Mike Wakefern Food Corporation Ball James Food Lion Kelleher Gillian Wegmans Food Markets Frappier Robert Ahold USA O'Donnell Kathleen Wegmans Food Markets Tryba Cas Big Y Foods Hancock Roger Albertsons Marra Paul Wegmans Food Markets Hedman Cory Hannaford Levee Terry Winn-Dixie Stores, Inc. Pasley Dianna Schnucks Market, Inc. Sherback George Stop & Shop Stoffer Chuck Safeway (may send Keith) Van Ostenbridge Mark Bi-Low Menke-Schaenzer Joan Wal-Mart Bacon Brenda Harris Teeter Martin Monica Farm Fresh, Inc. Hollingsworth Jill FMI Reimers Fred Consultant Edwards Larry FMI Lee Gwen FMI DiPersio Tricia Wild Oats Market

Lm Intervention Updates

July 25, 2006

Dr. Dan Engeljohn Deputy Assistant Administrator Office of Policy, Program and Employee Development USDA 1400 Independence Avenue SW Washington, DC 20250-3700

Re: New Technology Submission, FSIS Ref. 05-NT-0128

Dear Dr. Engeljohn:

Since early in 2005, the American Meat Institute (AMI) has been working with the New Technology Office (NTO) to use GRAS ingredients (calcium, potassium and or sodium sorbate, sorbic acid, propionic acid, sodium propionate, sodium benzoate and benzoic acid) individually or in combination for the prevention of growth of Listeria monocytogenes in various ready-to-eat (RTE) meat and poultry products to enhance food safety. Correspondence has gone back and forth between NTO and AMI on this issue. In the most recent letter (dated January 25, 2006) from the New Technology Office, Dr. Syed provided guidance on the following points: (1) FSIS has “received a written comment from FDA’s Office of Food Additive Safety that conveys an opinion that FDA has no objections with regard to the safety of sodium benzoate, benzoic acid, potassium sorbate, sodium sorbate and sorbic acid under the proposed conditions of use,” and (2) that “suitability data will be needed by FSIS to support the proposed use” of the aforementioned ingredients in meat and poultry products.

As follow-up to this letter, I responded to Dr. Syed on February 1, 2006 stating that in his letter to me he indicated that his office had received a written comment from FDA’s Office of Food Additive Safety that conveyed an opinion that FDA has no objections with regard to the safety of sodium benzoate, benzoic acid, potassium sorbate, sodium sorbate and sorbic acid under the proposed conditions of use. However, in the new technology notification, AMI had also requested the use of calcium sorbate, propionic acid and sodium propionate. These ingredients were left out of Dr. Syed’s reference to FDA’s opinion. AMI requested clarification on whether these omissions were an oversight by Dr. Syed in his response, or an oversight in the request for an opinion by FDA from his office, or whether FDA purposefully left these ingredients out of their opinion. On February 22, 2006 Dr. Syed responded that “FDA’s opinion on safety did include these ingredients in their response to FSIS” with the same conclusion on no objections with regard to their safety. Dr. Syed further stated that suitability data could be collected as per the previous new technology request made by AMI.

This letter provides the suitability data on sliced, cooked, uncured turkey breast (<1% fat) and cured ham (5-7% fat manufactured with 156 ppm sodium nitrite). The objectives of the suitability research were:

• Identify minimum levels of sorbate, benzoate and propionate, individually and in combination, which prevent growth of L. monocytogenes on sliced, cooked, uncured turkey breast and cured, cooked ham products stored at 4, 7 and 10 °C for up to 3 months; and • Determine the effect that selected antimicrobial treatments have on sensory qualities.

The ready-to-eat turkey and ham products (sixteen test formulations for each product type plus control formulations without antimicrobials) were manufactured by commercial establishments (see previous communications and approvals from the NTO on commercial establishments). Inoculations studies (triplicate samples), proximate and chemical analyses, and sensory panels were conducted at the appropriate University of Wisconsin laboratories (the entire report and the appendix describing the consumer preference evaluation are attached for your review).

Results of these suitability studies demonstrate that propionate, benzoate and sorbate will enhance the safety of ready-to-eat, cured and uncured meat and poultry products with minimal impact on sensory profiles when used at the lowest effective levels. Conclusions from these suitability studies include:

• The GRAS ingredients (calcium, potassium and or sodium sorbate, sorbic acid, propionic acid, sodium propionate, sodium benzoate and benzoic acid) individually or in combination, control the growth of L. monocytogenes on the surface of high-moisture, high-pH processed meat and poultry products when used at levels which are deemed safe and acceptable for other food products. • Propionate (> 0.2%) or the combination of propionate (> 0.1%) and sorbate (0.1%) prevented the growth of L. monocytogenes in uncured turkey stored at 4 °C for 12 weeks. • Benzoate (0.1%), propionate (> 0.2%), sorbate (0.3%), or the combination of propionate (0.1%) and sorbate (0.1%) prevented the growth of L. monocytogenes in ham stored at 4 °C for 12 weeks. • When used in conjunction with nitrite, lower levels of the tested antimicrobials were required to control the growth of L. monocytogenes. • Sensory analysis for products using the highest concentration of effective antimicrobials indicated that consumers preferred the flavor of ham with 0.3% propionate or 0.1% benzoate compared with product made with 1.6% lactate and 0.1% diacetate; and there was no significant difference compared with the control product made without the antimicrobials. • For the cooked, uncured turkey product, consumers rated the product made with 0.3% sorbate equivalent (rating and overall preference) to the control product formulated without antimicrobials. Flavor impact appeared to limit the use of propionate in turkey as consumers preferred control product to product formulated with propionate. • Products formulated with antimicrobials spoiled as did control products when stored under abuse conditions, i.e., the use of antimicrobials did not mask normal spoilage.

Dr. Engeljohn, this suitability research, in addition to other historical research previously shared with the NTO, demonstrates the value of adding the GRAS ingredients (calcium, potassium and or sodium sorbate, sorbic acid, propionic acid, sodium propionate, sodium benzoate and benzoic acid) individually or in combination to prevent the growth of L. monocytogenes in various ready-to-eat meat and poultry products to enhance food safety. As further evidence of the safety of these ingredients when used at the proposed levels in ready-to-eat meat and poultry products, as well as to satisfy the NTO, a review of published data for the last three years was done by the University of Wisconsin Food Research Institute on the safety of these ingredients. This report is attached for your review. The key conclusion is that “data indicate that these compounds are of low toxicity with little or no genotoxic or carcinogenic potential.”

Dr. Engeljohn, we look forward to your expediting the approval of these GRAS ingredients (calcium, potassium and or sodium sorbate, sorbic acid, propionic acid, sodium propionate, sodium benzoate and benzoic acid) individually or in combination to prevent the growth of L. monocytogenes in various ready-to-eat meat and poultry products to enhance food safety. This will help the establishments manufacturing these products further reduce the risks to consumers and assist the Food Safety and Inspection Service achieve their goals of reducing the potential for food borne illnesses associated with L. monocytogenes and ready-to-eat meat and poultry products. We believe this should satisfy all of the requests for additional information communicated to AMI since the notification and protocol for the new technology was initiated in March, 2005. We look forward to an expeditious process of approval on your end.

Sincerely,

Robert (Skip) A. Seward II, Ph.D. Vice President – Regulatory Affairs American Meat Institute 1150 Connecticut Avenue NW 12th Floor Washington, DC 20036 pc: M. Dopp - AMI August 30, 2006

MEMORANDUM FOR AMI INSPECTION COMMITTEE

FROM: SKIP SEWARD

SUBJECT: POTENTIAL USE OF ACIDS AND SALTS OF SORBATE, PROPIONATE & BENZOATE TO CONTROL GROWTH OF LISTERIA MONOCYTOGENES IN RTE MEAT AND POULTRY PRODUCTS

As you may recall, a task force was created to investigate the potential use of sorbic, propionic and benzoic acids and their salts in ready-to-eat meat and poultry products to prevent the growth of L. monocytogenes. After a year of communication with the FSIS New Technology Office (NTO), production of test products by select AMI members, and research conducted by the University of Wisconsin, it appears that the NTO is prepared to move forward with allowing the use of these ingredients on an extended waiver basis under particular oversight and with specified requirements while rulemaking is underway (according to Doug Palo, NTO, 8/29/06).

In order to ensure that the waiver instructions from the NTO are helpful to the industry, we have a brief period in which to supply our comments relative to the requirements (e.g., validation) that must be met before production can occur using these antimicrobials, and those associated with on-going verification. Specifically, the following are being proposed as requirements:

• Establishment will have validation data that demonstrate the efficacy of a formulation in preventing growth of L. monocytogenes • Establishment will have shelf life data that demonstrate that spoilage is not masked by the use of the antimicrobials in the formulation • Establishment will have on-going verification (e.g., quarterly) that demonstrates that the levels of antimicrobials proven effective in the validation studies are being achieved in the final product • Establishment will have labels that identify the use of these antimicrobials

I need to understand your perspective on these requirements as soon as possible if these seem reasonable (I assume they are since these are requirements for any such application).

Although the exact nature of the waiver and approval is unknown at this time, it is my understanding that the following information will be necessary (and reported to the NTO and shared with the local inspection personnel) in order to be recognized as an establishment approved to use these antimicrobials. It would be useful to begin the compilation of these data if the use of these antimicrobials in your formulations is of interest to you.

• Establishment # • List of specific products for which the formulae would include the antimicrobials • Validation data for the formulated products • Validation that spoilage is not masked by the formulations • Written plan for on-going verification that the formulae are being achieved • Labels to be used for the formulated products that identify the antimicrobials

There was mention of a requirement for labeling that specified the reason for use of the antimicrobials, e.g., benzoate used to prevent the growth of bacteria. I strongly objected to this requirement, indicating that this had been proposed by FSIS earlier for antimicrobials, and was subsequently dropped as a requirement.

The NTO indicated in conversations to me that FSIS recognized that there were no safety or public health issues associated with the use of these antimicrobials at the proposed levels. Their challenge appears to be one of allowing the use of these antimicrobials by an extended waiver process while rule making is underway, while justifying their actions to those stakeholders that may believe that these antimicrobials will be used to somehow mask spoilage or inferior products. The NTO also would like to better understand and develop the advantages that could be achieved through the use of these antimicrobials, e.g., reduced costs (thereby making the technology more available, e.g., vs. use of lactate-diacetate, to small and very small establishments); greater protection for products, < em>e.g., uncured turkey, where lactate- diacetate cannot be used because of organoleptic reasons; and protection against contamination that could occur at retail or in the home. If you have ideas or facts on the economics of other advantages, please send these ideas to me as soon as possible.

Thanks for your assistance and input. If you have any questions, please call me at 202- 587-4249 or e mail me at [email protected] .

cc: P.BOYLE M.DOPP L.MORRISSETTE R.HUFFMAN

August 30, 2006

MEMORANDUM FOR AMI INSPECTION COMMITTEE

FROM: SKIP SEWARD

SUBJECT: UPDATE: POTENTIAL USE OF ACIDS AND SALTS OF SORBATE, PROPIONATE & BENZOATE TO CONTROL GROWTH OF LISTERIA MONOCYTOGENES IN RTE MEAT AND POULTRY PRODUCTS

In a second call initiated by the FSIS New Technology Office (NTO) today, we learned that the NTO is considering the following requirements for the extended waiver process:

• Submit a protocol to the NTO for use of these antimicrobials • Have a procedure that ensures that what is written in the protocol is executed • Submit a label that identifies the antimicrobials used [Note: it appears that the NTO will not recommend that a reason for the antimicrobials be stated on the label] • State a sell-by (or use-by) date on the package • Submit validation documentation that the formula-product combination inhibits growth of L. monocytogenes [Note: there was discussion about including a log-reduction minimum; but I explained that this is not important for the application, and is covered by the regulations regarding alternative classifications. It remains to be seen whether a minimum will be included.] • Have a procedure to ensure that the validation data is reviewed annually or when a product-formula combination is altered [Note: there was discussion about re-validation and submission of data annually; but I explained that so long as the formulation-product combination remained fixed there would be no reason to repeat the validation and that the annual HACCP re-assessment would cover these products as well. It remains to be seen if they understood my point.] • On-going verification data that substantiate that the validated protocol is being executed

As you see, the NTO did not mentioned spoilage (i.e., masking spoilage) or shelf life in this initial draft; and I did not raise the issue. Two possible ways to address this for your records would be that the sell-by (use-by) date was not extended and that the number of organisms associated with spoilage in product with and without the antimicrobials was the same, independent of the time to achieve spoilage.

The NTO indicated that the draft extended waiver document would circulate to others in FSIS before being released which means we have a very brief time for input and that the requirements shown above could be altered.

Thanks for your assistance and input. If you have any questions, please call me at 202-587-4249 or e mail me at [email protected] .

cc: P.BOYLE M.DOPP L.MORRISSETTE R.HUFFMAN Glass and Claus, AMIF Final Report, June 2006 1 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products Using Food-Approved Antimicrobials Benzoate, Propionate, and Sorbate

Kathleen Glassa and James Clausb, Co-Principal Investigators Lindsey McDonnell, Rob Rassel, Kristine Zierke

University of Wisconsin-Madison aFood Research Institute, 1925 Willow Drive, Madison, WI 53706 bMeat Science and Muscle Biology Lab, 1805 Linden Drive, Madison, WI 53706 Phone: 608-263-6935 (KG); (608) 262-0875 (JC) E-mail: [email protected]; [email protected]

Summary: The primary objective of this study was to identify levels of potassium sorbate, sodium benzoate, and sodium propionate that prevent growth of Listeria monocytogenes on sliced, cooked, uncured turkey breast (<1% fat) and cured ham (5-7% fat manufactured with 156 ppm sodium nitrite) products. Data revealed that > 0.2% Propionate or combinations of >0.1% Proprionate+0.1% Sorbate will prevent listerial growth in uncured turkey stored at 4°C for 12 weeks. When used in conjunction with nitrite, lower concentrations of antimycotics are needed to control pathogen growth. Growth of L. monocytogenes was prevented in ham stored at 4°C for 12 weeks when formulated with 0.1% Benzoate, >0.2% Propionate, 0.3% Sorbate, or combinations of 0.1% Proprionate+0.1% Sorbate. Comparison with previous research in bologna suggests that relatively low moisture (55%) and pH (6.1) will also reduce the minimum concentration of antimycotic required to prevent listerial growth during 3 month storage at 4°C. Sensory analysis for products using the highest concentration of single, effective antimicrobials reported consumers preferred the flavor of ham with 0.3% Propionate or 0.1% Benzoate compared with the 1.6% Lactate + 0.1% Diacetate treatment, and no significant difference compared with the Control without antimicrobials added. The addition of 0.3% Sorbate rated lowest in consumer taste preference. For deli-style turkey, consumers rated 0.3% Sorbate treatments equivalent (P>0.05) to Controls without antimicrobials (rating and overall preference). In contrast, consumers preferred (P<0.05) the Control turkey over the turkey containing sodium propionate. Flavor appears to be a major limitation of the use sodium propionate in turkey. A literature review described the safety and sensitivities exhibited by individuals to the various antimicrobials. This research verified that propionate, benzoate, and sorbate will enhance the safety of high-moisture, RTE cured and uncured meat and poultry products and that addition of these antimicrobials should have minimal negative impact on consumer taste preference if used at the lowest effective levels. Data can be used to petition FSIS for approval of propionate, benzoate, and sorbate for use in product formulations to control L. monocytogenes.

Specific Objectives of Research Proposal: • Identify levels of sorbate, benzoate, and propionate, individually and in combination, which prevent growth of Listeria monocytogenes on sliced, uncured cooked turkey breast and cured cooked ham products stored at 4, 7, and 10°C for up to 3 months. • Determine the effect that select antimicrobial treatments have on sensory qualities. • Conduct a literature search on the effect that sorbate, benzoate, and propionate have on human health in the presence and absence of nitrite in ready-to-eat meats.

Glass and Claus, AMIF Final Report, June 2006 2 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

INTRODUCTION

As a means to prevent outbreaks of listeriosis, the USDA-FSIS 2003 Final Rule to Control of Listeria monocytogenes in Ready-to-Eat Meat and Poultry Products permits the use of growth inhibitors for Listeria on RTE meat and poultry products so there is no more than 1.0 log CFU/g increase during its shelf life (Anonymous, 2003). Currently, many manufacturers have incorporated lactate and diacetate into formulations of RTE products to prevent growth of L. monocytogenes in high-moisture, high-pH products and prevent additional outbreaks and recalls (Anonymous, 2001). However, while studies have verified that the addition of lactate and diacetate combinations inhibit growth of L. monocytogenes in cured meat and poultry products (with sodium nitrite), these organic acid salts are less effective in uncured products (Glass et al., 2002; Legan et al., 2004; Mdandi and Shelef, 2001; Seman et al., 2002). High levels of lactate and diacetate which prevent listerial growth in uncured meat and poultry products may have a negative affect on sensory attributes. Therefore, alternate antimicrobial ingredients are needed to provide safe and acceptable options for manufacturing without adversely affecting product quality, especially in nitrite-free products.

Three GRAS additives, sorbate, benzoate, and propionate, approved for use to control mold growth in a variety of food products other than meats (US FDA 2004), have also been shown to inhibit growth of several Gram-positive bacterial pathogens, such as Clostridium botulinum, Staphylococcus aureus, and Listeria monocytogenes in media, as well as in and on meat systems (El-Shenawy and Marth, 1988; Islam et al., 2002a, 2002b; Samelis et al., 2001, Tompkin et al., 1974; Wederquist et al., 1994).

Preliminary research in our laboratory demonstrated that L. monocytogenes will not grow in wiener or turkey slurries supplemented with 0.25% potassium sorbate, propionic acid, or benzoic acid and stored at 4 or 10°C for up to 4 weeks (Glass et al., 2004). Additional experiments revealed that beef-pork bologna with nitrite that incorporated combinations of Benzoate+Propionate or Benzoate+Sorbate (total 0.1%; 0.05% of each compound, w/w) into the formulation will not support growth of L. monocytogenes on bologna stored 12 weeks at 4°C, compared with a >3.5-log increase in listerial populations in the control bologna without antimicrobials (Preston et al., 2005). Low levels of antimycotic agents were less effective in uncured turkey products, but they still slowed listerial growth. When uncured turkey was stored at 4°C for four weeks, populations of L. monocytogenes increased 2.5 and 4.5 log cfu/package in the Benzoate+Propionate and Benzoate+Sorbate treatments, respectively, compared with a 6.5 log cfu/pkg increase for the turkey control without antimicrobials.

While these antimycotic agents are not yet approved in the U.S. for use within formulations of processed meats, confirming their efficacy and safety in a variety of products will be useful in a successful petition for regulatory approval. This study was designed to identify the minimum levels of sorbate, benzoate, and propionate, that prevent growth of Listeria monocytogenes on sliced, uncured cooked turkey breast (<1% fat) and cured cooked ham (5-7% fat manufactured with 156 ppm sodium nitrite) products stored at 4, 7, and 10°C for up to 3 months, and to determine the effect that ingredients have on consumer taste preference.

Glass and Claus, AMIF Final Report, June 2006 3 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

MATERIALS AND METHODS

Production of ready-to-eat turkey and ham: Sixteen test formulations plus control formulations without antimicrobials were manufactured for each product type by commercial producers to provide the formulations listed on Table 1.

Ingredient statement for Control Turkey included: Turkey breast, water, 2% or less of modified food starch, salt, dextrose, carrageenan, sodium phosphate, turkey flavor (maltodextrin, salt, flavor). Ingredient statement for Control Ham included: Ham cured with water, salt, less than 2% dextrose, sodium phosphate, sodium erythorbate, sodium nitrite. Sodium lactate, sodium diacetate, sodium propionate, potassium sorbate, and sodium benzoate were added as defined in the experimental design (Table 1).

Cooked, uncured turkey breast and cured smoked composite ham were produced under Good Manufacturing Practices in USDA-inspected commercial facilities according to industry standard practices, stuffed into casings, and cooked to the desired endpoint temperature ( >71.1°C, 160°F for ham and >73.9°C, 165°F for turkey). Cooked, chilled products were sliced on a commercial slicer, packaged (vacuum-package for turkey; nitrogen flush for ham) and stored at <4°C through transport to the Food Research Institute, UW-Madison, for inoculation and testing. The study was replicated twice. Replicate Control treatments were run for each inoculation day.

Sensory analysis: Consumer taste preference panels for ham and turkey were completed at the UW-Madison campus by the UW Department of Food Science Sensory Laboratory, Babcock Hall, and by the Meat Science and Muscle Biology Lab (ASTM, 1988; Berry et al., 1983), respectively. Freshly prepared products were sent directly from the manufacturer to the Meat Science and Muscle Biology Laboratory and used within 4 weeks after manufacture. For both product types, consumer preference was compared (pair-wise comparison) for Control without antimicrobials, with 0.3% Sorbate, and with 0.3% Propionate. In addition, ham treatments containing 0.1% Benzoate and 1.6% Lactate+0.1% Diacetate were also evaluated. The ballot used for consumer preference evaluation contained a structured 7-point hedonic scale (Amerine et al., 1965). The sensory anchors used were “dislike very much (1)” to “like very much (7). “Two hundred fifteen and 40 consumers were surveyed for the ham and turkey treatments, respectively. For the overall preference attribute, statistical analyses provided the mean score for each sample, the F-value for all samples, and the least significant difference (LSD) for making sample comparisons (significance level of P < 0.05).

Proximate and chemical analysis: Moisture (5 h, 100°C, vacuum oven method, 950.46; AOAC, 2000), pH (1:10 dilution for 10 g homogenized portion, Accumet Basic pH meter and Orion 8104 combination electrode), NaCl - (measured as % Cl , AgNO3 potentiometric titration, Brinkman Metrohm autotitrator), nitrite (Colorimetric Method, 973.31, AOAC, 2000), and water activity (Decagon AquaLab CX-2 water activity meter, Pullman, WA) were assayed by the Food Research Institute for triplicate samples for each formulation; values for protein and fat were provided by the manufacturers. Chemical analysis for distribution of sorbate, benzoate, and propionate in the product matrix were determined in treatments with the maximum concentrations of each antimycotic agent and in the Lactate-Diacetate control. For each trial, two samples were taken from different regions of the ham or turkey and assayed by commercial laboratories (gas chromatography method, 983.13, AOAC, 2000; concentrations of benzoic acid, sorbic acid, and propionic acid used to determine percentage of potassium sorbate, sodium benzoate, and sodium propionate).

Glass and Claus, AMIF Final Report, June 2006 4 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Table 1. Treatments for cured ham and uncured turkey manufactured with various levels of Sorbate, Benzoate, and Propionate. 1

Ham2 Potassium Sodium Sodium Sodium Sodium Total sorbate benzoate propionate lactate 3 diacetate antimycotic 1 0.05 0.05 2 0.05 0.05 3 0.05 0.05 4 0.1 0.1 5 0.1 0.1 6 0.1 0.1 7 0.05 0.05 0.1 8 0.05 0.05 0.1 9 0.05 0.05 0.1 10 0.075 0.075 0.15 11 0.1 0.1 0.2 12 0.2 0.2 13 0.2 0.2 14 0.3 0.3 15 0.3 0.3 16 1.6 0.1 0 Control No additional antimicrobials; positive growth control 0 Turkey Potassium Sodium Sodium Sodium Sodium Total sorbate benzoate propionate lactate diacetate antimycotic 1 0.1 0.1 2 0.1 0.1 3 0.1 0.1 4 0.05 0.05 0.1 5 0.05 0.05 0.1 6 0.05 0.05 0.1 7 0.15 0.15 8 0.15 0.15 9 0.075 0.075 0.15 10 0.1 0.1 0.2 11 0.2 0.2 12 0.2 0.2 13 0.15 0.15 0.3 14 0.3 0.3 15 0.3 0.3 16 3.2 0.2 0 Control No additional antimicrobials; positive growth control 0

1 All percentages are given on a finished weight basis without excessive moisture loss expect for nitrite, which is meat-block basis to be in compliance with federal regulations 2 Ham formulated with 156 ppm sodium nitrite and 550 ppm sodium erythorbate on meat-block basis in compliance with federal regulations 3 Calculated on an anhydrous basis

Glass and Claus, AMIF Final Report, June 2006 5 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Preparation of inocula: Listeria monocytogenes strains Scott A (clinical isolate, serotype 4b), LM 101 (hard salami isolate, 4b), LM 108 (hard salami isolate, 1/2a), LM 310 (goat milk cheese isolate, 4), and V7 (raw milk isolate, 4), were grown individually in 10 ml Trypticase® soy broth (BBL, Cockeysville, MD) at 37°C for 18 to 20 h. Cells were harvested by centrifugation (2,500 x g, 20 min) and suspended in 4.5 ml 0.1% buffered peptone water (pH 7.2). Equivalent populations of each isolate were combined to provide a five-strain mixture of L. monocytogenes to yield target level of 5-log CFU per 100-g package. Populations of each strain and the mixture were verified by plating on Trypticase® soy agar (TSA) and Modified Oxford agar (MOX; Listeria Selective Agar base, Difco).

Inoculation and testing: Slices were surface-inoculated with L. monocytogenes to provide approximately 5-log CFU per 100 g package (equivalent to 3-log CFU per ml rinse material), vacuum-packaged (Multivac AGW, Sepp Haggemuller KG, Wolfertschewenden, Germany) in gas-impermeable pouches (3 mil high barrier EVOH pouches, Deli 1 material, oxygen transmission 2.3 cm3 per cm2, 24 h at 24°C; water transmission 7.8 g per cm2, 24 h at 37.8°C and 90% relative humidity, WinPak, Winnepeg, Manitoba, Canada), and stored at 4, 7, and 10oC for up to 12 weeks.

Triplicate inoculated and duplicate uninoculated samples for each treatment were assayed for changes in L. monocytogenes populations, and changes in lactic acid bacteria and pH, respectively, at 0-time, and at 2, 4, 6, 8, 10 and 12 weeks storage at 4°C, and at 4, 8, and 12 weeks only at 7 and 10°C. Bacterial populations were determined in rinse material obtained after adding 100 ml of sterile Butterfield phosphate buffer to each package and massaging the contents externally by hand for about 3 minutes (Glass and Doyle, 1989; Glass et al., 2002). L. monocytogenes was enumerated by surface plating serial (1:10) dilutions of rinse material on MOX. Select colonies were confirmed as L. monocytogenes by Gram-stain, tumbling motility, CAMP test, hemolysis on Trypticase soy agar with sheep blood, and biochemical analysis using MICRO-ID® Listeria (Remel, Lenexa, KS). The pH was measured by removing representative 10 g of the uninoculated samples and homogenizing with 90 ml hot deionized water using a Stomacher. Homogenized sample was allowed to cool to room temperature and the pH measured on the slurry. Changes in populations of lactic acid bacteria were assayed for the remaining uninoculated sample by plating rinse material on plate count agar with bromcresol purple (25oC, 48-72 h). Testing of a treatment was discontinued if listerial growth (2-log increase) was confirmed for packages tested for two consecutive sampling intervals.

RESULTS AND DISCUSSION

Proximate Analyses: Proximate analyses demonstrate variability that can be expected in normal commercial production (Table 2). Average moisture (75.04+1.08%) and pH (6.42+0.1) values for turkey were typically greater than that found in ham (73.65+0.27% moisture and pH 6.39+0.02), and salt values were typically lower in turkey (1.71+0.20%) compared with ham (2.59+0.02%). However, no consistent correlation was found between proximate analysis and microbial growth for the treatments tested (data not shown).

Glass and Claus, AMIF Final Report, June 2006 6 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Table 2. Proximate analysis of turkey and ham.4

Turkey Ham Moisture % 75.04+1.08 (Range 71.9-77.1) 73.65+0.27 (Range 71.4-79.9) NaCl % 1.71+0.20 (Range 1.14-2.01) 2.59+0.02 (Range 2.24-2.90) pH 6.42+.01 6.39+0.02 Water Activity 0.972+0.001 0.967+0.000 Nitrite (ppm) Not added 34.6+13.8 Protein %5 17.9 17.9 Fat % 0.80 5.35

Levels of benzoate and sorbate were within expected target range for both the turkey and ham products, with little variation found between samples taken from different regions of the manufacturer’s package or between samples from the two trials (Table 3). In contrast, propionate levels reported by a commercial laboratory using gas chromatography method yield significantly lower than expected results. Given the consistent inhibition of L. monocytogenes in the treatments with 0.3% sodium propionate, reported results are considered invalid and are either due to the commercial lab error or decomposition of propionic acid during processing or storage.

Table 3. Analysis of antimycotic agents for select formulations6, 7

Formulation Sodium Benzoate % Potassium Sorbate % Sodium Propionate % Ham 58 0.11+0.01 Ham 149 0.31+0.03 Ham 1510 <0.0060 Ham 1611 <0.0045 <0.0040 <0.0060

Turkey 28 0.11+0.01 Turkey 149 0.29+0.07 Turkey 1510 0.10+0.02 Turkey 1611 <0.0045 <0.0040 <0.0060

4 Results are an average + standard deviation for analysis of triplicate samples for each formulation 5 Protein and fat levels reported by manufacturer

6 Gas chromatography method, 983.13, AOAC, 2000; measured as acid form and converted to a percentage basis of salt form. Benzoate and Sorbate analyzed by R-Tech Laboratories, Arden Hills, MN; Propionate analyzed by Covance Laboratories, Madison, WI. 7 Average for four samples + standard deviation. 8 Target 0.1% sodium benzoate 9 Target 0.3% potassium sorbate 10 Target 0.3% sodium propionate 11 No antimycotic agent added, Lactate-Diacetate control

Glass and Claus, AMIF Final Report, June 2006 7 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Control of L. monocytogenes: Results from this study confirmed that antimycotic agents control growth of L. monocytogenes on the surface of high-moisture, high-pH processed meat and poultry products when used at levels which are deemed safe and acceptable for other food products.

When used in conjunction with nitrite in ham, low concentrations of antimycotics are needed to inhibit pathogen growth. Growth of L. monocytogenes was consistently prevented in cured ham stored at 4°C for 12 weeks when formulated with 0.1% Benzoate, >0.2% Propionate, 0.3% Sorbate, or combinations of 0.1% Propionate+0.1% Sorbate or 1.6% Lactate + 0.1% Diacetate (Figure 1). Other treatment, including combinations of >0.1% total antimycotic agents, 0.1% Propionate, and 0.1 and 0.2 % Sorbate, delayed growth until 6 to 10 weeks, but permitted >1 log growth for sporadic samples in at least one trial at or prior to 12 weeks, even though the overall average for both trials appeared to have <1 log increase. The Control treatment without antimicrobials and treatments with only 0.05% of any individual antimycotic agents supported 0.6 to 1.1-log growth at the 2-week sampling interval.

Antimicrobial treatments that prevented growth at 4°C were less effective when the ham was stored higher temperatures. Products supplemented with 0.1% Sorbate+0.1% Propionate, >0.2% Propionate, >0.2% Sorbate, and 1.6% Lactate + 0.1% Diacetate prevented listerial growth for 4 weeks at 7°C, but all formulations supported >1-log increase at 8-weeks (data not shown). For products stored at 10°C, only 0.3% Propionate delayed pathogen growth for 4 weeks (average 1 log increase), whereas all the other treatments supported a 2-4 log increase in the same period (data not shown).

For uncured turkey, products supplemented with > 0.2% Propionate, combinations of >0.1% proprionate+0.1% Sorbate, and combination of 3.2% Lactate+0.2% Diacetate consistently inhibited growth of L. monocytogenes (<1-log increase) when stored at 4°C for 12 weeks (Figure 2). Combination of 0.075% Propionate+0.075% Sorbate prevented pathogen growth in samples from only one replicate for the duration of the study. Surprisingly, turkey with 0.3% Sorbate supported a 1-log increase starting at 8 weeks at 4°C, whereas 0.2% Sorbate delayed growth for 10 weeks. Chemical analysis of the Turkey with 0.3% Sorbate did not suggest insufficient addition or uneven distribution of sorbate within the product matrix. Additional study may be required before discounting the use of 0.3% Sorbate as an antilisterial agent in high-moisture, uncured products.

As observed for the ham treatments, inhibition was less pronounced when formulation was stored at abuse temperatures. Products formulated with 0.15% Sorbate+0.15% Propionate, 0.2% Propionate, 0.3% Propionate, and 3.2% Lactate+0.2% Diacetate delayed listerial growth 4 weeks (<1-log increase) when stored at 7°C, but supported significant growth (2- log increase) at 8 weeks. None of the treatments delayed listerial growth when stored at 10°C, with 1.5 to 5.5-log increase within 4 weeks.

Growth of spoilage lactic acid bacteria was inconsistent among samples within and between treatments. While most samples assayed contained populations fewer than detectable limit by direct plating (<1-log CFU/ml rinse), populations ranged to >8 log CFU/ml rinse (data not shown). The pH of the uninoculated samples tested did not decrease appreciably for any sample, with <0.15 pH unit decrease observed throughout the testing interval. No

Glass and Claus, AMIF Final Report, June 2006 8 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

correlation between growth of spoilage microflora and pH reduction was observed for the uninoculated samples assayed. Therefore, there is no clear evidence that the antimycotic treatments will also inhibit spoilage microflora and mask spoilage.

These results support previous studies indicating that antimycotic agents inhibit growth of L. monocytogenes on RTE meats, but demonstrate that they are more effective when used in combination with nitrite in cured meat products than in uncured products. Of the three antimycotic agents used, sorbate appears to have the least inhibitory effect on L. monocytogenes growth at any given concentration.

In addition to the presence of nitrite, product moisture and pH also appear to have a significant effect on efficacy of low concentrations of antimicrobials. Previous research (see Final Report to AMIF, Glass et al., Antimicrobial Combinations in RTE Meats, June 10, 2005) revealed that combinations of 0.05% Benzoate+0.05% Sorbate or 0.05% Benzoate+0.05% Propionate prevented growth of L. monocytogenes in cured beef-pork bologna (~57% moisture, pH 6.1), but the same concentrations have less inhibitory effect in the cured ham (~73-75% moisture, pH 6.3) or uncured turkey (~75-76% moisture, pH 6.3- 6.4) products evaluated in this study.

Sensory Analysis: Sensory analysis for products using the highest concentration of effective antimicrobials reported consumers preferred the flavor of ham with 0.3% Propionate or 0.1% Benzoate compared with the Lactate-Diacetate treatment, and no significant difference compared with the Control ham without antimicrobials added (Table 4). The addition of 0.3% Sorbate rated lowest in consumer preference (see Appendix A for details on ham sensory analysis). Consumers rated deli-style turkey containing potassium sorbate equivalent (P>0.05) to Controls turkey without antimicrobials (rating and overall preference; Table 5). In contrast, consumers did not prefer turkey containing sodium propionate over Control turkey (P<0.05). However, when a direct comparison of potassium sorbate was made to sodium propionate, no preference was noted (P>0.05).

Glass and Claus, AMIF Final Report, June 2006 9 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Table 4. Consumer evaluation of smoked ham containing sodium lactate+sodium diacetate blend (1.6+0.1%), sodium benzoate (0.1%) potassium sorbate (0.3%) or sodium propionate (0.3%). Sensory Response1 Treatment comparison Preference Rating Overall preference Control 5.8A 0.63 Lactate-Diacetate (1.6+0.1%) 5.5B 0.37

Control 6.0A 0.49 Sodium Benzoate (0.1%) 6.0A 0.51

Control 6.1A 0.56 Potassium Sorbate (0.3%) 5.9B 0.44

Control 6.0A 0.49 Sodium Propionate (0.3%) 6.0A 0.51

Lactate-Diacetate (1.6+0.1%) 5.9A 0.35 Sodium Benzoate (0.1%) 6.2B 0.65

Lactate-Diacetate (1.6+0.1%) 5.9A 0.56 Potassium Sorbate (0.3%) 5.8A 0.44

Lactate-Diacetate (1.6+0.1%) 5.9A 0.39 Sodium Propionate (0.3%) 6.2B 0.61

1Consumer response: Preference rating (1=dislike very much to 7= like very much; 4=neither like or dislike). For overall preference, consumers responses for preferred were entered as a 1 and not preferred as 0. More than 160 consumers per treatment comparison were used. A, B Means within a comparison and sensory response with unlike superscript letters are different (P<0.05)

Glass and Claus, AMIF Final Report, June 2006 10 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Table 5. Consumer evaluation of Deli-Style Turkey containing potassium sorbate (0.3%) or sodium propionate (0.3%). Sensory Response1 Treatment comparison Preference Rating Overall preference Control 5.6A 0.39A Potassium Sorbate (0.3%) 6.0A 0.59A Std. error 0.28 0.11

Control 6.1A 0.69A Sodium Propionate (0.3%) 5.3B 0.31B Std. error 0.26 0.11

Potassium Sorbate (0.3%) 5.5A 0.54A Sodium Propionate (0.3%) 5.5A 0.46A Std. error 0.28 0.11

1Consumer response: Preference rating (1=dislike very much to 7= like very much;4=neither like or dislike). For overall preference, consumers responses for preferred were entered as a 1 and not preferred as 0. Forty different consumers per treatment comparison were used. A, B Means within a comparison and sensory response with unlike superscript letters are different (P<0.05)

Literature Review: A literature review describing the safety and sensitivities exhibited by individuals to the various antimicrobials is detailed in Appendix B. Overall, data indicate that these compounds are of low toxicity with little or no genotoxic or carcinogenic potential.

CONCLUSIONS

This research verifies that propionate, benzoate, and sorbate will enhance the safety of high-moisture, RTE cured and uncured meat and poultry products and that addition of these antimicrobials will have little negative impact on consumer taste preference. Data can be used to petition FSIS for approval of propionate, benzoate, and sorbate for use in product formulations to control L. monocytogenes.

Glass and Claus, AMIF Final Report, June 2006 11 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

ACKNOWLEDGEMENTS

Thank you to Scott Wyman, Nicole Hardy, and Garret Kau for technical assistance in completing laboratory work. We greatly appreciate the donations of test product by Sara Lee Foods and Land O’Frost and helpful discussions with Dr. Andy Milkowski, Kraft-Oscar Mayer (retired). This work was funded by the American Meat Institute Foundation, University of Wisconsin-Madison College of Agriculture and Life Sciences and by contributions from the food industry.

REFERENCES

1. Amerine, M., Pangborn, R. M., and Roessler, E. B. 1965. "Principles of Sensory Evaluation of Food." Academic Press, Inc., New York, NY.

2. Anonymous. 2001. Summary report: results of interviews regarding lactate and diacetate use in RTE meat and poultry products - May 14, 2001. Office of Policy, Program Development and Evaluation (OPPDE), Food Safety and Inspection Service, US Department of Agriculture. www.fsis.usda.gov/oppde/peis/Reports/OPPDE/Summary_Rpt.pdf Accessed 21 March 2006.

3. Anonymous. 2003. Control of Listeria monocytogenes in Ready-to-Eat Meat and Poultry Products; Final Rule. Fed. Reg. 68: 34207-34254.

4. AOAC. 2000. Official Methods of Analysis. 17th ed. Association of Official Analytical Chemists, Gaithersburg, MD.

5. ASTM. 1988. ASTM Standards On Sensory Evaluation Of Materials And Products. ASTM Committee E-18, American Society for Testing and Materials, Philadelphia, PA.

6. Berry, B.W., Dethmers, A.E., Gillett, T.A., Marshall, W.H., Moody, W.G., Olson, D.G., Rainey, B.A., and Schwartz, W.C. 1983. AMSA guidelines for sensory, physical, and chemical measurements. Proceedings 36th Annual Reciprocal Meat Conference of the American Meat Sci. p. 221-228. Assoc. North Dakota State University, Fargo, North Dakota.

7. El-Shenawy, M.A., and E.H. Marth 1991. Organic acids enhance the antilisterial activity of potassium sorbate. J. Food Prot. 54:593-597.

8. Glass, K.A., and M.P. Doyle. 1989. Fate of Listeria monocytogenes in processed meat products during refrigerated storage. Appl. Environ. Microbiol. 55:1565-1569.

9. Glass, K.A., D. A. Granberg, A. L. Smith, A. M. McNamara, M. Hardin, J. Mattias, K. Ladwig, and E. A. Johnson. 2002. Inhibition of Listeria monocytogenes by sodium diacetate and sodium lactate on wieners and cooked bratwurst. J. Food Prot. 65:116–123.

Glass and Claus, AMIF Final Report, June 2006 12 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

10. Glass, K.A., J. Veesenmeyer, L.L. McDonnell, P. Eimerman, and E.A. Johnson. 2004. Control of Listeria monocytogenes in wiener and turkey slurries by combinations of antimicrobials J. Food Prot. 67 (Sup. A): 55.

11. Islam, M., J. Chen, M.P. Doyle, and M. Chinnan. 2002a. Control of Listeria monocytogenes on turkey frankfurters by generally-recognized-as-safe preservatives. J. Food Prot. 65:1411-1416.

12. Islam, M., J. Chen, M.P. Doyle, and M. Chinnan. 2002b. Effect of selected generally-recognized-as-safe preservative sprays on growth of Listeria monocytogenes on chicken luncheon meat. J. Food Prot. 65:794-798.

13. Legan, J. D., D. L. Seman, A. L. Milkowski, J. A. Hirschey, and M. H. Vandeven. 2004. Modeling the growth boundary of Listeria monocytogenes in ready-to-eat cooked meat products as a function of the product salt, moisture, potassium lactate, and sodium diacetate concentrations. J. Food Prot. 67:2195-2204.

14. Mbandi, E. and L. A. Shelef. 2002. Enhanced antimicrobial effects of combination of lactate and diacetate on Listeria monocytogenes and Salmonella Spp. In Beef Bologna. Int. J. Food Microbiol. 76:191-198.

15. Preston, D. J. Vessenmeyer, E. A. Johnson, and K. A. Glass. 2005. Inhibition of Listeria monocytogenes in processed meat and poultry by combinations of sorbate, benzoate, and propionate. J. Food Prot 68 (Sup. A):125

16. Samelis, J., J.N. Sofos, M.L. Kain, J.A. Scanga, K.E. Belk, and G.C. Smith. 2001. Organic acids and their salts as dipping solutions to control Listeria monocytogenes inoculated following processing of sliced pork bologna stored at 4 degrees C in vacuum packages. J. Food Prot. 64:1722-1729.

17. Seman, D.L., A.C. Borger, J.D. Meyer, P.A. Hall, and A.L. Milkowski. 2002. Modeling the growth of Listeria monocytogenes in cured ready-to-eat processed meat products by manipulation of sodium chloride, sodium diacetate, potassium lactate, and product moisture content. J. Food Prot. 65:651-658.

18. Tompkin, R. B., L. N. Christiansen, A. B. Shaparis, and H. Bolin. 1974. Effect of potassium sorbate on Salmonellae, Staphylococcus aureus, Clostridium perfringens, and Clostridium botulinum in cooked, uncured sausage. Appl. Microbiol. 28:262-264

19. United States Food and Drug Administration. 2004. Substances Generally Recognized as Safe. Code of Federal Regulations, Title 21, Part 182. Office of the Federal Register, U.S. Government Printing Office, Washington, DC.

20. Wederquist, H.J., J.N. Sofos, and G.R. Schmidt. 1994. Listeria monocytogenes inhibition in refrigerated vacuum packaged turkey bologna by chemical additives. J. Food Sci. 59:498-500, 516.

Glass and Claus, AMIF Final Report, June 2006 13 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Figure 1. Changes in populations of L. monocytogenes on cured ham prepared with various levels of sodium benzoate, sodium propionate, or potassium sorbate, and stored at 4°C for 12 weeks (averages for duplicate trials; standard deviations not shown).

6

5

e 4 s n i ml r /

U 3 F C

Log 2 nge a h

C 1 Weeks at 4°C 10 6 0 2

-1 l % % % % % % o 5 1% 1% 05% 3% 05% 3% . 05% . . . . . 0.1 075% 0.05 0.2% 0.2% 0.0 ontr e 0.1 e 0.1 0.05 te 0 C te 0 te 0 te 0. ate 0 ate 0 ate 0 ate nate te 0. rbat bate nate oate nate r ona rb rb io o o i ona ona z nzoat o on i o i o o p S n e acet ona p S op S S i i enzoa o e opi r B o op r opi opi p r r B r r r P D o B P P P P r P P + + + P + + % + 1% . 05% 05% 05% 5% . . e 1.6 e 0. t ate 0 ctat te 0 0.07 rb ate 0 a a rba o L o rb S S o bate r S enzo o B S

Glass and Claus, AMIF Final Report, June 2006 14 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison

Figure 2. Changes in populations of L. monocytogenes on uncured turkey prepared with various levels of sodium benzoate, sodium propionate, or potassium sorbate, and stored at 4°C for 12 weeks (averages for duplicate trials; standard deviations not shown).

6

5

e 4 ns i r ml /

U 3 F

2 Log C ange

h 1 C Weeks at 4°C 10 6 0 2

-1 r p % % l p % 1% o o 5% so . ro r 2% 2% - 05% 05% t p . . 0.1 0.1% - n -pr 0.2% 0.1 0.05 e 0 a a t e 0. 0.07 t te 0. Co a N 15 K ate 0.3% ate bate ate 0.3% o ate 0 ate 0 nate t nate r 0. oate o ate rba 15 N o rb o 15 ona nz on on on . i o o S 0. e on ace opi S S i opi enz r B op opi opi opi r r r r r P r / 0 D B opi P P P P r

P + + so + - P + + % % + 2% 5 . 15 K 05% 05% . . 0. te 3 e 0.1 e 0.0 075% e 0 t te 0 bat bat r r e 0. o Lacta t o rba S S o S rba enzoa o B S

Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

APPENDIX A

Report on the

Consumer Preference Evaluations of Sliced Hams

for

Dr. James R. Claus

University of Wisconsin-Madison Meat Science & Muscle Biology Laboratory 1805 Linden Dr. West Madison, WI 53706

by the

Sensory Analysis Laboratory Department of Food Science University of Wisconsin-Madison Madison, WI 53706

Date: June 28, 2006 revised Project No: #337 Sensory Laboratory Manager: Sungjoon Jang

1 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Introduction

The Sensory Analysis Laboratory, Department of Food Science, University of Wisconsin- Madison, was contacted by Dr. James R. Claus, UW-Madison, Meat Science and Muscle Biology Laboratory, Madison, WI., regarding consumer preference evaluations of sliced hams. The objective of the evaluations was to determine consumer preferences for four or three pairs of sliced ham samples in each of two series of testing, respectively. Standard sensory evaluation conditions were slightly modified with agreement between Mr. Sungjoon Jang (UW-Madison, Sensory Laboratory) and Dr. James Claus (UW-Madison, Meat Science & Muscle Biology) to meet the specific requests by Dr. Claus for his experimental purposes. The agreements included that each session was to be composed of four pairs of sliced hams for series I or three pairs of sliced hams for series II, the total numbers of panelists at a session to be served each pair of samples would be 50 or 66 for each series (I and II) of evaluation, after the number of panelist evaluating the first pair of samples was met, the second and third, or fourth pair of sliced hams would be consecutively evaluated. In a session, each panelist would evaluate only one of the three or four pairs, and each session should be completed within the testing for a day. This report summarizes the consumer preference evaluations that were conducted on February 7, 9, 14, and 20, 2006 for the first series and on February 23, 27, and 28, 2006 for the second series. Tabulated panel results were provided to Dr. Claus earlier as they were accumulated (March 15, 2006).

Materials and Methods Samples Sample lots of sliced hams were delivered to the Sensory Analysis Laboratory by a graduate student of Dr. Claus on each morning of sensory evaluation sessions. Sliced hams for each sample lot were individually sealed, and were designated by Dr. Claus as "Control", "LD", "AM1", "AM2", and "AM3" for first series evaluated on February 7, 9, 14, 20, and as "LD", "AM1", "AM2", and "AM3" for second series evaluated on February 23, 27, and 28, 2006. The Control did not have a test antimicrobial added. The test antimicrobials were: 0.1% sodium benzoate (AM1), 0.3% potassium sorbate (AM2), and 0.3% sodium propionate (AM3). The samples among lots were paired appropriately for each evaluation. Samples were in an excellent cold condition upon receipt. Sample pairs for first series of consumer preference evaluation were "Control vs LD", "Control vs AM1", "Control vs AM2", and "Control vs AM3", and each pair was evaluated (maximum 50 judgments each) on February 7, 9, 14, and 20, 2006. Sample pairs for second series of consumer preference evaluation were "LD vs AM1", "LD vs AM2", and "LD vs AM3", and each pair was evaluated (maximum 66 judgment each) on February 23, 27, and 28, 2006.

2 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Preparation of Samples for Serving Samples of sliced ham were served after further preparation to cut each slice to four pieces as requested by Dr. James Claus (UW-Madison, Meat Science & Muscle Biology). Each refrigerated sliced ham (oval-shape) was center-cut both vertically and horizontally into quadrants to obtain four similar slice-pieces (ca 5.0 cm X 6.0 cm, one-quarter of original oval shape; ca 0.1 cm thickness). Sample preparation was conducted ca 20 min before each test. The slice samples were placed in the original plastic zip-closure bags, and stored in a household refrigerator (ca 4 °C) until served.

Taste Panel Testing Conditions Consumer preference panels were held in the Consumer Testing Laboratory adjacent to the Dairy Products Salesroom located in Babcock Hall on the University of Wisconsin-Madison campus. For the first series of evaluations, unscreened participants (total of 215, 202, 217, and 164, respectively) evaluated each sample pair ("Control vs LD", "Control vs AM1", "Control vs AM2", and "Control vs AM3", respectively) during four consecutive sessions on February 7, 9, 14, and 20, 2006. For the second series of evaluations, unscreened participants (total of 197, 193, and 189, respectively) evaluated each sample pair ("LD vs AM1", "LD vs AM2", and "LD vs AM3", respectively) during three consecutive sessions on February 23, 27, and 28, 2006. At the time of each panel evaluation, samples were portioned by placing two pieces (ca 5.0 cm X 6.0 cm one- quarter of an oval shape; ca 0.1 cm thickness) of each sliced ham into separate 2 oz plastic portion cups which were each coded with a three-digit random number for identification of the sample. The sliced ham samples were placed on a serving tray as they were removed from the refrigerator immediately after each panelist was seated in a testing booth. Each panelist received a ballot (Appendixes I, II, III, and IV for first series and V, VI, and VII for second series) for recording responses, a tray containing two pieces of sliced ham samples, and two plastic forks. Cold drink cups and napkins were placed in a corner of each sensory booth for panelist free-choice use.

Ballot The ballot for consumer preference evaluation of sliced ham samples was specially designed for this project, and it was finalized in collaboration with Dr. James Claus. The top part of the consumer preference ballot contained a structured 7-point hedonic scale (Amerine et al., 1965). After tasting both samples, panelists were also asked to answer two exit questions which were requested by Dr. Claus. Question 1 was “Which of the two samples you tasted do you prefer?” Question 2 was “Why do you prefer the sample checked in question 1 better than the other sample?” The ballot also contained an “Other comments” section to collect voluntary comments by panelists. To minimize the sample presentation sequence effects, the ballots for each pair samples were constructed with two permutations of two code numbers. Each group contained an equal numbers of ballots, and each ballot contained instructions for the tasting sequence of the samples. Copies of the ballots used in the test are presented in Appendixes I, II, III, IV, V, VI, and VII.

Statistical Analysis For calculation of overall preference scores, a code value of 1.0 was assigned to the category of “dislike very much” and a value of 7.0 to the category of “like very much’ with appropriate whole- number code values assigned to the intermediate categories. The coded values from the panel sessions were subjected to analysis of variance appropriate for a randomized complete block design (Steel and Torrie, 1960). For the overall preference attribute, statistical analyses provided the mean score for each sample, the F-value for all samples, and the least significant difference (LSD) for making sample comparisons.

3 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

If the F-value is statistically significant, then the null hypothesis of no difference among population means is rejected. A significant F-value implies that the evaluation provided evidence of real differences among treatment means that would not be expected to occur by chance more than 5 % of the time. Because the F-value does not indicate which of theses differences can be considered significant, the LSD value computed for a 5 % level of significance is used for comparisons of paired means. If the difference between any pair of treatment mean scores within a sensory attribute exceeds the LSD, then that difference is considered statistically significant. If the F-value for treatments is not significant, then the evidence is not strong enough to indicate a real difference among treatment means at a 5 % level of significance, and specific treatment comparisons cannot be made.

Results and Discussion Results and discussion of the consumer preference evaluations of sliced ham samples are organized into a sequence providing the combined results for each pair of samples evaluated on four or three consecutive days in series I and II, respectively. The order of presentation is Sample pair I: Control vs LD, Sample pair II: Control vs AM1, Sample pair III: Control vs AM2, and Sample pair IV: Control vs AM3 for first series of evaluations and Sample pair V: LD vs AM1, Sample pair VI: LD vs AM2, and Sample pair VII: LD vs AM3 for second series of evaluations.

Sample Pair I: Control vs LD Sliced Ham (Evaluated on February 7, 9, 14, and 20, 2006; maximum 50 observations for a pair on each day)

Results of the consumer preference evaluation of the Control and LD sliced ham samples are summarized in Table 1. The results show that the preference mean score of the Control sliced ham sample was statistically significantly higher at the 5% level than that for the LD sliced ham sample (mean preference scores were 5.83 and 5.47, respectively). The panelist response distribution data (Table 1) revealed that the numbers of panelist response for the Control sliced ham sample (72) was much greater than for the LD sliced ham (52) in the top preference rating of "like very much" category. A relatively larger number of panelists also rated the LD sliced ham sample (85) in the "like moderately" category compared to the LD sliced ham sample (77). In contrast, larger numbers of panelists rated the LD sliced ham sample in the "like slightly" compared to responses for the Control sliced ham sample (47 vs 28; Table 1). A summary of panelist responses to exit question #1 (“Which of the two samples you tasted do you prefer?”) is shown in Table 2. One hundred thirty two (132) out of 208 panelists expressed their preferences for the Control sliced ham sample than those for the LD sliced ham sample (76). The responses to question #2 (“why do you prefer the sample checked in question #1...?”) were summarized into two groups (i.e., those who preferred the Control sliced ham sample, and those who preferred the LD sliced ham sample). The responses from the panelists who preferred the Control sample are listed in Table 3. The responses from the panelists who preferred the LD sample are listed in Table 4. Of the comments from panelists who preferred the Control sliced ham sample the most (Table 3), and who commented, a large number of panelist responses indicated that the Control sample tasted “less salty” (52), "better” (24), had "better flavor” (20), "better texture" (8), and tasted "sweeter" (8). Some panelists commented that the Control sliced ham sample tasted "saltier" (4), has "more smoked flavor” (4), and tasted "richer" (2). The major comments for panelists who preferred the LD sliced ham sample the most (Table 4) were that the LD sample had "better flavor" (24), “better texture” (11), tasted "better" (10), "saltier" (10), and "less salty" (8). Some comments from panelists included that the LD sliced ham

4 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

sample tasted "sweeter" (6), had "more smoked flavor” (6), tasted "richer" (5), "less tart" (1), and had " less smoked flavor" (1). Panelist optional voluntary comments for this pair of samples are summarized in Table 5. Some comments indicated that the Control sliced ham sample tasted "too bland" (2), had "bad texture" (1), tasted "too sweet" (1), and had "off-flavor" (1). On the other hand, a notable number of panelists commented that the LD sliced ham sample had "bad aftertaste" (5), tasted "bland" (3), and "too salty" (2). Several other minor comments for the LD sample are listed in Table 5.

Sample Pair II: Control vs AM1 0.1% Sodium benzoate Sliced Ham (Evaluated on February 7, 9, 14, and 20, 2006; maximum 50 observations for a pair on each day)

Results of the consumer preference evaluation of the Control and AM1 sliced ham samples are summarized in Table 6. The results show that there was no statistically significant preference for either the Control sample or the AM1 sample at the 5% level, and the mean preference scores were 6.03 and 6.02, respectively. The panelist response distribution data (Table 6) show that similar response distribution patterns were obtained for most of the preference rating categories for both the Control and the AM1 sliced ham samples. The summary of panelist responses to exit question #1 (“Which of the two samples you tasted do you prefer?”) is shown in Table 7. Of the 197 panelists, ninety seven (97) panelists preferred the Control sliced ham sample, and one hundred (100) preferred the AM1 sample. The responses to question #2 (“why do you prefer the sample checked in question #1...?”) are summarized in Tables 8 and 9 for the Control and the AM1 sliced ham samples, respectively. The responses from panelists who preferred the Control sliced ham sample (Table 8) and commented, showed that the Control sample had "better" flavor (26), tasted "less salty" (21), had "better texture" (15), and tasted "better" (8). Some minor responses from panelists who preferred the Control sample the most included that the Control sliced ham sample tasted "less fatty" (3), had "more smoked flavor" (3), "better aftertaste" (2), tasted "mild" (1), "sweeter" (1), had "less smoked flavor" (1), and "more moisture" (1) than those for the AM1 sliced ham sample (Table 8). In comparison, a number of panelists preferring the AM1sample (Table 9) responded that the AM1 sliced ham sample tasted "less salty" (33), had "better flavor" (29), tasted "better" (15), and tasted "sweeter" (9) than did those for the Control sample. Several comments from panelists who preferred the AM1 sliced ham sample the most included that the AM1 sample tasted "saltier" (7), had "better texture" (7), "more smoked flavor" (7), tasted "richer" (3), "milder" (2), "spicier" (1), "less sweet" (1), and had more moisture" (1). General voluntary comments for this pair of samples (Control vs AM1) are summarized in Table 10. A few panelists commented that the Control sample had "bad aftertaste" (3), "artificial taste" (1), "a tough casing" (1), had "too strong pork smell" (1), tasted "raw meat-like" (1), and "not fresh" (1). Only two panelists commented that the AM1 sliced ham sample had "off-flavor" (1), and tasted "too salty" (1).

Sample Pair III: Control vs AM2 0.3% Potassium Sorbate Sliced Ham (Evaluated on February 7, 9, 14, and 20, 2006; maximum 50 observations for a pair on each day)

Results of the consumer preference evaluation of the Control and the AM2 sliced ham samples are summarized in Table 11. The results indicated that the Control sliced ham sample was statistically significantly more preferred than the AM2 sliced ham sample at the 5% level, and the mean preference scores were 6.06 and 5.88, respectively.

5 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

The panelist response distribution data (Table 11) revealed that seventy four (74) out of 217 panelists scored the Control sliced ham sample in the category of "like very much" compared to sixty four (64) of 217 panelists for the AM2 sliced ham sample. The Control sliced ham sample received three (3) more responses from panelists for “like moderately” category than did the AM2 sample (95 and 92, respectively). In contrast, the Control sample received fewer panelist responses for the "like slightly" category than the AM2 sliced ham sample (38 and 46, respectively). Somewhat larger numbers of panelists rated the AM2 sliced ham sample in the lower preference categories compared to parallel ratings for the Control sliced ham sample (Table 11). The summary of panelist responses to exit question #1 (“Which of the two samples you tasted do you prefer?”) is shown in Table 12. One hundred nineteen (119) out of 213 panelists expressed their preferences for the Control sliced ham sample than those for the AM2 sliced ham sample (94). Of the comments from panelists who preferred the Control sliced ham sample the most (Table 13), and who commented, a large number of panelist responses indicated that the Control sample in Pair III had "better flavor” (25), tasted "saltier” (18), "better” (17), "less salty" (15), and had "more flavor" (13). Several additional comments for the Control sample also were given, and included that the Control sliced ham sample had "better texture" (7), tasted "sweeter" (5), had "less smoked flavor" (3), "better aftertaste" (2), tasted "juicer" (2), and "less sweet" (2). The comments from panelists who preferred the AM2 sliced ham sample the most are summarized in Table 14, and relatively fewer responses were obtained from panelists for question #2, "why do you prefer the sample...?" than those for the Control sliced sample. The comments from panelists who preferred the AM2 sliced ham sample the most indicated that the AM2 sample (Table 14) tasted "less salty" (19), "sweeter" (14), had "more flavor" (11), "better texture" (10), tasted "better" (9), "saltier" (8), and had "more smoked flavor" (8). Several comments also indicated that the AM2 sliced ham sample had "less aftertaste" (4), tasted "juicier" (2), "less buttery" (1), and "milder" (1). General voluntary comments for this session (Sample Pair III) are summarized in Table 15. Some panelists commented that the Control sliced ham tasted "too salty" (4), had "rubbery texture" (2), tasted "too sweet" (1), had "bad aftertaste" (1), and "too much smoked flavor" (1). The voluntary comments from panelists for the AM2 slice ham sample included that the AM2 sample had " off flavor" (7), tasted "too sweet" (4), "too salty" (3), had "slimy texture" (2), "slightly more smoked flavor" (1), was "tender" (1), and tasted "bland" (1).

Sample Pair IV: Control vs AM3 0.3% Sodium Propionate Sliced Ham (Evaluated on February 7, 9, 14, and 20, 2006; maximum 50 observations for a pair on each day)

Results of the consumer preference evaluation of the Control and AM3 sliced ham samples are summarized in Table 16. The results show that there was no statistically significant preference for either the Control or the AM3 sliced ham sample at the 5% level, and the mean preference scores were 5.99 and 6.02, respectively. The Control sliced ham sample received four (4) more panelist responses for both “like moderately” and "like slightly" categories than did the AM3 sample (Table 16). However, the Control sliced ham sample received fewer panelist responses for "like moderately" than did the AM3 sample (65 and 75, respectively). The summary of panelist responses to exit questions #1 (“Which of the two samples you tasted do you prefer?) is shown in Table 17. Panelists gave merely equal preference responses to the Control and AM3 sliced ham samples for exit question #1. Seventy eight (78) panelists stated they

6 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

preferred the Control sliced ham sample compared to eighty one (81) who expressed their preference for the AM3 sample. The responses to question #2 (“why do you prefer the sample checked in question #1...?”) were summarized into two groups (i.e., those who preferred the Control sliced ham sample, and those who preferred the AM3 sample). The responses from the panelists who preferred the Control sliced ham sample are listed in Table 18. The responses from the panelists who preferred the AM3 sample are listed in Table 19. Of the comments from panelists who preferred the Control sliced ham sample the most (Table 18), and who commented, a notable number of panelist responses indicated that the Control sample tasted “less salty” (24), tastes "better” (15), and had "better flavor” (15). Some comments for panelists who preferred the Control sliced ham sample the most included that the Control sample had "more smoked flavor" (6), tasted "saltier" (5), "sweeter" (4), had "more moisture" (4), and tasted "less fatty" (3). The major comment categories for panelists who preferred the AM3 sliced ham sample the most (Table 19) included that the AM3 sample had "better flavor" (23), tasted “less salty" (18), "better” (12), "saltier” (9), and "sweeter" (8). Several comments also indicated that the AM3 sliced ham sample had "more smoked flavor" (6), "better texture" (6), tasted "milder" (3), had "better aftertaste" (1), "better appearance" (1), and "less smoked flavor" (1). A few general voluntary comments for this pair of samples are summarized in Table 20. The comments for the Control sliced ham sample included that it tasted "greasy" (4), had "better aftertaste" (3), and tasted "bland" (2). Panelist comments for the AM3 sliced ham sample included that it tasted "bitter" (2), had "dry mouthfeel" (1), and "too strong aftertaste" (1).

7 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair V: LD vs AM1 0.1% Sodium benzoate Sliced Ham (Evaluated on February 23, 27, and 28, 2006; maximum 66 observations on each day)

Results of the consumer preference evaluation of the LD and AM1 sliced ham samples are summarized in Table 21. The results show that there was a statistically significant greater preference for the AM1 sliced ham sample than for the LD sample at the 5% level, and the mean preference scores for the AM1 and LD ham slice samples were 6.15 and 5.64, respectively. A larger number of panelists expressed that they preferred the AM1 sliced ham sample more than the LD sample in "like very much" preference category (77 and 47, respectively). Correspondingly, the LD sliced ham sample received fifteen (15) more panelist responses for "like slightly" category than the AM1 sliced ham (41 and 26, respectively). The summary of panelist responses to exit questions #1 (“Which of the two samples you tasted do you prefer?) is shown in Table 22. A larger number of panelists again expressed their preference for the AM1 sliced ham sample compared to the LD sliced ham sample (124 and 67 responses, respectively). Responses to question #2 (“why do you prefer the sample checked in question #1...?”) are summarized into two groups, i.e., those who preferred the LD sliced ham sample, and those who preferred the AM1 sliced ham sample. The responses from the panelists who preferred the LD sample are listed in Table 23. The responses from the panelists who preferred the AM1 sample are listed in Table 24. Of the comments from panelists who preferred the LD sliced ham sample the most, and who commented, a large number of panelist responses indicated that the LD sliced ham sample had “better flavor” (20), tasted "better” (10), "saltier” (10), "sweeter" (8), had "better texture" (6), tasted "milder" (5), and "less salty" (4). The major comment categories for panelists who preferred the AM1 sliced ham sample the most (Table 24) included that the AM1 sample tasted “less salty" (49), had “better flavor” (34), tasted "better" (16), had "better texture" (13), and tasted "sweeter" (6). A few panelists also commented that the AM1 sliced ham tasted "saltier" (3), had "more smoked flavor" (3), and "better aftertaste" (3). A summary of panelist voluntary comments for this pair of samples (LD vs AM1) is summarized in Table 25. The comments for the LD sliced ham sample included that the LD sample tasted "too salty" (11), "slightly bitter" (3), had "unclean off-flavor" (3), tasted "bland" (2), had "discolored edge" (1), and "soft/mushy texture" (1). Panelists also commented that the AM1 sliced ham sample tasted "too salty" (2), had "too chewy texture" (1), tasted "greasy" (1), and "slightly bland" (1).

Sample Pair VI: LD vs AM2 0.3% Potassium Sorbate Sliced Ham (Evaluated on February 23, 27, and 28, 2006; maximum 66 observations on each day)

Results of the consumer preference evaluation of the LD and AM2 sliced ham samples are summarized in Table 26. The results show that there was no statistically significant preference for either the LD or the AM2 sliced ham sample at the 5% level, and the mean preference scores were 5.90 and 5.82, respectively. The panelist response distribution data (Table 26) revealed that the numbers of panelist responses showed parallel distributions of patterns in most of the preference rating categories between the LD and AM2 sliced ham samples. However, a slightly larger number of panelists rated the LD sample in both the "like very much" and "like slightly" categories compared to those responses for the AM2 sample (Table 26). In contrast, a slightly larger number of panelists rated the

8 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

AM2 sliced ham sample (75) in the "like moderately" category compared to responses for the LD sliced ham sample (67). In addition, a slightly larger number of panelists rated the AM2 sliced ham sample (10) in the "neither like nor dislike" category compared to responses for the LD sliced ham sample (5). A summary of panelist responses to exit questions #1 (“Which of the two samples you tasted do you prefer?) is shown in Table 27. A relatively larger number of panelists expressed their preference for the LD sliced ham compared to those response for the AM2 sliced ham sample (Table 27). The responses to question #2 (“why do you prefer the sample checked in question #1...?”) were summarized into two groups (i.e., those who preferred the LD sliced ham sample, and those who preferred the AM2 sliced ham sample). The responses from the panelists who preferred the LD sliced ham sample are listed in Table 28. The responses from the panelists who preferred the AM2 sliced ham sample are listed in Table 29. Of the comments from panelists who preferred the LD sliced ham sample (Table 28) the most, and who commented, a notable number of panelist responses indicated that the LD sample had "better flavor” (36), tasted "saltier” (18), "better” (14), had "better texture" (13), and tasted "less salty" (7). The major comment categories for panelists who preferred the AM2 sliced ham sample (Table 29) the most included that the AM2 sample tasted “less salty" (20), had “better flavor” (18), tasted "better” (12), "sweeter" (10), had "more smoked flavor" (6), and "better texture" (6). Panelist voluntary comments for this pair (VI) of samples (LD vs AM2) are summarized in Table 30. The comments included that the LD sliced ham sample had "off-flavor" (5), "bad aftertaste" (3), tasted "too salty" (2), "too fatty" (1), and "bland" (1). Several panelists commented that the AM2 sliced ham sample had "bad aftertaste" (5), tasted "too salty" (4), "too fatty" (3), "too sweet" (1), "soapy" (1), and had "tough texture" (1).

Sample Pair VII: LD vs AM3 0.3% Sodium Propionate Sliced Ham (Evaluated on February 23, 27, and 28, 2006; maximum 66 observations on each day)

Results of the consumer preference evaluation of the LD and AM3 sliced ham samples are summarized in Table 31. The results show that the preference score of the AM3 sliced ham sample was statistically significantly higher at the 5% level than that for the LD sliced ham sample (mean preference scores were 6.17 and 5.86, respectively). The panelist response distribution data (Table 31) revealed that the numbers of panelist responses for the AM3 sliced ham sample (87) was much greater than for the LD sliced ham (56) in the top preference rating of "like very much" category. In contrast, larger numbers of panelists rated the LD sliced ham sample in both the "like moderately" and "like slightly" preference categories compared to responses for the AM3 sliced ham sample (Table 31). A somewhat notable numbers of panelists rated the LD sliced ham sample in the "dislike slightly" category compared to parallel rating for the AM3 sliced ham sample (Table 31). A summary of panelist responses to exit question #1 (“Which of the two samples you tasted do you prefer?”) is shown in Table 32. One hundred fifteen (115) out of 188 panelists expressed their preferences for the AM3 sliced ham sample while 73 preferred the LD sliced ham sample. The responses to question #2 (“why do you prefer the sample checked in question #1...?”) were summarized into two groups (i.e., those who preferred the LD sliced ham sample, and those who preferred the AM3 sliced ham sample). The responses from the panelists who preferred the LD sample are listed in Table 33. The responses from the panelists who preferred the AM3 sample are listed in Table 34.

9 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Of the comments from panelists who preferred the LD sliced ham sample the most (Table 33), and who commented, a notable number of panelist responses indicated that the LD sample had "better texture” (13), "better flavor” (12), tasted "better” (11), "less salty" (9), "saltier" (8), and had "more smoked flavor" (5). Several minor comments from panelists who preferred the LD sliced ham sample were also listed in Table 33. Relatively larger numbers of panelists who preferred the AM3 sliced ham sample the most (Table 34) commented that the AM3 sample tasted "less salty" (36), had "better flavor” (28), tasted "better" (15), had "better texture" (14), tasted "saltier" (7), and had "more smoked flavor" (6). Some minor comments from panelists included that the AM3 sliced ham sample tasted "sweeter" (3), "juicier” (2), "less fatty" (2), "tangier" (2), had "less smoked flavor" (1), "better color" (1), and "less chewy" (1). Panelist optional voluntary comments for this pair of samples (VII) are summarized in Table 35. Some comments indicated that the LD sliced ham sample tasted "too salty" (5), had "a little slimy texture" (3), "off-flavor" (2), and tasted "slightly bitter" (2). On the other hand, a few number of panelists commented that the AM3 sliced ham sample tasted "sour" (1), "too fatty" (1), and had "off-flavor" (1).

Summary

Results of consumer preference evaluations of sliced ham samples showed statistically significant differences between common Control sample and experimental sample (LD or AM2) in Pairs I and III, respectively, and also revealed that the common Control sample was statistically significantly more preferred than both the LD or AM2 sliced ham samples. Panelists perceived that the Control sample was notably or significantly less salty, better flavor, and better taste than the LD or AM2 sample in both Pairs I and II. For Sample Pair II: Control vs AM1 and Sample Pair IV: Control vs AM3 sliced ham samples, consumer preference evaluations did not show statistically significant preferences for either pair of samples. Greater numbers of panelists generally responded that saltiness and flavor were major influential factors for their preference choices of sliced ham products in the Sample Pair II and IV. The preference evaluations of Sample Pairs V (LD vs AM1) and VII (LD vs AM3) revealed that the LD sliced ham sample was statistically significantly less preferred than both the AM1 and AM3 sliced ham samples. Results of both Pairs V and VII of sliced ham samples indicate that saltiness intensity play a key role in the preference for the experiment sliced ham sample (AM1 and AM3) versus the common LD sliced ham sample. Besides responses for the saltiness attribute, panelists gave notably large responses for flavor and texture descriptors for their preference choice of sliced ham product in both Sample Pair V and VII sensory evaluations. For Sample Pair VI, panelists showed no statistically significant preference between the LD and AM2 sliced ham samples. Panelists commonly cited flavor, saltiness, and taste descriptors as the main reasons for their preference choices expressed in Sample Pair VI.

10 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Bibliography

Amerine, M., Pangborn, R. M., and Roessler, E. B. 1965. “Principles of Sensory Evaluation of Food.” Academic Press, Inc., New York, NY.

ASTM. 1988. ASTM Standards On Sensory Evaluation Of Materials And Products. ASTM Committee E-18, American Society for Testing and Materials, Philadelphia, PA, 75 pp.

Aust, L. B., Gacula, M. C., Beard, S. A. and Washam, R. W. 1985. Degree Of Difference Test Method In Sensory Evaluation Of Heterogeneous Product Types. J. Food Sci. 50:511.

Burgard, D. R., and Kuznicki, J. T. 1990. Chemometrics: Chemical and Sensory Data. CRC Press, Boca Raton, FL, 196 pp.

Chambers, E., and Wolf, M. B., eds. 1996. Sensory Testing Methods. 2nd Edition, ASTM manual series MNL 26, American Society for Testing Materials, West Conhohocken, PA. 115 pp.

Gacula, M. C. and Singh, J. 1984. Statistical Methods in Food and Consumer Research. Academic Press, Inc., Orlando, FL.

Hootman, R. C., ed. 1992. Manual on Descriptive Analysis for Sensory Evaluation. ASTM Manual Series MNL 13, American Society for Testing and Materials, Philadelphia, PA. 52 pp.

IFT. 1981. Sensory Evaluation Guide for Testing Food and Beverage Products. Sensory Evaluation Division, Institute of Food Technologists. Food Technol. 35(11): 50-59.

Lawless, H. T., and Klein, B. P. eds. 1991. Sensory Science: Theory and Applications in Foods. IFT Basic Symposium Series. Marcel Dekker, New York, 441 pp.

Meilgaard, M., Civille, G. V. and Carr, B. T. 1991.. Sensory Evaluation Techniques. 2nd ed.,.CRC Press, Inc., Boca Raton, FL.

Munoz, A. M., ed. 1997. Relating Consumer, Descriptive, and Laboratory Data. ASTM publication code 28-030097. American Society for Testing and materials, West Conshohocken, PA. 103 pp.

Moskowitz, H. R. 1983. Product Testing and Sensory Evaluation of Foods. Food and Nutrition Press, Inc., Westport, CT.

O’Mahony, M. 1986. Sensory Evaluation of Food: Statistical Methods and Procedures. Marcel Dekker, Inc., New York, NY.

Steel, R. G. D. and Torrie, J. H. 1960. Principles and Procedures of Statistics. McGraw-Hill Publishing Co., Inc., New York, NY.

Stone, H., Sidel, J. 1985. Sensory Evaluation Practices. Academic Press, Inc., New York, NY.

Wu, L. S., ed. 1989. Product Testing with Consumers for Research Guidance. STP 1035, American Society for Testing and Materials(ASTM), Philadelphia, PA, 90 pp.

11 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Appendix A Sample Descriptive Sensory Analysis Ballot used for Evaluation of Sliced Hams

PERSONS WITH KNOWN FOOD ALLERGIES MUST REQUEST SPECIFIC INFORMATION ABOUT FOOD INGREDIENTS AND PREPARATION. PLEASE ASK THE PANEL ATTENDANT FOR THIS INFORMATION BEFORE PARTICIPATING IN TASTING

SENSORY ANALYSIS LABORATORY Department of Food Science University of Wisconsin-Madison

Ham Date: February 7, 2005

Directions: 1. Taste Sample #824 2. Check the box below which best expresses your opinion of the sample 3. Repeat with Sample #563

Overall Preference 824 563

like very much

like moderately

like slightly

neither like nor dislike

dislike slightly

dislike moderately

dislike very much

1. Which of the two samples you tasted do you prefer? (check one box only) 824 563 2. Why do you prefer the sample checked in Question 1 better than the other sample?

3. Other comments:

STATEMENT OF CONDITIONS FOR VOLUNTARY PARTICIPATION IN TASTE TESTING OF FOODS IN THE SENSORY EVALUATION LABORATORY, DEPARTMENT OF FOOD SCIENCE, UNIVERSITY OF WISCONSIN-MADISON. The sensory evaluation laboratory is engaged by various food research groups to evaluate the flavor, texture, and color of foods. In most instances the food samples are similar to commercially available products. In other instances food samples are the result of new product development or product modification efforts. However foods served to taste panelists are considered safe, wholesome, and prepared under good manufacturing practices. Participation in taste panels is voluntary, and panelists may withdraw participation at any time. Information concerning product composition and manufacturing history will be provided upon specific request.

12 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair I (Control vs. LD): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 1. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Numerical Preference Rating Score Control LD (------Number of Responses------) Like very much 7 72 52 Like moderately 6 85 77 Like slightly 5 28 47 Neither like nor dislike 4 14 12 Dislike slightly 3 13 18 Dislike moderately 2 2 3 Dislike very much 1 1 6 Total number of responses N = 215 Mean Score 5.83A 5.47B Statistical Analysis F-value S LSD (at 5% level) (0.18) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Table 2. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) Control 132 LD 76

13 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair I (Control vs. LD): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 3. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the Control sample the most.

The Control Ham sample: Number of responses1 Tastes less salty 52 Tastes better 24 Has better flavor 20 Has better texture 9 Tastes sweeter 8 Tastes saltier 4 Has more smoked flavor 4 Tastes richer 2 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample.

Table 4. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the LD sample the most. The LD Ham sample: Number of responses1 Has better flavor 24 Has better texture 11 Tastes better 10 Tastes saltier 10 Tastes less salty 8 Tastes sweeter 6 Has more smoked flavor 6 Tastes richer 5 Tastes less tart 1 Has less smoked flavor 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample.

14 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair I (Control vs. LD): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 5. Summary of optional voluntary comments. Control Ham sample: Number of responses1 Tastes too bland 2 Has bad texture 1 Tastes too sweet 1 Has off-flavor 1 LD Ham sample: Has bad aftertaste 5 Tastes bland 3 Tastes too salty 2 Has bad texture 1 Has piggy off-flavor 1 Tastes fatty 1 Tastes sour 1 Tastes metallic 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample.

15 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair II (Control vs. AM1 0.1% Sodium benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 6. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Preference Rating Numerical Control AM1 0.1% Sodium Score Benzoate (------Number of Responses------) Like very much 7 76 71 Like moderately 6 79 88 Like slightly 5 34 30 Neither like nor dislike 4 5 5 Dislike slightly 3 7 6 Dislike moderately 2 1 1 Dislike very much 1 0 1 Total number of responses N = 202 Mean Score 6.03A 6.02A Statistical Analysis F-value NS LSD (at 5% level) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level. Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 7. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) Control 97 AM1 0.1% Sodium Benzoate 100 Date of evaluation: February 7, 9, 14, and 20, 2006.

16 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair II (Control vs. AM1 0.1% Sodium benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 8. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the Control sample the most. The Control Ham sample: Number of responses1 Has better flavor 26 Tastes less salty 21 Has better texture 15 Tastes better 11 Tastes saltier 6 Tastes less fatty 3 Has more smoked flavor 3 Has better aftertaste 2 Tastes milder 1 Tastes sweeter 1 Has less smoked flavor 1 Has more moisture 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 9. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the AM1 0.1% Sodium Benzoate sample the most. The AM1 0.1% Sodium Benzoate Ham Number of responses1 sample: Tastes less salty 33 Has better flavor 29 Tastes better 15 Tastes sweeter 9 Tastes saltier 7 Has better texture 7 Has more smoked flavor 7 Tastes richer 4 Tastes fresher 3 Tastes milder 2 Tastes spicier 1 Tastes less sweet 1 Has more moisture 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

17 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair II (Control vs. AM1 0.1% Sodium Benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 10. Summary of optional voluntary comments. Control Ham sample: Number of responses1 Has bad aftertaste 3 Has artificial taste 1 Has a tough casing 1 Has too strong pork smell 1 Taste raw meat-like 1 Tastes not fresh 1 AM1 0.1% Sodium Benzoate Ham sample: Has off-flavor 1 Tastes too salty 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample.

Date of evaluation: February 7, 9, 14, and 20, 2006.

18 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair III (Control vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 11. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Preference Rating Numerical Control AM2 0.3% Potassium Score Sorbate (------Number of Responses------) Like very much 7 74 64 Like moderately 6 95 92 Like slightly 5 38 46 Neither like nor dislike 4 7 5 Dislike slightly 3 2 8 Dislike moderately 2 1 1 Dislike very much 1 0 1 Total number of responses N = 217 Mean Score 6.06A 5.88B Statistical Analysis F-value S LSD (at 5% level) (0.17) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 12. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) Control 119 AM2 0.3% Potassium Sorbate 94 Date of evaluation: February 7, 9, 14, and 20, 2006.

19 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair III (Control vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 13. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the Control sample the most. The Control Ham sample: Number of responses1 Has better flavor 25 Tastes saltier 18 Tastes better 17 Taste less salty 15 Has more flavor 13 Has better texture 7 Tastes sweeter 5 Has less smoked flavor 3 Has better aftertaste 2 Tastes juicier 2 Tastes less sweet 2 Has more moisture 1 Has better color 1 Has more smoked flavor 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 14. Summary of the responses to exit question #3 (why do you prefer the sample....?) from panelists who preferred the AM2 0.3% Potassium Sorbate sample the most. The AM2 0.3% Potassium Sorbate Ham Number of responses1 sample: Tastes less salty 19 Tastes sweeter 14 Has more flavor 12 Has better flavor 11 Has better texture 10 Tastes better 9 Tastes saltier 8 Has more smoked flavor 8 Has less aftertaste 4 Tastes juicier 2 Tastes less buttery 1 Tastes milder 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

20 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair III (Control vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 15. Summary of optional voluntary comments. Control Ham sample: Number of responses1 Tastes too salty 4 Has rubbery texture 2 Tastes too sweet 1 Has bad aftertaste 1 Has too much smoked flavor 1

AM2 0.3% Potassium Sorbate Ham sample: Has off-flavor 7 Tastes too sweet 4 Tastes too salty 3 Has slimy texture 2 Has slightly more smoked flavor 1 Has tender 1 Tastes bland 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

21 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair IV (Control vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 16. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Preference Rating Numerical Control AM3 0.3% Sodium Score Propionate (------Number of Responses------) Like very much 7 60 56 Like moderately 6 65 75 Like slightly 5 26 22 Neither like nor dislike 4 6 5 Dislike slightly 3 4 3 Dislike moderately 2 3 3 Dislike very much 1 0 0 Total number of responses N = 164 Mean Score 5.99A 6.02A Statistical Analysis F-value NS LSD (at 5% level) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 17. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) Control 78 AM3 0.3% Sodium Propionate 81 Date of evaluation: February 7, 9, 14, and 20, 2006.

22 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair IV (Control vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 18. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the Control sample the most. The Control Ham sample: Number of responses1 Tastes less salty 23 Tastes better 15 Has better flavor 15 Has more smoked flavor 6 Tastes saltier 5 Tastes sweeter 4 Has more moisture 4 Tastes less fatty 3 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

Table 19. Summary of the responses to exit question #2 (why do you prefer the sample from panelists who preferred the AM3 0.3% Sodium Propionate sample the most. The AM3 0.3% Sodium Propionate Ham Number of responses1 sample: Has better flavor 23 Tastes less salty 18 Tastes better 12 Tastes saltier 9 Tastes sweeter 8 Has more smoked flavor 6 Has better texture 6 Tastes milder 3 Has better aftertaste 1 Has better appearance 1 Has less smoked flavor 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

23 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair IV (Control vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 7, 9, 14, and 20, 2006.

Table 20. Summary of optional voluntary comments. Control Ham sample: Number of responses1 Tastes greasy 4 Has bad aftertaste 3 Tastes bland 2 Tastes watery 1 Tastes bitter 1 Has dry mouthfeel 1 AM3 0.3% Sodium Propionate Ham sample: Tastes bitter 3 Has dry mouthfeel 1 Has too strong aftertaste 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 7, 9, 14, and 20, 2006.

24 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair V (LD vs. AM1 0.1% Sodium Benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 21. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Preference Rating Numerical LD AM1 0.1% Sodium Score Benzoate (------Number of Responses------) Like very much 7 47 77 Like moderately 6 80 82 Like slightly 5 41 26 Neither like nor dislike 4 11 6 Dislike slightly 3 9 3 Dislike moderately 2 3 0 Dislike very much 1 3 0 Total number of responses N = 194 Mean Score 5.64A 6.15B Statistical Analysis F-value S LSD (at 5% level) (0.19) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Date of evaluation: February 23, 27, and 28, 2006.

Table 22. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) LD 67 AM1 0.1% Sodium Benzoate 124 Date of evaluation: February 23, 27, and 28, 2006.

25 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair V (LD vs. AM1 0.1% Sodium Benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 23. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the LD sample the most. The LD Ham sample: Number of responses1 Has better flavor 20 Tastes better 10 Tastes saltier 10 Tastes sweeter 8 Has better texture 6 Tastes milder 5 Tastes less salty 4 Tastes juicier 2 Tastes less sweet 1 Has better aroma 1 Has more smoked flavor 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

Table 24. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the AM1 0.1% Sodium Benzoate sample the most. The AM1 0.1% Sodium Benzoate Ham Number of responses1 sample: Tastes less salty 49 Has better flavor 34 Tastes better 16 Has better texture 13 Tastes sweeter 6 Tastes saltier 3 Has more smoked flavor 3 Has better aftertaste 3 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

26 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair V (LD vs. AM1 0.1% Sodium Benzoate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 25. Summary of optional voluntary comments. LD Ham sample: Number of responses1 Tastes too salty 11 Tastes slightly bitter 3 Has unclean off-flavor 3 Tastes bland 2 Has discolored edge 1 Has soft/mushy texture 1 AM1 0.1% Sodium Benzoate Ham sample: Tastes too salty 2 Has too chewy texture 1 Tastes greasy 1 Tastes slightly bland 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

27 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VI (LD vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 26. Response frequency and mean scores for the consumer preference evaluation of ham. Assigned Ham Preference Rating Numerical LD AM2 0.3% Potassium Score Sorbate (------Number of Responses------) Like very much 7 70 62 Like moderately 6 67 75 Like slightly 5 39 32 Neither like nor dislike 4 5 10 Dislike slightly 3 9 13 Dislike moderately 2 1 0 Dislike very much 1 2 1 Total number of responses N = 193 Mean Score 5.90A 5.82A Statistical Analysis F-value NS LSD (at 5% level) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Date of evaluation: February 23, 27, and 28, 2006.

Table 27. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) LD 107 AM2 0.3% Potassium Sorbate 83 Date of evaluation: February 23, 27, and 28, 2006.

28 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VI (LD vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 28. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the LD sample the most. The LD Ham sample: Number of responses1 Has better flavor 36 Tastes saltier 18 Tastes better 14 Has better texture 13 Tastes less salty 7 Has more smoked flavor 4 Tastes sweeter 4 Has better aftertaste 2 Tastes fresher 1 Tastes less sweet 1 Tastes spicier 1 Has better appearance 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

Table 29. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the AM2 0.3% Potassium Sorbate sample the most. The AM2 0.3% Potassium Sorbate Ham Number of responses1 sample: Tastes less salty 20 Has better flavor 18 Tastes better 12 Tastes sweeter 10 Has more smoked flavor 6 Has better texture 6 Has better color 2 Has better aftertaste 2 Tastes less sweet 1 Tastes juicier 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

29 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VI (LD vs. AM2 0.3% Potassium Sorbate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 30. Summary of optional voluntary comments. LD Ham sample: Number of responses1 Has off-flavor 5 Has bad aftertaste 3 Has too salty 2 Tastes too fatty 1 Tastes bland 1 AM2 0.3% Potassium Sorbate Ham sample: Has bad aftertaste 5 Tastes too salty 4 Tastes too fatty 3 Tastes too sweet 1 Tastes soapy 1 Has tough texture 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

30 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VII (LD vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 31. Response frequency and mean scores for the consumer preference evaluation of ham.

Assigned Ham Numerical Preference Rating Score LD AM3 0.3% Sodium Propionate (------Number of Responses------) Like very much 7 56 87 Like moderately 6 83 65 Like slightly 5 31 25 Neither like nor dislike 4 7 8 Dislike slightly 3 11 2 Dislike moderately 2 0 2 Dislike very much 1 1 0 Total number of responses N = 189 Mean Score 5.86A 6.17B Statistical Analysis F-value S LSD (at 5% level) (0.19) S= significant at the 5% level; NS = not significant at the 5% level. A,B Mean scores in the same row with the same superscript are not significantly different at the 5% level.

Date of evaluation: February 23, 27, and 28, 2006.

Table 32. Summary of panelist responses to exit questions #1. Exit questions Number of response #1: Which of the two samples you tasted do you prefer? (check one box only) LD 73 AM3 0.3% Sodium Propionate 115 Date of evaluation: February 23, 27, and 28, 2006.

31 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VII (LD vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 33. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the LD sample the most. The LD Ham sample: Number of responses1 Has better texture 13 Has better flavor 12 Tastes better 11 Tastes less salty 9 Tastes saltier 7 Has more smoked flavor 5 Has more moisture 3 Tastes less sweet 2 Tastes juicier 2 Tastes less fatty 1 Tastes less piggy 1 Tastes sweeter 1 Tastes less processed 1 Has better color 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

Table 34. Summary of the responses to exit question #2 (why do you prefer the sample....?) from panelists who preferred the AM3 0.3% Sodium Propionate sample the most. The AM3 0.3% Sodium Propionate Ham Number of responses1 sample: Tastes less salty 36 Has better flavor 28 Tastes better 15 Has better texture 14 Tastes saltier 7 Has more smoked flavor 6 Tastes sweeter 3 Tastes juicier 2 Tastes less fatty 2 Tastes tangier 2 Has less smoked flavor 1 Has better color 1 Has less chewy 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

32 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix A Consumer Taste Preference for Ham

Sample Pair VII (LD vs. AM3 0.3% Sodium Propionate): Consumer Preference Sensory Evaluation for Ham Evaluated on February 23, 27, and 28, 2006.

Table 35. Summary of optional voluntary comments. LD Ham sample: Number of responses1 Tastes too salty 5 Has a little slimy texture 3 Has off-flavor 2 Tastes slightly bitter 2 Has too strong aftertaste 1 Has metallic aftertaste 1 AM3 0.3% Sodium Propionate Ham sample: Tastes sour 1 Tastes too fatty 1 Has off-flavor 1 1 Data reflect total number of comments offered for each descriptor category; some panelists commented about more than one descriptor category for a given sample. Date of evaluation: February 23, 27, and 28, 2006.

33 Glass and Claus, AMIF Final Report, June 2006 Controlling L. monocytogenes in RTE Meats Using Benzoate, Propionate and Sorbate University of Wisconsin-Madison Appendix B Literature Review by M. Ellin Doyle, Ph.D.

APPENDIX B

Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products Using Food-Approved Antimicrobials Benzoate, Propionate, and Sorbate

M. Ellin Doyle, Ph.D. Food Research Institute University of Wisconsin–Madison Madison, WI 53706 [email protected]

Objectives for Literature Search: • Conduct a thorough search on the effects that sorbate, benzoate, and propionate in ready-to-eat meats, in the presence and absence of nitrite, might have on human health • Summarize recent (last 3 years) articles on effects of sorbate, benzoate, and propionate on Listeria monocytogenes and on spoilage organisms

Effects of Organic Acids on L. monocytogenes in RTE Meats: Recent Scientific Literature Numerous recent research papers (2003–2006) have presented more data on the inhibitory effects of organic acids on growth and survival of L. monocytogenes in RTE meats. Important factors to consider in challenge testing were also investigated and discussed (87). Frankfurters Sodium or potassium lactate with or without sodium diacetate were used in frankfurter formulations to control growth of L. monocytogenes in frankfurters stored at 4–7ºC. When used alone, lactate concentrations were 3–3.3% (14, 70). When used in combination with diacetate, lactate levels were in the range of 1.4–1.8% and diacetate levels ranged from 0.05–0.25% (6, 25, 48). In addition, organic acids have been used as dipping solutions to suppress growth of L. monocytogenes that may have contaminated the surface of frankfurters. The following dipping solutions were reported to aid in pathogen control on frankfurters that contained sodium or potassium lactate (with or without sodium diacetate) in their formulation: i) lactic acid (2.5%) and acetic acid (2.5%) (6, 25); ii) potassium benzoate (5%) (25); iii) mixture containing 2% acetic, 1% lactic, 0.1% benzoic, and 0.1% propionic acids (60); iv) mixture containing 3% sodium diacetate and 6% sodium lactate (86); v) 3% or 6% sodium diacetate, potassium benzoate, sodium lactate, alone or in combination (46, 47); vi) acidic calcium sulfate with propionic and lactic acids (14). If organic acids were not included in the formulation, the dipping solutions delayed growth of L. monocytogenes somewhat but did not effectively retard growth during the expected shelf life of the product. Dipping solutions containing 0.125–0.5% sodium diacetate reduced the irradiation dose required to suppress listerial growth on frankfurters (80). Lactate and diacetate, incorporated into casings or coatings for frankfurters, also help to retard growth of L. monocytogenes (48, 50). However, sorbate in a coating did not significantly inhibit listerial growth (51).

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison 2 FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products

Bologna Organic acids have also been used to control listerial growth on bologna. Incorporation of 1.8% sodium lactate and 0.25% sodium diacetate into pork bologna was reported to be the best combination for inhibiting bacterial growth during 20 days at 4 or 10ºC (7). Dipping solutions containing 2.5% acetic acid, 2.5% lactic acid or 5% potassium benzoate all reduced populations of listeria on bologna slices during storage at 10ºC for 48 days. These slices did have lower sensory scores, however (24). A plastic (polyvinylidene chloride) film containing 1.5% or 3% sorbic acid, placed between slices of inoculated beef bologna, prevented growth of L. monocytogenes for 28 days at 4ºC (44). Sodium lactate (1–2%) and sodium diacetate (0.07–0.15%) decreased the dose of irradiation required to effectively suppress listerial growth on beef bologna during refrigerated storage. Use of these compounds allowed a lower dose of irradiation which decreased oxidation of lipids and minimized adverse effects on flavor (81). Ham Dipping solutions containing 2.5% acetic acid, 2.5% lactic acid or 5% potassium benzoate all reduced populations of listeria on ham slices during storage at 10ºC for 48 days. These slices did have lower sensory scores, however (24). Lactates were found to act synergistically with high hydrostatic pressure and low storage temperature to inhibit growth of L. monocytogenes (3). Sodium lactate (2%) and sodium diacetate (0.1%) in combination with 1 kGy of irradiation effectively suppressed listerial growth on turkey ham for 6 weeks at 4ºC. Use of these compounds allowed a lower dose of irradiation which minimized adverse effects of irradiation on flavor of the ham. Although potassium benzoate also had inhibitory effects, some benzene was detected in the irradiated ham containing benzoates, indicating that this is not an appropriate preservative to use in combination with irradiation (94). Sausage Addition of 0.125–0.5% sodium diacetate to fine-emulsion sausages reduced the irradiation dose required to suppress listerial growth (80). Sodium lactate (3.3%) was found to have anitlisterial effects similar to 0.05% or 0.1% potassium sorbate and sodium benzoate in sausage and delayed lag phase growth of L. monocytogenes by at least two weeks at 4ºC (13). Poultry Some organic acids decrease listerial populations on raw poultry. L monocytogenes levels on chicken breast were reduced by 3.88 logs following a 15 min dip in a solution of 2.5% sodium lactate (26). A solution of 1.54% potassium lactate and 0.11% sodium diacetate only somewhat inhibited listerial growth on cook in bag turkey breasts (49). Addition of 4.8% sodium lactate to ground chicken actually increased the heat resistance of L. monocytogenes added to the meat (61). Predictive Models Some recent research has also described broth-based predictive models with data on the listeriocidal or listeriostatic effects of lactate and diacetate (36, 93). Evaluations of predictive models for effective control of L. monocytogenes in meat emphasize that such models need to be validated in meat or predictive errors will result (45, 53). Effects of sodium lactate (0–4.8%) and sodium diacetate (0–0.25%) on heat resistance of L. monocytogenes in ground beef were tested at 60–73.9ºC (37). Sodium lactate alone was found to increase heat resistance of the bacteria (similar to the effect in ground chicken; 61) while combinations of the two organic acids reduced heat resistance. Data were used to construct a predictive model with D values indicating the safety of different combinations of heat and organic acids for controlling L. monocytogenes. Another predictive model was constructed with data on the effects of sodium chloride (0.8–3.6%), potassium lactate (0.2–9.25%), and sodium diacetate (0–0.2%) on growth of L. monocytogenes at 4ºC in cured and uncured RTE meats (42).

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products 3

Effects of Organic Acids on Spoilage Organisms Spoilage organisms may affect growth of L. monocytogenes on RTE meats. Recently published research indicated that the population of listeriae inoculated on to frankfurters was highest at the lowest concentration of spoilage bacteria. However, the growth rate of listeriae at 10ºC was similar at all concentrations of spoilage bacteria (71). Experiments in laboratory media have demonstrated that sodium lactate inhibits the growth of spoilage bacteria (33, 34, 66). This has also been demonstrated in several meat models. Sodium lactate (3%) and a combination of sodium lactate (2, 3, or 4%), 0.5% sodium acetate, and 0.15% potassium sorbate were most effective in retarding growth of a spoilage bacterium (Lactobacillus curvatus) on frankfurters and a "pariza" type meat product (17). Sodium lactate also effectively inhibited growth of Lactobacillus sake and Lactobacillus curvatus in frankfurters and ham at 4ºC (82, 83) and of Brochothrix thermosphacta, a spoilage bacterium, in poultry sausage (43). Growth of aerobic psychrotrophic bacteria and lactic acid bacteria on poultry sausage were inhibited by 1–2% sodium lactate (11). Shelf life of cooked meat products has been extended by sodium lactate and this has been modeled (15). Combinations of 1.8–3.0% sodium lactate and 0.2–0.30% sodium diacetate were reported to be effective at retarding growth of spoilage bacteria on pork bologna (7), beef bologna (54, 69), and cook-in-bag turkey (55). Sodium lactate has also been reported to inhibit germination and growth of Clostridium perfringens (38, 39) and some, but not all, psychrotrophic clostridia (41). When used as dipping solutions, a solution of 2.5% sodium lactate and 0.2% sodium diacetate inhibited growth of spoilage organisms on beef (76) while with bologna and ham, sequential dipping in 2.5% acetic acid, 2.5% lactic acid, and 5% sodium benzoate was reported to inhibit growth of spoilage bacteria (24). In laboratory media, acetic and propionic acids (0.1–1.0%) inhibited growth of six common meat spoilage bacteria. Because of their low solubility in broth sorbic and benzoic acids could be used only at low concentrations (<0.15%) and were not effective against these spoilage organisms (66). Sorbic acid containing plastic films inhibited growth of common spoilage organisms on beef bologna (44). Sodium diacetate appears to be ineffective in controlling spoilage bacteria on cook-in-bag turkey (55), frankfurters (82) and ham (83). However, sodium diacetate (0.3%) did inhibit some Gram negative spoilage bacteria (77).

Toxicology Studies with Organic Acids Since organic acids have been widely used as preservatives in some foods for many years, numerous tests have been conducted to determine their possible toxic effects in rodents and cell cultures. In addition, there have been clinical studies to assess possible allergenicity and there are some reports of adverse reactions to these compounds in humans. Overall, data indicate that these compounds are of low toxicity with little or no genotoxic or carcinogenic potential. Reports of significant adverse reactions are rare. However, four potentially significant issues should be mentioned: i) Benzoate, more than any of the other organic acids, appears to provoke hypersensitivity reactions in sensitive individuals. These are not common but one should be aware of this possibility. ii) Several reports indicate that under acidic conditions and during irradiation small amounts of the known carcinogen benzene can form from benzoates. Therefore, benzoates should probably not be used in some acidic foods or in foods that will also be irradiated. iii) Under some conditions, sorbates have been reported to degrade during long storage times to form genotoxic compounds and sorbates were reported to form mutagens with nitrites. This is considered unlikely using current procedures of meat processing but should be reassessed if processes change. iv) At very high dietary levels (4% of diet) over extended feeding periods, propionates have caused forestomach cancers in rodents. Again this is very unlikely to occur in humans,

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison 4 FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products

but one should be aware that this has been reported in the literature. Currently available information is summarized below. Sodium Diacetate Sodium diacetate is an approved GRAS substance for use as an antimicrobial in several foods. No adverse reactions have been reported for humans or animals and no recent acceptable daily intake (ADI) has been established. In 1973, FAO stated that up to 15 mg/kg body weight/ day was acceptable (16). Sodium diacetate was approved by FSIS in 2000 for use as an antimicrobial in meat and poultry products up to a concentration of 0.25% by weight of total formulation (Federal Register 65: 3121–3123 and 65:17128– 17129) Higher concentrations (up to 0.4%) are used in some bread and bakery products (16). Lactate Lactic acid is an approved GRAS substance for use in various foods. No acceptable daily intake has been established. Oral LD50 doses of lactic acid in rodents exceed 1.8 g/kg body weight (16). Sodium and potassium lactates were approved by FSIS in 2000 for use as antimicrobials in meat and poultry products, singly or in combination, up to a concentration of 4.8% by weight of total formulation (Federal Register 65: 3121–3123 and 65:17128–17129) A 2-yr. study demonstrated no toxic or carcinogenic effects in rats given water containing 2.5 or 5% calcium lactate in drinking water (52). Propionic Acid Propionic acid and its salts are approved GRAS substances for use in various foods, including cheese and bakery products. Sodium propionate is made from the reaction of sodium hydroxide with propionic acid. Propionic acid can be prepared by a variety of methods, but occurs naturally as the result of metabolic processes and can be obtained from fermentation by Propionibacterium. Use of sodium propionate has no limitations in a variety of food products (including cheeses, soft candies, baked goods, jams, jellies, nonalcoholic beverages) other than current good manufacturing practices (21CFR 184.1784). Acceptable daily intakes have not been established. Oral LD50 doses of propionates in rodents exceed 3 g/kg body weight (16). Genotoxicity. Tests with propionic acid in the following genotoxicity assays were negative: SOS chromotest, Salmonella/microsome mutagenicity assay, sister chromatid exchange in vitro, and micronucleus test in vivo. Some positive results were obtained with the E. coli DNA repair assay. These data reinforce other evidence that propionic acid is not mutagenic (8). Carcinogenicity. Rats fed diets containing 4% propionic acid develop hyperplastic lesions and tumors in the forestomach; tumors were not formed when propionic acid constituted 0.4% of the diet. Damage and cellular proliferation were detectable in the forestomach of rats, mice, and hamsters after 7 days of consuming diets containing 4% (but not 0.4%) propionic acid (28, 29, 68). In some other experiments, rats developed hyperplasia and severe inflammatory lesions only when fed propionic acid in a powdered diet but not when fed pellets containing propionic acid. Propionic acid accumulated in inflammatory lesions (10). Humans do not have forestomachs but they do have similar epithelial tissue in the esophagus and pharynx. Because food contact time in the forestomach is much longer than that in the pharynx and esophagus, propionic is not considered a carcinogenic risk for humans (28). Benzoic Acid and Sodium Benzoate Benzoic acid occurs naturally in some fruits and fermented products and has been used as a preservative since the early 20th century. Sodium benzoate, approximately 200 times more water soluble than benzoic acid, is produced by neutralization of benzoic acid with sodium bicarbonate, sodium carbonate, or sodium hydroxide. Both benzoic acid and sodium benzoate are GRAS in the U.S. and are permitted in certain foods as antimicrobial or flavoring agents, with current maximum usage level of 0.1% (21CFR 184.1733). CODEX specifies higher permitted levels in some foods such as liquid eggs (0.5%) and semi- preserved fish (0.2%) (http://www.codexalimentarius.net).

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products 5

Benzoates are readily absorbed from the intestine but do not accumulate in the body since they are rapidly metabolized in the liver and excreted in the urine as hippuric acid. An oral dose of 250 mg benzoic acid was quantitatively excreted within four hours by a healthy male. An acceptable daily intake (ADI) of up to 5 mg/kg body weight has been established. Benzoates and benzoic acid scavenge hydroxyl free radicals (12, 63). Under some conditions benzoic acid or benzoates may form small amounts of benzene, a volatile compound with known toxic and carcinogenic effects. This has been reported in some fruit drinks containing sodium benzoate and ascorbic acid (12) and during irradiation of a turkey breast roll (95) and ham (94) containing potassium benzoate. Hypersensitivity reactions. Reported human reactions to benzoic acid and benzoates can be categorized as hypersensitivity reactions. These appear to be non-immunological reactions and may involve production of histamine and prostaglandins. In most cases, reactions appear to be mild-moderate but a few cases of anaphylaxis have been reported. Dermatitis/urticaria. A number of challenge tests indicated that oral doses of 20–250 mg sodium benzoate can cause urticaria in some individuals including children (23, 35, 56, 65, 84). Sodium benzoate caused a significant increase in production of leukocyte inhibitory factor (a compound involved in cell- mediated immune responses) by mononuclear cells from persons who reported developing urticaria in response to benzoates (89). Sodium benzoate was reported to stimulate release of histamine and prostaglandin from human gastric mucosa in some patients (12). Oral reactions to toothpastes containing sodium benzoate have also been reported (1, 59). Another study which involved careful testing of 47 patients who reported allergic reactions after consuming foods containing sodium benzoate, revealed that only one reacted to a challenge with 75 mg sodium benzoate but did not react to a placebo (64). Asthma. Some reports indicate that benzoate induces asthmatic symptoms in some susceptible people (23, 32, 91). Four of 14 patients tested, reacted with asthmatic symptoms to sodium benzoate (20– 120 mg) added to an orange drink (21). Two asthmatic children, whose symptoms worsened when treated with some antiasthmatic drugs and antibiotics, were found to be sensitive to the benzoates contained in these drugs. Treatment with benzoate-free versions of these drugs did not elicit respiratory symptoms (4). Rhinitis. Sodium benzoate (50 mg oral dose) was implicated in a rigorously investigated case of rhinitis (2). A double-blind placebo-controlled trial investigating the effects of food additives demonstrated that sodium benzoate (100 and 200 mg oral doses) induced or aggravated symptoms of rhinitis in 8.8% of 226 subjects tested (67). Still another challenge study demonstrated that some people developed rhinitis after ingesting 250 mg benzoate (91). Anaphylaxis. A few cases of severe anaphylactic reactions have been attributed to benzoate. One involved a patient undergoing general anesthesia. A challenge test later demonstrated that an oral dose of 100 mg sodium benzoate decreased peak expiratory flow rate (57). Another case patient apparently reacted to sodium benzoate used as a preservative in mustard and cheese with a severe systemic reaction. A later oral challenge with 20 mg sodium benzoate induced urticaria, but after the patient was treated for a sinus infection, she tolerated up to 160 mg benzoate before developing symptoms (56). Genotoxicity. Benzoic acid and sodium benzoate tested negative in bacterial and most in vivo mammalian genotoxicity assays. They have also tested negative in most, but not all, in vitro mammalian assays (63). DNA adducts with benzoate were detected in liver and kidney after oral doses of 100, 500, and 1000 µg/kg body weight. Adduct levels declined by 75% within a day as the benzoate was metabolized to hippurate and excreted in the urine (92). Carcinogenicity. No evidence for the carcinogenicity of sodium benzoate was detected in studies with rats (2% in pellets) (79) or mice (2% in drinking water) (85). Toxicity. Some early studies (1907, 1908, 1924) reported adverse effects in some (but not all) human volunteers consuming very high levels of sodium benzoate (33 g/kg dose or a liter of juice containing 0.2% or 0.3%) or benzoic acid (2.5 g/day for 5 days) (63). Consumption of a diet containing sodium benzoate (2.4% for rats, 3% for mice) for ten days caused some significant changes in hepatocytes and liver and kidney weights in male rodents. Some effects were

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison 6 FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products

observed in females but these were less pronounced and often not significant (22). Benzoic acid appeared to depress weight gain in mice (3 months at 80 mg/kg/day or 17 months at 40 mg/kg/day) and in rats (18 months at 40 mg/kg/day) (78). In subchronic toxicity tests with rats, decreased feed intake and growth were only observed in animals fed diets containing >1% benzoic acid. Other studies with mice showed that they did respond to benzoic acid concentrations <1% (63). Neurotoxicity tests with rodents were negative. In teratogenicity tests, benzoic acid was negative in rats (up to 500 mg/kg/day) but positive in hamsters (600 mg/kg). Sodium benzoate caused no teratogenic effects in rats and mice at 175 mg/kg/day (63). Oral LD50 values for sodium benzoate and benzoic acid in rodents, rabbits, cats, and dogs were all >1.5 g/kg body weight (63). Sorbic Acid, Sodium Sorbate, Potassium Sorbate Sorbic acid and potassium sorbate, are generally recognized as safe in the U.S. when used in accordance to good manufacturing practices (21CFR 182.3640). Allowable usage levels this antimycotic agent vary depending on the target food and range between 0.1% for fruit preserves (21CFR Part 150) and 0.3% for certain cheeses (21CFR Part 133). Allergy. Some challenge tests with humans indicate that sorbic acid or sorbates can cause non- immunological contact urticaria in susceptible people (27, 88). A few other reports indicate that a small number of people may be sensitive to oral doses of sorbic acid (74). Genotoxicity. Results from a number of in vitro and in vivo tests indicate that sorbic acid (alone) and potassium sorbate do not produce genotoxic effects (5, 40, 62, 73). Some other genotoxicity tests using Chinese hamster cells in vitro indicated that sorbic acid and sorbates exert weak genotoxic effects (31). Freshly prepared sodium sorbate also tested negative in a number of experiments (90). However, solutions of sodium sorbate stored for several weeks did exhibit weak genotoxic effects (62, 73, 90). This may be due to the formation of the degradation product, 4,5-oxohexenoate, which is a mutagen (31, 40). Some other experiments have demonstrated genotoxic effects (increased levels of sister chromatid exchanges and micronuclei in mice) of sorbic acid at high concentrations (75–150 mg/kg body weight) in mice (58). Carcinogenicity. Neither sorbic acid nor sorbates appear to be carcinogenic in rodent studies with animals fed diets containing as much as 10% sorbic acid in the diet (88). In the presence of sodium nitrite, potassium sorbate was reported to have cytotoxic but not mutagenic effects (9). However, other experiments demonstrated that nitrite and potassium sorbate formed direct acting mutagens (30). Japanese researchers reported that interactions between sorbate and nitrite may induce production of mutagens when concentrations of 2200 ppm sorbic acid and 1300 ppm nitrites are present in high acid conditions (72). However, mutagenicity was blocked when pH values were higher than 3.5, or in the presence of ascorbate. Cured meat products produced in the United States typically contain erythorbate or ascorbate to reduce nitrosoamine formation, and this would prevent mutagen formation at pH values that mimic gastric conditions should otherwise optimal conditions exist (30, 75). Sorbic acid and nitrite exerted additive or synergistic effects in several in vivo genotoxicity assays in mice (5, 58). Sorbic acid can react with amines (that may be present in foods) but the products do not appear to be mutagenic or genotoxic (18–20). Toxicity. Sorbic acid and sorbates are reported to exert a very low level of mammalian toxicity. Median lethal doses of sorbic acid and its potassium and sodium salts in rodents range from 4 to 10 g/kg body weight. This corresponds to about 500 g (over 1 pound) for an adult human. Short term toxicity assays with rodents and dogs demonstrated that dietary levels of 2–10% sorbic acid/sorbate had no adverse effects. No adverse reproductive or developmental effects have been observed in multigenerational rodent studies (88).

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products 7

Reference List 1. Aguirre A, Izu R, Gardeazabal J, and Diaz-Perez JL. 1993. Edematous allergic contact cheilitis from a toothpaste. Contact Dermatitis 28:42. 2. Asero R. 2001. Perennial rhinitis induced by benzoate intolerance. J Allerg Clin Immunol 107:197. 3. Aymerich T, Jofre A, Garriga M, and Hugas M. 2005. Inhibition of Listeria monocytogenes and Salmonella by natural antimicrobials and high hydrostatic pressure in sliced cooked ham. J Food Prot 68:173–177. 4. Balatsinou L, Di Gioacchino G, Sabatino G, Cavallucci E, Caruso R, Gabriele E, Ramondo S, Di Giampaolo L, Verna N, and Di Gioacchino M. 2004. Asthma worsened by benzoate contained in some antiasthmatic drugs. Int J Immunopathol Pharmacol 17:225–256. 5. Banerjee TS and Giri AK. 1986. Effects of sorbic acid and sorbic acid-nitrite in vivo on bone marrow chromosomes of mice. Toxicol Lett 31:101–106. 6. Barmpalia IM, Geornaras I, Belk KE, Scanga JA, Kendall PA, Smith GC, and Sofos JN. 2004. Control of Listeria monocytogenes on frankfurters with antimicrobials in the formulation and by dipping in organic acid solutions. J Food Prot 67:2456–2464. 7. Barmpalia IM, Koutsoumanis KP, Geornaras I, Belk KE, Scanga JA, Kendall PA, Smith GC, and Sofos JN. 2005. Effect of antimicrobials as ingredients of pork bologna for Listeria monocytogenes control during storage at 4 or 10 degrees C. Food Microbiol 22:205–211. 8. Basler A, von der Hude W, and Scheutwinkel M. 1987. Screening of the food additive propionic acid for genotoxic properties. Food Chem Toxicol 25:287–290. 9. Budayová E. 1985. Effects of sodium nitrite and potassium sorbate on in vitro cultured mammalian cells. Neoplasma 32:341–350. 10. Bueld JE and Netter KJ. 1993. Factors affecting the distribution of ingested propionic acid in the rat forestomach. Food Chem Toxicol 31:169–176. 11. Cegielska-Radziejewska R and Pikul J. 2004. Sodium lactate addition on the quality and shelf life of refrigerated sliced poultry sausage packaged in air or nitrogen atmosphere. J Food Prot 67:601–606. 12. Chipley JR. 2005. Sodium benzoate and benzoic acid, p. 11–48. In Davidson PM, Sofos JN, and Branen AL (eds.), Antimicrobials in Food. Taylor & Francis, New York. 13. Choi SH and Chin KB. 2003. Evaluation of sodium lactate as a replacement for conventional chemical preservatives in comminuted sausages inoculated with Listeria monocytogenes. Meat Sci 65:531–537. 14. de Gonzalez MTN, Keeton JT, Acuff GR, Ringer LJ, and Lucia LM. 2004. Effectiveness of acidic calcium sulfate with propionic and lactic acid and lactates as postprocessing dipping solutions to control Listeria monocytogenes on frankfurters with or without potassium lactate and stored vacuum packaged at 4.5 degrees C. J Food Prot 67:915–921. 15. Devlieghere F, Geeraerd AH, Versyck KJ, Bernaert H, Van Impe JF, and Debevere J. 2000. Shelf life of modified atmosphere packed cooked meat products: addition of Na-lactate as a fourth shelf life determinative factor in a model and product validation. Int J Food Microbiol 58:93–106. 16. Doores S. 2005. Organic acids, p. 91–142. In Davidson PM, Sofos JN, and Branen AL (eds.), Antimicrobials in Food. Taylor & Francis, New York. 17. Drosinos EH, Mataragas M, Kampani A, Kritikos D, and Metaxopoulos I. 2006. Inhibitory effect of organic acid salts on spoilage flora in culture medium and cured cooked meat products under commercial manufacturing conditions. Meat Science 73:75–81. 18. Ferrand C, Marc F, Fritsch P, Cassand P, and Blanquat GD. 2000. Chemical and toxicological studies of products resulting from sorbic acid and methylamine interaction in food conditions. Amino Acids 18:251–263. 19. Ferrand C, Marc F, Fritsch P, Cassand P, and Blanquat GD. 2000. Genotoxicity study of reaction products of sorbic acid. J Agr Food Chem 48:3605–3610. 20. Ferrand C, Marc F, Fritsch P, Cassand P, and Blanquat GD. 2000. Mutagenicity and genotoxicity of sorbic acid-amine reaction products. Food Additiv Contamin 17:895–901. 21. Freedman BJ. 1977. Asthma induced by sulphur dioxide, benzoate and tartrazine contained in orange drinks. Clin Allerg 7:407–415. 22. Fujitani T. 1993. Short-term effect of sodium benzoate in F344 rats and B6C3F1 mice. Toxicol Lett 69:171–179. 23. Genton C, Frei PC, and Pécoud A. 1985. Value of oral provocation tests to aspirin and food additives in the routine investigation of asthma and chronic urticaria. J Allerg Clin Immunol 76:40–45.

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison 8 FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products

24. Geornaras I, Belk KE, Scanga JA, Kendall PA, Smith GC, and Sofos JN. 2005. Postprocessing antimicrobial treatments to control Listeria monocytogenes in commercial vacuum-packaged bologna and ham stored at 10°C. J Food Prot 68:991–998. 25. Geornaras I, Skandamis PN, Belk KE, Scanga JA, Kendall PA, Smith GC, and Sofos JN. 2006. Postprocess control of Listeria monocytogenes on commercial frankfurters formulated with and without antimicrobials and stored at 10°C. J Food Prot 69:53–61. 26. Goncalves AC, Almeida RCC, Alves MAO, and Almeida PF. 2005. Quantitative investigation on the effects of chemical treatments in reducing Listeria monocytogenes populations on chicken breast meat. Food Control 16:617–622. 27. Hannuksela M and Haahtela T. 1987. Hypersensitivity reactions to food additives. Allergy 42:561–575. 28. Harrison PT. 1992. Propionic acid and the phenomenon of rodent forestomach tumorigenesis: a review. Food Chem Toxicol 30:333–340. 29. Harrison PT, Grasso P, and Badescu V. 1991. Early changes in the forestomach of rats, mice and hamsters exposed to dietary propionic and butyric acid. Food Chem Toxicol 29:367–371. 30. Hartman PE. 1983. Review: putative mutagens and carcinogens in foods. II: sorbate and sorbate-nitrite interactions. Environl Mutagen 5:217–222. 31. Hasegawa MM, Nishi Y, Ohkawa Y, and Inui N. 1984. Effects of sorbic acid and its salts on chromosome aberrations, sister chromatid exchanges and gene mutations in cultured Chinese hamster cells. Food Chem Toxicol 22:501–507. 32. Hong SP, Park HS, Lee MK, and Hong CS. 1989. Oral provocation tests with aspirin and food additives in asthmatic patients. Yonsei Med J 30:339–345. 33. Houtsma PC, Dewit JC, and Rombouts FM. 1993. Minimum inhibitory concentration (MIC) of sodium lactate for pathogens and spoilage organisms occurring in meat products. Int J Food Microbiol 20:247–257. 34. Houtsma PC, Dewit JC, and Rombouts FM. 1996. Minimum inhibitory concentration (mic) of sodium lactate and sodium chloride for spoilage organisms and pathogens at different pH values and temperatures. J Food Prot 59:1300–1304. 35. Ibero M, Eseverri JL, Barroso C, and Botey J. 1982. Dyes, preservatives and salicylates in the induction of food intolerance and/or hypersensitivity in children. Allergologia et Immunopathologia 10:263–268. 36. Jin SS, Khen BK, Yoon KS, Woo GJ, Hwang IG, and Oh DH. 2005. Effects of temperature, pH, and potassium lactate on growth of Listeria monocytogenes in broth. Food Sci Biotechnol 14:847–853. 37. Juneja VK. 2003. Predictive model for the combined effect of temperature, sodium lactate, and sodium diacetate on the heat resistance of Listeria monocytogenes in beef. J Food Prot 66:804–811. 38. Juneja VK and Thippareddi H. 2004. Control of Clostridium perfringens in a model roast beef by salts of organic acids during chilling. J Food Safety 24:95–108. 39. Juneja VK and Thippareddi H. 2004. Inhibitory effects of organic acid salts on growth of Clostridium perfringens from spore inocula during chilling of marinated ground turkey breast. Int J Food Microbiol 93:155–163. 40. Jung R, Cojocel C, Müller W, Böttger D, and Lück E. 1992. Evaluation of the genotoxic potential of sorbic acid and potassium sorbate. Food Chem Toxicol 30:1–7. 41. Kalinowski RM and Tompkin RB. 1999. Psychrotrophic clostridia causing spoilage in cooked meat and poultry products. J Food Prot 62:766–772. 42. Legan JD, Seman DL, Milkowski AL, Hirschey JA, and Vandeven MH. 2004. Modeling the growth boundary of Listeria monocytogenes in ready-to-eat cooked meat products as a function of the product salt, moisture, potassium lactate, and sodium diacetate concentrations. J Food Prot 67:2195–2204. 43. Lemay MJ, Choquette J, Delaquis PJ, Gariepy C, Rodrigue N, and Saucier L. 2002. Antimicrobial effect of natural preservatives in a cooked and acidified chicken meat model. Int J Food Microbiol 78:217– 226. 44. Limjaroen P, Ryser E, Lockhart H, and Harte B. 2005. Inactivation of Listeria monocytogenes on beef bologna and cheddar cheese using polyvinylidene chloride films containing sorbic acid. J Food Sci 70:M267–M271. 45. Lu Z, Sebranek JG, Dickson JS, Mendonca AF, and Bailey TB. 2005. Application of predictive models to estimate Listeria monocytogenes growth on frankfurters treated with organic acid salts. J Food Prot 68:2326–2332. 46. Lu Z, Sebranek JG, Dickson JS, Mendonca AF, and Bailey TB. 2005. Effects of organic acid salt solutions on sensory and other quality characteristics of frankfurters. J Food Sci 70:S123–S127.

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products 9

47. Lu Z, Sebranek JG, Dickson JS, Mendonca AF, and Bailey TB. 2005. Inhibitory effects of organic acid salts for control of Listeria monocytogenes on frankfurters. J Food Prot 68:499–506. 48. Luchansky JB and Call JE. 2004. Evaluation of nisin-coated cellulose casings for the control of Listeria monocytogenes inoculated onto the surface of commercially prepared frankfurters. J Food Prot 67:1017– 1021. 49. Luchansky JB, Cocoma G, and Call JE. 2006. Hot water postprocess pasteurization of cook-in-bag turkey breast treated with and without potassium lactate and sodium diacetate and acidified sodium chlorite for control of Listeria monocytogenes. J Food Prot 69:39–46. 50. Lungu B and Johnson MG. 2005. Fate of Listeria monocytogenes inoculated onto the surface of model turkey frankfurter pieces treated with zein coatings containing nisin, sodium diacetate, and sodium lactate at 4 degrees C. J Food Prot 68:855–859. 51. Lungu B and Johnson MG. 2005. Potassium sorbate does not increase control of Listeria monocytogenes when added to zein coatings with nisin on the surface of full fat turkey frankfurter pieces in a model system at 4°C. J Food Sci 70:M95–M99. 52. Maekawa A, Matsushima Y, Onodera H, Shibutani M, Yoshida J, Kodama Y, Kurokawa Y, and Hayashi Y. 1991. Long-term toxicity/carcinogenicity study of calcium lactate in F344 rats. Food Chem Toxicol 29:589–594. 53. Martino KG, Marks BP, Campos DT, and Tamplin ML. 2005. Quantifying the robustness of a broth- based model for predicting Listeria monocytogenes growth in meat and poultry products. J Food Prot 68:2310–2316. 54. Mbandi E and Shelef LA. 2002. Enhanced antimicrobial effects of combination of lactate and diacetate on Listeria monocytogenes and Salmonella spp. in beef bologna. Int J Food Microbiol 76:191–198. 55. Meyer JD, Cerveny JG, and Luchansky JB. 2003. Inhibition of nonproteolytic, psychrotrophic clostridia and anaerobic sporeformers by sodium diacetate and sodium lactate in cook-in-bag turkey breast. J Food Prot 66:1474–1478. 56. Michils A, Vandermoten G, Duchateau J, and Yernault JC. 1991. Anaphylaxis with sodium benzoate. Lancet 337:1424–1425. 57. Moneret-Vautrin DA, Moeller R, Malingrey L, and Laxenaire MC. 1982. Anaphylactoid reaction to general anaesthesia: a case of intolerance to sodium benzoate. Anaesthes Intens Care 10:156–157. 58. Mukherjee A, Giri AK, Talukder G, and Sharma A. 1988. Sister chromatid exchanges and micronuclei formations induced by sorbic acid and sorbic acid-nitrite in vivo in mice. Toxicol Lett 42:47–53. 59. Munoz FJ, Bellido J, Moyano JC, Alvarez M, and Fonseca JL. 1996. Perioral contact urticaria from sodium benzoate in a toothpaste. Contact Dermatitis 35:51. 60. Murphy RY, Hanson RE, Johnson NR, Chappa K, and Berrang ME. 2006. Combining organic acid treatment with steam pasteurization to eliminate Listeria monocytogenes on fully cooked frankfurters. J Food Prot 69:47–52. 61. Murphy RY, Osaili T, Duncan LK, and Marcy JA. 2004. Effect of sodium lactate on thermal inactivation of Listeria monocytogenes and Salmonella in ground chicken thigh and leg meat. J Food Prot 67:1403–1407. 62. Münzner R, Guigas C, and Renner HW. 1990. Re-examination of potassium sorbate and sodium sorbate for possible genotoxic potential. Food Chem Toxicol 28:397–401. 63. Nair B. 2001. Final report on the safety assessment of benzyl alcohol, benzoic acid, and sodium benzoate. Int J Toxicol 20:23–50. 64. Nettis E, Colanardi MC, Ferrannini A, and Tursi A. 2004. Sodium benzoate-induced repeated episodes of acute urticaria/angio-oedema: randomized controlled trial. Brit J Dermatol 151:898–902. 65. Ortolani C and Ispano M. 1992. Foods and chronic urticaria. Clin Rev Allerg 10:325–47. 66. Ouattara B, Simard RE, Holley RA, Piette GJP, and Begin A. 1997. Inhibitory effect of organic acids upon meat spoilage bacteria. J Food Prot 60:246–253. 67. Pacor ML, Di Lorenzo G, Martinelli N, Mansueto P, Rini GB, and Corrocher R. 2004. Monosodium benzoate hypersensitivity in subjects with persistent rhinitis. Allergy 59:192–197. 68. Parke DV and Lewis DF. 1992. Safety aspects of food preservatives. Food Additiv Contamin 9:561–77. 69. Peirson MD, Guan TY, and Holley RA. 2003. Thermal resistances and lactate and diacetate sensitivities of bacteria causing bologna discolouration. Int J Food Microbiol 86:223–230. 70. Porto ACS, Wonderling L, Call JE, and Luchansky JB. 2003. Use of pulsed-field gel electrophoresis to monitor a five-strain mixture of Listeria monocytogenes in frankfurter packages. J Food Prot 66:1465– 1468.

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison 10 FRI LITERATURE REVIEW: Controlling Listeria monocytogenes on Ready-to-Eat Meat and Poultry Products

71. Radin D, Niebuhr SE, and Dickson JS. 2006. Impact of the population of spoilage microflora on the growth of Listeria monocytogenes on frankfurters. J Food Prot 69:679–681. 72. Robach MC and Sofos JN. 1982. Use of sorbates in meat products, fresh poultry and poultry products: a review. J Food Prot 45:374–383. 73. Schiffmann D and Schlatter J. 1992. Genotoxicity and cell transformation studies with sorbates in Syrian hamster embryo fibroblasts. Food Chem Toxicol 30:669–672. 74. Schultz-Ehrenburg U and Gilde O. 1987. Results of studies in chronic urticaria with special reference to nutritional factors [German]. Zeitschrift fur Hautkrankheiten 62:88–95. 75. Scotter MJ and Castle L. 2004. Chemical interactions berween additives in foodstuffs: a review. Food Additiv Contam 21:93–124. 76. Serdengecti N, Yildirim I, and Gokoglu N. 2006. Effects of sodium lactate, sodium acetate and sodium diacetate on microbiological quality of vacuum-packed beef during refrigerated storage. J Food Safety 26:62–71. 77. Shelef LA and Addala L. 1994. Inhibition of Listeria monocytogenes and other bacteria by sodium diacetate. J Food Safety 14:103–115. 78. Shtenberg AJ and Ignat'ev AD. 1970. Toxicological evaluation of some combinations of food preservatives. Food Cosmet Toxicol 8:369–380. 79. Sodemoto Y and Enomoto M. 1980. Report of carcinogenesis bioassay of sodium benzoate in rats: absence of carcinogenicity of sodium benzoate in rats. J Environ Pathol Toxicol 4:87–95. 80. Sommers C and Fan XT. 2003. Gamma irradiation of fine-emulsion sausage containing sodium diacetate. J Food Prot 66:819–824. 81. Sommers C, Fan XT, Niemira BA, and Sokorai K. 2003. Radiation (gamma) resistance and postirradiation growth of Listeria monocytogenes suspended in beef bologna containing sodium diacetate and potassium lactate. J Food Prot 66:2051–2056. 82. Stekelenburg FK. 2003. Enhanced inhibition of Listeria monocytogenes in frankfurter sausage by the addition of potassium lactate and sodium diacetate mixtures. Food Microbiol 20:133–137. 83. Stekelenburg FK and Kant-Muermans MLT. 2001. Effects of sodium lactate and other additives in a cooked ham product on sensory quality and development of a strain of Lactobacillus curvatus and Listeria monocytogenes. Int J Food Microbiol 66:197–203. 84. Supramaniam G and Warner JO. 1986. Artificial food additive intolerance in patients with angio- oedema and urticaria. Lancet 2:907–909. 85. Toth B. 1984. Lack of tumorigenicity of sodium benzoate in mice. Fundam Appl Toxicol 4:494–496. 86. Uhart M, Ravishankar S, and Maks ND. 2004. Control of Listeria monocytogenes with combined antimicrobials on beef franks stored at 4 degrees C. J Food Prot 67:2296–2301. 87. Uyttendaele M, Rajkovic A, Benos G, Francois K, Devileghere F, and Debevere J. 2004. Evaluation of a challenge testing protocol to assess the stability of ready-to-eat cooked meat products against growth of Listeria monocytogenes. Int J Food Microbiol 90:219–236. 88. Walker R. 1990. Toxicology of sorbic acid and sorbates. Food Additiv Contamin 7:671–676. 89. Warrington RJ, Sauder PJ, and McPhillips S. 1986. Cell-mediated immune responses to artificial food additives in chronic urticaria. Clin Allerg 16:527–533. 90. Würgler FE, Schlatter J, and Maier P. 1992. The genotoxicity status of sorbic acid, potassium sorbate and sodium sorbate. Mutat Res 283:107–111. 91. Wüthrich B and Fabro L. 1981. Acetylsalicylic acid and food additive intolerance in urticaria, bronchial asthma and rhinopathy [German]. Schweiz Med Wsch 111:1445–1450. 92. Xue B, Wang HF, Xu LH, Sun HF, Liu YF, Peng SX, Liu KX, and Guo ZY. 2004. Benzoate derived DNA adduction and its decay in mice measured by AMS. Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials & Atoms 223–24:765–769. 93. Yoon KS, Burnette CN, and Whiting RC. 2003. Effects of pH and agitation on the growth of Listeria monocytogenes Scott A in brain heart infusion broth containing combined potassium lactate and sodium diacetate during storage at 4 or 10 degrees C. J Food Prot 66:1469–1473. 94. Zhu MJ, Mendonca A, Ismail HA, Du M, Lee EJ, and Ahn DU. 2005. Impact of antimicrobial ingredients and irradiation on the survival of Listeria monocytogenes and the quality of ready-to-eat turkey ham. Poultry Sci 84:613–620. 95. Zhu MJ, Mendonca A, Min B, Lee EJ, Nam KC, Park K, Du M, Ismail HA, and Ahn DU. 2004. Effects of electron beam irradiation and antimicrobials on the volatiles, color, and texture of ready–to-eat turkey breast roll. J Food Sci 69:C382–C387.

Prepared by M. Ellin Doyle, Ph.D., [email protected] March 2006 Food Research Institute, www.wisc.edu/fri/ ©University of Wisconsin–Madison http://www.cfsan.fda.gov/~dms/opabacqa.html

FDA Approval of Listeria-specific Bacteriophage Preparation on Ready-to-Eat (RTE) Meat and Poultry Products In the Federal Register of August 18, 2006, FDA announced that it had approved the use of a bacteriophage preparation made from six individually purified phages to be used on RTE meat and poultry products as an antimicrobial agent against Listeria monocytogenes. The rule is in response to a food additive petition submitted in 2002 from Intralytix, Inc. Questions & Answers 1. What is a bacteriophage? Bacteriophages (phages) are viruses that infect only bacteria and do not infect mammalian or plant cells. Phages are ubiquitous in the environment, and humans are routinely exposed to them at high levels through food and water without adverse effect. 2. What is the additive that was approved? The additive that was approved is a mixture of equal proportions of six phages specific against L. monocytogenes. The petitioner's rationale for incorporating six phages in one formulation is to minimize the possibility of L. monocytogenes developing resistance to the additive. The approved phage preparation is reported to be effective against 170 strains of L. monocytogenes. 3. How is the bacteriophage used in the additive produced? The phage is cultured using L. monocytogenes as the host organism. Because the phage infects and destroys the host organism during its life cycle, the final product does not contain any viable Listeria. In addition, because of the purification steps that are used in making the phage, any toxic components from the host organism are effectively removed from the final phage product. To ensure safety, the regulation specifies that the additive must test negative for L. monocytogenes, and the L. monocytogenes toxin, listeriolysn O, must not be present at detectable levels, where the limit of detection is 5 hemolytic units per ml. 4. What is the purpose of the additive? The phage preparation will be used in meat and poultry processing plants for spray application to the surface of RTE meat and poultry products, such as lunch meats and hot dogs, to kill Listeria. The phage preparation will be applied to the surface of RTE meat and poultry products at a level not to exceed 1 ml per 500 cm2 food surface just prior to packaging. These foods can become contaminated with Listeria during production, but unlike fresh meat and poultry, these foods are consumed without additional cooking that would kill the bacteria, thereby increasing the risk to Listeriosis, an infection caused by the bacteria. Furthermore, L. monocytogenes is capable of multiplying at low temperatures, so it can multiply in food over time even when the food is kept in the refrigerator. Persons with increased risk to Listeriosis are pregnant women, newborns, those with a compromised immune system, and the elderly. 5. Is this additive novel? This is the first time FDA has regulated the use of a phage preparation as a food additive. However, phages are currently approved in the U.S. for pesticide applications, including uses on crops. Although not currently permitted in the U.S., phages are used in other countries in antibiotic therapy. 6. What was the basis of FDA's safety decision? To support the safety of the proposed use of the Listeria-specific phage preparation, the petitioner submitted published reports from animal studies and results from the use of phages as therapies in the treatment and prevention of infections in humans caused by bacteria. Based on this and other information, FDA concluded that the additive does not pose any safety concerns, providing that it complies with the identity and specifications in the regulation. 7. What was the process of deciding that the phage was suitable for use in meat and poultry under the conditions of use? Under the Memorandum Of Understanding (MOU) for the joint USDA Food Safety and Inspection Service (FSIS) and FDA ingredient approval process, FDA is authorized to determine the safety of substances, as well as prescribe safe conditions of use. FSIS is responsible for evaluating the suitability of the use of new ingredients (when the use involves meat or poultry products). "Suitability" relates to the effectiveness of the substance in performing the intended technical purpose of use, at the lowest level necessary, and the assurance that the conditions of use will not result in an adulterated product or one that misleads consumers. In regard to L. Monocytogenes control, the guidance has been at least a 1 log reduction (for an antimicrobial agent used to reduce or eliminate Listeria) and no more than 2 logs growth during a product's shelf life (for an antimicrobial agent that is intended to suppress the growth of Listeria). Therefore, to satisfy FSIS' concerns, the petitioner included data that showed that the use of the phage preparation was effective in reducing/controlling Listeria without resulting in an adulterated product or one that misleads consumers. 8. Will the additive be declared on the food label? Use of this additive must comply with the Federal Meat Inspection Act or the Poultry Products Inspection Act, which are both administered by USDA. According to USDA, the use of the phage preparation will need to be declared as an ingredient on the label of the treated meat or poultry product, e.g., "bacteriophage preparation." In addition, when the phage is applied to meat and poultry products with standards of identity that do not permit the addition of antimicrobial agents, the products will need to be descriptively named, e.g., "beef steak treated with an antimicrobial solution to reduce microorganisms."

Retail Deli Task Force on Lm

Project Title: An Industry Survey For the Presence of Listeria monocytogenes in Grocery Deli Operations

Principle Investigator: Ann Marie McNamara, Sc.D., Silliker, Inc., 900 Maple Road, Homewood, IL. 60430, Tel : 630-715-2416, Fax : 708-957-8446, email : annmarie,[email protected]

Project Summary: This study will survey 100 grocery deli operations throughout the U.S. The design will incorporate the collection of 15 environmental samples, cook-in-bag product sliced at the deli, verification of temperature recording devices, and a survey of deli housekeeping and SOPs by interviewing grocery staff at each deli visit. The intent is to determine if, and to what extent, Listeria monocytogenes (Lm) is present in delis and to assess the possibility of Listeria transfer through the slicing of cook-in-bag meats and in the contamination of the environment. Verification of temperature recordings and surveys of retail food handling practices may provide insight into any contamination found and may provide possible mitigation factors lacking in the NIFSI study (described below). This study will be blinded on multiple levels so that the identity of groceries, manufacturers of cook-in-bag products to be sliced, and identity of surveyors will not be known to the laboratories processing the samples and to the principal investigator analyzing the data. Data will be blinded in a way to make traceback impossible.

This project directly relates to AMI goals on understanding the environmental niches, transfer, and potential control measures for Listeria monocytogenes in a segment of our industry, grocery deli operations. Results will allow AMI manufacturers and grocery store members to respond to USDA questions on control practices and mitigation measures for Listeria monocytogenes. In addition, these study results will allow for proactive industry responses, documented by independent data collection, when NIFSI study results are presented to the public.

Project Description:

Background: USDA is currently funding a NIFSI study in which deli and manufacturer packaged, RTE meat and poultry products from grocery stores will be tested for the presence of Listeria monocytogenes (Lm). The design of this study contains numerous flaws that will not elicit the necessary data needed by the industry to counteract consumer criticism and to generate mitigation strategies. This project proposes a survey of deli operations that is designed to answer questions about the deli environment, people practices, temperature controls and deli-sliced products that the NIFSI study lacks.

The design of the NIFSI study allows only for the sampling of meat and poultry products at deli counters. Since the investigators are visiting these stores surreptitiously, investigators are unable to take concomitant environmental samples, independently verify temperature readings, or question store personnel on product handling practices. These omissions are serious flaws in this study since any Lm positive findings cannot be traced back to a root cause. At best, this study will be able to conclude with certainty that only a given number of samples were positive. These results may incite concern by the general public and consumer activists and provide no answers to retail operators for the cause of the contamination or possible mitigation measures.

The limitations of this NIFSI study and the potential for sensationalism of the results without linkages to causation, was discussed at the March 2-3, 2005 AMI Scientific Affairs Advisory Committee Meeting, Atlanta, GA. Committee members believed that a properly designed study would address the question of the impact of deli operations on product contamination. SAC members identified the following data gaps in the current NIFSI study that will be addressed by this new survey: · Environmental deli samples should be considered. · The exterior of intact packages should be sampled. · Listeria inhibitors in the product sliced at deli should be known if possible. · Retail cooperators are needed since deli access will be required to conduct the study appropriately. · Use Cook-in-Bag product. Consider obtaining a list of products (brands) that are known to have been cooked in bag. This could determine if inhibitors were present in the products and determine if Lm transfer is occurring at the retail slicer. · Design of study should be geared to assess Listeria transfer. · General deli housekeeping and SOPs within the deli should be recorded.

Description of Project: This study will survey 100 grocery deli operations throughout the U.S. The design will incorporate the collection of 15 environmental samples, cook-in-bag product sliced at the deli, verification of temperature recording devices, and a survey of deli housekeeping and SOPs at each deli visit. The intent is to determine if, and to what extent, Lm is present in delis and to assess the possibility of Listeria transfer through the slicing of cook-in- bag meats. Verification of temperature recordings and surveys of retail food handling practices may provide insight into any contamination found and may provide possible mitigation factors lacking in the NIFSI study. This study will be blinded on multiple levels so that the identity of groceries, manufacturers of cook-in-bag products to be sliced, and identity of surveyors will not be known to the laboratories processing the samples and to the principal investigator analyzing the data. Data will be blinded in a way to make traceback impossible.

Sliced deli meat and environmental sponges of the slicer, food contact surfaces, and the deli case will be collected and analyzed for the presence of Lm. Sponges will be collected in DE Neutralizing broth and shipped to participating Silliker laboratories for processing. Sample collection kits will be prepared by the Silliker Corporate Research Laboratory and sent to Silliker sample collectors. Verification of temperature recording devices at retail will be conducted using calibrated thermometers.

This study will require the cooperation of AMI processed meat and poultry manufacturers and grocery operators to be successful. AMI meat and poultry manufacturers on the AMI Scientific Affairs Advisory Committee have supported participation in this project. At the time of AMI proposal request, at least one grocery chain had agreed to participate and four other grocery chains were reviewing the proposal with their legal staff, Participating manufacturers will need to provide the names of cook-in-bag products found in a typical deli and the additives/post-packaging interventions used to make these safe products. All cook-in-bag products will be considered to be free of Listeria from the manufacturer and sliced at deli to show the potential for cross-contamination at the deli and transfer of Listeria through slicing. Grocery operators will need to provide a letter of introduction for Silliker sample collectors, complete with contact information of a designated food safety/regulatory person at the grocery chain, to assure cooperation during unannounced visits.

Benefit of Research to Industry: The flawed design of the NIFSI study may produce results that will only incite concern by the general public and consumer activists and provide no answers to retail operators for the cause of the contamination or possible mitigation measures. This survey is designed to quickly ascertain the degree to which deli environments and people practices may contribute to potential Lm contamination and to suggest possible mitigation practices for retail operators. The data generated may help counteract any negative press generated by the NIFSI study as well as contribute to the general understanding of Listeria contamination, transfer, and possible control measures in grocery operations..

Experimental Plan: This proposal tests the deli environment and sliced product for Lm. This same study can be designed to detect Listeria species if there is a concern about regulatory issues. All samples will be blinded so no trace back will be possible. The NFPA retail survey and now the NIFSI study have been, or are being conducted, without concomitant regulatory action on Lm positive products.

A. Collection of Deli-Sliced Product Samples Upon each site visit, the sample collector will request that 0.5 lb (minimum of 15 slices) of a cook-in-bag product identified by participating manufacturers will be sliced in their presence at the deli. The purpose will be to assess potential Lm transfer. Samples will be shipped refrigerated overnight to a participating Silliker laboratory.

B, Collection of Environmental Samples

Environmental sponges of the slicer, food contact surfaces, and the deli case will be collected for the presence of Lm. Sponges will be collected in DE Neutralizing broth and shipped to participating Silliker laboratories for processing. Sample collection kits will be prepared by the Silliker Corporate Research Laboratory and sent to Silliker sample collectors.

Tentatively identified environmental sample sites include the following. Further identification/modification environmental sites by the AMI Foundation members are welcomed.

1. Slicer-blade, back plate, meat collection areas, guard, handle.

2. Food contact surfaces – weighing station, cutting boars/preparation tables, trays for transporting products.

3. Deli Case – trays or flooring of case, door handles, door track, drains beneath/around deli case.

4. Outside packaging of cook-in-bag product sliced at deli.

C,Laboratory Methodology

Lab methodology will follow that of D.E. Gombas et al1. In brief, 25g of product will be enriched in UVM broth for 24h, transferred to MOPS-BLEB for 24h, and analyzed by Dupont Qualicon’s ABAX PCR test for Lm. Environmental samples will be processed using current Qualicon ABAX recommendations: UVM for 24h incubation then MOPS-BLEB for 24h incubation. (Note: Neither the NFPA nor the NIFSI study collected environmental samples.) Positive samples will be quantitated by MPN and confirmed biochemically.

D.Verification of Temperature Recordings

The sample collector using a calibrated thermometer will perform verification of temperature readings by temperature recording devices in the deli area. Readings will be taken of the deli case and product storage refrigerators.

E.Deli Practices and SOP Survey.

Upon acceptance of this proposal by the AMI Foundation members, a complete survey instrument will be developed. The sample collector will conduct this survey at the time of the grocery store visit. Deli personnel will be interviewed on deli practices and not general store managers. This survey is intended to determine current deli food handling practices and SOPs that may correlate to any positive Listeria environmental or product samples. This survey will include questions on products offered at deli counters (deli salads, blue-veined or soft-mold- ripened cheeses, cooked products such as rotisserie chickens) separation of raw and RTE products and preparation areas, shelf-life dating of products, age and manufacturers of slicers, certification of refrigeration thermometers, cleaners and sanitizers used, sanitation schedule for food utensils and food contact surfaces, and safe food handling practices.

Silliker is uniquely positioned to conduct such a study quickly and competently. We have sixteen consultants and technical sales managers trained in the collection of environmental and product samples and their shipping requirements. These Silliker personnel are located in Illinois, Wisconsin, Iowa, Minnesota, Pennsylvania, New Jersey, New York, Georgia, Texas, and Northern and Southern California. They can be mobilized to collect samples in adjoining states. Silliker has conducted large marketbasket surveys for trade organizations, such as AMI and NCBA, and individual companies.

References:

1. Gombas DE, Chen Y, Clavero RS, Scott VN. Survey of Listeria monocytogenes in ready-to- eat foods. J Food Prot. 2003 Apr;66(4):559-69.

Budget: Cost of swab kits, collection of 15 environmental and 1 product samples, shipment of samples to lab, processing of samples by PCR, results generation, survey administration to deli personnel, and temperature verifications = $703.00 per site visit or $70,300.00 per 100 visits (number of visits may be modified by AMI Foundation members)

MPN analysis for hypothesized 25% positive environemental/productsamples = $6,250.00 Note: If less than 25% of samples need MPN analyses, only cost for number of samples tested will be charged to AMI

Survey preparation, review by AMI Foundation, correlation meeting for sample collectors =$4,750.00

Data analysis and report = $6,000.00

Total for 100 site visits = $87,300.00

Note: Silliker, Inc has conducted research projects for the AMI Foundation and agrees to the terms of the AMIF policies on research funding and execution.

Approximate Timetable: 5 months

. Processing of samples will follow the NFPA protocol used in their retail survey (D.E. Gombas et al1). In brief, 25g of product will be enriched in UVM broth for 24h, transferred to MOPS-BLEB for 24h, and analyzed by Dupont Qualicon’s ABAX PCR test for Lm. The remaining portion will be held for quantitation by MPN analysis and biochemical confirmation of isolates. One product will be sliced at each visit (100 products, 100 visits total).

Proposed Blinding Scheme for AMI-FMI Retail Listeria Study

Goal: To protect the identity of the retail stores participating in the retail Listeria study.

Protocol: Sample collectors will be assigned sufficient numbers to represent the number of stores being visited. Each sample collector will randomly assign a number to a retail store at the time of the visit. That number will be used on each sample collection form, retail survey, and sample labels to protect the identity of the retail store. This number will follow the materials through the laboratory to the reports to the statistician. The statistician will assign a new number to the data collected from each store at the time that the data is being entered into the database. Original data containing the original, randomly assigned number is then destroyed to prevent any traceback. All data containing the new number will be in numerical order based on the time of receipt of the data.

An example of this procedure is as follows: Sample collector A is assigned responsibility for sampling 10 retail stores. He will be assigned the numbers 40 to 50 to represent these stores. The sample collector will randomly assign one of these numbers to each retail store at the time of the visit. Only he will know the identity of the store. That number, for example number 43, will be used on each sample collection form, retail survey, and sample labels from that store. This number will follow the materials through the laboratory to the laboratory reports to the statistician. Neither the laboratory nor the statistician will know the identity of the store. The statistician will assign a new number to the data collected from each store at the time that the data is being entered into the database. Original data containing the original, randomly assigned number is then destroyed to prevent any traceback. For instance, if the data labeled 43 is the 67th data set to arrive at the statistician, it will now become 67, the data is entered into the database using this new number, and the original number and documents discarded. In this manner, the original store has been protected through two assignments of new identifiers (numbers) that do not link back in any way to the original store.

FSIS Response to AMI Comments on Listeria Risk Assessment FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006.

FSIS Response to Supplemental Public Comments on The 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat

by Daniel Gallagher, Ph.D. Virginia Polytechnic University Under SAIC Contact 43-3A94-4-5115

and

The Risk Assessment Division Office of Public Health Science Food Safety and Inspection Service U.S. Department of Agriculture

March 15, 2006 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006.

On February 26, 2003, FSIS held a public meeting to present the draft peer-reviewed 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat. At this meeting, FSIS announced that it would like to receive additional public input and information through Docket 03-005N. In addition to having the risk assessment report on the FSIS website, the risk assessment model and supporting data were publicly available in Docket 03-0005N. As a result, FSIS received a number of comments and additional information that strengthened this final risk assessment and it was used to guide the development of the FSIS Interim Final Rule to Control Listeria monocytogenes in Ready-to-Eat Meat and Poultry Products (9 CFR 430). FSIS published the interim final rule on June 6, 2003 and sought further public comment on both the rule and the risk assessment, with a public comment period of more than one year.

The following are eight supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat received in January 2006 from the American Meat Institute (AMI). FSIS has provided a response to each of AMI’s comments (see http://www.fsis.usda.gov/Science/Risk_Assessments/index.asp for a complete text of AMI’s submitted supplemental comments). References to page numbers, table numbers, figure numbers, etc. correspond to the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat provided on the FSIS web site.

1. The model assumes that the L. monocytogenes contamination comes from a reservoir (a niche, or harborage site) in the plant, without consideration for contamination from sporadic positives or contamination arising at retail.

The model does assume that Listeria monocytogenes contamination comes from a reservoir. However, the commenter needs to provide a physically plausible explanation of sporadic positives. “Sporadic” Listeria species positives may appear because the sampling frequency is too low, because the concentrations are near the detection limit, because the duration of a contamination event is short, or due to a lag in sampling product after a Listeria species positive. The presence of Listeria monocytogenes on a food contact surface (FCS) clearly increases the risk of product contamination.

The in-plant model generates transient and sporadic positives, as well as positives clustered in time. This is not inconsistent with stochastic contamination events that move Listeria monocytogenes from a harborage site to FCS. The duration and frequency of the contamination events are stochastic, and thus shorter events can produce positive findings that appear transient and sporadic. Longer events together with high frequencies of sampling appear as a process no longer in control. The model produces both situations.

The current version of the model does not consider contamination at retail, and published data (Gombas et al 2003) indicate that such contamination does occur. FSIS is currently assessing recently completed research by the National Alliance for Food Safety and Security to understand more fully the degree of contamination that occurs at retail establishments versus processing plants.

2 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006.

2. The draft risk assessment fails to consider the operational parameters associated with processing deli meats and other RTE meat and poultry products. These factors are significant to the discussions of product contact surfaces and other such issues raised as major considerations in the draft risk assessment. Failure to examine the operational factors in detail greatly reduces the value of the draft risk assessment in delivering an appropriate and useful risk estimate.

The 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat was designed to answer specific risk management questions. Rather than simulate each possible process configuration, a range of model input values are used that capture the variability within the industry. This eliminates the need to simulate every possible individual operation or process, while still allowing valid conclusions to be drawn. Stochastic modeling is also appropriate in cases such as this where data are limited.

FSIS is not aware of any published data which indicates large plants are actually less risky than small or very small plants. Current risk based plant sampling may eventually be able to answer this question, but for now the model assumes that plant size does not directly correlate with risk.

3. The draft risk assessment makes unrealistic estimates of the efficacy of sanitation and corrective actions that are critical to the success of on-going control of Listeria in processing environments. The efficacy of post-packaging treatments is also unrealistically low.

FSIS agrees that sanitation effectiveness is one of the more uncertain parameters. However, as currently configured the model results are not very sensitive to sanitation effectiveness. Because of the range of transfer coefficients used in the model, most of the bacterial cells are transferred from the FCS to the product before the surface is sanitized at the end of the shift. This is clearly mentioned on page 29 of the report.

When additional data are made available regarding transfer coefficients, they will be evaluated for use. If the transfer coefficient is lowered, then the specific value of sanitation effectiveness becomes more important.

The routine daily sanitation effectiveness in the model is 87.5%, and this is within the range of published data.

If AMI has data to indicate 99-99.9% sanitation effectiveness is more likely, they should make this data available in the open literature. As discussed on page 14 of the report, one interpretation of the Lunden et al. 2002 paper is that it can be very difficult to sanitize harborage sites.

These efficiencies are not assumptions in the model. As shown in Figure 23, the lethality effectiveness ranged from 70 to 99%, with industry participation ranging from 50 – 100%. At no point were they tied to specific interventions. A range of values was

3 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006. chosen so that risk managers could relate public health improvements to post processing efficiencies. See risk management question #2 on page 8 of the report.

This risk assessment was not asked and did not attempt to identify specific types of interventions other than general post processing lethality and use of growth inhibitors. There was no need to model specific treatments such as sodium lactate, irradiation, or steam pasteurization.

Note: the AMI comments themselves indicate the very high efficiencies suggested are not always applicable. Post processing lethality is effective only when “environmental contamination” levels exist. But in some, possibly rare, instances the process is not fully in control or the product is abused, resulting in greater than “environmental levels” of contamination. Requirements for post processing lethality treatments are provided in the Compliance Guidelines accompanying the Interim Final Rule to Control Listeria monocytogenes in Ready to Eat Meat and Poultry Products (9 CFR 430), available at www.fsis.usda.gov.

4. All current, relevant scientific literature and industry data have not been integrated into the draft risk assessment. There is an over-reliance on single sets of data to develop the draft risk assessment when, in some cases, additional data were available. The draft risk assessment does not provide all references cited in the document.

The 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat made use of all available and appropriate data sets. The report discusses the limitations associated with a relative paucity of particular types of data. The report includes data gathered by FSIS during In-Depth Verification (IDV) reviews following a non-compliance event, such as a product sample which tested positive for Listeria monocytogenes. IDV data was used to estimate contamination frequency. The report clearly discusses this limitation on page 23.

FSIS disagrees with the comment that “all current relevant scientific literature and industry data have not been integrated”. FSIS used both industry and published scientific data as shown in the citations of the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat report. As just one of several examples, FSIS used industry data provided by Dr. Bruce Tompkin that was published in the peer reviewed scientific literature in order to estimate the duration of a contamination event. The AMI comment cited the need for specific industry data with regards to the effectiveness of specific interventions for which the model accommodates through the use of sensitivity analyses. If specific data on the effectiveness of specific interventions is known, then users can simply use the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat model and evaluate the estimated public health impact for specific interventions.

4 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006.

“A statement on page 14 indicates that the data from this IDV do not tend to exhibit the duration seen in other data, but it is not clear what these other data are (the Tompkin data?)” This statement included in the comments does not appear in the FSIS published report. A similar statement concerning the IDV data appears on page 23-24 “Nor does the data provide sufficient sampling evidence to estimate the duration of contamination in comparison to other data (i.e., Tompkin, 2002).

With regards to the inappropriate use of the term indicator organism, as opposed to index organism, ICMSF 7 (Microbiological Testing in Food Safety Management) provides a discussion of indicators and a table describing factors to consider when selecting an indicator. The FSIS use appears consistent with this reference. The term “index” organism does not appear in the index.

5. In many cases the draft risk assessment fails to provide adequate support for the assumptions, variability and uncertainty for the model parameters. In some cases the draft risk assessment appears to use unrelated and inappropriate data as bases for it mathematical calculations, greatly decreasing the potential validity of the draft risk assessment, particularly in relation to the transfer coefficient. Furthermore, data and opinions unrelated to the scope of the draft risk assessment are included.

The suggestion that contaminant event timing should change based upon plant interventions is probably accurate. However, the Lunden et al. 2002 paper illustrates how difficult it can be to control Listeria monocytogenes contamination once a harborage site is established. Given the limited contamination event timing data, treating the contamination event parameters as independent of interventions is conservative in protecting public health.

With regard to plant size not being related to FCS area, the Exponent report conducted for AMI clearly shows that FCS is related to plant size. Conceptually, there may be better variables to consider such as process line configuration and packaging technology, but plant size in a stochastic model appears adequate.

With regard to product configuration (stacked, shingled, etc) affecting the transfer coefficient, the level of detail required to simulate each product configuration is not warranted. The stochastic nature of the model should account for these variations.

With regard to the assumption in the risk assessment that concentrations distributions were similar to prevalence distributions, FSIS agrees that this assumption requires further data. But this assumption is discussed in detail on Page 33-34 of the report, and further evaluated by sensitivity analysis in Table 28.

With regard to the assumption that bacteria are uniformly distributed spatially on the FCS and throughout a product lot, FSIS agrees that this is a major limitation, and is discussed as such on Page 37 of the report. From a sampling perspective however, such a limitation

5 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006. could be overcome by an appropriately designed composite sampling design rather than simple random samples.

With regard to the fixed growth of 1 log from production to retail, FSIS agrees that it is a simplification. However, it is a consistent approach to the model used in the 2003 FDA USDA Quantitative Assessment of Relative Risk to Public Health from Foodborne Listeria monocytogenes Among Selected Categories of Ready-to-Eat Foods, which used 2 log growth as a constant. The limitation is clearly discussed on page 39 of the report and Appendix B is used to discuss the appropriate value.

The comment that the University of Georgia data were not considered is simply incorrect. It is referenced as Deaver (2002). These data were used as the basis of Figure 9 dealing with sequential testing and therefore heterogeneity, and an example of why this prevalence data could not be used to generate a transfer coefficient is given in Table 17. These data were carefully considered in the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, not ignored as the comments suggested.

6. The draft risk assessment should describe in more detail the limitations of sampling and testing programs to detect low level prevalence of Listeria, whether on food contact surfaces or in RTE products. Oversimplification leads to unscientific conclusions relative to sampling and testing as a means to control Listeria, particularly in operations where Listeria control programs are very effective in reducing the likelihood of Listeria being present, or persisting, in the processing environment.

FSIS feels these comments do not present the full benefit of food contact surface and product sampling. It is true, as discussed in the report, that sampling has limitations especially as prevalence decreases. The modeling results showed that even extensive sampling only identified lots with Listeria monocytogenes concentrations above the detection limit. If product abuse occurred prior to consumption, even low concentrations could grow to cause public health impacts. So sampling is limited by prevalence and detection limits. Most of the criticisms of sampling assume a binomial distribution because a sample is found to be either positive or negative. However, the binomial distribution assumes homogeneity and independence from one sample to the next. The risk assessment model does not make these assumptions. Because of contamination events and mass balance principles, there is a correlation in the Listeria monocytogenes concentrations from one lot to the next. Thus there is a degree of clustering in the product lots that are contaminated. In this case, something like a negative binomial distribution is more appropriate than a binomial. Even low frequency of sampling can produce public health benefits if response is taken after a Listeria monocytogenes positive.

With regard to the statement in the AMI comments that “In order to define sampling and testing program, it is necessary to define prevalence of pathogen, sensitivity & selectivity of assay, and number of samples from lot… These inputs will provide a probability of excluding defective lots.”

6 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006.

In effect the risk assessment did exactly this, but in a more physically and biologically plausible mathematical framework than a simple binomial table. The model was also designed to answer risk management questions, not merely a probability of excluding defective lots. The risk assessment simulates Listeria monocytogenes concentrations rather than prevalences. Note the concentrations are model outputs, not inputs. Also note that prevalences can be calculated if concentrations and detection limits are known, but not the reverse. Concentrations are also required if mortality and morbidity are to be predicted (as through a dose response curve). An example of the predicted Listeria monocytogenes concentrations is given in Table 20. The assay sensitivity is discussed in pages 34-36. The various sampling frequencies are discussed throughout (e.g. 4-2-1 indicates 4 FCS samples per line per month for large plants, 2 for small, and 1 for very small). An example of the probability of excluding lots is provided in Table 27.

The table from the International Commission on Microbiological Specifications for Foods (ICMSF) (Table 7-1) referenced in the AMI comments is a simple binomial table. For the use of this table to be valid, Listeria monocytogenes concentrations must be homogeneous and samples must be independent. In the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, FSIS did not want sampling to be independent. Rather, the Agency approach is to trigger additional sampling (and corrective actions) whenever a positive is found.

Thus, the binomial distribution of the ICMFS table is not a valid point of comparison, and the table does not apply in the context of the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat and its associated risk management questions. If the degree of clustering was quantifiable a priori, a table based on a negative binomial distribution would be more appropriate. For the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, the dynamic model generates the clustering over time, and degree of clustering does not need to be specified beforehand. (It is implied from the contamination event duration & frequency, as well as the plant production levels and sanitation effectiveness, i.e. anything that affects the mass balance.)

Because Listeria monocytogenes contamination is clustered over time, even limited sampling can produce public health benefits if rapid actions are taken when positive samples are found. This also suggests that waiting until a fixed number of sequential positives are found greatly reduces the effectiveness of sampling.

7. The draft risk assessment should provide more consideration to the numerous intervention technologies in use to help control Listeria, particularly where L. monocytogenes is not a hazard reasonably likely to occur because of control procedures addressed in the Sanitation SOPs and other programs, as acknowledged in the draft risk assessment by FSIS.

Rather than simulate each possible intervention or technology possible, a range of input values are used that capture the variability within the industry. This eliminates the need

7 FSIS response to supplemental comments on the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat, (Docket #03-005N), submitted by the American Meat Institute on January 03, 2006. to simulate every possible individual operation or process, while still allowing valid conclusions to be drawn.

8. The draft risk assessment was not released for “use and experimentation” by interested stakeholders, providing no opportunity for further, “hands-on” analysis of the draft risk assessment before the comment period was over. The FSIS draft risk assessment needs to be reviewed by an independent, expert third-party.

The 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat report and model were formally peer reviewed in accordance with the guidelines issued by the Office of Management and Budget and made publicly available through Docket 03- 0005N prior to the February 2003 public meeting. Subsequently, the risk assessment report and model were updated based on public input and have been publicly available through the Docket (03-0005N) since May 2003. The 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat model and report are currently posted to the FSIS website (www.fsis.usda.gov ) for increased public accessibility.

In addition, AMI’s contractor Exponent requested and received the 2003 FSIS Risk Assessment for Listeria monocytogenes in Deli Meat model directly from FSIS in 2004. Attached to these comments from AMI was a report by Exponent that was funded by AMI evaluating the model with industry data.

8 Petition for Regulatory Limit for Lm NACMCF Report on Campylobacter Methods NATIONAL ADVISORY COMMITTEE ON MICROBIOLOGICAL CRITERIA FOR FOODS

Analytical Utility of Campylobacter Methodologies

Adopted September 28, 2005 Coral Gables, FL

1 TABLE OF CONTENTS

Executive Summary

I. Introduction

II. Purpose of the Document

III. Original Work Charge and Background

IV. Response to Questions

1. What additional circumstances should be considered in order

for the Food Safety and Inspection Service (FSIS) to conclude that the

poultry baseline study should address more than the two principal

Campylobacter species of Campylobacter jejuni and Campylobacter coli?

2. How can the Agricultural Research Service (ARS) method be most

successfully used for high volume analysis in the conduct of a baseline

study of Campylobacter presence and enumeration on poultry

(chicken, turkey, goose, etc.) carcasses, parts, and

ground product that may lead to a potential performance

standard or guideline for the regulated industry? What if

any modifications should be made as a result of

discussing this method in comparison with others

presented to the Committee? Please consider whether

the above described atmospheric conditions, media,

pre-enrichment, and storage media are acceptable

for the objective of this baseline study.

3. To utilize FSIS resources efficiently and effectively,

2 FSIS expects to maintain as much continuity as possible

between the current broiler rinse sampling for Salmonella

and the proposed sampling for Campylobacter spp.

What concerns regarding the Campylobacter spp.

sampling method need to be attended to in order

to properly address post-chill injured Campylobacter spp.

cells as well as viable non-culturable (coccoid) cells?

4. What further subtyping methods should be performed

on confirmed cultures (restriction fragment length

polymorphism (RFLP), amplified fragment-length

polymorphism (AFLP), pulsed-field gel electrophoresis (PFGE),

ribosomal DNA sequencing, antibiotic susceptibility, etc.),

and what, if any, limitations do any of these

methods pose?

5. What effect would in situ Campylobacter spp. cell

aggregation have on the accuracy and reproducibility of

enumeration counts and is there any remedy to

address this issue?

6. Occasionally non-thermophilic Campylobacter species

cause human illness. It is unclear whether livestock and

poultry are reservoirs for these species, or if they are

present on meat and poultry products following

slaughter and processing. Current methodologies

3 use selective agents and incubation conditions which

may reduce their detection. If a pilot study was conducted

to ascertain the presence of these species on meat and

poultry products, what methodologies would best

detect these species?

V. Conclusions

VI. Acknowledgments

VII. References

Appendix I. ARS/FSIS Broiler Rinse Study Campylobacter spp.

Enumeration Method

Figure I. NACMCF Suggested Campylobacter Analysis Protocol for FSIS

Baseline Studies for Poultry

4 EXECUTIVE SUMMARY

The National Advisory Committee on Microbiological Criteria for Foods

(NACMCF, or the Committee) was asked to address six questions posed to the

Committee regarding Campylobacter. These questions relate to the analytical utility of

Campylobacter methodologies in preparation for an upcoming Food Safety and

Inspection Service (FSIS) baseline study to enumerate Campylobacter spp. from broiler carcass rinse samples.

To address the FSIS questions, the Committee reviewed the available literature regarding Campylobacter spp. methodologies, consulted four US experts on

Campylobacter research, and examined the current method being used in an on-going

United States Department of Agriculture (USDA) collaborative study between FSIS and the Agricultural Research Service (ARS), hereinafter referred to as the ARS/FSIS Broiler

Rinse Study, for possible use in the upcoming FSIS baseline study on broilers.

As described more fully in this report, the Committee acknowledges that

Campylobacter spp. are a leading cause of foodborne illness in the United States and that poultry is a primary reservoir of this pathogen. In addition, the Committee recognizes that the majority of human campylobacteriosis cases are caused by C. jejuni, followed by

C. coli, and other species. Finally, the Committee understands that FSIS is awaiting

scientific recommendations from NACMCF prior to initiating a nationwide baseline

study to determine the prevalence and numbers of Campylobacter spp. in broiler carcasses at Federally-inspected establishments as a basis for developing risk management strategies to reduce human exposure to Campylobacter spp.

5 A general summary of the recommendations of the Committee, based on the six questions posed by FSIS, follows:

• The FSIS microbiological baseline study design for broiler carcasses should be

based on the species of Campylobacter causing the majority of human illness,

namely C. jejuni and C. coli;

• FSIS should partner with other researchers to develop methodologies and

conduct surveillance studies to sample poultry products for Campylobacter

species other than C. jejuni and C. coli since their prevalence and human illness

impact is presently unknown;

• FSIS must clearly state the objectives and potential uses of the baseline data;

• FSIS should determine if analyses from a single carcass rinse for generic

Escherichia coli, Salmonella, and Campylobacter would strengthen the

evaluation of process control for FSIS and industry;

• The direct plating enumeration methodology currently used in the ARS/FSIS

Broiler Rinse Study should be adapted for the upcoming FSIS baseline study,

with modifications as indicated throughout this report;

• Individuals conducting sample collection and microbial analyses should be

adequately trained; and

• FSIS should speciate Campylobacter spp. to differentiate C. jejuni and C. coli.

6 I. INTRODUCTION

Campylobacter spp. are major bacterial agents of human foodborne

gastroenteritis. Poultry is a primary reservoir of Campylobacter spp. and studies show

that prevalence may be more than 80% in commercial chicken carcasses (19, 25). Data show that 95% of human illnesses associated with campylobacteriosis are caused by C. jejuni, followed by C. coli comprising 4%, and other species comprising 1% (23).

For clarity, NACMCF defines the term “broilers” in the same manner as the 2004

NACMCF report “Response to the Questions Posed by FSIS Regarding Performance

Standards with Particular Reference to Broilers (Young Chickens)” (15). In this report, a

broiler is defined as a young chicken of either sex usually under 13 weeks of age. FSIS

has proposed to reduce this age requirement to under 10 weeks.

In the past, FSIS has conducted baseline studies of Campylobacter spp. enumeration from broiler carcass rinses using a labor intensive Most Probable Number

Method (MPN) (25, 26); however, not all of these studies have been published. FSIS is planning to initiate a new Campylobacter spp. baseline study and asked NACMCF to advise FSIS in developing the upcoming methodology. The Committee was asked to evaluate a direct plating method currently being used for enumerating Campylobacter spp. in a joint ARS/FSIS Broiler Rinse Study (Appendix I.), for its utility in the upcoming FSIS baseline study.

II. PURPOSE OF THE DOCUMENT

The purpose of this document is to summarize the past Committee deliberations

on Campylobacter, including previous questions posed to the Committee, and to address new questions posed by FSIS to the Committee at the July 12, 2005, public meeting.

7 Since a formal report specifically addressing Campylobacter methods had not previously been developed by NACMCF, the Committee is using this report to compile all past and current activities associated with Campylobacter.

III. ORIGINAL WORK CHARGE AND BACKGROUND

Background

Previous NACMCF Activity: The Committee deliberated on the issue of Campylobacter as an emerging pathogen in 1993, and in December 1994, published a review of

Campylobacter jejuni/coli (13). In May 1999, the National Advisory Committee on Meat and Poultry Inspection (NACMPI) asked NACMCF to evaluate options for defining a

Campylobacter performance standard or, in the alternative, a performance standard that would accomplish the same public health objective. In response, a subcommittee of

NACMCF (i.e., at that time, the Meat and Poultry Subcommittee worked on the issue) made a series of recommendations that were never formally adopted by the full

Committee but were pursued for further action by both NACMPI and FSIS (12).

At the August 28, 2002 NACMCF meeting, a work charge relating to

Campylobacter was introduced by FSIS. The presentation consisted of three talks outlining the following: the methods and resulting data from previous Campylobacter baseline studies performed in 1994-1995 and 1999-2000 by FSIS; available laboratory methods for Campylobacter spp.; and Campylobacter cell aggregation (14). The charge to NACMCF at the 2002 public meeting included three elements:

• Review and compare the methodologies used for Campylobacter detection in the

1994-1995 and 1999-2000 baseline studies in young chickens;

8 • Evaluate them for the accuracy and precision that they provide in assessing the

prevalence and quantity of Campylobacter on chicken carcasses; and

• Compare the methodologies used in the two studies with recent methodological

advances for their ability to provide data on the presence and quantity of

Campylobacter for application in risk assessment and the establishment of baselines.

Although the 2002 NACMCF Campylobacter Subcommittee met and discussed

the FSIS charge, no formal report was adopted. In addition, the FSIS baseline data from

1999-2000 have never been released due to Campylobacter methodology concerns expressed by FSIS.

FSIS Activity: Prior to 2004, FSIS used a labor intensive and resource consuming MPN method for the detection and enumeration of Campylobacter spp. (26). Current literature indicates that when Campylobacter spp. are present, numbers per ml of carcass rinse can vary from 1 to 3 log CFU (19) and thus direct enumeration on agar plates can be an alternative to MPN methods. The Food and Drug Administration (29), the Centers for

Disease Control and Prevention (5), and industry constituents currently use methods incorporating selective plating to detect and/or enumerate Campylobacter spp. from clinical and food samples. Since 2004, FSIS, in cooperation with ARS researcher, Dr. J.

Stanley Bailey, is conducting a special project that includes Campylobacter spp. detection and enumeration. The current ARS method (Appendix I) calls for collection of aseptic whole bird rinses with 100 ml of Buffered Peptone Water (BPW) followed by storage and overnight transport of rinses at 4° ± 4° C to the laboratory for analysis. The rinses are serially diluted and plated onto Campy-Cefex agar (with Bolton broth enrichment and selective agar plate streaks as a backup) for presence/absence

9 determination using a customized atmosphere of 5% O2, 10% CO2, and 85% N2 in sealed bags. Presumptive colonies are examined microscopically and confirmed using a serological latex agglutination test confirmatory for C. jejuni, C. coli, and C. lari.

Confirmed cultures are then stored at -80°C in Brucella broth with 15% glycerol for possible subsequent subtyping.

Present Charge to the Committee: In the near future, FSIS will conduct a baseline study to determine the prevalence of and to enumerate Campylobacter spp. of known importance on poultry (possibly including carcasses, parts, and ground product).

It is currently proposed that the study will focus on thermotolerant species, C. jejuni and C. coli, because these human pathogens account for the vast majority of laboratory-confirmed Campylobacter infections. An additional justification for this focus is supported by the numerous microbiological studies of poultry products that indicate that these two species are the only species of Campylobacter routinely isolated from chickens. Although some of the other 16 named Campylobacter species are reported to rarely cause human illness, the burden of human illness is low and poultry have not been shown to be a reservoir. Many of these other Campylobacter species require specialized growth conditions, such as atmospheres containing 5% H2 (i.e., non-thermotolerant

campylobacters) or growth media other than Campy-Cefex agar, which contains the

antimicrobial cefoperazone that inhibits growth of C. upsaliensis.

FSIS seeks advice on the proposed Campylobacter methodology, as well as any other relevant methodology that may be of equal or greater value which should be considered for the upcoming baseline study.

10 The questions to be addressed are:

1. What additional circumstances should be considered in order for FSIS

to conclude that the poultry baseline study should address more than

the two principal Campylobacter species of C. jejuni and C. coli?

2. How can the ARS method be most successfully used for high volume

analysis in the conduct of a baseline study of Campylobacter presence

and enumeration on poultry (chicken, turkey, goose, etc.) carcasses,

parts, and ground product that may lead to a potential performance

standard or guideline for the regulated industry? What if any

modifications should be made as a result of discussing this method in

comparison with others presented to the Committee? Please consider

whether the above described atmospheric conditions, media, pre-

enrichment, and storage media are acceptable for the objective of this

baseline study.

3. To utilize FSIS resources efficiently and effectively, FSIS expects to

maintain as much continuity as possible between the current broiler

rinse sampling for Salmonella and the proposed sampling for

Campylobacter spp. What concerns regarding the Campylobacter spp.

sampling method need to be attended to in order to properly address

post-chill injured Campylobacter spp. cells as well as viable non•

culturable (coccoid) cells?

4. What further subtyping methods should be performed on confirmed

cultures (restriction fragment length polymorphism (RFLP), amplified

11 fragment-length polymorphism (AFLP), pulsed-field gel

electrophoresis (PFGE), ribosomal DNA sequencing, antibiotic

susceptibility, etc.), and what, if any, limitations do any of these

methods pose?

5. What effect would in situ Campylobacter spp. cell aggregation have

on the accuracy and reproducibility of enumeration counts and is there

any remedy to address this issue?

6. Occasionally non-thermophilic Campylobacter species cause human

illness. It is unclear whether livestock and poultry are reservoirs for

these species, or if they are present on meat and poultry products

following slaughter and processing. Current methodologies use

selective agents and incubation conditions which may reduce their

detection. If a pilot study was conducted to ascertain the presence of

these species on meat and poultry products, what methodologies would

best detect these species?

IV. RESPONSE TO QUESTIONS

Question 1. What additional circumstances should be considered in order for FSIS to conclude that the poultry baseline study should address more than the two principal Campylobacter species of C. jejuni and C. coli?

The Committee stated that Campylobacter species that cause the majority of human illness from meat and poultry products should drive the testing for particular species in developing a baseline study. Those two target species are C. jejuni and C. coli

at present. However a certain percentage of samples should also be analyzed in a

12 separate surveillance research project to estimate the prevalence of other Campylobacter

spp. No etiological agent is ever attributed to the vast majority of foodborne illnesses;

this fact underscores the importance of such an additional study, as this would provide

valuable information.

To strengthen the case that FSIS should focus the baseline study on C. jejuni and

C. coli, an analogy was made using the verification testing of certain meat products for E. coli O157:H7. E. coli O157:H7 was first recognized in 1982 following two outbreaks of hemorrhagic colitis (20). In the years following, E. coli O157:H7 has been associated with numerous cases of hemorrhagic colitis and has rapidly become the most studied member of the enterohemorrhagic group (7). Over 25 non-O157 Shiga-like toxin- producing E. coli serotypes have been isolated but E. coli O157:H7 remains the most common enterohemorrhagic serotype in the United States (30). The epidemiological association of E. coli O157:H7 with ground beef along with its low infectious dose

necessitated that E. coli O157:H7 be the focus of FSIS’ intervention efforts to reduce

illness due to hemorrhagic colitis. At present, C. jejuni and C. coli are the leading causes

of human campylobacteriosis from poultry; therefore, the baseline should address these

two species.

Question 2. How can the ARS method [used presently in the ARS/FSIS Broiler

Rinse Study] be most successfully used for high volume analysis in the conduct of a

baseline study of Campylobacter presence and enumeration on poultry (chicken,

turkey, goose, etc.) carcasses, parts, and ground product that may lead to a potential

performance standard or guideline for the regulated industry? What if any

modifications should be made as a result of discussing this method in comparison

13 with others presented to the Committee? Please consider whether the above

described atmospheric conditions, media, pre-enrichment, and storage media are

acceptable for the objective of this baseline study.

The Committee chose to alter the question to reflect the specific ARS

Campylobacter enumeration method presently being used in the joint ARS/FSIS Broiler

Rinse Study (Appendix I.) since there are several methods being used by various ARS

researchers. In the above question, the Committee inserted the clarification within the

brackets.

In initial discussions regarding a baseline study for Campylobacter on poultry, the

Committee recommends that FSIS must clearly state the objectives and potential uses of the data. Specifically, the Committee suggests that FSIS consider whether the results of the baseline study will be used to examine multiple points along the poultry processing line and identify interventions that the industry could further develop into points in the process where interventions are needed, and or “best practices”; whether FSIS will look at overall numbers of Campylobacter spp. on products in the inspected plants to ascertain the success of intervention strategies; and whether FSIS will use the data in a future risk assessment.

The Committee also suggests that FSIS consider testing for generic E. coli,

Salmonella, and Campylobacter from the same carcass rinse to obtain information in

relation to the utility of an indicator organism for the poultry industry. This would

require that a standardized protocol with a neutralizing rinse broth be determined for

quantitative and qualitative analysis of selected microorganisms. To ensure that data

can be utilized for evaluating the suitability of indicator organisms, the sensitivities of the

14 methods for both indicators and target pathogens must be addressed. Methods for

indicator organisms and pathogens should have equal sensitivities.

The Committee stated that FSIS should choose a Campylobacter method for its

upcoming poultry baseline study that can be validated and easily used with a high sample

throughput. The method chosen should be widely available to industry constituents for

comparison sample analysis. The Committee recommends that for enumeration of

campylobacters a direct plating method would presently best fit expected criteria for a

baseline study over the previous labor intensive MPN method. The Committee

acknowledges that a traditional method validation usually entails comparative evaluation

against a “gold standard” method. In this case, it appears there is no “gold standard”

Campylobacter enumeration method and there would be little value in comparing the new method against the FSIS MPN method. However, if a direct plating method is used, well trained technicians, proficient in colony identification are needed since identification of

Campylobacter spp. can be difficult; they are non-fermenters and produce translucent colonies on Campy-Cefex agar. It was pointed out by the Committee that properly

trained technicians would be essential no matter what direct plating method was chosen.

The Committee also recommends that FSIS consult with other entities, such as other

national governments, other US Federal agencies, and other private and state research

institutions to correlate Campylobacter methodologies when possible. For example, the

European Union is currently designing a monitoring scheme for Campylobacter in

broilers (4), and the Nordic Committee on Food Analysis has recently developed methodological standards for the detection and enumeration of thermotolerant

Campylobacter in foods (18).

15 The Committee discussed the current ARS/FSIS Broiler Rinse Study

Campylobacter methodology at length with Dr. J. Stanley Bailey, the principal ARS researcher whose laboratory is performing the analyses. Direct plating was discussed as a method of choice. The Committee saw value in the 1-ml inoculation over four agar plates to achieve plating of a 100 dilution.

In further discussions with Dr. Bailey, the Committee determined that a back-up enrichment is not recommended. This was based on this principal researcher’s description of preliminary data, using 100-ml carcass rinses, which indicates that back-up enrichments would only provide a 1-2% increase in the number of positive samples, and would also present additional challenges to assigning values to samples that were negative by direct plating (i.e., not detected) but were positive by back-up enrichment. It is noted that previous FSIS work indicated that a back-up enrichment in conjunction with

a 400 ml rinse and MPN enumeration procedure increased positives by approximately

17% (27). The need for back-up enrichment to supplement direct plating for

Campylobacter spp. was further analyzed by the FSIS, Office of Public Health Science,

Risk Assessment Division staff members. This internal work acknowledged the above

NACMCF recommendation to not perform back-up enrichment with direct plating of

Campylobacter spp., and agreed with it, based on their FSIS evaluation (6).

In addition to the Campy-Cefex media used in the current ARS/FSIS Broiler

Rinse Study, the Committee discussed at length the advantages and disadvantages of

various direct plating agars available for Campylobacter with the subject matter experts,

and their review of literature, with particular attention given to method comparison studies from Line et. al. (8); Oyarzabal et al. (19); and Siragusa et. al. (21). The

16 Committee discussed comparisons between Campy-Line and Campy-Cefex agars

concluding that even though colonies on Campy-Line agar are easier to distinguish, it has

additional selective agents which could reduce positive samples by up to 20% over

Campy-Cefex agar. The Committee also pointed out that other organisms can produce

colonies on Campy-Line agar with the same morphology as Campylobacter spp. The

Committee also discussed various other media such as Modified Campylobacter Charcoal

Differential Agar and a commercial Simplate ® method for enumeration. As a result of extensive discussion, comparing media preparation, costs of media, and comparable recoveries on available solid plating media (including a modification to the Campy-

Cefex (m-Campy-Cefex) media using lysed horse blood in the place of laked horse blood, and a substituted antifungal (19), the Committee recommends that either Campy-Cefex agar or m-Campy-Cefex would be a sensitive, cost effective choice.

Incubation temperatures were also discussed and 42 ± 1°C for 48 h as used in the current ARS/FSIS Broiler Rinse Study was recommended. The optimal growth temperature for C. jejuni ranges between 42 and 45°C. Under appropriate atmospheric and nutritional conditions, C. jejuni will grow at temperatures above 30 and at or below

45°C (22). A two-stage 37/42°C incubation was discussed, but these methods were deemed cumbersome and were originally used with broth media. The Committee charged itself with following up on whether there were any documented studies regarding staging incubation temperatures with solid media, and will make the updated information known to FSIS.

NACMCF has, in the past, addressed the parameters important in designing baseline studies and the Committee recommends that FSIS consult the NACMCF reports

17 entitled: “Response to the Questions Posed by FSIS Regarding Performance Standards

with Particular Reference to Broilers (Young Chickens)” (15), “Response to the

Questions Posed by FSIS Regarding Performance Standards with Particular Reference to

Raw Ground Chicken” (16), and “Response to the Questions Posed by FSIS Regarding

Performance Standards with Particular Reference to Raw Ground Turkey” (17).

NACMCF is aware that FSIS has received funding for ongoing baseline studies

and that FSIS intends to begin a broiler baseline study in 2006. In any scientific study,

the sampling and data collection methods employed, as well as the study design

parameters, are critical in assessing the validity, interpretability and generalizability of

the results. Therefore, in addition to addressing study parameters, it is important that

NACMCF address statistical and data collection issues that should be considered in

designing any future baseline studies. NACMCF recommends that the agency come back

with a charge to the Committee to review the statistical aspects as well as the data

collection methodologies of any future baseline study designs.

Question 3. To utilize FSIS resources efficiently and effectively, FSIS expects to

maintain as much continuity as possible between the current broiler rinse sampling

for Salmonella and the proposed sampling for Campylobacter spp. What concerns

regarding the Campylobacter spp. sampling method need to be attended to in order to properly address post-chill injured Campylobacter spp. cells as well as viable non• culturable (coccoid) cells?

As discussed previously, sampling and data collection methods are critical in

assessing the validity, interpretability and generalizability of the study results. Therefore,

in determining the sampling and data collection methods used in the baseline studies,

18 several statistical considerations should be addressed. Foremost, the study objective(s) should be clearly stated, the population of interest should be identified and the sampling unit selected should be representative of that population. Sampling methods should also consider other potential factors such as seasonal and regional differences as well as inter- flock and inter-plant correlation. In addition, there should be some statistical justification to the sample size selected for the study. The Committee recommends that FSIS consider the statistical power in selecting the number of plants, number of carcasses and frequency of sampling for the baseline study, and FSIS should create a power calculation matrix to determine the optimal sample size. Further, insofar as possible, samples should be randomly selected and the sampling and data collection methods should be consistent throughout the study. Specifically, FSIS should define how carcasses will be randomly chosen at establishments and at what point(s) in the process they will be selected for rinsing. Sample handling factors such as rinse methods (i.e. type of neutralizing diluent, rinse solution), temperature conditions during shipment, and microbiological testing procedures, should be specified and uniform throughout the study. To assure consistency in sample as well as data collection, it is recommended that a sample and data collection protocol is developed and those involved in carrying out the protocol are trained with a common format.

The choice of validated neutralizing diluent for carcass rinsing and rinse volume

(400 ml versus 100 ml) is important when designing a baseline study for Campylobacter spp. The desirable features of a rinse diluent include: gives maximum buffering capacity, aids in injured cell recovery but does not promote cell growth during refrigerated transportation and/or result in false negatives due to improper neutralization and

19 sampling. When carcasses are chemically treated as an intervention, there is a need to document this information on the sampling form using standardized language.

Information related to such chemical treatments must be collected to ensure sample integrity and would not be used to measure the effect of the treatments; although, the information may be used for generating hypotheses or informing the design of future studies specifically addressing interventions. If chemical treatments are used, proper neutralization procedures need to be performed with sampling. Proper training and supplies are essential for sample collectors. Post-chill antimicrobial carcass dipping is a practice currently being utilized in industry. Therefore proper carcass draining practices in addition to using non-antimicrobial neutralizing additives, tailored to each chemical treatment, should be developed to maximize Campylobacter spp. recovery, as well for generic E. coli and Salmonella being tested for under the current regulations.

Presently, FSIS uses BPW for Salmonella rinse sample collection and has used it for Campylobacter sample collection (25), even though it is considered a pre-enrichment broth for Salmonella (2). FSIS could also use Butterfield’s phosphate diluent, which is not considered to be a pre-enrichment broth, for collecting carcass rinse samples. The

Committee recommends using the smallest rinse volume needed to cover all surfaces of the broiler and to perform microbial analysis of Campylobacter and other organisms.

Researchers conducting the present ARS/FSIS Broiler Rinse Study determined a 100 ml volume of BPW was sufficient, and NACMCF recommends this volume of rinse be validated. Note, it was pointed out by Dr. Bailey during discussions with the NACMCF subcommittee, that based on preliminary results from the ARS/FSIS Broiler Rinse Study, the higher volume of rinse used in the FSIS HACCP verification program (FSIS uses 400

20 ml BPW, ARS method calls for a 100 ml) may contribute to a lower observed

Campylobacter spp. count for broiler rinses, as compared to what is being observed in

the ARS project.

FSIS should determine the specific volume and type of rinse to be used, taking

into account any additional microbiological assays being performed as part of the

baseline, and provide scientific justification for the volume chosen. Referencing

statistically valid studies/documents comparing different rinse volumes should be included. Rinse solutions should be at 4°C before rinsing, and rinsate should be immediately placed on ice.

In addition, sample shipment temperature conditions were discussed. FSIS baseline studies and the current ARS/FSIS Broiler Rinse Study require a temperature of between 0°-10°C for samples on arrival at the FSIS laboratories. The Committee recommends overnight shipping, and suggests a study be performed to determine the number of ice packs and/or volume of ice needed to maintain temperature, given

anticipated ambient temperature extremes.

The Committee discussed microaerobic conditions needed for incubation. For a

large volume of samples, as would be generated in a large long-term study such as a

baseline, a tri-gas incubator is recommended. However, guidance should be issued on

alternative ways to achieve microaerobic conditions if such an incubator is unavailable.

Specific details of any gas-filled bag protocols, such as whether bags are to be heat

sealed, should be provided, and FSIS should validate the specific methodology for using

gas-filled bags.

21 The Committee recommends that FSIS speciate especially between C. jejuni and

C. coli. Methods such as latex agglutination and multiplex PCR can be used. In collaborating with research partners, FSIS should explore developing and validating molecular technologies such as microarray for speciation and subtyping of

Campylobacter.

The Committee recognizes that the advantages of phase contrast microscopic examination of a wet-mount for characteristic morphology/motility, are that the test is quick, and that it provides instant feedback. However, a wet-mount exam is not a confirmatory test. FSIS should address the training of laboratory technicians to achieve a high level of proficiency in identifying presumptive Campylobacter colonies. A minimum of five colonies, up to a total of 10% of the typical colonies on a countable (or lowest dilution) plate, representing each colony morphology, should be picked for semi- confirmatory testing by cellular morphology and motility on a wet-mount using phase contrast microscopy. Each isolate demonstrating typical Campylobacter morphology and motility will be further confirmed using latex agglutination, and then also speciated.

The Committee recommends that FSIS use consistent microbiological methods and procedures, outlining defined parameters for drying agar plates, storage and shelf life of plates, and report enumeration data as CFU/ml rinse when whole carcass rinsates are tested. A subject matter expert noted that a number of researchers from industry and other laboratories have been trained by ARS laboratories in Athens, GA in Campylobacter methodology, indicating that this laboratory would be a good resource.

If FSIS determines that classes of poultry other than broilers will be assessed in the future (e.g., turkeys), FSIS should partner with appropriate researchers to develop

22 methodologies and conduct surveillance studies to sample these products possibly for

other Campylobacter species in addition to C. jejuni and C. coli. Turkeys, due to their size and weight, also pose unique sample collection challenges beyond a simple broiler rinse. FSIS should consult research studies such as those by McEvoy et al. (9) and

Bodnaruk et al. (1), along with research partners to optimize turkey sample collection

techniques. This is a topic that could possibly be brought before the Committee in the

future, should FSIS require more guidance.

For ground product, the Committee recommends the use of 25g of product in a

filtered stomacher bag with diluent of choice with stomaching for 1 min followed by

serial dilution and plating.

The possible importance of viable nonculturable Campylobacter strains is not

known. This topic could be brought before the Committee again by FSIS when more

information becomes available.

Question 4. What further subtyping methods should be performed on confirmed

cultures (restriction fragment length polymorphism (RFLP), amplified fragment-

length polymorphism (AFLP), pulsed-field gel electrophoresis (PFGE), ribosomal

DNA sequencing, antibiotic susceptibility, etc.), and what, if any, limitations do any of these methods pose?

The Committee discussed a number of subtyping methods. To maximize

correlation among government entities, the utility of PFGE was recognized since the

method is used by the Centers for Disease Control and Prevention PulseNet (28) to track

human illness isolates and by ARS VetNet (24) to track animal diagnostic isolates. The

Committee recognizes that PFGE is more readily available than some of the other

23 methods discussed, such as multilocus sequence typing (MLST). The Committee, in

consultation with subject matter experts, discussed the MLST method. Data from the

ARS Campylobacter research laboratory in Albany, CA using MLST have shown that C.

jejuni and C. coli exchange genetic material (swap genes) making speciation difficult

(11). The Committee pointed out that in certain circumstances where PFGE has not

provided useful information, MLST has been used successfully.

In addition, the Committee recommends that FSIS explore the feasibility and

value of serotyping Campylobacter as well as investigate the feasibility of flaA sequence comparisons in subtyping Campylobacter which has been used at ARS in Athens, GA

(10).

Because antibiotic resistance among Campylobacter spp. is a public health problem, and there are interagency-established protocols for resistance testing, the

Committee recommends that a defined subset of isolates be tested for antibiotic resistance. The results can be used in analyses to help develop hypotheses about how resistant Campylobacter spp. enter a facility and move through production lines, and whether some resistant strains are maintained in facilities.

Finally, while a number of subtyping methods have been performed on

Campylobacter spp. (e.g., serotyping, antibiotic resistance, MLST, PFGE, flaA

sequencing, etc.), none have yet been sufficiently discriminatory to be generally

applicable as a gold standard. A combination of two or more subtyping methods can

often increase discriminatory power. However, continued subtyping studies are essential, since with refinement these methods have been of crucial importance in tracking other pathogens to their source. Therefore, the Committee recommends that research on these

24 methods be continued because of their value in gaining epidemiologically significant

information. As part of ongoing sampling, isolates should be preserved in storage for

further molecular characterization, but such characterization should not be part of an

initial baseline study.

Question 5. What effect would in situ Campylobacter spp. cell aggregation have on

the accuracy and reproducibility of enumeration counts and is there any remedy to

address this issue?

The Committee acknowledges that Campylobacter spp. cell aggregation is a real phenomenon, but whether it causes significant differences in counts has not been shown.

Further research is necessary in this area.

Question 6. Occasionally non-thermophilic Campylobacter species cause human

illness. It is unclear whether livestock and poultry are reservoirs for these species,

or if they are present on meat and poultry products following slaughter and

processing. Current methodologies use selective agents and incubation conditions

which may reduce their detection. If a pilot study was conducted to ascertain the

presence of these species on meat and poultry products, what methodologies would

best detect these species?

The Committee recommends that FSIS partner with appropriate researchers to

conduct surveillance studies to sample poultry products for analysis of Campylobacter

species other than C. jejuni and C. coli. FSIS should examine the findings of the

European Campycheck research initiative (3) and consult with other research partners in development of protocols to analyze for other Campylobacter species as part of a surveillance study. The surveillance data could inform FSIS regarding whether, in the

25 future, to expand species testing if other Campylobacter species become significant with regard to human illness from FSIS-regulated products. The Committee suggests that

FSIS may be able to benefit from the geographical proximity of the FSIS Western Field

Service Laboratory and the ARS Campylobacter research laboratory in Albany, CA, in that these two facilities can split and share collected rinse samples to maximize the testing performed on these rinses.

V. CONCLUSIONS

Target organisms selected for a microbiological baseline study should be based on species causing the majority of human illness. In designing the FSIS Campylobacter spp. from broiler carcass baseline study, presently the two target species are C. jejuni and C. coli. FSIS should however partner with appropriate researchers to develop methodologies and conduct surveillance studies to sample poultry products for other

Campylobacter species. Surveillance data could then be used to direct expansion of

Campylobacter testing in the future if necessary.

In designing the upcoming FSIS baseline for Campylobacter spp. enumeration of broilers, and any future baseline studies, FSIS must clearly state the objectives and potential uses of the data. FSIS can achieve maximum data utilization if the objectives are set before sample collection begins. Possible objectives may be to sample at multiple points along the poultry processing line and identify interventions that the industry can use as “best practices” and/or determine the overall numbers of Campylobacter spp. leaving establishments to ascertain if regulatory policies are successful and/or develop data to be used in future risk assessments. Consideration should be given to the need to also collect generic E. coli and Salmonella data simultaneously from whole bird carcass

26 rinses. To ensure the validity, interpretability and generalizability of the study results,

sampling and data collection methods should be evaluated, and a document that details

the study protocol should be developed and made available.

The direct plating Campylobacter spp. enumeration methodology currently being performed in the ARS/FSIS Broiler Rinse Study should be the basis for developing an upcoming baseline study with modifications as indicated throughout this report. See

Figure 1. for the NACMCF suggested Campylobacter analysis protocol for FSIS baseline studies for poultry. This method would be widely available to industry constituents and easily used with high numbers of samples that is impractical with MPN methods. It would be paramount to develop and adequately train individuals conducting sample collection and microbiological analysis to maximize data accuracy thus allowing the creation of a data set that could be used to develop FSIS risk management policy with regard to Campylobacter spp. contamination on poultry products.

VI. ACKNOWLEDGEMENTS

The NACMCF acknowledges and thanks the listed Campylobacter research experts for sharing their expertise with the NACMCF subcommittee in support of this report.: Dr. J. Stanley Bailey, ARS, Athens, GA, ARS/FSIS Broiler Study , Dr. Eric Line,

ARS, Athens, GA, Dr. Robert Mandrell, ARS, Albany, CA, and Dr. Mark Berrang, ARS,

Athens, GA.

The NACMCF acknowledges the hard work and efforts and thanks the following

individuals in the creation of this document: William Shaw, Celine Nadon, and Evelyne

Mbandi.

27 VII. REFERENCES

1. Bodnaruk, P. W., J. L. Schmelder, P. J. Krakar, and R. B. Tompkin. 1998. A

comparison of Escherichia coli levels at four sampling sites on turkey

carcasses. Poul. Sci. 77:1531-1533.

2. Difco Manual. 1998. Culture media and ingredients, dehydrated, p. 99-100.

Difco Laboratories, Division of Becton Dickinson and Company, Sparks, MD.

3. European Commission. CAMPYCHECK Research Project. Available at:

http://www.campycheck.org/. Accessed 10 April 2006.

4. European Union. 2005. Draft technical specifications for an EU monitoring

scheme for Campylobacter in broiler chickens. Unpublished report.

5. Fitzgerald, C. 14 July 2005. Personal communication [E-mail:

[email protected]].

6. Golden, N. 13 September, 2005. Personal communication [E-mail:

[email protected]].

7. Griffin, P. M. and R. V. Tauxe. 1991. The epidemiology of infections caused

by Escherichia coli O157:H7, other enterohemorrhagic E. coli, and the

associated hemolytic uremic syndrome. Epidemiol. Rev. 13:60-98.

8. Line, J., N. Stern, C. Lattuada, and S. Benson. 2001. Comparison of methods

for recovery and enumeration of Campylobacter from freshly processed

broilers. J. Food Prot. 64:982-986.

9. McEvoy, J. M., C. W. Nde, J. S. Sherwood, and C. M. Logue. 2005. An

evaluation of sampling methods for the detection of Escherichia coli and

Salmonella on turkey carcasses. J. Food Prot. 68:34-39.

28 10. Meinersmann, R.J., Phillips, R.W., Hiett, K.L., and Cray, P.J. 2005.

Differentiation of Campylobacter populations demonstrated by flagellin short

variable region sequences. Appl. Environ. Microbiol. 71:6368-6374.

11. Miller, W.G., On, S., Wang, G., Fontanoz, S., Lastovica, and A., Mandrell,

R.E. 2005. Extended multilocus sequence typing system for Campylobacter

coli, C. lari, C. upsaliensis, and C. helveticus. J. Clin. Microbiol.43:2315-

2329.

12. National Advisory Committee on Meat and Poultry Inspection. 2000.

Briefing on Agency’s Efforts with Respect to Campylobacter: National

Advisory Committee on Meat and Poultry Inspection. Available at:

http://www.fsis.usda.gov/OPPDE/NACMPI/May2000/nacmpicampy.html.

Accessed October 12, 2005.

13. National Advisory Committee on Microbiological Criteria for Foods. 1994.

Campylobacter jejuni/coli. J. Food Prot. 57:1101-1121.

14. National Advisory Committee on Microbiological Criteria for Foods. 2002.

Plenary Session (Transcript). Available at:

http://www.fsis.usda.gov/ophs/nacmcf/2002/transcript082802.pdf. Accessed

October 12, 2005.

15. National Advisory Committee on Microbiological Criteria for Foods 2004.

Response to the Questions Posed by FSIS Regarding Performance Standards

with Particular Reference to Broilers (Young Chickens). Available at:

http://www.fsis.usda.gov/OPHS/NACMCF/2004/NACMCF_broiler_4_13_04

.pdf. Accessed October 12, 2005.

29 16. National Advisory Committee on Microbiological Criteria for Foods. 2004.

Response to the Questions Posed by FSIS Regarding Performance Standards

with Particular Reference to Raw Ground Chicken. Available at:

http://www.fsis.usda.gov/ophs/nacmcf/2004/NACMCF_Ground_Chicken_08

2704.pdf. Accessed October 14, 2005.

17. National Advisory Committee on Microbiological Criteria for Foods 2004.

Response to the Questions Posed by FSIS Regarding Performance Standards

with Particular Reference to Raw Ground Turkey. Available at:

http://www.fsis.usda.gov/ophs/nacmcf/2004/NACMCF_Ground_Turkey_082

704.pdf. Accessed October 14, 2005.

18. Nordic Committee on Food Analysis. April 2005. Thermotolerant

Campylobacter. Detection and enumeration in foods.

19. Oyarzabal, O., K. Macklin, J. Barbaree, and R. Miller. 2005. Evaluation of

agar plates for direct enumeration of Campylobacter spp. from poultry carcass

rinses. Appl. Environ. Microbiol. 71:3351-3354.

20. Riley L.W., R.S. Remis, S.D. Helgerson, H.B. McGee, J.G. Wells, B.R.

Davis, R.J. Hebert, E.S. Olcott, L.M. Johnson, N.T. Hargrett, P.A. Blake, and

M.L. Cohen. 1983. Hemorrhagic colitis associated with a rare Escherichia

coli serotype. N. Engl. J. Med. 308:681-685.

21. Siragusa, G., J. Line, L. Brooks, T. Hutchinson, J. Laster, and R. Apple. 2004.

Serological methods and selective agars to enumerate Campylobacter from

broiler carcasses: data from inter- and intralaboratory analyses. J. Food Prot.

67:901-907.

30 22. Stern, N. J., and S. U. Kazmi. 1998. Campylobacter jejuni. p 77. In M. P.

Doyle, (ed), Foodborne Bacterial Pathogens. Marcel Dekker, Inc. New York.

23. Tauxe, Robert V. 2005. Human Campylobacteriosis Risk Factors. At South

East Poultry Association Annual Meeting. Atlanta, GA February 23, 2005.

24. U.S. Department of Agriculture, Agricultural Research Service. 2005.

Antimicrobial Resistance in Pathogenic and Commensal Bacteria of Food

Animals: Narms, Vetnet and Cahfse. Available at:

http://www.ars.usda.gov/research/publications/publications.htm?SEQ_NO_11

5=191019&pf=1. Accessed 19 April 2006.

25. U.S. Department of Agriculture, Food Safety and Inspection Service. 1994•

1995. Nationwide Broiler Chicken Microbiological Baseline Data Collection

Program (July 1994-June 1995. Available at:

www.fsis.usda.gov/OPHS/baseline/broiler2.pdf. Accessed at 10 April 2006.

26. U.S. Department of Agriculture, Food Safety and Inspection Service. 1998.

Isolation, Identification, and Enumeration of Campylobacter jejuni/coli from

Meat and Poultry Products. Available at:

http://www.fsis.usda.gov/ophs/Microlab/Mlgchp6.pdf.

Accessed 10 April 2006.

27. U.S. Department of Agriculture, Food Safety and Inspection Service. 1998.

Preliminary summary report for enumerating Campylobacter spp., E. coli, and

Salmonella spp. in broiler carcass rinses before and after simulated transport

in artificial ice for 24 h. Unpublished report.

31 28. U. S. Department of Health and Human Services. Centers for Disease Control

and Prevention. PulseNet. Available at: http://www.cdc.gov/pulsenet/.

Accessed 10 April 2006.

29. U.S. Food and Drug Administration, Center for Food Safety and Applied

Nutrition. 2001. Bacteriological Analytical Manual Online: Campylobacter.

Available at: http://www.cfsan.fda.gov/~ebam/bam-7.html. Accessed 10

April 2006.

30. Woody, J., J. Stevenson, R. Wilson, and S. Knabel. 1998. Comparison of the

Difco EZ Coli™ Rapid Detection System and Petrifilm™ Test Kit-HEC for

detection of Escherichia coli O157:H7 in fresh and frozen ground beef. J.

Food Prot. 61:110-112.

32 APPENDIX I. ARS/FSIS BROILER RINSE STUDY CAMPYLOBACTER

SPP. ENUMERATION METHOD

Aseptically collect whole bird rinses with 100 ml of BPW, and then ship overnight at 4° ± 4° C to the laboratory for analysis. Direct plate carcass rinsates to enumerate Campylobacter spp. To obtain CFU/ml of rinse from carcasses (where low numbers (countable plates) are expected) apply 1 ml to a total of four Campy-Cefex agar plates (Hardy Diagnostics, Santa Maria, CA), 0.25 ml per plate. Spread each aliquot on the agar surface with a sterile plastic hockey stick; allow plates to remain upright until dried (approximately 15 min). Incubate Campy-Cefex agar plates at 42°C for 48 h in sealable bags flushed to produce a microaerobic gas environment (5% O2, 10% CO2, 85%

0 N2). As a group of four, designate these plates as the 10 dilution per ml. (Note: Because these plates receive a 0.25-ml volume, the agar needs to be pre-dried to facilitate soaking up the inoculum. Achieve dried agar plates by leaving them on a laboratory bench at ambient temperature and humidity (protected from light) for 24 h prior to plating.)

To allow enumeration of higher numbers of Campylobacter spp. per ml, prepare

10-fold serial dilutions of carcass rinsates in phosphate buffered saline. Spread the

undilute rinsates (0.1 ml) and aliquots from serial dilutions (0.1 ml) on the surface of

duplicate Campy-Cefex agar plates with a sterile plastic hockey stick. Incubate Campy-

Cefex agar plates at 42°C for 48 h in sealable bags flushed to produce a microaerobic gas

environment (5% O2, 10% CO2, 85% N2).

Following incubation, count colonies characteristic of Campylobacter spp.

Calculate the number of Campylobacter spp. CFU/ml rinsate using either a total number

33 from all four 100 dilution plates, or an average of the duplicate counts at higher dilutions, as appropriate. Characterize each colony type counted as Campylobacter spp. from each sample as a member of the genus Campylobacter by examination of cellular morphology and motility on a wet-mount under phase contrast microscopy. Further confirm each colony type as species jejuni, coli, lari by a latex agglutination test kit (Med-Ox

Diagnostics, Inc.; Ogdensburg, NY).

34 Figure 1. NACMCF suggested Campylobacter analysis protocol for FSIS baseline studies for poultry

Ship overnight maintaining 4° ± 4°C

Choose Campy-Cefex or m-Campy Cefex agar Ground Turkeys Broilers Product

Plating: Rinsates: Direct plate – 1 ml r insate over 4 Select buffer: plates for 100 (0.25 ml per plate) FSIS needs to Buffered peptone For higher numbers p late 0.1 ml develop custom water or 25 g sample of the undiluted rinsate and 0.1 sampling and Butterfield’s ml of the 10-fold serial dilutions in analysis phosphate d iluent duplicate

Ground Product: Stomach 1 min in filter bag in Select appropriate a diluent of choice, serially dilute, rinse volume plate 400 ml or 100 ml @ 4°C

Incubate: Microaerobic (5% O2 , 10% CO2 , 85% N 2) Rinse carcass 42° C +/- 1°C , 48 hrs according to procedures

Pick a minimum of 5 colon ies with proper morpholog ies on countable plate Recover solution and immediately place on ice Prepare wet-mount and examine for characteristic morphology and motility

Confirm cultures of C. jejuni, C. coli, a An appropriate neutralizing diluent must be included if C. lari by latex agglutination the carcasses are chemically treated

Report as CFU/ml ri nsate or CFU/g ground product

C. jejuni, C. coli speciation by PCR, validated micr oarray, etc...

Preserve isolates for further molecular character ization – Perform additional subtyping studies NACMCF Recommendations on Campylobacter Test Procedure

Wafa Birbari Sara Lee Corporation AMI SAC Meeting Hollywood, FL Oct 3, 2006

1 Background • Campylobacter species are a leading cause of foodborne illness in the US • Majority of campylobacteriosis cases are caused by C. jejuni followed by C. coli and other species • NACMCF reviewed available literature for methodologies and examined current method being used in the FSIS/ARS Broiler Rinse Study for possible use in the FSIS Baseline Study on broilers 2 NACMCF Report “Analytical Utility of Campylobacter Methodologies

• FSIS microbiological baseline study for broiler carcasses should be based on the species of Campylobacter causing the majority of human illnesses, C. jejuni & C. coli • FSIS should partner with researchers to develop methodologies and conduct surveillance studies to sample poultry products for other Campylobacter since their prevalence is unknown • FSIS must clearly state the objectives and potential uses of the baseline data

• FSIS should determine if analysis from a single carcass rinse (generic E. coli, Salmonella, Campylobacter) would strengthen the evaluation control for FSIS and industry • The direct Plating currently used in the ARD/FSIS Broiler Rinse Study should be adapted for the upcoming FSIS Baseline study, with modifications

• FSIS should speciate Campylobacter spp. to differentiate C. jejuni & C. coli 3 FSIS/ARS Campylobacter Method • Rinsate Ø 100 ml carcass rinse of BPW (better recovery than 400 ml) Ø Direct plate 1 ml onto 4 plates of Campy-CEFEX agar (less selective than Campy-Line but allows 20% more growth of stressed cells) Ø For higher numbers plate 0.1 ml of undiluted rinsate and 0.1 ml of 10-fold serial dilutions, in duplicate Ø Trained technicians proficient in colony identification • Ground Product Ø Stomach 1 min in diluent, serially dilute, and plate on

Campy-CEFEX 4 FSIS/ARS Campylobacter Method

• Incubate plates microaerobic in 5% O2, 10% CO2, 85% N2) @ 42°C, 48 h • Pick 5 colonies with proper morphogies on a countable plate • Examine wet-mount for characteristic morphology & motility • Confirm cultures of C. jejuni, C. coli, C. lari by latex agglutination • Report as CFU/ml rinsate or CFU/g

product 5 FSIS/ARS Campylobacter Method • Enrichment in Bolton Broth for a presence/ absence result as a backup if undetectable counts on agar plates • Was not recommended by NACMCF, however, FSIS added it to their Baseline Study • Isolates will be stored, subtyped (PFGE for correlation with human illnesses), and tested for antibiotic resistance (subset of isolates) 6 Other NACMCF Recommendations • Take samples at multiple points along poultry processing line, identify interventions • Sampling methods should consider seasonal & regional differences • Test for generic E. coli & Salmonella from the same rinsate in addition to Campy-might give information on the utility of an indicator organism • Sample handling factors such as rinse methods, temperature conditions during shipment & microbiological testing procedures should be

specified 7 Other NACMCF Recommendations • Choice of a validated neutralizing broth for carcass rinses that gives max. buffering, aids in injured cell recovery & does not promote cell growth during refrigerated transportation • Rinse solution should be at 4°C before rinsing and rinsate should be immediately placed on ice • Samples should arrive at the lab at 0-10°C • Use consistent micro methods & procedures outlining defined parameters such as drying agar plates, storage and shelf life of plates

8 Other NACMCF Recommendations

• Turkeys, due to their size and weight, pose a unique sample collection challenges • Ground product, use 25g in diluent of choice with stomaching for 1 min followed by serial dilution and plating

9 NAS Dioxin Review

IARC Update and Nitrite Alternative Research FightFight forfor thethe Cure:Cure: SodiumSodium NitriteNitrite StatusStatus ReportReport AgendaAgenda

The Controversy about Sodium Nitrite Andrew L. Milkowski, Ph.D., Adjunct Professor, University of Wisconsin

Update on International Agency for Research Monograph on Ingested Nitrite and Nitrate James R. Coughlin, Ph.D., President, Coughlin & Associates TThehe ControversyControversy aboutabout SodiumSodium NitriteNitrite

Andrew Milkowski, Ph.D

University of Wisconsin TopicsTopics

•• ChronologyChronology ofof IssuesIssues •• TheThe NationalNational ToxicologyToxicology ProgramProgram studystudy •• TheThe PropositionProposition 6565 reviewreview •• SomeSome factsfacts aboutabout nitritenitrite && relatedrelated compoundscompounds ChronologyChronology ofof SafetySafety IssuesIssues NitriteNitrite ControversyControversy AncientAncient HistoryHistory

1960’s - Nitrosamine formation potential discovered

1970’s - 1980 - Risk assessment - Curing research and product reformulation 1977 - USDA proposals to ban nitrite 1978 -1981 - Peak of activity Functions of nitrite reviewed Nitrite surveys Nitrosamine surveys 1981 - 1982 - NAS reports 1984 - USDA final rule published NitriteNitrite ControversyControversy ModernModern HistoryHistory

• 1994 Preston-Martin Publications – Epidemiological associations – Hot dog consumption related to childhood leukemia and brain cancers – Carried in popular media - – Category had 8% decline

• 1997 Georgetown University Conference - Nitrite as a Food Additive – Review of science and regulatory history – NTP testing for nitrite safety disclosed

• 1998 California Proposition 65 – Proposed Listing as Developmental and Reproductive Toxicant

• 2000 Resolution of NTP and Proposition 65 activities

• 2006 IARC Monograph – June working group session in Lyon, France NTPNTP -- TheThe NationalNational ToxicologyToxicology ProgramProgram NationalNational ToxicologyToxicology ProgramProgram (NTP)(NTP)

• NTP was established by the Secretary of Health and Human Services (HHS) in 1978 • The NTP consists of the relevant toxicological activities of:

– National Institute of Environmental Health Sciences (NIH/NIEHS) – National Institute for Occupation Safety and Health (CDC/NIOSH) – National Center for Toxicological Research (FDA/NCTR)

• NTP is obligated to provide Congress a Biennial Report on Carcinogens (mandated by Section 301 (b) (4) of the Public Health Services Act as amended) NTPNTP CancerCancer BioassaysBioassays

• The “gold standard” of cancer tests

• Standard protocols

• Multiple doses

• Extensive tissue pathology and statistical analysis

• Draft report prepared for peer review committee

• Results affect regulatory actions NTPNTP CancerCancer BioassaysBioassays

• Carcinogenicity Determination Categories

• no evidence ☺

• equivocal evidence ☺

• some evidence

• clear evidence NTPNTP TestingTesting ofof SodiumSodium NitriteNitrite • Nominated for testing by FDA

• Studies initiated in 1989

• Long-term Carcinogenicity – 2yr in rats and mice, dosed by drinking water – sodium nitrite doses: 750, 1500, or 3000 ppm • (35-150 mg/kg/day in rats) • (45-220 mg/kg/day in mice) – in-life portion completed 8/97 – extensive histopathology and statistical evaluation

• Draft Technical Report No. 495 was issued in April 2000 for peer review PeerPeer ReviewReview byby TechnicalTechnical ReportsReports ReviewReview SubcommitteeSubcommittee onon MayMay 18,18, 20002000

• Arguments for changes to conclusions in draft report presented by L. Borchert, D. Archer, G. Williams

• Unanimous decision: MALE RATS NO EVIDENCE ☺ FEMALE RATS NO EVIDENCE ☺ MALE MICE NO EVIDENCE ☺ FEMALE MICE EQUIVOCAL EVIDENCE ☺

• Strongest statistical finding was p < 0.001 decrease in “all organs’ mononuclear cell leukemia” in male and female rats. PropositionProposition 6565 -- thethe CaliforniaCalifornia environmentalenvironmental protectionprotection lawlaw CaliforniaCalifornia ““PropositionProposition 6565””

(Safe Drinking Water and Toxic Enforcement Act of 1986, passed by voter initiative 2:1) • Hazard Identification and Governor’s List of: – carcinogens – reproductive & developmental toxicants • Consumer product, occupational and environmental warnings [Right-to-Know Warnings] • Enforcement through government and private citizen [“bounty hunter”] suits PropositionProposition 6565 -- NitriteNitrite ChronologyChronology

• April 1998 -- “Request for Relevant Information” on 15 chemicals for possible listing, including sodium nitrite as a developmental and reproductive toxicant (DART)

• June 1998 -- AMI’s written objections submitted

• December 1998 -- “Notice of Intent to List Sodium Nitrite “

• February 1999 -- AMI led coalition submits additional objections

• May 1999 -- Agency passes listing decision to “DART” Committee for scientific review PropositionProposition 6565 -- NitriteNitrite ChronologyChronology

• July 1999 -- March 2000

– Draft Hazard Identification Document (HID) prepared by California agency

• March 3, 2000 -- Draft HID released for 60-day public comment

• May 2, 2000 -- AMI-led coalition submits detailed scientific comments on Draft HID

• June 2, 2000 -- AMI oral presentation to DART Committee DARTDART CommitteeCommittee ReviewReview -- JuneJune 2,2, 20002000

• 9 Member Board (8 present) • Presentation of HID by California Agency Staff • Presentation of arguments by AMI representatives – J. Murray, M. Corash • Open discussion and questions • Vote on recommendation to list CaliforniaCalifornia PropositionProposition 6565 NitriteNitrite ArgumentsArguments

–– IncompleteIncomplete literatureliterature reviewreview

–– NTPNTP reproductivereproductive teststests inin earlyearly 19901990’’ss showedshowed nono reproductivereproductive toxicitytoxicity –– MixedMixed prepre--natalnatal andand postpost natalnatal exposureexposure risksrisks

–– DidDid notnot separateseparate maternalmaternal fromfrom fetalfetal toxicitytoxicity

–– FailedFailed toto meetmeet legallegal standardstandard ofof ““clearlyclearly shownshown”” DARTDART CommitteeCommittee VotesVotes NotNot toto ListList SodiumSodium NitriteNitrite (June(June 2,2, 2000)2000)

• Male reproductive toxicity: 8 No, 0 Yes • Female reproductive toxicity: 8 No, 0 Yes • Developmental toxicity: 7 No, 1 Yes (epidemiologist) SomeSome factsfacts aboutabout thethe benefitsbenefits ofof nitritenitrite && relatedrelated compoundscompounds NitriteNitrite –– TheThe multimulti--functionalfunctional foodfood additiveadditive

•• DevelopmentDevelopment ofof curedcured meatmeat colorcolor •• ModifiesModifies freshfresh meatmeat flavorflavor •• FunctionsFunctions asas anan antioxidantantioxidant •• AntimicrobialAntimicrobial –– especiallyespecially anan antianti--ClostridialClostridial TheThe otherother significantsignificant antimicrobialantimicrobial effecteffect ofof nitritenitrite

(and(and itit’’ss important!)important!) NitriteNitrite isis naturallynaturally producedproduced inin salivasaliva andand thethe nitritenitrite youyou swallowswallow alsoalso protectsprotects againstagainst manymany pathogenspathogens

• Salivary glands extract nitrate from the blood plasma

• Salivary nitrate is chemically reduced by bacteria on the tongue to nitrite

• 92% of swallowed nitrite comes from saliva, <5% from cured meat

• Salivary nitrite is swallowed and enters the acidic stomach

• Acidified nitrite is bactericidal for a number of gastrointestinal pathogens

McKnight GM et al. 1999. Br. J. Nutr. 81:349-358 TheThe nitritenitrite youyou swallowswallow alsoalso protectsprotects againstagainst manymany pathogenspathogens

AcidifiedAcidified nitritenitrite -- sensitivesensitive gastrointestinalgastrointestinal pathogenspathogens

• Yersinia enterocolitica • Helicobacter pylori

• Salmonella enteritidis • Candida albicans

• S. typhimurium

• Shigella sonnei

• E. coli O157:H7 NitricNitric OxideOxide -- thethe bodybody’’ss internalinternal sourcesource ofof nitritenitrite andand nitratenitrate •• ““NormalNormal”” nitratenitrate produced=1mg/kgproduced=1mg/kg bwbw/day/day •• MetabolicMetabolic functionsfunctions firstfirst recognizedrecognized inin midmid 19801980’’ss •• SynthesizedSynthesized fromfrom argininearginine •• Neurotransmitter,Neurotransmitter, smoothsmooth musclemuscle relaxation,relaxation, involvedinvolved inin immuneimmune responseresponse andand woundwound healinghealing •• ScienceScience MagazineMagazine 19921992 ““moleculemolecule ofof thethe yearyear”” •• 19981998 NobelNobel PrizePrize forfor nitricnitric oxideoxide researchresearch •• DrugsDrugs havehave targetedtargeted nitricnitric oxideoxide releaserelease TheThe humanhuman nitrogennitrogen oxideoxide cyclecycle

Source: Gladwin, M. T., A. N. Schechter, D. B. Kim-Shapiro, R. P. Patel, N. Hogg, S. Shiva, R. O. Cannon 3rd, M. Kelm, D. A. Wink, M. G. Espey, E. H. Oldfield, R. M. Pluta, B. A. Freeman, J. R. Lancaster Jr, M. Feelisch, and J. O. Lundberg. 2005. The emerging biology of the nitrite anion. Nat. Chem. Biol. 1:308-314. UPDATEUPDATE ONON INTERNATIONALINTERNATIONAL AGENCYAGENCY FORFOR RESEARCHRESEARCH ONON CANCERCANCER (IARC)(IARC) MONOGRAPHMONOGRAPH ONON INGESTEDINGESTED NITRITENITRITE ANDAND NITRATENITRATE

James R. Coughlin, Ph.D.

Coughlin & Associates

Observer for the AMI Foundation at IARC Monograph Meeting WhatWhat isis IARC?IARC?

••InternationalInternational AgencyAgency forfor ResearchResearch onon CancerCancer

••HeadquarteredHeadquartered inin Lyon,Lyon, France;France; foundedfounded inin 19651965 asas partpart ofof thethe UNUN’’ss WorldWorld HealthHealth OrganizationOrganization

••ConsistsConsists ofof 1818 membermember countriescountries andand aa staffstaff ofof 150.150. WhatWhat isis IARC?IARC?

•• IARCIARC’’ss mission:mission:

•• ToTo coordinatecoordinate andand conductconduct humanhuman epidemiologicalepidemiological andand laboratorylaboratory researchresearch onon thethe causescauses ofof humanhuman cancercancer andand mechanismsmechanisms ofof carcinogenesiscarcinogenesis

•• ToTo developdevelop scientificscientific strategiesstrategies forfor cancercancer controlcontrol

•• ToTo disseminatedisseminate scientificscientific informationinformation throughthrough publications,publications, meetings,meetings, courses,courses, fellowshipsfellowships

http://www.iarc.fr/ IARCIARC’’ss CarcinogenCarcinogen IdentificationIdentification andand EvaluationEvaluation GroupGroup

•• ConductsConducts 33 MonographMonograph WorkingWorking GroupGroup meetingsmeetings perper year,year, wherewhere selectedselected groupsgroups ofof independentindependent scientificscientific expertsexperts ““classifyclassify”” chemicalchemical substancessubstances oror exposureexposure circumstances;circumstances; industrialindustrial organizationsorganizations andand NGOsNGOs cancan participateparticipate asas Observers.Observers.

•• Dr.Dr. VincentVincent CoglianoCogliano (ex(ex--EPA)EPA) headsheads thisthis group,group, andand hishis staffstaff scientistsscientists serveserve asas coordinatorscoordinators ofof subgroupssubgroups

•• WGWG membersmembers decidedecide allall texttext ofof eacheach IARCIARC MonographMonograph reportreport byby meetingmeeting’’ss end.end. ““IngestedIngested NitratesNitrates andand NitritesNitrites””

•• IARCIARC convenedconvened anan ExpertExpert WorkingWorking GroupGroup inin LyonLyon France,France, JuneJune 1414--2121

•• PurposePurpose waswas toto developdevelop aa monographmonograph toto bebe publishedpublished 1212--1515 monthsmonths fromfrom now,now, withwith backgroundbackground science,science, evaluatuonevaluatuon ofof thethe publishedpublished literatureliterature andand conclusionsconclusions onon thethe potentialpotential ofof thesethese compoundscompounds toto causecause cancer.cancer. IngestedIngested NitratesNitrates andand NitritesNitrites

•• IARCIARC classifiesclassifies compounds/exposurescompounds/exposures asas follows:follows:

Group 1: The agent is carcinogenic to humans

Group 2A: The agent is probably carcinogenic to humans

Group 2B: The agent is possibly carcinogenic to humans

Group 3: The agent is not classifiable as to its carcinogenicity to humans

Group 4: The agent is probably not carcinogenic to humans IngestedIngested NitratesNitrates andand NitritesNitrites

•• WorkingWorking GroupGroup (18(18 ScientistsScientists fromfrom manymany countries)countries) evaluatedevaluated thethe literatureliterature inin 44 separateseparate subgroups:subgroups:

•• ExposureExposure

•• CancerCancer inin HumansHumans

•• CancerCancer inin ExperimentalExperimental AnimalsAnimals

•• OtherOther RelevantRelevant DataData (and(and MechanismsMechanisms ofof Action)Action)

http://monographs.iarc.fr/ENG/Meetings/vol94-participants.pdf FinalFinal ClassificationClassification ofof NitriteNitrite andand NitrateNitrate

•• HumanHuman EvaluationEvaluation::

•• ThereThere isis limitedlimited evidenceevidence inin humanshumans forfor thethe carcinogenicitycarcinogenicity ofof nitritenitrite inin food.food. NitriteNitrite inin foodfood isis associatedassociated withwith anan increasedincreased incidenceincidence ofof stomachstomach cancer.cancer.

•• ThereThere isis inadequateinadequate evidenceevidence inin humanshumans forfor thethe carcinogenicitycarcinogenicity ofof nitratenitrate inin food.food.

•• ThereThere isis inadequateinadequate evidenceevidence inin humanshumans forfor thethe carcinogenicitycarcinogenicity ofof nitratenitrate inin drinkingdrinking water.water. ClassificationClassification ofof NitriteNitrite andand NitrateNitrate

•• AnimalAnimal EvaluationEvaluation::

•• ThereThere isis sufficientsufficient evidenceevidence inin experimentalexperimental animalsanimals forfor thethe carcinogenicitycarcinogenicity ofof nitritenitrite inin combinationcombination withwith aminesamines oror amides.amides.

•• ThereThere isis limitedlimited evidenceevidence inin experimentalexperimental animalsanimals forfor thethe carcinogenicitycarcinogenicity ofof nitritenitrite perper sese..

•• ThereThere isis inadequateinadequate evidenceevidence inin experimentalexperimental animalsanimals forfor thethe carcinogenicitycarcinogenicity ofof nitrate.nitrate. OverallOverall EvaluationEvaluation ofof NitriteNitrite andand NitrateNitrate

•• ““IngestedIngested nitratenitrate oror nitritenitrite underunder conditionsconditions thatthat resultresult inin endogenousendogenous nitrosationnitrosation isis probablyprobably carcinogeniccarcinogenic toto humanshumans (Group(Group 2A).2A).””

•• NoteNote:: IARCIARC’’ss decisiondecision rulesrules (stated(stated inin itsits Preamble)Preamble) definedefine GroupGroup 2A2A asas thethe DefaultDefault classificationclassification forfor limitedlimited evidenceevidence inin humanshumans combinedcombined withwith sufficientsufficient evidenceevidence inin experimentalexperimental animals;animals; oneone scientistscientist actuallyactually wantedwanted aa GroupGroup 11 classification.classification. WhatWhat doesdoes thisthis ClassificationClassification Mean?Mean?

• The ingestion of food and water that contain nitrate or nitrite (e.g., spinach, root vegetables, bread, beer, cured meats), in combination with naturally occurring amines and amides found in food, can react in the stomach to form nitrosamines and nitrosamides, known animal carcinogens. This was already very well known.

• Nitrosamines and nitrosamides are also well-known animal carcinogens, and most are already classified as carcinogens by IARC and other global regulatory and public health bodies.

• This Working Group added the conclusion that nitrite and nitrate combined with amines and amides probably contribute to an increase in human stomach cancer incidence. WhatWhat doesdoes thisthis ClassificationClassification Mean?Mean?

••IARCIARC’’ss conclusionconclusion standsstands inin sharpsharp contrastcontrast toto thethe conclusionsconclusions ofof numerousnumerous studiesstudies demonstratingdemonstrating nitritenitrite’’ss safetysafety overover thethe pastpast 3030 yearsyears

••WeWe inin thethe meatmeat industryindustry dodo notnot agreeagree withwith IARCIARC’’ss probablyprobably carcinogeniccarcinogenic classificationclassification ofof nitrite.nitrite. WhatWhat areare thethe ImmediateImmediate Consequences?Consequences?

• These conclusions were published on July 31, 2006 in a British journal, Lancet Oncology, in its “Policy Watch” column (2 pages).

• IARC did not choose to issue a simultaneous Press Release stating these conclusions.

• IARC uploaded its detailed Summary and Evaluation onto its website (~10 pages).

• Researchers and governments all over the world routinely look to IARC’s carcinogen classification decisions as THE source of information on carcinogens. WhatWhat areare thethe ImmediateImmediate Consequences?Consequences?

• Even though cured/processed meats were not specifically evaluated by the Working Group, they will probably be a major focus of IARC’s future public conclusions.

• Dr. Cogliano and his staff tried repeatedly to force the Epidemiology subgroup to evaluate the epidemiology studies on processed meats, but the members refused; they did look carefully at these studies, but did not include them in their final evaluation because nitrite was not quantified in these studies.

• The published epidemiology studies usually indict both red and processed meats as possible causes of stomach and colorectal cancer, but there are deficiencies in the design and conduct of these studies. WhatWhat MajorMajor ProblemsProblems DoDo WeWe Face?Face?

•• NearNear termterm -- globalglobal mediamedia attentionattention onon thethe cancercancer--causingcausing potentialpotential ofof nitritenitrite--processedprocessed meats,meats, followedfollowed byby consumerconsumer activistactivist (CSPI,(CSPI, others)others) pressurepressure toto limitlimit oror banban sodiumsodium nitritenitrite’’ss useuse asas aa foodfood additive.additive.

•• LongerLonger rangerange -- demonstratingdemonstrating andand continuingcontinuing thethe supportsupport forfor thethe safetysafety ofof nitritenitrite andand processedprocessed meatsmeats toto foodfood regulatoryregulatory agenciesagencies (FDA,(FDA, USDA,USDA, HealthHealth Canada,Canada, EUEU’’ss EFSA,EFSA, etc.)etc.) andand publicpublic healthhealth agenciesagencies (National(National CancerCancer Institute,Institute, AmericanAmerican CancerCancer Society,Society, Codex,Codex, etc.)etc.) aroundaround thethe world.world. WhatWhat MajorMajor ProblemsProblems DoDo WeWe Face?Face?

• "In 2007, WCRF/AICR will publish the second expert report containing a rigorous review and evaluation of an estimated 10,000 pieces of research on diet, physical activity and weight management and their effect on cancer risk.“ • The literature linking red meats to colon cancer and processed meats to stomach cancer will surely be highlighted.

• By late 2007, California ProposProposiitiontion 65 is likely to propose the listing of nitrite as a carcinogcarcinogen,en, andand ifif nitritenitrite isis listed,listed, ththis would require a battle over cancer warnings on nitrite- processed meats. UpdateUpdate onon thethe InternationalInternational AgencyAgency forfor ResearchResearch onon CancerCancer ReviewReview ofof NitrateNitrate andand NitriteNitrite

ThankThank youyou forfor youryour attention!attention!

Questions?Questions? 10/2/2006 1 of 15

White paper

The Latest Chapter in the Nitrite Ingredient Safety Saga

Prepared for the American Meat Institute

Andrew L. Milkowski University of Wisconsin August, 2006

10/2/2006 2 of 15

Introduction

Since the 1960’s when the potential to form nitrosamines from the reaction of nitrous acid with secondary amines was recognized, the use of nitrite salts as food preservatives has been under intense scrutiny. The concern for the use in processed meats has waxed and waned as numerous studies were published and independent review panels convened to make determinations about nitrite for use in public policy on food regulations.

In 2000 the National Toxicology Program completed its work on a two year lifetime feeding study of rats and mice with a peer review of the study and conclusions. The final report (40) indicated that at best there was only equivocal evidence that nitrite was carcinogenic in the forestomach of female mice. Shortly thereafter nitrite was reviewed for potential listing as a developmental and reproductive toxicant under the Proposition 65 Statute in California. The review by a panel of independent scientists concluded it was not a developmental or reproductive toxicant.

The literature and activities associated with the above events were reviewed and published by Archer (3).

The latest review was recently conducted by a working group convened by the International Agency for Research on Cancer (IARC); a part of the United Nations sponsored World Heath Organization headquartered in Lyon, France.

IARC's mission is to “coordinate and conduct research on the causes of human cancer, the mechanisms of carcinogenesis, and to develop scientific strategies for cancer control. The agency is involved in both epidemiological and laboratory research and disseminates scientific information through publications, meetings, courses, and fellowships” (67). The agency conducts its work with four main objectives: 1. Monitoring global cancer occurrence; 2. Identifying the causes of cancer; 3. Elucidation of mechanisms of carcinogenesis; 4. Developing scientific strategies for cancer control.

It was under the second of these objectives that IARC convened an expert working group to evaluate the carcinogenicity of ingested nitrates and nitrites, and cyanobacterial peptide toxins (65, 66).

The results of this review are to be published in a summary monograph. During an eight day period the working group composed of epidemiologists, toxicologists and cancer researchers reviewed the literature and made a decision on classifying nitrate and nitrite for their potential as human carcinogens. Their final conclusion was:

10/2/2006 3 of 15

Ingested nitrate or nitrite under conditions that result in endogenous nitrosation is probably carcinogenic to humans (Group 2A).

Thus, we are now back to controversy around “is nitrite safe as a food additive”? History of use, endogenous production and natural metabolism argue that nitrite as currently used is a safe food additive, despite the above conclusion which in a very narrow sense is accurate, but in a broader biological sense has little practical application.

Antibacterial properties of nitrite - an ingredient critical to food safety

The antibotulinal properties of nitrite have long been recognized. The use of nitrite to preserve meat has been employed either indirectly or directly for thousands of years. In the last 20 years however, the role of nitrite in protecting public health with respect to other pathogens has begun to be better recognized.

Nitrite produced from nitrate in saliva by oral bacteria (3) has significant antimicrobial benefits when it is swallowed and converted to nitrous acid and other nitrogen oxides in the gut. Not withstanding the potential for nitrosation of amines, the bacteriocidal effects of gastric fluids are significantly enhanced by the presence of ingested nitrite. This has been demonstrated for known foodborne pathogens such as Escherichia coli 0157:H7 and Salmonella (16, 17, 19). Nitrite and nitric oxide are also effective bacteriocidal agents against other microorganisms associated with diseases such as Helicobactor pylori , which has been linked to gastric cancer(18), organisms associated with dental caries, (2) and skin pathogens (63, 64).

As a food additive, nitrite is a key in controlling potential growth of Listeria monocytogenes in processed meats. Models that estimate the effects of ingredients on microbial growth show dramatic reductions when nitrite is included (7, 15, 22, 33, 37, 45, 49). The use of such models have enabled formulations of nitrite cured processed meat products that will not support growth of Listeria monocytogenes. To date, this has not been achieved for uncured counterparts where the only ingredient difference is nitrite. The U.S. Department of Agriculture (USDA) has done extensive research to develop models to predict growth of pathogens under a variety of conditions (59). In these models incorporation of nitrite at currently used levels significantly inhibits growth of Listeria, E. coli, and Salmonella.

10/2/2006 4 of 15

Nitrite and nitrate as part of the nitrogen oxide metabolic cycle

The concern about nitrite and nitrate and their potential to react with secondary amines to form carcinogenic nitrosamines resulted in intensive nutritional, biochemical and metabolic research in the 1970’s and 1980’s. One of the basic approaches in these studies was to measure consumption and excretion of nitrite and nitrate. The results produced an anomaly. Excretion always seemed to exceed consumption (26, 27, 55). This implied synthesis of these compounds in the body and led to a search for the pathways. In 1987, nitric oxide was identified as the transient factor that caused smooth muscle relaxation (29, 30). It was determined that synthesis of nitric oxide was synthesized from the amino acid arginine as a substrate and that it was degraded into nitrite and nitrite as a biological control mechanism and for excretion (8, 24, 25). Since that time there has been an explosion of research around the biological functions of nitric oxide. Nitric oxide is a profoundly biochemically active molecule, being involved in control of smooth muscle relaxation, neurotransmission, wound healing, immune response and a host of other biological functions. In 1998 the Nobel Prize for medicine was awarded to researchers for their efforts in the discovery of the basic functions and synthetic pathways of nitric oxide (51).

Typically endogenous nitric oxide is produced at about 1milligram per kilogram of body weight per day in humans (56). Being a very reactive molecule it is quickly bound to heme and oxidized to nitrite and nitrate. The nitrate is circulated in the blood and can be excreted in the urine, sweat or saliva of the individual.

Additionally, emerging evidence indicates that nitrite itself has a biological function as a signaling molecule independent of nitric oxide (5, 21). Curiously this function involves nitrosyl and s-nitroso heme species – the very same compounds involved in production of cured meat pigments.

In a recent publication a German researcher proposed classifying nitrite as a “prodrug” based on its many newly discovered physiological functions (53).

A 2005 symposium at the U.S. National Institutes of Health highlighted advances in the understanding of nitrite biochemistry, physiology and therapeutics (23).

Figure 1 below summarizes the interrelationships of nitric oxide, nitrite and nitrate. Clearly, nitrite is a metabolite and is naturally made in significant quantities. Exogenous nitrite ingestion is small by comparison, being at a residual level of approximately 10 ppm in commercial cured processed meats (10, 11, 31, 44).

10/2/2006 5 of 15

Figure 1. The Human Nitrogen Cycle - From Figure 4 of Gladwin et. al. 2005 (23) Abbreviations: NO3 – nitrate, NO2 – Nitrite, NO nitric oxide, NOS – nitric oxide synthase

Medical uses of nitrogen oxides

Since the discovery of nitric oxide, numerous therapeutic uses of targeted delivery of nitric oxide to specific tissues have been identified. Newborn infants, both full term and premature, with a variety of pulmonary problems are now treated with nitric oxide gas to relieve respiratory distress (32, 34, 48). Heart medications utilize delivery of nitric oxide to dilate smooth muscles and improve oxygenation of heart tissue. The long utilized cardiac drug, nitroglycerin, is a nitric oxide delivery vehicle (6, 13).

10/2/2006 6 of 15

A particularly active area of research involves skin. Psoriasis, acne and other skin disorders are treated with nitrite and nitric oxide delivery drugs (46, 47, 61, 62, 63, 64). The response of skin tissue to UV light and wound healing are beneficially affected by nitric oxide and nitrite. (4, 12, 35, 52, 53, 54)

Additionally, measurement of plasma or exhaled nitric oxide and nitrite is now being used in medical diagnosis. The severity and type of asthma can now be estimated by measurements of exhaled nitric oxide (1, 20, 28).

Although it is not widely realized, potassium nitrate is used in many toothpaste formulations. Toothpaste designed for sensitive teeth is formulated with up to 5% potassium nitrate, (which is 5000 ppm) (14, 36, 42, 43, 50, 57, 60). These products can be purchased at any drugstore.

NTP, California Proposition 65 and IARC Reviews

After the 1981 and 1982 National Academy of Sciences reports on nitrite and nitrate (38, 39), US regulations were changed for the usage of these ingredients. Upper ingoing limits were reduced for all cured meat products and in the case of bacon, specific requirements for curing and nitrosamine monitoring were incorporated (58). Only bacon was chosen for regulatory monitoring because it was found that volatile nitrosamine formation required high temperature (>300oF) and minimal quantities were found in most cured processed meats. The regulatory limit for nitrosamines in bacon fried by the USDA monitoring protocol was 17 parts per billion (ppb) (58). The result was a regulatory policy balance between the need to maintain botulinal safety and address concerns about nitrosamine formation risk in cured meats. The 1982 National Academy of Sciences report also called for a more thorough evaluation of nitrite in cancer bioassays and thus it was nominated by U.S. Food and Drug Administration (FDA) for study in the National Toxicology Program. This study was completed in the 1990’s and the results were peer reviewed by a panel at a meeting held on May 18, 2000. The panel heard comments from both staff managers of the study and external scientists sponsored by the American Meat Institute. They concluded that there was no evidence for nitrite induced carcinogenicity in any tissues of male and female rats and male mice. There was only equivocal evidence for carcinogenicity in the forestomach of female mice (41). This “gold standard” of cancer bioassays was, at the time, the definitive statement of safety for nitrite as an ingredient.

Shortly thereafter, the California Office of Environmental Health Hazard Assessment (OEHHA) completed a review of nitrite as a potential Developmental and Reproductive Toxicant (DART) under the Proposition 65 law enacted in that state. The staff produced a Hazard Identification Document which proposed

10/2/2006 7 of 15 listing nitrite as a DART (9), however, at a public hearing on June 2, 2000 the state’s review committee of independent technical experts voted 8-1 against listing. Surprisingly the originally published minutes of the meeting are no longer found on the OEHHA web site although all of the staff documents prepared for the meeting are available.

As discussed in the Introduction, IARC convened a working group to produce a monograph including nitrite and nitrate in June of 2006. The structure of a monograph has six sections: exposure data; studies of cancer in humans; studies of cancer in experimental animals; mechanistic and other relevant data; summary; evaluation and rationale. The first four sections are drafted by subgroups meeting concurrently while the last two sections are developed in plenary sessions where all members are included. There are strict requirements covering input to working group members which are intended to prevent undue influence by any commercial or industrial interests. Observers can be approved to attend and are only allowed limited opportunity to participate (65). The American Meat Institute sponsored Dr. James Coughlin as an official observer at the 2006 IARC monograph review of nitrite and nitrate (66). His function was to provide information to the reviewers from public peer reviewed literature or official sources. He also was able to watch the proceedings and provide perspective to the process. To those who disagree with the conclusion of the IARC working group, there are a number of examples of what could be considered irregular activities during the IARC process.

Many IARC sponsored studies correlating processed meat intake to colorectal cancer were rejected by the monograph’s epidemiology subgroup because the processed meat classification was confounded by inclusion of both cured and uncured products and they lacked any measurement or estimation of nitrite consumption. Despite significant pressure from IARC leadership, the epidemiologists remained rigorous in only considering studies where there was measurement of nitrite intake. Based on this, they concluded in a draft evaluation and rationale:

“There is limited evidence in humans for the carcinogenicity of nitrite in food. Nitrite in food is associated with an increased incidence of stomach cancer. [unanimous]”

The exposure section of the monograph was not done in draft form prior to the meeting and considerable information was gathered at the meeting to develop the draft. There was careful work done by IARC staff to frame statements about nitrite ingestion in terms of exogenous intake from food with limited inclusion of the endogenous production and salivary ingestion.

Furthermore, the toxicology subgroup deviated from scientific rigor by considering nitrite plus amines and amides and then reinterpreted the 2000 NTP

10/2/2006 8 of 15 study resulting in the following conclusions in this portion of the draft evaluation and rationale:

“There is sufficient evidence in experimental animals for the carcinogenicity of nitrite in combination with amines or amides. [unanimous]

There is limited evidence in experimental animals for the carcinogenicity of nitrite per se. [unanimous]”

The former statement, by referring to combination with amines and amides is outside the bounds of a review of just nitrite in the monograph. One wonders if the same conclusion would be reached if the IARC monograph had been focused on ingestion of foods containing amines or amides (which is virtually all food).

The latter statement also raises questions because the NTP nitrite study leader, Dr. Po Chan, was also the animal cancer subgroup leader in the IARC process (66), although he did not explicitly reveal that during their working sessions. At the 2000 NTP peer review his preliminary observation had been overturned. In effect this statement amounts to a reinterpretation of the NTP results back to his preliminary conclusions.

The IARC monograph preamble (65) refers to means of making classifications and the combination of the three conclusions above, led automatically to classification in Group 2A as a probable human carcinogen.

The net result was a statement that is truly accurate but very narrow in scope. Under certain conditions ingested amines and amides can be nitrosated to form carcinogenic nitrosamines and nitrosamides. However, one must question the practical ability to use this statement in any type of public policy or regulatory activity. As stated, swallowing one’s own saliva in combination with any food could be considered a potentially carcinogenic event. This simply lacks common sense.

The ingestion of nitrate from vegetables, grains and water or use of some toothpaste formulations followed by nitrite formation in saliva would also qualify as probably causing cancer according to the statement. Thus, to eat is to ingest nitrite, nitrate, amines and amides regardless of the diet. Ingestion of cured meats would be no more or less risky given the low amount of residual nitrite in cured meats relative to other sources in the diet.

Conclusions

There will very likely be considerable debate about the need for regulatory review and action on nitrite as a food ingredient as a result of the IA RC classification.

10/2/2006 9 of 15

However, when considered in the context of evolving research about the biological function of all nitrogen oxides and their metabolism, any changes to current regulations are unwarranted.

The food additive use of nitrite represents a small addition to the body burden of endogenously produced nitrogen oxides. It is hard to conceive that ingestion of nitrite from cured meats could have an impact on any potential toxic outcomes.

Due to the intense social desire to identify causes of cancer we have come to a point where technically accurate but narrow findings can mislead us. I believe that as more is understood about the human metabolic nitrogen oxide cycle it will become apparent that nitrite is a safe and appropriate food additive which provides many more benefits to society than risks.

References

1. Anonymous. 2006. Exhaled nitric oxide monitoring as a guide to treatment decisions in chronic asthma. Technol. Eval. Cent. Asses Program. Exec. Summ. 20:1-3.

2. Allaker, R. P., L. S. Silva Mendez, J. M. Hardie, and N. Benjamin. 2001. Antimicrobial effect of acidified nitrite on periodontal bacteria. Oral Microbiol. Immunol. 16:253-256.

3. Archer, D. L. 2002. Evidence that ingested nitrate and nitrite are beneficial to health. J. Food Prot. 65:872-875.

4. Broughton, G.,2nd, J. E. Janis, and C. E. Attinger. 2006. The basic science of wound healing. Plast. Reconstr. Surg. 117:12S-34S.

5. Bryan, N. S., B. O. Fernandez, S. M. Bauer, M. F. Garcia-Saura, A. B. Milsom, T. Rassaf, R. E. Maloney, A. Bharti, J. Rodriguez, and M. Feelisch. 2005. Nitrite is a signaling molecule and regulator of gene expression in mammalian tissues. Nat. Chem. Biol. 1:290-297.

6. Buch, A. N., S. Chowdhary, J. H. Coote, and J. N. Townend. 2004. Effects of nitroglycerin treatment on cardiac autonomic control in heart failure. Clin. Auton. Res. 14:9-14.

10/2/2006 10 of 15

7. Buchanan, R. L., M. H. Golden, and J. G. Phillips. 1997. Expanded models for the non-thermal inactivation of Listeria monocytogenes. J. Appl. Microbiol. 82:567-577.

8. Buga, G. M., M. E. Gold, K. S. Wood, G. Chaudhuri, and L. J. Ignarro. 1989. Endothelium-derived nitric oxide relaxes nonvascular smooth muscle. Eur. J. Pharmacol. 161:61-72.

9. California Office of Environmental Health Hazard Assessment, M. Campbell. 2000. Evidence on Developmental and Reproductive Toxicity of Sodium Nitrite: Draft Hazard Identification Document.

10. Cassens, R. G. 1997. Residual nitrite in cured meat. Food Technol. 51:53- 55.

11. Cassens, R. G. 1997/8. Composition and safety of cured meats in the USA. Food Chemistry 59:561-566.

12. Childress, B. B., J. K. Stechmiller. 2002. Role of nitric oxide in wound healing. Biol. Res. Nurs. 4:5-15.

13. Clermont, G., S. Lecour, C. Vergely, M. Zeller, C. Perrin, V. Maupoil, O. Bouchot, and L. Rochette. 2003. Direct demonstration of nitric oxide formation in organs of rabbits treated by transdermal glyceryl trinitrate using an in vivo spin trapping technique. Fundam. Clin. Pharmacol. 17:709-715.

14. Cohen, S., T. Schiff, J. McCool, A. Volpe, and M. E. Petrone. 1994. Anticalculus efficacy of a dentifrice containing potassium nitrate, soluble pyrophosphate, PVM/MA copolymer, and sodium fluoride in a silica base: a twelve-week clinical study. J. Clin. Dent. 5 Spec No:93-96.

15. Duffy, L. L., P. B. Vanderlinde, and F. H. Grau. 1994. Growth of Listeria monocytogenes on vacuum-packed cooked meats: effects of pH, aw, nitrite and ascorbate. Int. J. Food Microbiol. 23:377-390.

16. Duncan, C., H. Li, R. Dykhuizen, R. Frazer, P. Johnston, G. MacKnight, L. Smith, K. Lamza, H. McKenzie, L. Batt, D. Kelly, M. Golden, N. Benjamin, and C. Leifert. 1997. Protection against oral and gastrointestinal diseases: importance of dietary nitrate intake, oral nitrate reduction and enterosalivary nitrate circulation. Comp. Biochem. Physiol. A Physiol. 118:939-948.

17. Duncan, C., H. Li, R. Dykhuizen, R. Frazer, P. Johnston, G. MacKnight, L. Smith, K. Lamza, H. McKenzie, and L. Batt. 1997/12. Protection against oral and gastrointestinal diseases: Importance of dietary nitrate intake, oral nitrate reduction and enterosalivary nitrate circulation. Comparative Biochemistry and Physiology Part A: Physiology 118:939-948.

10/2/2006 11 of 15

18. Dykhuizen, R. S., A. Fraser, H. McKenzie, M. Golden, C. Leifert, and N. Benjamin. 1998. Helicobacter pylori is killed by nitrite under acidic conditions. Gut 42:334-337.

19. Dykhuizen, R. S., R. Frazer, C. Duncan, C. C. Smith, M. Golden, N. Benjamin, and C. Leifert. 1996. Antimicrobial effect of acidified nitrite on gut pathogens: importance of dietary nitrate in host defense. Antimicrob. Agents Chemother. 40:1422-1425.

20. Ehrs, P. O., B. M. Sundblad, and K. Larsson. 2006. Quality of life and inflammatory markers in mild asthma. Chest 129:624-631.

21. Feelisch, M., T. Rassaf, S. Mnaimneh, N. Singh, N. S. Bryan, D. Jourd'Heuil, and M. Kelm. 2002. Concomitant S-, N-, and heme-nitros(yl)ation in biological tissues and fluids: implications for the fate of NO in vivo. FASEB J. 16:1775-1785.

22. Gill, A. O., R. A. Holley. 2003. Interactive inhibition of meat spoilage and pathogenic bacteria by lysozyme, nisin and EDTA in the presence of nitrite and sodium chloride at 24 degrees C. Int. J. Food Microbiol. 80:251-259.

23. Gladwin, M. T., A. N. Schechter, D. B. Kim-Shapiro, R. P. Patel, N. Hogg, S. Shiva, R. O. Cannon 3rd, M. Kelm, D. A. Wink, M. G. Espey, E. H. Oldfield, R. M. Pluta, B. A. Freeman, J. R. Lancaster Jr, M. Feelisch, and J. O. Lundberg. 2005. The emerging biology of the nitrite anion. Nat. Chem. Biol. 1:308-314.

24. Gold, M. E., P. A. Bush, and L. J. Ignarro. 1989. Depletion of arterial L- arginine causes reversible tolerance to endothelium-dependent relaxation. Biochem. Biophys. Res. Commun. 164:714-721.

25. Gold, M. E., K. S. Wood, G. M. Buga, R. E. Byrns, and L. J. Ignarro. 1989. L- arginine causes whereas L-argininosuccinic acid inhibits endothelium- dependent vascular smooth muscle relaxation. Biochem. Biophys. Res. Commun. 161:536-543.

26. Green, L. C., S. R. Tannenbaum, and P. Goldman. 1981. Nitrate Synthesis in the Germfree and Conventional Rat. Science 212:56-58.

27. Green, L. C., K. R. D. Luzuriaga, D. A. Wagner, W. Rand, N. Istfan, V. R. Young, and S. R. Tannenbaum. 1981. Nitrate Biosynthesis in Man. Proc. Natl. Acad. Sci. U. S. A. 78:7764-7768.

28. Heinen, V., M. Claeys, and R. Louis. 2006. Exhaled nitric oxide: a new biomarker for respiratory pathologies. Rev. Med. Liege 61:37-42.

29. Ignarro, L. J. 1999. Nitric oxide: a unique endogenous signaling molecule in vascular biology. Biosci. Rep. 19:51-71.

10/2/2006 12 of 15

30. Ignarro, L. J., R. E. Byrns, G. M. Buga, and K. S. Wood. 1987. Endothelium- derived relaxing factor from pulmonary artery and vein possesses pharmacologic and chemical properties identical to those of nitric oxide radical. Circ. Res. 61:866-879.

31. Kilic, B., R. G. Cassens, and L. L. Borchert. 2001. Influence of turkey meat on residual nitrite in cured meat products. J. Food Prot. 64:235-239.

32. Kinsella, J. P., S. H. Abman. 2005. Inhaled nitric oxide therapy in children. Paediatr. Respir. Rev. 6:190-198.

33. Legan, J. D., D. L. Seman, A. L. Milkowski, J. A. Hirschey, and M. H. Vandeven. 2004. Modeling the growth boundary of Listeria monocytogenes in ready-to-eat cooked meat products as a function of the product salt, moisture, potassium lactate, and sodium diacetate concentrations. J. Food Prot. 67:2195-2204.

34. Lonnqvist, P. A., B. Jonsson. 2005. Premature infants benefit from inhaled nitric oxide, too. Not only full-term infants with severe hypoxic respiratory failure. Lakartidningen 102:3880-3882.

35. Luo, J. D., A. F. Chen. 2005. Nitric oxide: a newly discovered function on wound healing. Acta Pharmacol. Sin. 26:259-264.

36. Manochehr-Pour, M., M. Bhat, and N. Bissada. 1984. Clinical evaluation of two potassium nitrate toothpastes for the treatment of dental hypersensitivity. Periodontal Case Rep. 6:25-30.

37. McClure, P. J., A. L. Beaumont, J. P. Sutherland, and T. A. Roberts. 1997/3/3. Predictive modelling of growth of Listeria monocytogenes The effects on growth of NaCl, pH, storage temperature and NaNO2. International Journal of Food Microbiology 34:221-232.

38. National Academy of Sciences. 1982. Alternatives to the current use of nitrite in foods.

39. National Academy of Sciences. 1981. The health effects of Nitrate, Nitrite and N-Nitroso Compounds.

40. National Toxicology Program. 2001. NTP Technical Report on the toxicology and carcinogenesis studies of sodium nitrite (CAS NO. 7632-00-0)IN F344/N rats and B6C3F1 mice(drinking water studies), NTP TR 495. NIH Publication No. 01-3954:

41. National Toxicology Program. 2001. Toxicology and carcinogenesis studies of sodium nitrite (CAS NO. 7632-00-0) in F344/N rats and B6C3F1 mice (drinking water studies). Natl. Toxicol. Program. Tech. Rep. Ser. 495:7-273.

10/2/2006 13 of 15

42. Orchardson, R., D. G. Gillam. 2000. The efficacy of potassium salts as agents for treating dentin hypersensitivity. J. Orofac. Pain 14:9-19.

43. Poulsen, S., M. Errboe, O. Hovgaard, and H. W. Worthington. 2001. Potassium nitrate toothpaste for dentine hypersensitivity. Cochrane Database Syst. Rev. (2):CD001476.

44. Saccani, G., E. Tanzi, S. Cavalli, and J. Rohrer. 2006. Determination of nitrite, nitrate and glucose-6-phosphate in muscle tissues and cured meat by IC/MS. J. A. O. A. C Int. 89:712-719.

45. Schlyter, J. H., K. A. Glass, J. Loeffelholz, A. J. Degnan, and J. B. Luchansky. 1993/9. The effects of diacetate with nitrite, lactate, or pediocin on the viability of Listeria monocytogenes in turkey slurries. International Journal of Food Microbiology 19:271-281.

46. Schwentker, A., Y. Vodovotz, R. Weller, and T. R. Billiar. 2002. Nitric oxide and wound repair: role of cytokines? Nitric Oxide 7:1-10.

47. Seabra, A. B., A. Fitzpatrick, J. Paul, M. G. De Oliveira, and R. Weller. 2004. Topically applied S-nitrosothiol-containing hydrogels as experimental and pharmacological nitric oxide donors in human skin. Br. J. Dermatol. 151:977- 983.

48. Sehgal, A., I. Callander, J. Stack, T. Momsen, and K. Sterling-Levis. 2005. Experience with inhaled nitric oxide therapy in hypoxic respiratory failure of the newborn. Indian J. Chest Dis. Allied Sci. 47:245-249.

49. Seman, D. L., A. C. Borger, J. D. Meyer, P. A. Hall, and A. L. Milkowski. 2002. Modeling the growth of Listeria monocytogenes in cured ready-to-eat processed meat products by manipulation of sodium chloride, sodium diacetate, potassium lactate, and product moisture content. J. Food Prot. 65:651-658.

50. Silverman, G. 1985. The sensitivity-reducing effect of brushing with a potassium nitrate-sodium monofluorophosphate dentifrice. Compend. Contin. Educ. Dent. 6:131-3, 136.

51. Smith, O. 1998. Nobel Prize for NO research. Nat. Med. 4:1215.

52. Suschek, C. V., A. Paunel, and V. Kolb-Bachofen. 2005. Nonenzymatic nitric oxide formation during UVA irradiation of human skin: experimental setups and ways to measure. Methods Enzymol. 396:568-578.

53. Suschek, C. V., T. Schewe, H. Sies, and K. D. Kroncke. 2006. Nitrite, a naturally occurring precursor of nitric oxide that acts like a 'prodrug'. Biol. Chem. 387:499-506.

10/2/2006 14 of 15

54. Suschek, C. V., P. Schroeder, O. Aust, H. Sies, C. Mahotka, M. Horstjann, H. Ganser, M. Murtz, P. Hering, O. Schnorr, K. D. Kroncke, and V. Kolb- Bachofen. 2003. The presence of nitrite during UVA irradiation protects from apoptosis. FASEB J. 17:2342-2344.

55. Tannenbaum, S. R., J. P. Witter, S. J. Gatley, and E. Balish. 1979. Nitrate and Nitrite: Origin in Humans. Science 205:1332+1334-1337.

56. Tricker, A. R. 1997. N-nitroso compounds and man: sources of exposure, endogenous formation and occurrence in body fluids. Eur. J. Cancer Prev. 6:226-268.

57. Tzanova, S., Z. Ivanova, and S. Velinova. 2005. Clinical evaluation of dentinal hypersensitivity treatment with 5% potassium nitrate dentifrice. Folia. Med. (Plovdiv) 47:65-69.

58. United States Code of Federal Regulations. 2006. Title 9 Animals and Animal Products Chapter III:Food Safety and Inspection Service. Department of Agriculture. Sections 319.2 Products and nitrates and nitrites; 424.21 Curing Agents; 424.22(b) Use of nitrite and sodium ascorbate or sodium erythorbate (isoascorbate) in bacon..

59. USDA Eastern Regional Research Laboratory. 03/02/2006 Pathogen Modeling Program (PMP). Available at: http://ars.usda.gov/Services/docs.htm?docid=6786.

60. Wara-aswapati, N., D. Krongnawakul, D. Jiraviboon, S. Adulyanon, N. Karimbux, and W. Pitiphat. 2005. The effect of a new toothpaste containing potassium nitrate and triclosan on gingival health, plaque formation and dentine hypersensitivity. J. Clin. Periodontol. 32:53-58.

61. Weller, R. 2003. Nitric oxide donors and the skin: useful therapeutic agents? Clin. Sci. (Lond) 105:533-535.

62. Weller, R., A. D. Ormerod, R. P. Hobson, and N. J. Benjamin. 1998. A randomized trial of acidified nitrite cream in the treatment of tinea pedis. J. Am. Acad. Dermatol. 38:559-563.

63. Weller, R., R. Dykhuizen, C. Leifert, and A. Ormerod. 1997. Nitric oxide release accounts for the reduced incidence of cutaneous infections in psoriasis. J. Am. Acad. Dermatol. 36:281-282.

64. Weller, R., R. J. Price, A. D. Ormerod, N. Benjamin, and C. Leifert. 2001. Antimicrobial effect of acidified nitrite on dermatophyte fungi, Candida and bacterial skin pathogens. J. Appl. Microbiol. 90:648-652.

65. World Health Organization. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Preamble to the IARC Monographs

10/2/2006 15 of 15

(amended January 2006). Available at: http://monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdf.

66. World Health Organization. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 94: Ingested nitrates and nitrites, and cyanobacterial peptide toxins Lyon, France: 14-21 June 2006 , List of Participants. Available at: http://monographs.iarc.fr/ENG/Meetings/vol94- participants.pdf.

67. World Health Organization. July 3, 2006 IARC's Mission: Cancer research for cancer control. Available at: http://www.iarc.fr/; http://www.iarc.fr/ENG/General/index.php.

Acta Physiol Scand 2000, 168, 623±634

Deciphering the mysteries of myoglobin in striated muscle

K.E. CONLEY,1 G.A. ORDWAY2 andR.S. RICHARDSON3 1 Departments of Radiology, Physiology & Biophysics, Bioengineering, University of Washington Medical Center, Seattle, WA, USA 2 Department of Physiology, University of Texas Southwestern Medical Center, Dallas, USA 3 Department of Medicine, University of California, San Diego, La Jolla, California, USA

ABSTRACT

Myoglobin (Mb) is a large protein that reversibly binds oxygen in the muscle cell and is thought to be

critical for O2 supply to the mitochondria during exercise. The role of Mb in aerobic function is

evaluated based on the physical properties of Mb as an O2 carrier and experimental evidence of Mb

function in vivo. This role depends on the reversible binding of O2 by Mb depending on PO2, which

results in: (1) storage of O2; (2) buffering of PO2 in the cell to prevent mitochondrial anoxia; and (3)

parallel diffusion of O2 (so-called, `facilitated diffusion'). The storage role is well established in diving

mammals and buffering of cell PO2 above anoxic levels is shown here by in vivo magnetic resonance

spectroscopy (MRS). However, the quantitative role of Mb in `facilitated' or parallel diffusion of O2 is controversial. Evidence in support of this role is from MRS analyses, which reveal rapid Mb

desaturation with exercise, and from the proportionality of Mb content of a muscle to the O2 diffusion limitation. Recent experiments with myoglobin knockout mice demonstrating high levels of aerobic function in normal and myoglobin-free mice argue against a link between Mb and oxidative

phosphorylation. Thus, the current evidence supports the role of Mb in the physical diffusion of O2; however, the unimpaired aerobic function of Mb knockout mice indicates that this role may not be

critical to O2 supply in active muscle.

Keywords body size scaling, facilitated diffusion, magnetic resonance spectroscopy, myoglobin

knockout mouse, O2 diffusion, oxidative phosphorylation.

Received 18 November 1999, accepted 9 December 1997

This review considers the role of myoglobin (Mb) in to be diminished by the recent ®nding that the oxygen delivery in active muscle. Many excellent myoglobinless mouse has aerobic function similar to reviews have covered the purported mechanisms of the wild-type (Garry et al. 1998); however, these animals

function, mathematics of diffusion and key roles in O2 appear to have both cellular and cardiovascular storage, PO2 buffering and facilitated diffusion of O2 adaptations that might compensate for the lack of Mb (Kreuzer 1970, Wittenberg 1970, Wittenberg & Wit- (Godecke et al. 1999). The goal of this paper is to

tenberg 1989). Despite this extensive treatment in the provide a simple analysis of Mb's role in O2 supply to literature, the role of Mb in muscle function remains understand how Mb functions in normal tissue in vivo controversial with many papers claiming a small or and how muscle and cardiovascular adaptations could

non-existent role of Mb in O2 delivery (JuÈrgens et al. result in unimpaired aerobic function in Mb-knockout 1994, Papadopoulos et al. 1995) and others demon- muscle. To achieve this goal, the authors present their strating a relation between Mb content and aerobic work on Mb function: (1) Mechanisms of myoglobin function function (Conley & Jones 1996). In support of its presents a simple diffusion model as a guide to Mb

classical role in O2 delivery is evidence that Mb-O2 function in the muscle cell; (2) Myoglobin measurements: shows desaturation with exercise in vivo (Richardson Access to intracellular PO2 presents magnetic resonance 1et al. 1995a, Mole et al. 1999), which suggests that a measurements of Mb-O2 saturation during exercise; Mb-O2 gradient is present for Mb-mediated O2 diffu- and (3) Myoglobinless mice reports on studies of the sion. The importance of this classical role would appear myoglobin knockout mouse.

Correspondence: Kevin E. Conley PhD, Department of Radiology, Box 357115, University of Washington Medical Center, Seattle, WA 98195- 7115, USA.

Ó 2000 Scandinavian Physiological Society 623 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

MECHANISMS OF MYOGLOBIN means that binding will occur at the high PO2s char- FUNCTION acteristic of the capillary (>10 Torr) and O2 unloading will occur at the lower PO s in the muscle ®bre Paradoxes of O diffusion in muscle 2 2 (<10 Torr in active muscle). Several additional criteria The ®rst step in this analysis is to consider the role of necessary for Mb to signi®cantly contribute to oxygen diffusion distance in O2 supply. Diffusion is thought to ¯ux (Wittenberg 1970) have been satis®ed: (1) Mb is be the key pathway for O2 supply from capillary to the found in concentrations suf®cient for carrier-mediated respiring mitochondria and is often invoked as a limit diffusion (Armstrong et al. 1992); (2) Mb diffuses in vivo to muscle respiratory capacity (Wagner 1993). However, (JuÈrgens, Peters et al. 1994); and (3) Mb-O2 is desatu- several examples show no evidence for diffusion rated in exercising muscle indicating a Mb-O2 gradient distance in limiting oxygen delivery (see Conley & Jones to drive Mb-mediated O2 ¯ux (Richardson et al. 1995a, 1996). For example, muscles with different intercapil- Mole et al. 1999). lary distances have been found to have either similar Myoglobin is a good candidate for supplementing O2 muscle or similar mitochondria-speci®c metabolic rates supply because it provides a parallel pathway for diffu- (Hudlicka et al. 1988, Kayar & Weiss 1992). Thus, we sion in the muscle ®bre. Fig. 1 shows the factors gov- _ are faced with the paradox that diffusion distance does erning O2 ¯ux (M O2) in a simple electrical circuit with not appear to signi®cantly limit O2 delivery or conductances (inverse of resistance) through the ®bre consumption in vivo despite the key role played by de®ned for O2 (Df) and Mb diffusion (DMb). Later we distance in diffusion. consider the full O2 diffusion pathway from the capillary Could parallel diffusion of O2 and O2-Mb be the through the ®bre to its core using the Fick principle (see mechanism by which the key role of distance in O2 Myoglobin measurements: Access to intracellular PO2), but here diffusion is minimized? Myoglobin has been shown to we consider only the ®bre and examine diffusion be an important mediator of O2 ¯ux (Wittenberg 1970, resulting from gradients set by PO2 (PO2) and Mb sat- Wittenberg & Wittenberg 1990) and extensive model- uration (SMb) between the sarcolemma and ®bre core: ling literature has shown how Mb-mediated diffusion M_ O ˆ D PO †‡D S † 1† could work in muscle tissue (Meyer et al. 1984, Popel 2 f 2 Mb Mb 1989). Several results con®rm this important role for The conductances re¯ect the ®bre size (r±2), content

Mb in O2 supply. Mb is one of the most concentrated (oxygen solubility, aO2 or myoglobin concentration, proteins in the cell and modelling analyses have shown [Mb]) and the diffusivities of O2 (DO2) and Mb (DMb) that Mb content is proportional to the shortfall in O2 for diffusion into a cylinder: supply at maximum O demand (Conley & Jones 1996). 2 D ˆ 4 Á aO Á DO Á r 2 2† Further, Mb content is very high in diving animals, f 2 2 indicating that the storage of O2 afforded by a protein in mM levels provides an important source of O2 during dives. However, there is evidence that appears incon- sistent with a primary role of Mb-mediated diffusion. Mb concentration is not closely related to oxidative capacity or ®bre radius as the highest Mb level is found in ®bres of intermediate size and mitochondrial content (Conley & Jones 1996). In fact, the highly aerobic heart has relatively low levels of Mb [i.e. 0.25 mM (Armstrong et al. 1992)]. In addition, recent measurements in muscle tissue indicate that O2 diffusivity is considerably higher and Mb diffusivity considerably lower than previously thought (Papadopoulos et al. 1995, Bentley & Pittman 1997). The low Mb diffusivity led JuÈrgens et al. (1994) to conclude that Mb plays a minor role in mediating O2 ¯ux. Clearly, an analysis of the Mb- mediated O2 diffusion is needed to resolve these para- doxes.

A simple analysis of diffusion Several characteristics make Mb well suited for diffu- Figure 1 Diagram of a simple diffusion model of O2 ¯ux in muscle sion. The equilibrium binding nature of O2 to Mb ®bres.

624 Ó 2000 Scandinavian Physiological Society Acta Physiol Scand 2000, 168, 623±634 K E Conley et al. Á Myoglobin function in striated muscle

Figure 2 Myoglobin concentration ([Mb]) and diffusion conductance

(DMb) as a function of mitochondrial volume density of horse and steer muscle. n, From skeletal muscle; n, from the heart. The original data are reported in Armstrong et al. (1992) and Conley & Jones (1996).

diffusion in the ®bre ± ®bre radius and [Mb] ± in 2 DMb ˆ 4 Á [Mb] Á DMb Á r 3† combination result in a diffusion conductance that is closely correlated with mitochondrial content and Thus, Mb provides a parallel pathway for O2 diffusion therefore O ¯ux. driven by a separate gradient (i.e Mb-O2 gradient). 2 These conductances represent only the capacity for O supply while the actual ¯ux via O and Mb-O High Mb content trades-off for low diffusivity 2 2 2 diffusion depends on the PO2 levels and gradients in the The large size of the Mb molecule (17 500 mw) com- muscle. A simple diffusion model in the ®bre (Eqn 1) is evaluated for these O ¯uxes and conductances. This pared with O2 would appear to make Mb a poor can- 2 simple model permits determination of how much of didate for parallel diffusion of O2. The 100-fold this potential of Mb-mediated diffusion is used in O difference in the diffusivity of O2 and Mb would sug- 2 delivery. Figure 3 shows the fraction of O carried by gest that Mb-mediated O2 ¯ux would make a trivial 2 Mb vs. free diffusion as a function of [Mb] for three contribution to O2 supply. However, the tissue con- ductances depend both on the diffusivities and tissue ratios of the diffusivities for Mb (DMb)toO2 (DO2). Clearly, a larger fraction of the O ¯ux is carried by Mb concentrations of O2 and Mb as de®ned in Eqns 2 and 2 3. The tissue concentrations of these molecules are also as its concentration in the cell increases. Also, a larger nearly 100-fold different, but in the opposite direction, fraction is carried by Mb-O2 as diffusivity of Mb relat- ive to O increases. The value of both diffusivities have with the low solubility of O2 in tissues resulting in [O2] 2 in the lM range compare to typical [Mb] in the mM recently changed as new techniques have emerged to range. Thus, the slow diffusivity of the Mb molecule is better measure this property. Recently, the DO2 value compensated for by large tissue concentrations that was revised upward by directly measuring the value at tissue temperatures expected in exercising muscle (i.e. result in a similar conductance (DMb) as compared with 40 °C) (Bentley & Pittman 1997). The D value was O2 alone (Df). Mb revised downwards based on in situ spectroscopic measurements at physiological temperature (JuÈrgens Myoglobin conductance, not concentration, is important et al. 1994, Papadopoulos et al. 1995), but measure- ments by magnetic resonance in intact, beating cardiac Mb content alone is a poor predictor of its role in O2 delivery as illustrated in Fig. 2, which compares [Mb] muscle support the higher DMb values (Wang et al. 1997). Direct measurements of the quantitative role of with the mitochondrial volume density (VV(mt,f))of muscle samples from horses and steers. The heart has Mb in O2 delivery are needed to resolve this discrep- ancy in D and determine which Mb diffusivity value the highest VV(mt,f) and therefore maximal O2 demand, Mb yet this tissue has [Mb] less than half the level found in governs Mb-mediated O2 diffusion. skeletal muscles with a third the VV(mt,f). Equation 3 shows that concentration alone does not determine the Mb content vs. diffusion limitation conductance to Mb, but that ®bre radius is also a critical factor. Figure 2 also shows that the combination of Demonstration of a diffusion limitation to O2 supply ®bre radius and Mb content contained in the Mb dif- would indicate that additional O2 delivery via Mb is fusion conductance results in a continuous relation of needed to meet the muscle O2 demand. We previously DMb with VV(mt,f). Thus, the two factors critical to O2 reported an analysis of O2 delivery in muscle tissue in

Ó 2000 Scandinavian Physiological Society 625 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

1981) and muscle mitochondrial volume density (Mat-

hieu et al. 1981). Thus the maximum O2 ¯ux per muscle mass is many-fold higher in a 0.05 kg mouse as com- pared with a 500-kg horse. If myoglobin content were proportional to total ¯ux, the content would be expected to increase as the maximum ¯ux per mass rises (and body mass decreases). Instead, the opposite is apparent in Fig. 4 where Mb content is shown to increase with body mass in the Psoas muscle in animals ranging in size from the pigeon to the blue whale (Lawrie 1953). Elimination of the one diving mammal ± the blue whale ± has little effect on the strong rela- tionship between Mb content and body mass shown in this ®gure. How do we explain the increase in Mb content in Figure 3 Myoglobin contribution to O supply as a function of 2 the face of reduced maximal O2 demand? The myoglobin concentration and the ratio of the myoglobin to oxygen answer lies in the changing muscle ®bre and car- diffusivity (D /DO ). Mb 2 diovascular properties with body size. First, the horse and steers running at rates that evoke the maxi- driving force for O2 delivery from the blood mum oxygen uptake (Conley & Jones 1996). Our increases as body size decreases. More than 40 years analysis found that many ®bres had an O2 demand that ago, the P50 of blood was found to increase with exceeded the capacity for O2 supply by diffusion as the lower body size (Schmidt-Neilsen & Larimer 1958) muscle reached its maximum oxygen uptake. The and, more recently, this P50 scaling was found to lowest Mb content was found in those ®bres without a result in a higher mixed venous PO2 in smaller ani- diffusion limitation (O2 demand/O2 diffusion ˆ 1) mals exercising at VO2max (Kaya et al. 1994). Thus, and was highest in the ®bres with the greatest diffusion smaller mammals have a larger PO2 driving O2 dif- limitation (O2 demand/O2 diffusion > 1). These ®nd- fusion into the muscle cell, especially at the end of ings indicated that Mb was needed as a parallel source the capillary where the magnitude of the gradient is of O2 supply and that the Mb content increased in critical to O2 and Mb-mediated diffusion. proportion to the size of the diffusion limitation to At the same time that the O2 pressure head is rising, mitochondrial O2 demands. ®bre diameter is decreasing and capillary density is increasing with smaller body size (Hoppeler et al. 1981). The consequence of these changes is a large increase in Mb scaling: Paradox of body size on diffusion O conductance (Eqn 2) as well as a decrease in and oxidative capacity 2 intercapillary distances. The result is that smaller

Nature has provided us with a good experiment to test mammals not only have a larger pressure head for O2 the role of myoglobin in O2 delivery using the scaling diffusion into the muscle cell but an increased con- of muscle and physiological properties with body size. ductance for O2 diffusion as well. Thus, one way to The classic mouse-to-elephant curve illustrates the explain the decrease in Mb content with decreasing point that resting O2 uptake increases per body mass body size is by the reduced need for supplemental O2 with decreasing size. A similar increase per muscle mass supply because of an increased O2 diffusion capacity in is seen in the maximum O2 uptake (Seeherman et al. smaller vs. larger animals.

Figure 4 Scaling of myoglobin content (Mb content) in muscles as a function of body mass. (a) [Mb] in a variety of cardiac and skeletal mus- cles from animals ranging from pigeons to the blue whale. (b) The scaling of Mb content in a single muscle (psoas) in mammals with body mass. Data are taken from Lawrie (1953).

626 Ó 2000 Scandinavian Physiological Society Acta Physiol Scand 2000, 168, 623±634 K E Conley et al. Á Myoglobin function in striated muscle

The two examples reported here not only demon-

strate the important role of Mb in O2 supply, but also the interaction of muscle properties in the balance of

O2 supply to O2 demand. First, Mb content was found to be higher in ®bres with an O2 diffusion limitation in horse and steer muscle exercising at the maximum oxygen consumption (Conley & Jones 1996). Second, lower Mb content in smaller animals accompanied the

reduction in the O2 diffusion limitation resulting from changes in cellular and cardiovascular properties with

body size. Key to this modelling is that Mb-O2 desat- urates during exercise and generates a gradient in the

muscle ®bre down which parallel Mb-O2 diffusion can occur. In the following section, Richardson reports

in vivo magnetic resonance measurements of Mb-O2 saturation levels in exercising human limb muscle. The ®nal section considers life without myoglobin. Could

the reduced O2 diffusion limitation in small mammals and adaptations that balance O2 supply to O2 demand permit normal aerobic function in the absence of Figure 5 The percentage of maximum cuff deoxy-Mb signal in hypoxia and normoxia as a function of the percentage of maximum myoglobin? The aerobic properties and adaptations of normoxic leg VO during knee-extensor exercise. the myoglobin knockout mouse presented in the ®nal 2 section will help to answer this question. deoxy-Mb signal (re-saturation) when exercise is ceased for less than 45 s, for instance to adjust a subject's shin MYOGLOBIN MEASUREMENTS: ACCESS brace during knee-extensor exercise. This rapid Mb TO INTRACELLULAR PO 2 desaturation to 50±60% is indicative of the immediate

The recent ability to detect myoglobin (Mb) desatura- use of at least half of the Mb O2 stores. Thus, the speed tion utilizing proton magnetic resonance imaging and magnitude of this response may have several (MRS) has made it possible to measure intracellular functions. First, the immediate availability of 50% of

PO2, the ®nal and somewhat elusive last step in the the stored Mb-associated O2 may be an important O2 oxygen cascade from air to muscle (Wang et al. 1990, source for the increased oxidative metabolism at the Noyszewski et al. 1997). Proton MRS, unlike most start of exercise. Second, this Mb desaturation reduces previous techniques that address tissue oxygenation, is the carrier-depleted region, maximizing the PO2 gradi- non-invasive, is without deleterious effects, and there- ent from blood to cell (Fig. 6). Consequently, the pas- fore very well suited for in vivo human studies (Thomas sive transport system responsible for O2 in¯ux into the & Morris 1981). The functional isolation of the quad- muscle cell is facilitated by this rapid desaturation, even riceps muscle group has been a useful model to study during light exercise. The greater deoxy-Mb signal in exercising muscle in man (Andersen et al. 1985, Rich- hypoxia (Fig. 5) and the concurrent elevation in the O2 ardson et al. 1993, Radergran 1997) and allows the diffusional conductance in hypoxia throughout sub-

measurement of ef¯uent (femoral) venous PO2 and the maximal exercise (Fig. 7), supports these concepts. The calculation of mean capillary PO2 (Richardson et al. effective diffusivity at submaximal WR's in hypoxia was 21995b). In combination, this proton MRS technique to markedly elevated in comparison with normoxic values. determine Mb saturation and the functionally isolated At »65% of VO2max the diffusional conductance in human quadriceps muscle model (Andersen et al. 1985, hypoxia was already 85% of the maximum recorded Richardson et al. 1995a, b, Richardson et al. 1998) have value, whereas in normoxia at a comparable VO2 con- provided the opportunity to study the relationships ductance was only 59% of the maximum value. This between intracellular and intravascular events in difference may be the result of the increased concen- humans during exercise. tration of available O2 carrier within the muscle tissue (deoxy-Mb) (Honig & Gayeski 1993) (Fig. 7). Rapid Mb desaturation at submaximal exercise Blood to cell PO gradients The maximum deoxy-Mb signal is apparent even at 2

50% of VO2max (Fig. 5) and this desaturation occurs The use of proton MRS and the knee-extensor model very rapidly. There is a rapid disappearance of the have provided the ®rst in vivo experimental evidence of

Ó 2000 Scandinavian Physiological Society 627 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

cytoplasm at submaximal WR's (50% leg VO2max) and remained 8±10-fold higher even at WRMAX. This equates to a PO2 gradient from mean capillary blood to Mb of 39 (normoxia, N) and 29 mmHg (hypoxia, H) at submaximal WR's, which fell slightly to 35 (N) and

27 mmHg (H) at WRMAX.

Constant Mb desaturation despite progressively intense exercise Both intuition and theoretical modelling (Severinghaus

1994) suggest it is reasonable to expect intracellular PO2 to fall as the intensity of exercise increases to enhance

the O2 ¯ux into the muscle cell. Repeated investigations utilizing the isolated knee-extensor model have reported an increase in Mb desaturation from rest to exercise, but have not documented a progressive fall in

intracellular PO2 from moderate to maximum exercise (Richardson et al. 1995b, Richardson et al. 1998) (Figs 5 and 6). However, as Fick's law of diffusion illustrates

[VO2 ˆ DO2 (PcapO2 ± PmitoO2], DO2 also regulates O2 ¯ux. A major determinant of DO2 is the effective sur- Figure 6 Calculated mean capillary PO2 and cellular Mb-associated face area available for diffusion (Federspiel & Popel PO in relation to the percentage of normoxic VO in both normoxia 2 2 1986). If DO2 increases in response to capillary and hypoxia. Note the large difference between the PO2 available in recruitment and increased dynamic haematocrit (Gor- the capillaries and the PO2 at the cellular level, even at submaximal work rates. zynski & Duling 1978, Lindbom et al. 1980, Honig et al. 1982), Mb-associated PO2 need not change with increasing exercise intensity. In fact, consistent with this

interpretation, the calculated DO2 for the present data did increase with each increased WR (Fig. 7) thus,

indicating that O2 conductance is recruited proportionally as VO2 increases, but PMbO2 remains constant after an initial fall during exercise eliciting submaximal VO2. Data that refute this interpretation have recently 3been published (Mole et al. 1999), indicating that, dur- ing progressive plantar ¯exion exercise, Mb appears to desaturate linearly with increasing work rate. Using

these data, the authors conclude that DO2 remains constant and it is in fact the gradient from blood to cell

that increases allowing a greater O2 ¯ux. However, there are several important distinctions that need to be recognized when comparing MRS data collected in two exercise paradigms as different in physiological response as knee-extensor exercise and plantar ¯exion exercise (Richardson 1999). For example, at maximal exercise Mole et al. (1999) reported a pH indicative of a [H+] of 135 nEq L±1 vs. knee-extensor study values of Figure 7 The relationship between VO and muscle O conductance. 2 2 269 nEq L±1 at maximal exercise. This is undoubtedly a Until the measurement of Mb-associated PO2 was achieved this result of the muscle groups studied and not NMR analysis of diffusivity was only possible at VO2max, where mitoc- hondrial PO2 was assumed to be very close to zero and hence could be methodology as we too have found similar pH values in ignored. the gastrocnemius during maximal effort of around 6.9 (Hogan et al. 1999), while we have recently again a steep O2 gradient from blood to intracellular sites in recorded pH values of 6.47 ‹ 0.16 in untrained human skeletal muscle in vivo. These independent subjects during maximal knee-extensor exercise. With numerical estimates reveal that O2 pressure in the the size of surface coil used by Mole et al. (1999) blood during exercise was 11±12-fold higher than in the (5 in. diameter), it is most likely that a signi®cant

628 Ó 2000 Scandinavian Physiological Society Acta Physiol Scand 2000, 168, 623±634 K E Conley et al. Á Myoglobin function in striated muscle contribution to the signal was attained from the soleus PO2 appears to result in diminishing returns with as well as the gastrocnemius (both myoglobin and respect to an increase in VO2max. These observations phosphorus). In this regard, it has previously been are consistent with cellular metabolism that is moving demonstrated that line width for the PI, which re¯ects toward a transition between O2 supply as a determinant relative homogeneity of the pH values within the tissue of VO2max and O2 demand as a determinant of VO2max. being observed, can alter signi®cantly during progres- Further increases in intracellular PO2, beyond those sive plantar ¯exion exercise indicating functional het- recorded in hyperoxia, would be anticipated to have erogeneity (Barstow et al. 1994). This heterogeneity has smaller effects upon VO2max until a plateau is reached previously been reported as an important limitation in and VO2max becomes invariant with intracellular PO2. the interpretation of MR data in the calf muscle From this point intracellular PO2 is no longer a deter- (Vanderbourne et al. 1991). This potential overlap into minant of skeletal muscle VO2max. This hyperbolic a muscle rich in slow-twitch ®bres (soleus) is addi- relationship, originating from the origin, between O2 tionally supported by the observation that at maximum tension and cellular respiration is in agreement with work levels the percentage phosphagen shifts reported data previously described by Wilson et al. (1977) in by Mole et al. (1999) were the same as previously kidney cells. We again suggest (Richardson et al. 1995a), observed at only 40±50% of aerobic maximum although now with more conclusive data, that these (Richardson et al. 1995b, Allen et al. 1997). ®ndings may represent the hyperbolic relationship

between in vivo muscle VO2 and intracellular PO2, sup- porting the concept that maximal respiratory rate Intracellular PO2 as a determinant of VO2max (VO2max) in trained skeletal muscle is normally limited Intracellular PO2 measurements have contrasted the by O2 supply. linear relationship between mean capillary and intra- It is also important to recognize that although the cellular PO2 and the hyperbolic relationship between magnitude of the intracellular PO2 changes reported intracellular PO2 and VO2max (Richardson et al. 1999) here are small, they appear to have biological signi®- (Fig. 8). This suggests that in hyperoxia there is the cance, based upon the observation that VO2max fell expected rise in intracellular PO2 (owing to increased 35% from hyperoxia to hypoxia (Fig. 8). This raises the mean capillary PO2), but this elevated O2 availability is issue of the critical PO2 (PO2crit), below which maximal now in excess of mitochondrial capacity (Fig. 8). Evi- mitochondrial rate is compromised. Previously, using dence that intracellular PO2 is a determinant of VO2max Mb cryomicrospectroscopy in dog gracilis muscle, when FiO2 is manipulated from 0.12 to 0.21 and 1.0. Connett et al. (1983, 1984, 1986) were unable to ®nd However, in the latter case the increased intracellular loci with a PO2 of less than 2 mmHg, but elevated blood lactate levels in the muscle ef¯uent. As previous investigations (Chance & Quistorff 1978) suggested

that PO2crit may be between 0.1 and 0.5 mmHg, Con- nett et al. (1984) concluded that elevated blood lactate concentrations must be caused by factors other than

simply O2-limited mitochondrial ATP synthesis rate. We have recently supported this conclusion by pro- viding in vivo data in man indicating that average

intracellular PO2 remains above these values even at maximal exercise in hypoxia, with rapidly rising lactate 4production (Richardson et al. 1998a) (Fig. 9). With the recognition that muscle lactate production may not be 5the result of cellular hypoxia (Richardson et al. 1998a),

the present data are suggestive of a much higher PO2crit in vivo in exercise trained human skeletal muscle as maximal mitochondrial metabolic rate appears to be

signi®cantly compromised when intracellular PO2 falls from a level around 4 mmHg (Fig. 8)

Figure 8 In vivo measurements of the relationship between mitoc- Intracellular PO vs. mitochondrial PO hondrial O2 uptake and intracellular PO2. Mitochondrial VO2 was 2 2 calculated based upon an assumed mitochondrial ®bre volume of At least two thought provoking questions are raised 7.5%, myo®bril volume of 80%, and muscle density of 1.06 g cc±1. Muscle mass was 2.5 kg, calculated from anthropometric measure- based upon the recent assessment of intracellular PO2 ments. using Mb desaturation: (1) Why at normoxic or

Ó 2000 Scandinavian Physiological Society 629 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

sation of high intensity exercise. In this scenario, a gradient exists from capillary to cytosol (»30 mmHg) and from cytosol to mitochondria (»2 mmHg). It should be noted that although the gradient is vastly different in each case the physiological signi®cance may well be of equal importance.

Lactate production and intracellular oxygen availability There is considerable circumstantial evidence to sup- port the notion that lactate production is related to

inadequate O2 availability during exercise (Wasserman 1984, Katz & Sahlin 1988). However, recent 31P MRS demonstrate no difference in glycolytic ¯ux, and therefore lactate production, in exercising human muscle under normoxic vs. ischaemic conditions 6(Conley et al. 1998). This result and the recent use of proton MRS to detect Mb desaturation indicates that

Figure 9 Net muscle lactate ef¯ux and intracellular PO2 as a function lactate generation occurs under fully aerobic conditions of O2 consumption in normoxia and hypoxia. (Richardson et al. 1998a). Earlier evidence in support of aerobic lactate generation (Jobsis & Stainsby 1968) is + hyperoxic VO2max did the Mb-associated PO2 not fall to the ®nding of the similar oxidation of NADH/NAD the level reached in hypoxia, and (2) why in each of the at rest and in lactate-producing muscle. One would three varied FiO2 was the Mb desaturation far less at expect a reduction in the members of the respiratory maximum exercise than under conditions of cuff chain (including the NADH/NAD+ pair) to coincide occlusion? A possible explanation may be found by with increased lactate production if lactate output was attempting to reconcile our Mb-associated data with the caused by O2-limited oxidative phosphorylation. With recent measurements of cytochrome a,a3 oxidation± another approach, Mb cryomicrospectroscopy in dog reduction state in exercising skeletal muscle (Duhay- 7gracilis muscle, Connett et al. (1983, 1984, 1986) were longsod et al. 1993). These data illustrated a progressive unable to ®nd loci with a PO2 of less than 2 mmHg. As decrease in the concentration of oxidized cytochrome previous investigations (Chance & Quistorff 1978) a,a3 (which correlated highly with rising muscle lactate suggested that the critical PO2 (PO2crit), below which ef¯ux), with increasing muscle O2 extraction. At maximal mitochondrial rate is compromised (between VO2max the magnitude of this redox response was 0.1 and 0.5 mmHg), Connett et al. (1984) concluded the equivalent to that observed at death or complete elevated La concentration must be caused by factors anoxia, suggesting that near depletion of O2 at the other than simply O2-limited mitochondrial ATP syn- mitochondrial level accompanies maximal exercise thesis rate. Our Mb-associated PO2 data support and intensities (Duhaylongsod et al. 1993). These ®ndings extend these latter observations by providing in vivo data are in stark contrast to our Mb-associated PO2 data in man and suggest that average intracellular PO2 which revealed a constant intracellular PO2 (which remains above these previously reported values for did not correlate with rising muscle lactate ef¯ux PO2crit even at maximal exercise in hypoxia (Richardson (Richardson et al. 1998) (Fig. 9), with increasing work et al. 1998). intensity. However, a reconciliation of these data is With respect to the concept of the `anaerobic' possible by approaching them with similar logic threshold these data demonstrate that during incre- employed to explain the observation that there is a large mental exercise skeletal muscle cells do not become gradient from blood to cell and venous PO2 (repre- `anaerobic' as lactate levels suddenly rise, because sentative of end capillary PO2), does not fall to zero intracellular PO2 is well preserved at a constant level, even at VO2max (Wagner et al. 1990, Richardson et al. even at maximal exercise (Fig. 9). Thus our data illus- 1995b): The concept that there is a ®nite O2 conduc- trate the lack of a relationship between intracellular tance (DO2) and that this may limit O2 transport PO2, lactate ef¯ux and muscle pH. However, the (Wagner 1992). Thus a mitochondrial DO2, which limits observation that in hypoxia intracellular PO2 and O2 conductance from the cytosol to mitochondria, may muscle VO2max are reduced and muscle lactate ef¯ux is explain both the difference in oxygen availability out- accelerated leaves open the possibility that intracellular side and within the mitochondria as well as the inability PO2 may still play a role in modulating muscle metab- for Mb-PO2 to fall to a greater extent before the ces- olism and ultimately muscle fatigue.

630 Ó 2000 Scandinavian Physiological Society Acta Physiol Scand 2000, 168, 623±634 K E Conley et al. Á Myoglobin function in striated muscle

Following birth, myoglobinless mice had no appar- MYOGLOBINLESS MICE ent phenotype (Garry et al. 1998, Godecke et al. 1999). Transgenic technologies have provided scientists They grew normally, and reached sexual maturity and with unique opportunities to study the effects of gain- were fertile at an age comparable with that of their wild- of-function or loss-of-function on animal development type and heterozygous littermates. Myoglobin knockout and performance. In the case of the former, the mice displayed no obvious signs of cardiac or skeletal introduction of a transgene that results in the over- muscle dysfunction, and survived at rates similar to expression of a particular protein can provide impor- those for normal mice. tant information not only about the normal function of that protein, but also about its potential role in the Stressing the systems development of or protection against a particular dis- ease process. As an example, over-expression of the The question then arose as to whether myoglobinless inducible 70-kDa heat shock or stress protein, HSP72, mice would be able to meet the demands of stresses in transgenic mice has been shown to have a cardio- involving a need for increased oxygen transport. To protective effect against ischaemia and reperfusion address this, myoglobin-de®cient and wild-type mice injury (Radford et al. 1996). Likewise, eliminating a were evaluated in a facility in the Department of protein through gene targeting or knockout techniques Molecular and Cellular Physiology at the University of can provide valuable information about the predicted Cincinnati that was designed especially for the testing role of the protein, as well as unexpected ®ndings of of transgenic mice (Garry et al. 1998). The mice were unanticipated functions. For instance, a recent study by subjected to two protocols, one involving exercise Chemelli et al. (1999) found that targeted disruption of performance on a motor-driven treadmill and the other the orexin gene, originally being studied for its effects on a challenge to a hypoxic gas mixture. In both cases, the energy homeostasis, resulted in a phenotype with myoglobin knockout mice responded in a manner that marked similarity to that of human patients with nar- was indistinguishable from their wild-type counterparts. colepsy. Thus, although the results of studies of Endurance performance at high levels of exercise transgenic or knockout animals are often `predictable' intensity was identical between the two groups, as were or con®rmatory in nature, many others have been the ventilatory responses to the hypoxic challenge surprising, providing valuable insights into mechanisms (13.5% O2). regulating normal and abnormal physiology. Recent Studies of isolated cardiac and skeletal muscle reports describing the effects of the loss of myoglobin function yielded similar results. Hearts isolated from in mice are examples of the latter (Garry et al. 1998, myoglobin-de®cient mice displayed similar indices of Godecke et al. 1999). They present intriguing ®ndings function as those from wild-type mice when challenged that depict the apparent lack of effect of the loss of this with increases in preload (Garry et al. 1998) or haemoprotein, as well as potential adaptive responses increasing doses of dobutamine (Godecke et al. 1999 that provide clues regarding the role of myoglobin in and unpublished results). Likewise, soleus and extensor normal muscle function. digitorum longus skeletal muscles isolated from myo- globin knockout and wild-type mice had identical pat- terns of fatigue (Garry et al. 1998). Interestingly, the No apparent phenotype for myoglobin knockouts similarity of responses for both cardiac and skeletal `Myoglobinless' mice were created using targeting muscle was independent of the degree of oxygenation strategies that deleted exon 2 of the myoglobin gene of the solutions perfusing or bathing the isolated hearts (Garry et al. 1998, Godecke et al. 1999). The deleted or muscles (Garry et al. 1998). region encodes nearly one-half of the 154 amino acids that comprise the myoglobin protein, including the Some adaptations and explanations essential heme-binding domain. Disruption of the myoglobin gene was con®rmed by Southern blot Given the characteristics and regulation of myoglobin analysis of digested genomic DNA and by the absence described in the ®rst section of this paper, these of RNA transcripts that included exon 2 as detected by results from myoglobinless mice appear to hinder rather reverse-transcription/polymerase chain reaction assay. than enable a deciphering of the mysteries of this In addition, myoglobin protein could not be detected in protein. The surprising and perhaps unexpected results cardiac or skeletal muscles of myoglobin-de®cient mice seem to indicate that myoglobin plays little if any role either by immunohistochemical assays or by Western in normal cardiac and skeletal muscle function. analysis. Further, the muscles of the myoglobin-de®- Alternatively, there may be subtle adaptive mecha- cient mice were pale, re¯ecting the complete lack of nisms that promote and permit viability and normal pigmentation normally provided by this protein. function in the absence of myoglobin. Supporting the

Ó 2000 Scandinavian Physiological Society 631 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

latter, there also were a number of adaptive changes might reveal impairment in exercise capacity or in noted that might help explain the apparent lack of cardiac or skeletal muscle function in the absence of

effect of eliminating this O2-binding protein. First, this protein. Finally, the apparent adaptive responses myoglobin-de®cient mice were reported to have displayed by the myoglobin-de®cient mice allow these increased haemoglobin levels (Godecke et al. 1999). animals to function normally, and demonstrate the Although modest, the altered haemoglobin levels remarkable ability of organisms to compensate for

would increase the arterial O2 carrying capacity by disruptions in normal structure or function. The tissue, »8%. Second, the myoglobin knockout mice appeared cellular, and molecular differences already noted to have increased myocardial capillary density between the myoglobin null and wild-type mice indicate (Godecke et al. 1999). This increase in capillary density that, although perhaps not necessary for normal muscle

would thus reduce the O2 diffusion distance between function, this conserved haemoprotein may contribute red blood cell and muscle ®bre membrane. Third, the to O2 delivery and/or storage. Equally importantly, myoglobin-de®cient mice were reported to have myoglobin may have other essential functions such as increased coronary ¯ow and increased coronary ¯ow serving as a buffer for oxygen-derived free radicals or reserve compared with their wild-type counterparts nitric oxide. (Godecke et al. 1999). This would result in an increase in overall blood ¯ow and O delivery both at rest and 2 The value of transgenic and knockout animals during periods of elevated cardiac function. Taken together, these adaptive changes would increase the Clearly, transgenic technologies will not provide

PO2 gradient from red blood cell to mitochondrion answers to all fundamental questions in biology. What and thus perhaps offset the absence of any intracel- they do provide, however, is a powerful tool with which

lular myoglobin-mediated O2 diffusion. In addition, scientists can begin to address important biological microarray analysis of cardiac muscle from myoglobin questions within an integrative framework. Transgenic null and wild-type mice showed dysregulation of gene and knockout animals are valuable models that can be

expression supporting a molecular adaptive process in used to answer new and old questions alike about O2 mice that lack myoglobin (D. Garry, unpublished delivery and utilization. In addition, they allow scientists 8results). opportunities to re-examine previous studies and There are a number of possible conclusions that experiments in a new setting. For example, earlier might be drawn from the ®ndings of these studies of studies using carbon monoxide or chemical reagents myoglobinless mice. One, which is based on the results (Wittenberg & Wittenberg 1975, Doeller & Wittenberg of the exercise and isolated heart and skeletal muscle 1990) that showed an important functional role for studies, is that myoglobin plays little or no role in myoglobin might be revisited using myoglobin-de®cient normal cardiac and skeletal muscle function. Clearly, mice. Regardless of which of the above conclusion(s) the intact knockout mice, as well as the hearts and concerning myoglobinless mice is correct, these unique skeletal muscles isolated from them, responded animals have provided new insights and impetus to our

robustly to the stresses imposed upon them and were studies of O2 transport. able to function normally. A second possibility is that the knockout mice in these studies were not stressed CONCLUSIONS suf®ciently to see the effects of a lack of myoglobin. Studies in which myoglobin-de®cient mice are exposed A simple diffusion model provides insight into the role

acutely or chronically to extreme altitude and/or are of Mb in intracellular O2 supply. First, the need for exercised at altitude would test this hypothesis. How- parallel diffusion by O2 and Mb-O2 was demonstrated ever, the fact that isolated heart and skeletal muscle by showing that muscle ®bres with a diffusion limita-

function was identical between knockout and wild-type tion to O2 supply typically have higher Mb content than mice both in the presence and absence of O2 provides do ®bres with no O2 diffusion limitation. Thus, the strong evidence against this possibility. A third con- higher capillary PO2, smaller muscle ®bre size, and clusion is that the ®ndings are unique to mice, a small larger capillary density in small vs. large mammals

mammal endowed with relatively low levels of myo- results in a greater capacity for O2 diffusion alone with globin (see Mechanisms of myoglobin function). This is not less need for parallel diffusion by Mb. Direct meas- unique to the myoglobin protein, but rather is a con- urement of Mb saturation by magnetic resonance

cern for the vast majority of transgenic and knockout demonstrates that Mb-O2 does desaturate with exercise. experiments in which mice are the animal model. Very These results show that the cellular PO2 is low enough simply, mice may not be a good model for the study of for unloading of O2 from Mb-O2 and therefore that O2 transport. Studies of larger mammals with greater parallel diffusion of O2 is possible. Further, these MRS O2 diffusion distances and higher levels of myoglobin results show that, on average, the muscle does not

632 Ó 2000 Scandinavian Physiological Society Acta Physiol Scand 2000, 168, 623±634 K E Conley et al. Á Myoglobin function in striated muscle become anoxic even at exercise levels eliciting maxi- Conley, K.E., Kushmerick, M.J. & Jubrias, S.A. 1998. mum oxygen consumption. Thus, the lactate generation Glycolysis in independent of oxygenation state in characteristic of exercise at the aerobic maximum does stimulated human skeletal muscle in vivo. J Physiol 511 (3), not re¯ect cellular anoxia, but rather a higher rate of 935±945. glycolytic ¯ux than is needed to sustain oxidative Connett, R.J., Gayeski, T.E.J. & Honig, C.R. 1983. Lactate production in a pure red muscle in absence of anoxia: phosphorylation. Finally, the maintenance of aerobic Mechanisms and signi®cance. Adv Exp Med Biol 159, function in the myoglobin knockout mouse indicates 327±335. that there is no obligate link between Mb and oxidative Connett, R.J., Gayeski, T.E.J. & Honig, C.R. 1984. Lactate phosphorylation. The scaling of muscle and cardio- accumulation in fully aerobic, working, dog gracilis muscle. vascular properties indicates that Mb is expected to Am J Physiol 246, H120±H128. show a smaller role in O2 supply in mice than in the Connett, R.J., Gayeski, T.E.J. & Honig, C.R. 1986. Lactate muscles of larger mammals. More interesting, however, ef¯ux is unrelated to intracellular PO2 in working red is the possibility that changes in the underlying muscle skeletal muscle in situ. J Appl Physiol 61, 402±408. ®bre properties during development without myoglo- Doeller, J. & Wittenberg, B. 1990. Intracellular calcium and high-energy phophates in isolated cardiac myocytes. bin may eliminate an O2 supply de®cit and minimize the need for Mb-mediated O supply. Thus, a simple Am J Physiol 259, H1851±H1859. 2 Duhaylongsod, F.G., Griebel, J.A., Bacon, D.S., Wolfe, W.G. analysis of the intracellular diffusion problem not only & Piantadosi, C.A. 1993. Effects of muscle contraction on helps to understand the physical basis of Mb-mediated cytochrome a,a3 redox state. J Appl Physiol 75, 790±797. O2 supply but also to predict the changes in muscle Federspiel, W.J. & Popel, A.S. 1986. A theoretical analysis of properties necessary for full aerobic function in the the effect of the particluate nature of blood on oxygen absence of Mb. release in capillaries. Microvasc Res 32, 164±189. Garry, D.J., Ordway, G.A., Lorenz, J.N. et al. 1998. Mice Thanks go to Cliff Jones, John S. Leigh, Marie Matty, Elizabeth Noyszewski, and Peter Wagner for their contributions. This work was without myoglobin. Nature 395 (6705), 905±908. supported by US NSF grants IBN 93±06596 and IBN 96±04698; US Godecke, A., Flogel, U., Zanger, K. et al. 1999. Disruption of Public Health Service grants AR-41928, AG-10853, AR-45184, HL- myoglobin in mice induces multiple compensatory 17731 and RR-02305; and Parker B. Francis Fellowship. Travel and mechanisms. Proc Natl Acad Sci USA 96 (18), 10495±10500. accommodations were supported by the Nordic Physiological Society. Gorzynski, R.J. & Duling, B.R. 1978. Role of oxygen in functional vasodilation in hamster striated muscle. Am J Physiol 235, H505±H515. REFERENCES Hogan, M.C., Richardson, R.S. & Haseler, L. 1999. Human Allen, P., Matheson, G., Zhu, G. et al. 1997. Simultaneous 31P muscle performance and PCr hydrolysis with varied MRS of soleus and gastrocnemius in Sherpes during graded inspired oxygen fractions: A 31P-MRS study. J Appl Physiol calf muscle exercise. Am J Physiol 273, R999±R1007. 86, 1367±1374. Andersen, P., Adams, R.P., Sjogaard, G., Thorbe, A. & Saltin, Honig, C.R. & Gayeski, T.E.J. 1993. Resistance to O2 B. 1985. Dynamic knee extension as a model for study of diffusion in anemic red muscle: Role of ¯ux density and cell isolated exercising muscle in humans. J Appl Physiol 59 (5), PO2. Am J Physiol 265, H868±H875. 1647±1653. Honig, C.R., Odoroff, C.L. & Frieson, J.L. 1982. Active and Armstrong, R.B., EsseÂn-Gustavsson, B., Hoppeler, H. et al. passive control in red muscle at rest and in exercise. Am J Physiol 243, H196±H206. 1992. O2 delivery at VO2max and oxidative capacity in muscles of Standardbred horses. J Appl Physiol 73, Hoppeler, H., Mathieu, O., Krauer, R., Claassen, H., 2274±2282. Armstrong, R. & Weibel, E.R. 1981. Design of the Barstow, T., Buchthal, S., Zanconato, S. & Cooper, D. 1994. mammalian respiratory system. VI. Distribution of Changes in potential controllers of human skeletal muscle mitochondria and capillaries in various muscles. respiration during incremental calf exercise. J Appl Physiol Respir Physiol 44, 87±111. 77, 2169±2176. Hudlicka, O., Egginton, S. & Brown, M.D. 1988. Capillary Bentley, T.B. & Pittman, R.N. 1997. In¯uence of temperature diffusion distances-their importance for cardiac and skeletal on oxygen diffusion in hamster retractor muscle. muscle performance. NIPS 3, 134±138. Am J Physiol 272, H1106±H1112. Jobsis, F.F. & Stainsby, W.N. 1968. Oxidation of NADH Chance, B. & Quistorff, B. 1978. Study of tissue oxygen during contractions of circulated skeletal muscle. gradients by single and multiple indicators. Adv Exp Med Respir Physiol 4, 292±300. Biol 94, 331±338. JuÈrgens, K.D., Peters, T. & Gros, G. 1994. Diffusivity of Chemelli, R.M., Willie, J.T., Sinton, C.M. et al. 1999. myoglobin in intact skeletal muscle cells. Proc Natl Acad Sci, Narcolepsy in orexin knockout mice: Molecular genetics of USA 91, 3829±3833. sleep regulation. Cell 98 (4), 437±451. Katz, A. & Sahlin, K. 1988. Regulation of lactic acid Conley, K.E. & Jones, C. 1996. Myoglobin content and production during exercise. J Appl Physiol 65, 509±518. oxygen diffusion: Model analysis of horse and steer muscle. Kayar, S.R. & Weiss, H.R. 1992. Diffusion distances, total Am J Physiol 271, C2027±C2036. capillary length and mitochondrial volume in pressure-

Ó 2000 Scandinavian Physiological Society 633 Myoglobin function in striated muscle Á K E Conley et al. Acta Physiol Scand 2000, 168, 623±634

overloaded myocardial hypertrophy. J Mol Cell Cardiol 24, Richardson, R.S., Noyszewski, E.A., Leigh, J.S. & Wagner, P.D. 1115±1166. 13 1998a. Lactate ef¯ux from exercising human skeletal muscle:

Kayar, S.A., Hoppeler, H., Jones, J.H. et al. 1994. Capillary Role of intracellular PO2. J Appl Physiol 85, 627±634. blood transit time in muscles in relation to body size and 14Richardson, R.S., Frank, R.L. & Haseler, L.J. 1998b. Dynamic aerobic capacity. J Exp Biol 194, 69±81. knee-extensor and cycle exercise: Functional MRI of Kreuzer, F. 1970. Facilitated Diffusion of Oxygen and its muscular activity. Int J Sports Med 19, 182±187. Possible Signi®cance; A Review. Respir Physiol 9, 1±30. Richardson, R.S., Poole, D.C., Knight, D.R. et al. 1993. High

Lawrie, R. 1953. The activity of the cytochrome system muscle blood ¯ow in man: Is maximal O2 extraction in muscle and its relation to myoglobin. Biochem J 55, compromised? J Appl Physiol 75 (4), 1911±1916. 298±305. Schmidt-Neilsen, K. & Larimer, J.L. 1958. Oxygen Lindbom, L., Tuma, R.F. & Arfors, K.E. 1980. In¯uence dissociation curves of mammalian blood in relation to body of oxygen on perfused capillary density and capillary red size. Am J Physiol 195, 424±428. cell velocity in rabbit skeletal muscle. Microvasc Res 19, Seeherman, H.J., Taylor, C.R., Maloiy, G.M. & Armstrong, 197±208. R.B. 1981. Design of the mammalian respiratory system. II Mathieu, O., Krauer, R., Hoppeler, H. et al. 1981. Design of Measuring maximum aerobic capacity. Respir Physiol 44 (1), the mammalian respiratory system. VII. Scaling 11±23.

mitochondrial volume in skeletal muscle to body mass. Severinghaus, J.W. 1994. Exercise O2 transport model Respir Physiol 44 (1), 113±128. assuming zero cytochrome PO2 at VO2max. J Appl Physiol 77, Meyer, R.A., Sweeney, H.L. & Kushmerick, M.J. 1984. A 671±678. simple analysis of the `phosphocreatine shuttle' Am J Physiol Thomas, A. & Morris, P.G. 1981. The effects of NMR 246, C365±C377. exposure on living organisms. A microbial assay. Br J Radiol MoleÂ, P.A., Chung, Y., Tran, T.K., Sailasuta, N., Hurd, R. & 54, 615±621. Jue, T. 1999. Myoglobin desaturation with exercise intensity Vanderbourne, K., McCully, K., Kakihira, H. et al. 1991. in human gastrocnemius muscle. Am J Physiol 277, Metabolic heterogeneity in human calf muscle during 9 R173±R180. maximal exercise. Proc Natl Acad Sci USA 88, 5714±5718. Noyszewski, E.A., Chen, E.L., Reddy, R., Wang, Z. & Leigh, Wagner, P.D. 1992. Gas exchange and peripheral diffusion J.S. 1997. A simpli®ed sequence for observing limitation. Med Sci Sports Exer 24 (1), 54±58. deoxymyoglobin signals in vivo: Myoglobin excitation with Wagner, P. 1993. Algebraic analysis of the determinants of

dynamic unexcitation and saturation of water and fat. Mag VO2max. Respir Physiol 93, 221±237. Res Med 38, 788±792. Wagner, P.D., Roca, J., Hogan, M.C., Poole, D.C., Debout, Papadopoulos, S., Jurgens, K.D. & Gros, G. 1995. Diffusion D.E. & Haab, P. 1990. Experimental support of the theory of myoglobin in skeletal muscle cells ± dependence on ®bre of diffusion limitation of maximum oxygen uptake. In: type, contraction and temperature. P¯ugers Arch 430 (4), Oxygen Transport to Tissue, pp. 825±833. XII. Plenum Press, 519±525. New York, NY. Popel, A.S. 1989. Theory of oxygen transport to tissue. Wang, Z., Noyszewski, E.A. & Leigh, J.S. 1990. In vivo MRS Crit Rev Biomed Eng 17, 257±321. measurement of deoxymyoglobin in human forearms. Mag Radergran, G. 1997. Ultrasound Doppler estimates of femoral Res Med 14, 562±567. artery blood ¯ow during dynamic knee-extensor exercise in Wang, D., Kreutzer, U., Chung, Y. & Jue, T. 1997. Myoglobin humans. J Appl Physiol 83, 1383±1388. and hemoglobin rotational diffusion in the cell. Biophys J 73 Radford, N.B., Fina, M., Benjamin, I.J. et al. 1996. (5), 2764±2770. Cardioprotective effects of 70-kDa heat shock protein in Wasserman, K. 1984. The anaerobic threshold measurement transgenic mice. Proc Natl Acad Sci USA 93 (6), 2339±2342. to evaluate exercise performance. Am Rev Resp Dis 61,

Richardson, R.S. 2000. Intracellular PO2 and bioenergetic 260±270. measurements in skeletal muscle: The role of exercise Wilson, D.F., Erecinska, M., Drown, C. & Silver, I.A. 1977. 10 paradigm. Am J Physiol 278, in press. Effect of oxygen tension on cellular energetics. Am J Physiol Richardson, R.S., Leigh, J.S., Wagner, P.D. & Noyszewski, E. 233, C135±C140.

1999. Cellular PO2 as a determinant of maximal Wittenberg, J.B. 1970. Myoglobin-facilitated oxygen diffusion: mitochondrial O2 consumption in trained skeletal muscle. role of myoglobin in oxygen entry into muscle. Physiol Rev J Appl Physiol 87, 325±331. 50, 559±636. Richardson, R.S., Noyszewski, E.A., Kendrick, K.F., Leigh, Wittenberg, B.A. & Wittenberg, J.B. 1975. Role of Myoglobin

11 J.S. & Wagner, P.D. 1995a. Myoglobin O2 desaturation in the oxygen supply to red skeletal muscle. J Biol Chem 250, during exercise: Evidence of limited O2 transport. J Clin 9038±9043. Invest 96, 1916±1926. Wittenberg, J. & Wittenberg, B. 1989. Transport of oxygen in 12Richardson, R.S., Knight, D.R., Poole, D.C. et al. 1995b. muscle. Annu Rev Physiol 51, 857±878.

Determinants of maximal exercise VO2 during single leg Wittenberg, J.B. & Wittenberg, B.A. 1990. Mechanisms of knee extensor exercise in humans. Am J Physiol 268, cytoplasmic hemoglobin and myoglobin function. Annu Rev H1453±H1461. Biophys Chem 19, 217±241.

634 Ó 2000 Scandinavian Physiological Society Biophysical Chemistry 98 (2002) 127–148

Myoglobin as a model system for designing heme protein based blood substitutes

Yi Doua,b,1a,ba,b , David H. Maillett , Raymund F. Eich,2a,b , John S. Olson, *

aDepartment of Biochemistry and Cell Biology, Rice University, Houston, TX 77005, USA bW.M. Keck Center for Computational Biology, Rice University, Houston, TX 77005, USA

Received 13 September 2001; received in revised form 19 December 2001; accepted 19 December 2001

Abstract

The ligand binding properties and resistances to denaturation of )300 different site-directed mutants of sperm whale, pig, and human myoglobin have been examined over the past 15 years. This library of recombinant proteins has been used to derive chemical mechanisms for ligand binding and to examine the factors governing holo- and apoglobin stability. We have also examined the effects of mutagenesis on the dioxygenation of NO by MbO2 to form y NO3 and metMb. This reaction rapidly detoxifies NO and is a key physiological function of both myoglobins and hemoglobins. The mechanisms derived for O2 binding and NO dioxygenation have been used to design safer, more ( efficient, and more stable heme protein-prototypes for use as O2 delivery pharmaceuticals in transfusion therapy i.e. blood substitutes). An interactive database is being developed (http:yyolsonnt1.bioc.rice.eduywebymyoglobinho- me.asp) to allow rapid access to the ligand binding parameters, stability properties, and crystal structures of the entire set of recombinant myoglobins. The long-range goal is to use this library for developing general protein engineering principles and for designing individual heme proteins for specific pharmacological and industrial uses. ᮊ 2002 Elsevier Science B.V. All rights reserved.

Keywords: Mb engineering; Heme proteins; Blood substitutes

1. Introduction protein in striated muscle to provide a continuous

supply of O2 to the terminal mitochondrial oxidase. Mammalian myoglobin functions as a storage During muscle relaxation, red blood cells circulat-

ing through capillary beds provide O2 for both *Corresponding author. Department of Biochemistry and oxidative phosphorylation and oxygenation of Cell Biology, MS 140, Rice University, 6100 S. Main, Houston, TX 77005, USA. Tel.: q1-713-348-4910; fax: q1-713-348- myoglobin. During contraction, blood flow is inter- 5154. rupted and O2 is supplied to the mitochondria by ( ) E-mail address: [email protected] J.S. Olson . release from myoglobin. In order to perform this 1 Present address: Tanox, Inc., 10301 Stella Link Rd., Suite 500, Houston, TX 77025, USA. storage–delivery function, myoglobin has evolved 2 Present address: Williams Morgan & Amerson, 7676 a relatively high affinity for O250 . It has a P or f Hillmont, Ste. 250, Houston, TX 77040, USA. Kd 1 mM, which is intermediate between that of

0301-4622/02/$ - see front matter ᮊ 2002 Elsevier Science B.V. All rights reserved. PII: S0301-4622Ž 02. 00090-X 128 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

( f ) hemoglobin in red cells P50 20–30 mM and elevation, vascular lesions, and gastrointestinal ( ) w x the KM of cytochrome c oxidase ;0.01 mM . dysmotility 23–26 . We and the group at Baxter Myoglobin discriminates in favor of O2 binding Hemoglobin Therapeutics have argued that all of and against CO binding in order to function in the these side effects are due to interference with NO presence of the low levels of carbon monoxide signaling pathways w18,27–29x. Using the mecha- that are produced as a result of heme catabolism, nism of NO dioxygenation as a guide, we were neuronal function, and cell signaling w1x. Oxymy- able to design Mb prototypes with 20-fold lower oglobin also serves to detoxify NO by rapidly rates of NO scavenging. These ideas were trans- oxidizing it to nitrate w2–4x. Brunori and others ferred to Baxter Hemoglobin Therapeutics (for- have argued that this function is required to prevent merly Somatogen, Inc.) for testing in vivo, using NO inhibition of both aconitase and cytochrome recombinant human hemoglobins with the same oxidase activity in myocytes and in the last year sets of mutations w3,28x. A similar strategy for has promoted the idea that myoglobin is a ‘pseudo- reducing the NO reactivity of recombinant human enzymatic’ scavenger of nitric oxide w5–7x. Oxy- hemoglobin has been reported by Brunori’s group w x hemoglobin has been shown to perform a similar 30 . The O2 transport properties of these mutants detoxification function for inhaled nitric oxide w8x. have to be optimized by second site replacements

Between 1985 and 1994, five groups developed in order to increase O2 release rates and decrease vectors for expressing large quantities of recom- O2 affinity. The strategies for reducing NO scav- binant human hemoglobin and human, sperm enging and for increasing O2 transport are pre- whale, pig, and horse heart myoglobins in Escher- sented in the second half of Section 3. The ichia coli w9–13x. These expression systems have compromises and problems involved in the design been used to construct site-directed mutants that and production of heme protein-based blood sub- are designed to evaluate the structural factors that stitutes are presented in Section 4. regulate O2 affinity, ligand discrimination, autoox- idation, and the rate constants for oxygen associ- 2. Methods ation and disassociation (for an early review see Springer et al. w14x). This approach has also been 2.1. Recombinant myoglobin expression systems used to examine the oxidative reaction of NO with

MbO22 and HbO . A summary of the mechanisms Three different strategies have been used to for O2 binding and NO dioxygenation are present- express mammalian myoglobins in E. coli. The ed in the first part of Section 3. first strategy was based on methods developed by Mutagenesis approaches have also been used to Nagai’s group to express and reconstitute the determine the factors that govern holomyoglobin subunits of human hemoglobin w9x. Boxer’s group denaturation, heme loss, apoglobin unfolding and used this strategy to express human myoglobin as expression levels of intact myoglobin in E. coli a fusion protein with the N-terminal 31 amino w15–20x. These stability studies led to the devel- acids of cII, a regulatory protein of l phage w10x. opment of Mb-based reagents for measuring the This fusion protein system was also used by rate of hemin loss from holoproteins (i.e. see Wilkinson’s group to produce recombinant pig w21,22x). myoglobin w12x and by Ikeda-Saito, Yi Dou, and Our work on the mutagenesis of myoglobin has coworkers to produce sperm whale myoglobins been driven strongly by commercial development with an N-terminal Val residue w31x. In this system, of extracellular hemoglobin as a blood substitute. the protein is expressed in inclusion bodies, solu- The library of structural and functional data for bilized, reconstituted with hemin, and then the the Mb mutants has served as a resource for leader sequence cleaved with trypsin or Factor Xa. developing general heme protein engineering strat- The second system was developed by Springer egies and algorithms for optimizing multiple prop- and Sligar w11x and involves constitutive expres- erties in product candidates. The major side effects sion of sperm whale holomyoglobin from a plas- of extracellular hemoglobins are blood pressure mid containing the myoglobin gene plus the codon Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 129

Table 1 Ligand binding and NO-induced oxidation parameters for position 29(B10),64(E7), and 68(E11) mutants of sperm whale myoglobin at pH 7, 20 8C

( ) Myoglobin k9O2 kO2 1yKO2P 50 k9NO,ox (mMsy1 y1) (sy1) (mM)(mMsy1 y1) Wild-type 17 15 0.91 34 H64Q 24 130 5.4 ;60a H64F 75 10,000 130 60 L29F 21 1.4 0.067 8.1 L29W 0.25 8.5 34 3.2 V68I 3.2 14 4.4 33 V68L 23 6.8 0.30 24 V68F 1.2 2.5 2.1 9.4 V68W 0.17 0.26 1.5 4.1 L29FyV68F 13 0.17 0.013 2.9 H64FyV68F 13 1700 130 1.8 H64QyV68F 3.0 12 4.0 7.5 H64QyV68W 2.6 12 4.6 2.0 L29FyH64Qb 30 65 2.2 17 L29WyH64Q 12 67 5.6 3.0 L29FyH64QyV68I 5.1 79 15 1.5 L29FyH64QyV68L 61 14 0.23 5.1 L29FyH64QyV68Fc 13 3.4 0.26 13

k9NO,ox is the bimolecular rate constant for NO dioxygenation by MbO2 . Parameters for the single mutants were taken from the literature w3,14,55x. Parameters for some of the multiple mutants were first reported in Olson et al. w18x and Eich w55x. a It has proved difficult to get an accurate value of the rate of NO dioxygenation for Gln64(E7) myoglobin because of the rapid rate of NO binding to the ferric derivative. The value of 60 mMy1 sy1 is our current best estimated, based on reactions in which deoxyMb was varied and mixed with sub-stoichiometric amounts of NO. b Elephant Mb has naturally occurring Gln64 and Phe29 residues in its distal pocket w96x. c Lucina Hb I contains Gln(E7), Phe(B10), and Phe(E11) residues w97x. for an N-terminal methionine residue. In this sys- 2.2. Kinetics of ligand binding and NO tem, E. coli TB1 cells were used and are missing dioxygenation the gene for the lac repressor protein. The bacteria synthesize the required heme, and the holoprotein The new rate constants for the multiple mutants is obtained directly from the bacterial cytoplasm. listed in Table 1 were measured as described in

This constitutive expression system only works previous publications on O2 and NO binding to well with sperm whale myoglobins, which have deoxymyoglobin w32x and NO induced oxidation w x w x an unusually stable apoglobin 20 . In the third 3,33,34 . In general, rate constants for O2 and system, a pET vector is used to overexpress the NO association are measured on microsecond and myoglobin gene containing the starting N-Met milliseconds time scales using room temperature codon. Both holoprotein and unfolded apoglobin laser photolysis techniques and, when possible, are made after induction with IPTG. The latter checked in rapid mixing experiments. Oxygen system was first used to make recombinant horse dissociation rate constants are determined by anal- heart myoglobin w13x, and similar systems have ysis of CO replacement reactions w35x. The equi- ( been developed for labeling sperm whale myoglo- librium dissociation constant listed as P50 in units bin with13 C and 15 N amino acids. of mM) were calculated from the ratio of the 130 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 dissociation, kO2, and association, k9 O2, rate con- greater noise in the data shown in Fig. 8. In Page ( s s ) w x stants i.e. Kd50O2O2P k yk9 . et al.’s 38,40 and our previous measurements w41,42x, the Hb concentration was 1–5 mM. Com- 2.3. Mean arterial blood pressure or total periph- plete descriptions of the 25-30 mm capillary flow eral resistance experiments system and the microspectrophotometer are given in Boland et al. w41x and Page et al. w38,40x.A Blood pressure measurements were taken from 100-mm slab of clear, highly gas permeable sili- the literature (Fig. 5) and are based on work by cone rubber is cast around a 20–30 mm tungsten the research group at Baxter Hemoglobin Thera- wire. After polymerization, the wire is removed peutics (formerly Somatogen, w28,29x) and by and the input and output sides of the capillary are Winslow, Intaglieta, and coworkers at the Univer- cannulated with glass micropipettes. The pipette sity of California, San Diego w36,37x. All experi- on the input side is connected to a reservoir ments were carried out with rats and involved containing a concentrated Hb or Mb solution. The either 10% top load experiments or 40–50% iso- output side is connected to a syringe pump which volemic exchange, and when possible the data are draws the sample through the capillary at well- reported as the change in blood pressure compared defined flow rates. The silicone wafer is placed on to serum albumin controls. In the case of 10% top a stage which can be translated with fine precision load experiments, the albumin controls show small across the light beam of the microscope. The ;5% increases in MAP due to the initial increase transmitted light is split into two beams by fiber blood volume. When controls using human serum optic cables and sent to two photomultiplers, which albumin (HSA) are carried out for 40–50% iso- measure absorbance changes at 415 and 430 nm volemic exchanges, there are significant decreases simultaneously. Release experiments are initiated ( ) in both blood pressure ;10% and peripheral by flowing HbO22 or MbO through the capillary ( ) w x resistance ;20% 29,37 due to large decreases as it is being flushed with N2 . In uptake experi- in blood viscosity. The latter effect could mask the ments, deoxyHb or deoxyMb is used, and the pharmacological activity of the hemoglobin being capillary bed is flushed with air and O22yN examined. Unfortunately, Rohlfs et al. (1998) did mixtures. Absorbance change readings are made at not report serum albumin controls in their 50% 5 to 10 positions along the length of the capillary exchange experiments and only the uncorrected and used to calculate fractional degree of satura- changes in mean arterial blood pressure are report- tion. The residence time is calculated from the ed in Figs. 5 and 6. Regardless, the data for simple longitudinal position and the flow rate w38,40x. cross-linked hemoglobin tetramers (without distal ) pocket mutations show no obvious dependence 3. Results of the change in mean arterial blood pressure on

P50 , in either 10% top load or 40–50% isovolemic exchange experiments (Fig. 6a). 3.1. Mechanisms for O2 binding

2.4. O2 transport More than 300 mutants have been constructed at 50 different positions in myoglobin. The loca-

Measurements of O2 uptake and release by tions of these mutants are shown in Fig. 1 where sperm whale myoglobin mutants in a 25 mm the residues are color-coded by their helix position. diameter artificial capillary were made as described Many of the mutant proteins have multiple replace- by Page, Hellums, and colleagues in work with ments (e.g. Table 1). The substitutions were bovine hemoglobin and red cells w38–40x. The designed to test mechanisms of ligand binding and myoglobin samples were more dilute (; 0.5 mM to improve the properties of the molecule with ) heme due to the large volume required for these respect to O2 transport, stability, and NO sca- experiments (;5ml). The lower absolute absorb- venging. A database for the functional and struc- ance values for the Mb samples accounts for the tural properties of these mutants is being Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 131

binding w14,45,47x. Amino acids with good hydro- gen atom donors (i.e. His, Gln, Asn, Tyr) favor the binding of dioxygen by preferentially stabiliz- ing the highly polar Fed()q –O–Od ()y complex. We have quantified these effects by assigning values to the equilibrium constants for non-cova- lent water binding to deoxymyoglobin, ligand entry into the distal pocket, internal coordination to the iron atom, and electrostatic stabilization of the bound ligand w45x. In the case of mammalian myoglobin, the distal His64(E7) side chain forms a strong H-bond that selectively stabilizes bound

O2 by a factor of ;1000. Maps for O2 movement into and out of the myoglobin have been constructed based on analy- ses of time courses for geminate and bimolecular

O2 binding to 90 myoglobin mutants at 27 different positions w48,49x. We have argued that the protein acts as a baseball glove to ‘catch’ and then ‘trap’ Fig. 1. Structure of sperm whale myoglobin showing the posi- tions that have been mutated and examined. The colors are grouped according to helical positions. Deep green, A-helix; blue, B-helix; bright violet, C- and D-helices and CD corner; orange, E-helix; rose, EF corner, F-helix, and FG corner; light green (background), G-helix; and purple, H-helix. Lists of sin- gle mutants are given in various publications and the website. constructed at http:yyolsonnt1.bioc.rice.eduyweby myoglobinhome.asp.

The structural factors regulating O2 affinity in both hemoglobins and myoglobins are summarized in Figs. 2 and 3. The reactivity of the heme iron atom can be altered dramatically by stereochemical constraints on the proximal side of the heme group. Large changes in ligand affinity are observed when comparing equilibrium constants for ligand binding to the high affinity (R) and low affinity (T) quaternary states of human hemoglobin. The R to T conformational change causes the E and F helices to restrict the ability of the iron atom to move into the plane of the porphyrin ring and Fig. 2. Regulation of O2 affinity and CO discrimination in Mb w x and Hb. The structure of the active side of recombinant sperm react with ligands 43,44 . As a result, this quater- w x whale MbO2 was taken from Quillin et al. 98 . Schematic nary transition causes 700, 1200, and 1000-fold chemical drawings of the active site are shown in Fig. 3A,B. decreases in the equilibrium constants for O2 , CO, The amino acid numbers 28, 29, 32, 46, 64, 68, 93, and 107 and NO binding to human hemoglobin, respective- correspond to the helical positions, B9, B10, B13, CD4, E7, ly (see w45,46x and references therein). E11, F8, and G8, respectively. The amino acids labeled in blue comprise the distal histidine gate, and the amino acids labeled In contrast to the proximal effects, electrostatic in green circumscribe the Xe4 binding site. The factors gov- interactions with polar side chains on the distal erning O2 affinity are labeled in red and taken from Olson and w x side of the heme group selectively regulate O2 Phillips 45 . 132 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 . x 5,6 w ns rature. There . Oxygen binding can be x 3,34,45,48 w 2 22 2 Fig. 3. Comparison ofbroken kinetic down pathways into for fiveare O discrete four steps distinct and binding steps those andFig. for labeled NO 4. NO in dioxygenation. NO dioxygenation. bold, The signalingbased The ligand mechanisms in on first entry were the the step, and taken ideas NO endothelium bond from entry formation, developed and the limit into over the literature the the the roles rate distal past of constant pocket, 5 HbO for is years O rate on and limiting the association MbO at physiological at role neutral room in of pH tempe rapid detoxifying and NO room NO dioxygenation temperature. that by escapes the into oxygenatedforms the hemoglobins blood and stream myoglobi and muscle tissue. This scheme is Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 133 incoming ligand molecules long enough to allow The cavities in the interior of myoglobin are bond formation with the iron atom. Opening of also important for ligand release. These spaces the glove occurs by outward movements of the allow the dissociated O2 molecules to remain distal histidine. Ligands are captured in the interior unattached to the iron atom long enough for escape ‘webbing’ of the distal pocket, which is circum- out through the distal histidine gate. If these spaces scribed by residues 28(B9),29(B10),32(B13), are filled with large residues, the overall rate ( 68(E11), and 107(G8)(Fig. 2) and includes the constants for O2 dissociation decrease see V68F, Xe4 binding site discovered in Mb crystals equil- V68W entries Table 1, and Scott et al. w48x). ibrated with xenon gas w50x. The rate constants for ligand capture are reduced if the pocket of the 3.2. The mechanism of NO dioxygenation glove is reduced in size with Phe or Trp substitu- tions at positions 28, 29, 32, 68, and 107 w48x. The major NO scavenging reaction in mammals An exception occurs for the Leu29(B10) to Phe is NO dioxygenation by MbO and HbO , which mutation where the rate of O binding, k9 , 22 2O2are the predominant species found in aerobic mus- increases ;25% (Table 1). The large size of the cle and in arterial blood. When either MbO or phenyl ring displaces the water molecule that is 2 HbO is mixed with NO, the solution changes found attached to the distal His64(E7) residue in 2 immediately from red to brown, and this color native deoxyMb w51x. Loss of distal pocket water change has served as a simple assay for NO by replacement of His(E7) with Leu or Phe causes synthase activity for over 20 years. Before 1995, ;10-fold increases in the rate of O binding 2 the only detailed mechanistic study of this reaction (H64F, Table 1). The favorable effect of displacing was carried out by Doyle and Hoekstra w2x. They distal pocket water by Phe(B10) is offset by the showed that NO is rapidly and stoichiometrically inhibition of ligand entry due to the large size of converted to NOy in the presence of oxyhemoglo- the benzyl side chain. As a result, there is little 3 bin. In 1996, Wade and Castro w4x re-examined difference between 9 for Phe(B10) and that for k O2 the stoichiometry of this reaction, and at approxi- wild-type Mb. The remarkably low value of kO2 mately the same time, we examined the mechanism ( ) for Phe B10 is due to favorably electrostatic of NO-induced oxidation by wild type and mutant interactions between bound O2 and the positive w x w x HbO22 and MbO 3,18,55 . In the past three years, edge of the phenyl multipole 52 . Herold and coworkers have reported even more ( ) In the Trp29 B10 mutant, the large indole side definitive kinetic and chemical characterizations of chain not only fills the distal pocket, but also this reaction w33,34x. The mechanism is given in sterically hinders access to the iron and the bound Eq. 1 and Fig. 3B. ligand. The net result for Trp(B10) Mb is a dramatic 100-fold decrease in k9O2 compared with the wild-type protein (L29W, Table 1). These results point out the unique properties of the B10 residue, which can affect electrostatic interactions with the bound ligand, the size of the distal pocket for ligand capture, and the steric accessibility of the iron atom. Brunori and coworkers have exam- ined the effects of Tyr(B10) mutations in various myoglobin backgrounds that were originally designed to mimic the distal pocket structure of Ascaris suum hemoglobin w30,53x. They found that all three effects, pocket size, steric hindrance, and H-bonding, play important roles in regulating ligand binding by the Tyr(B10) side chain w54x. At neutral pH, the rate limiting step is bimolec- 134 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 ular entry of NO into the distal pocket. The non- Similar NO dioxygenase (NOD) activities have covalent capture of NO is followed by rapid been observed for Yeast and Alcaligenes flavoHbs reaction with bound oxygen, which resembles w57,59,60x. In principle, all hemoglobins and myo- superoxide anion (Fig. 3B). The resultant Fe– globins have NOD activity (i.e. NO scavenging). yy OONO complex isomerizes to NO3 in an inter- The physiological relevance depends on presence nal, metal catalyzed reaction that is too fast to be and speed of the corresponding reductase activity. detected at pH 7(i.e. G;500 sy1). Herold and For example, Stamler, Goldberg, and coworkers coworkers were able to identify the peroxynitrite have suggested that Ascaris suum Hb serves to intermediate by increasing the pH of the reaction scavenge both O2 and NO in the body wall of this solution to slow the rate of isomerization and to intestinal parasite in order to keep it anaerobic enhance the spectral differences between the inter- w61x. However, proof will require the discovery mediate and methemoglobin w33,34x. However, no and characterization of a cognate, highly active release of free peroxynitrite was observed at either flavoprotein reductase. In vivo, mammalian metMb low or high pH. Herold, Watanabe, and coworkers is rapidly reduced by the cytochrome b5 flavopro- have shown that H64G and H64A mutants of tein reductase system in myocytes, and human myoglobin can efficiently catalyze the isomeriza- metHb is re-reduced by erythrocyte methemoglo- tion of exogeneously added ONOOy salts w56x. bin reductase. This more recent work confirms that rapid isom- Myoglobin serves to detoxify any NO escaping erization of peroxynitrite is an intrinsic property the endotheliumysmooth muscle signaling system. of the heme group as long as the large anion can Brunori w5,6x has pointed out that this secondary gain access to the iron atom. function of myoglobin is needed to prevent inhi- In 1998 Paul Gardner w57x discovered that E. bition of muscle respiration. Submicromolar levels coli flavohemoglobin catalyzes the oxidation of of NO can irreversibly inhibit aconitase, shutting NO to nitrate by a reaction which is identical to down the TCA cycle w62x, and low levels of NO that involved in NO scavenging by HbO and inhibit cytochrome c oxidase w63x. Moncada and 2 w x MbO2 . Gardner named this process NO dioxygen- coworkers 64,65 have shown that NO greatly ation, which we now prefer to the older term, NO- inhibits respiration in blood vessel endothelial and 15 18 induced oxidation. Using NO and O2 , he has smooth muscle cells, presumably to allow O2 shown that both atoms of bound oxygen are transport through these tissues without its incorporated into the nitrate product when the consumption. reaction is carried out either by E. coli flavoHb or Gladwin et al. w8x have shown that in healthy human hemoglobin A w58x. In the case of flavo- human volunteers intracellular hemoglobin reacts hemoglobin, the ferric form of the protein is rapidly and stoichiometrically with inhaled NO to rapidly re-reduced by NADH through a flavin- produce metHb and nitrate (;80 mM increases in containing reductase domain. blood after 2 h of inhalation of 80 ppm NO). The level of nitrosylHb is small and increases by an amount (1–2 mM HbNO) predictable by the frac- tion of deoxyHb in blood passing through the lungs. The absolute amount of S-nitroso-b 93 S- nitrosocysteine is very small initially, 0.2 mM, and only increases to ;0.4 mM w8x. Thus, the NO dioxygenation activity of hemoglobin will rapidly remove and detoxify any NO entering the blood stream, either from signaling pathways, inflam- mation responses, or inhalation. A summary of the secondary physiological roles

of HbO22 and MbO in NO detoxification is shown in Fig. 4. This scheme is based on the studies Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 135 discussed in the preceding paragraph, our own and toxicity of blood substitutes w23,24,26x. The work on NO dioxygenation reactions, and ideas first generation Hb-based products were designed presented by Brunori, Lancaster, and their cowork- to prevent tetramer dissociation, using specific ers over the past several years w5,6,66,67x. Nitric chemical cross-linking, polymerization with alde- oxide generated by NO synthase activity in endo- hydes, or genetic fusion of a chains, because thelial cells diffuses into adjacent smooth muscle dimer formation leads to rapid clearance, oxidative cells and activates guanylyl cyclase causing a stress, and renal damage w72x. Second generation cascade of events that leads to relaxation and hemoglobin preparations are being developed to arterial dilation. This NO also represses respiration address issues of physiological efficacy and to in the vessel walls, facilitating O2 transport through eliminate side effects associated with this tissue. Myoglobin serves to scavenge any NO vasoconstriction. that escapes into muscle tissue, converting it rap- idly to nitrate and preventing inhibition of respi- 3.3.1. The hypertensive (pressor) side effect ( f w x ration t1y22for NO 10 ms, when MbO is ;2 Although the severity varies, all first generation mM w3,5x). Hb-based blood substitute products cause an ele- Hemoglobin removes NO from the blood stream vation of mean arterial blood pressure, an increase by the same reaction. The half-life for free NO in in total peripheral resistance, some gastrointestinal an oxygenated red cell is F1 ms because of the discomfort and loss of motility, and muscle lesions w x w x high, ;20 mM concentration of HbO2 3 . The adjacent to capillaries 24–27 . Although a causal reaction of external NO with red cells is ;1000 relationship with these side effects was not estab- ( f ) times slower t1y2 1ms due to unstirred layers lished, Baxter stopped development of their dias- surrounding the surface w66x. The rate of con- pirin cross-linked tetramer product due to adverse sumption of NO in blood vessels is reduced further clinical outcomes in phase III trauma trials and a by cell free plasma layers that form adjacent to safety study of acute ischemic stroke patients the endothelium as the red cells stream in the w73,74x. Simple genetically stabilized tetramers center of lumen w68x. Thus, during blood flow in were also dropped in favor of developing second vivo, HbO2 molecules in red cells do not interfere generation products, when Baxter purchased with NO signaling in the vessel walls but will Somatogen in 1998. remove excess NO that enters the blood stream The hypertensive effect reflects interference either by endogenous synthesis or inhalation. The with vasoregulation. A summary of recent blood biochemical and physiological evidence that intra- pressure measurements in rat models is shown in cellular hemoglobin acts as an NO scavenger in Figs. 5 and 6. The addition of a 10% top load (on vivo, converting it to nitrate, is overwhelming and per iron basis) of simple cross-linked human Hb argues strongly against any significant role in NO tetramers causes a very rapid, ;30 mmHg rise in transport via S-nitrosation as proposed by Stam- blood pressure (Fig. 6a; for time courses see w28x). ler’s group w69–71x. Three theories have been proposed to explain this effect. 3.3. Designing extracellular hemoglobin-based blood substitutes 3.3.1.1. Autooxidation and radical formation.In the early 1990s, Alayash, Cashon, and others The driving force behind all blood substitute suggested that extracellular hemoglobin causes research is the need for an unlimited and safe oxidative stress leading to an inflamation-like w x supply of an O2 delivery fluid for use in emergen- response 24,75 . There must be some oxidative cies and elective surgeries that would eliminate stress caused by administration of large doses of current problems with supply and storage of donat- extracellular hemoglobin since the protein is ed blood. A large number of reviews and com- cleared from the blood stream in 8–24 h, which mentaries have been published over the past two requires the removal of large amounts of heme years summarizing the clinical progress, efficacy, and iron. However, it is unlikely that hemoglobin 136 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

levels of hydrogen peroxide to form strong oxi- dants that irreversibly damage endothelial cells. However, there is no evidence that administration of cross-linked or polymerized hemoglobins by themselves cause oxidative stress in control ani- mals or humans which show the blood pressure effect. Most investigators see no evidence for rapid (F10 min) autooxidation of the hemoglobin in vivo w79–81x. Tsuchida’s w81x and Chang’s w82x groups have incorporated superoxide dismutase and catalase into their blood substitute preparations in efforts to make the hemoglobin molecules more resistant to the oxidative stress associated with reperfusion injury.

3.3.1.2. Excessive O2 transport. Winslow, Intag- lietta, and colleagues w23,36,37,83,84x and refer- ences therein) have suggested that extracellular hemoglobins are too efficient at transporting oxy- gen, particularly in small arterioles. They argue

that excessive O2 delivery triggers autoregulatory responses that cause closure of capillary beds by Fig. 5. Lack of dependence of the hypertensive effect on P50 . All experiments were carried out with rats and involved either uncontrolled constriction. Extracellular hemoglo- 10% top load experiments (circles, squares) or 50% isovolemic bins are two to three times more efficient at exchange (diamonds). Symbols: closed circles, D%MAP w x delivering O2 than red blood cells 38,39,41 and changes compared to HSA controls in 10% top load experi- as a result, Winslow and coworkers have argued ments for 3 rHbs with allosteric mutations to vary P50 and 3 rHbs with distal pocket mutations to reduce NO scavenging that extracellular Hb-based blood substitutes w28x. Open diamonds, chemically cross-linked Hb tetramers, should have low P50 values to limit the amount of and closed diamonds, Hbs decorated with PEG, POE, or o- oxygen delivered w23,36,83,84x. poly-raffinose, D%MAP changes in 50% isovolemic exchanges In our opinion, excess O delivery also does not w x 2 uncorrected for HSA controls 36 . Closed square, D%MAP appear to cause the initial increase in blood pres- changes compared with HSA controls in 10% top load exper- iments with glutaraldehyde treated rHb (29). Half-filled sure. First, the amount of extracellular hemoglobin squares, D%MAP changes compared with HSA controls for used in many of the experiments shown in Figs. 5 polyethylene glycol (PEG)-conjugated Hb, hydroxyethyls- and 6 is only 10% of the total present and should tarch-conjugated XLHb, polymerized XLHb, and PEG-modi- not have a major impact on oxygen delivery since fied Hb vesicles w37x. most of the O2 transport will be carried out by red cells. Second, there is no obvious dependence of oxidation causes the hypertensive effects shown in the blood pressure effect on the oxygen transport Fig. 6a. The increase in blood pressure is imme- characteristics of the hemoglobin sample. diate and occurs seconds to minutes after infusion A summary of all current data on the blood of extracellular hemoglobin w28,37x. Autooxidation pressure effect of human hemoglobin preparations of oxyhemoglobin and subsequent hemin loss are in rat models is shown in Fig. 5. The percent much slower processes that occur on time scales change in mean arterial blood pressure is plotted of hours to days even in the presence of facilitating vs. the P50 of the sample. Oxygen delivery in w x anions such as azide or cyanide 21,76,77 .In capillaries should be proportional to the P50 of the their most recent work, D’Agnillo and Alayash sample w42,85x (results in Fig. 8a). However, the w78x showed convincingly that simple cross-linked experimental results for both 10% top load and hemoglobin tetramers do react rapidly with low 40–50% isovolemic exchange show no correlation Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 137

Fig. 6. Dependence of the hypertensive effect on NO scavenging rate and molecular weight. All experiments were carried out with rats and involved either 10% top load experiments (circles, squares) or 50% isovolemic exchange (diamonds). Symbols are the ( ) same as in Fig. 5. a Effects of P50 on %DMAP for a series of simple cross-linked tetramers with very similar rates of NO scavenging; (closed circled) D%MAP changes compared with HSA controls in 10% top load experiments for 3 rHbs with allosteric w x w x ( ) mutations to vary P50 28 , open diamonds, chemically cross-linked Hb tetramers 36 . b Effects of k9NO,ox on %DMAP for a series of simple cross-linked tetramers and recombinant hemoglobins with identical sizes but widely different NO scavenging rates; closed circles, D%MAP changes compared with HSA controls in 10% top load experiments for 3 rHbs with allosteric mutations to vary w x P50 and 3 rHbs with distal pocket mutations to reduce NO scavenging 28 ; open diamonds, chemically cross-linked Hb tetramers uncorrected for HSA controls w36x. (c) Effects of molecular weight (size) on %DMAP; closed diamonds, Hbs decorated with PEG, POE, or o-poly-raffinose, D%MAP changes in 50% isovolemic exchanges uncorrected for HSA controls w36x; closed squares, D%MAP changes compared to HSA controls in 10% top load experiments with glutaraldehyde treated rHb w29x. Half-filled squares, D%MAP changes compared to HSA controls for polyethylene glycol (PEG)-conjugated Hb, hydroxyethylstarch-conjugated XLHb, polymerized XLHb, and PEG-modified Hb vesicles w37x.

between the hypertensive effect and oxygen affin- on the P50 of simple cross-linked tetramers that ity, and initially the data appear scattered and have the same rate constants for NO scavenging. undecipherable. In an attempt to understand these The data were taken from both the BaxterySoma- results, the various hemoglobin samples were togen group w28x and Winslow’s group w36x, and grouped as simple tetramers with cross-links and the range of P50 values is from ;3 to 40 mmHg. w x modifications that affect P50 by altering the R to Doyle et al. 86 reported the same lack of effect T transition, simple tetramers with distal pocket of P50 on total peripheral resistance in 40% iso- mutations to reduce NO scavenging rates, and volemic exchange reactions with hemoglobin tet- polymerized or decorated proteins with increased ramers showing even lower oxygen affinities. molecular size. These groupings were used to look Rohlfs et al. w36x suggested that their experimental individually at the dependence of the pressor effect results show an inverse relationship between O2 ( ) on: a P50 when reactivity with NO and size were affinity and the pressor response, particularly if ( ) ( kept constant; b NO reactivity when size was the data point at the highest P50 value ;50 kept constant; and (c) size when NO reactivity mmHg) is neglected (diamonds in Fig. 5). How- was kept constant (Fig. 6). ever, when all comparable data in the literature are

As shown in Fig. 6a, there is no dependence of examined, there is no obvious dependence on P50 the percent change in mean arterial blood pressure (Fig. 5). 138 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

Fig. 7. NO scavenging and its effects on smooth muscle function by extracellular hemoglobin. NO generated by endothelial cells binds to guanylyl cyclase (GC) in the smooth muscles lining arteriole walls, causing a cascade of events that leads to muscle relaxation. Consumption of NO by hemoglobin causes maintenance of the constricted state. DeoxyHb does bind NO rapidly, but oxyHb is the major species present in arterial blood where saturation levels are G90%. Thus, the major cause of NO depletion is the oxidative reaction of NO with bound dioxygen. The relative importance of luminal NO scavenging vs. albuminal reactions with extravasated hemoglobin is controversial. At present, most evidence suggests that both reactions affect NO signaling in vivo.

3.3.1.3. NO scavenging by oxyhemoglobin. Cur- making infrequent contact with the vessel walls. rently, the most widely accepted interpretation of However, all workers have reported contraction of the hypertensive effect is that extracellular hemo- arterial rings and isolated vessels when they are globin rapidly consumes NO, preventing activation exposed to extracellular hemoglobin, regardless of of smooth muscle guanylyl cyclase, subsequent flow rate w68,87x. The potency of simple stabilized relaxation, and vasodilation of the blood vessel. A tetramers appears to be increased significantly by model for NO scavenging is shown in Fig. 7 and rapid extravasation into the spaces between the is based on studies with whole animals, isolated endothelium and the smooth muscle layer w88x. blood vessels, arterial rings, and intestinal cells This model provides simple interpretations of using extracellular hemoglobin, red blood cells, the effects seen by blood substitute products that and NO synthase inhibitors. Liao and coworkers have been chemically or genetically modified to w68x have shown that red blood cells do not reduce the hypertensive side effect. Inhibiting the interfere significantly with NO gradients when they rate of NO dioxygenation by site-directed muta- are streaming through capillaries and presumably genesis should decrease interference with NO sig- Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 139 naling by allowing NO diffusion to compete with increasing molecular weight is less clear. We effectively with chemical reaction. Doherty et al. have argued (see Fig. 7) that increasing the radius w28x showed that there is a strong, linear correla- of the hemoglobin molecule, either by polymeri- tion between the rate of NO dioxygenation meas- zation or decoration with PEG-like molecules, ured in vitro and the percent change in mean slows the rate of extravasation and inhibits the arterial blood pressure in 10% top load experi- direct consumption of NO in the spaces between ments for a series of six distal pocket mutants endothelial and smooth muscle cells. If the PEGy- (Fig. 6b). In more recent work, Doyle et al. w28x lated hemoglobins become highly hydrated and have shown a similar linear dependence of the interact with each other, they may even tend to percent change in total peripheral resistance on stream in the center of the blood vessels and begin k9NO,ox for distal pocket mutants of human hemo- to show ‘slug’ flow in microcapillaries, both of globin that have P50 values in the range 30–50 which would inhibit NO scavenging at the lumen mmHg. These results confirm the lack of depend- wall. ence on P50 when size and reactivity with NO are Winslow and coworkers have suggested more kept constant. Those molecules that have low complex explanations of the benefits of decoration reactivity with NO show much reduced pressor with polyethers. Their interpretations involve alter- effects, regardless of their P50 values. ations in oncotic pressure, blood viscosity, and O2 The gastrointestinal discomfort and motility delivery, all of which must certainly occur for very problems associated with extracellular hemoglo- large PEGylated hemoglobins w23,36,83,84,90x. bins also appear to be related to NO scavenging However, these ideas do not provide an explana- and interference with smooth muscle relaxation. tion for the effects of mutagenesis on the pressor The Baxter group has shown that lowering the rate effects seen for the simple genetically cross-linked of NO scavenging in recombinant hemoglobin tetramers shown in Fig. 6b and those presented in mutants leads to a substantial decrease in this side Doyle et al.’s w84x most recent work. In the latter effect w27x. NO scavenging could also explain the cases, the only structural differences between the abnormally high rates of O2 consumption in micro- samples are internal aromatic amino acid substi- vascular walls that were reported by Intaglietta’s tutions; all the rheological properties of the hemo- group w89x after administration of cell-free hemo- globin preparations are identical. Even though globin. Low levels of free NO are known to inhibit possibly over-simplified, our scheme in Fig. 7 almost completely respiration in endothelial and attempts to explain all the size and mutagenesis smooth muscle cells w64x. Scavenging of NO by experiments reported in the literature. extracellular HbO2 would remove this inhibition, Regardless of the exact mechanism, two basic causing O2 to be consumed in blood vessel walls strategies have been used successfully to reduce rather than transported passively to contracting NO scavenging and the hypertensive side effect. striated muscle tissue. As pointed out by Tsai et The first is to keep Hb from extravasating into the al. w89x, the combination of vasoconstriction and endothelium and, if possible, away from the vessel premature O2 consumption could lead to serious walls by polymerization, PEGylation, or encapsu- oxygen deficits in actively contracting muscles. lation into artificial cells. The second is to re-

Increasing the size of the Hb-base O2 carrier engineer the active site of hemoglobin to inhibit (G128 000 Da) also causes substantial decreases the NO scavenging reaction. Our work has been and almost elimination of the pressor effect as is focused on the basic principles and implementation seen when reactivity NO is reduced G30-fold of the latter protein engineering approach, using (Fig. 6b,c). This effect is seen for glutaraldehyde recombinant myoglobin as a model system. It is cross-linked polymers of recombinant hemoglobin important to note, however, that the chemically w29x and when hemoglobin tetramers are coated simpler strategies of polymerization with glutaral- with large polyethylene glycol or sugar-like poly- dehyde (strategies adopted by Northfield Labora- mers to increase their radius of gyration w36,37x. tories and Biopure, Inc.) and decoration with The underlying cause of the decrease in %DMAP polyethylene glycol or sugar molecules (strategies 140 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 adopted by Hemosal, Ltd. and Sangart, Inc.) have lumen of the vessel, both of which would reduce also been successful in reducing the side effects interference with NO signaling (Fig. 7). However, associated with hypertension (Fig. 6c). the latter interpretation lacks firm experimental support and is speculative. 3.3.2. Strategies for inhibiting the NO scavenging in MbO22 and HbO 3.3.3. P50 and the efficiency of O 2 transport The rate limiting step for NO dioxygenation by The initial Mb and Hb mutants constructed to

MbO2 is movement of nitric oxide into the protein reduce NO dioxygenation have P50 values and and capture in the interior of the distal pocket oxygen dissociation rate constants that are low (Figs. 2 and 3B). Thus, one strategy to decrease enough to cause concern about their efficacy for ( the rate of NO entry is to reduce the capture O2 transport see L29F, V68F, and V68W entries, volume. Examples of this approach in recombinant Table 1 w28x). Winslow and others used these sperm whale myoglobin are shown in Table 1. initial results to argue that it is impossible to

Filling the back of the distal pocket by inserting reduce NO scavenging without also increasing O2 Phe or Trp residues at the B10 position reduces affinity and decreasing the rate of ligand dissoci- the rate of NO dioxygenation four- and 10-fold, ation w83x. Rohlfs et al. w36x argued that the respectively (Table 1). Similar reductions in mutagenesis approach would fail since they k9NO,ox are seen for Phe and Trp insertions at the observed a correlation between NO and O2 affini- E11 position and to a lesser extent at the G8 ties for a series of cross-linked tetramers, polymers, position in Mb (data not shown). and decorated hemoglobins. However, equilibrium Doherty, Mathews, Lemon, and coworkers at NO binding is not the key parameter for evaluating Somatogen constructed Phe and Trp replacements NO scavenging, which is a kinetic phenomenon at the same three positions in human hemoglobin that competes with diffusion of the gas to its target w3,28x. They discovered that Phe and Trp mutations in smooth muscle cells. In oxygenated blood sub- were most effective at the B10 position in a stitute preparations, the bimolecular rate constant subunits and at the E11 and G8 positions in b for NO oxidation to nitrate determines the speed subunits w3x. We explained these differences and extent of scavenging compared to signaling. between the subunits in terms of the proximity of Rohlfs et al.’s w36x data show no correlation the native Leu(B10),Val(E11), and Leu(G8) side between the bimolecular rate constants for NO ( f chains to ligand atoms in the crystal structures of dioxygenation, which are invariant k9NO,ox 30 y1 y1) a and b subunits containing bound ethyl isocyan- mMs, and P50 , which ranges from 10 to 53 _ _ ide. The Fe C N–CH23 –CH complexes serve as mmHg for their samples. To address the problem transition state analogues of the non-covalently of P50 vs. NO scavenging, we turned to our library bound NO intermediate shown in Fig. 3 and the of Mb mutants to obtain a more quantitative w x peroxynitrite intermediate shown in Eq. 1 3 . understanding of the factors governing O2 transport Using this approach, the Somatogen group con- and to devise engineering strategies to raise P50 ( ) structed a set of four hemoglobins in which the and kO2while keeping k9 NO,ox small. k9NO,ox values of the a and b subunits were Measurements of O2 uptake and release by matched and progressively decreased from ;60 recombinant sperm whale myoglobin prototypes mMstoy1 y1 ;2 mMsy1 y1 w28x. As shown in were made in a 25 mm capillary system, and the Fig. 6b, there is linear dependence of %DMAP on results are shown in Fig. 8. This system was used w x k9NO,ox for this series of test proteins. This result by Page et al. 38,39 to characterize the transport ( ) and the lack of dependence on P50 Fig. 6a argue properties of polymerized bovine hemoglobin strongly that the hypertensive effect is governed preparations that are being developed by for use by the rate of NO scavenging. In our view, the as blood substitutes by Biopure, Inc. The myoglo- favorable effects of increased molecular weight or bin mutants were selected for their widely different encapsulation are the result of reduced rates of oxygen affinities (Table 1). As shown in Fig. 7a, extravasation and perhaps even streaming in the the rate of O2 release in the capillary depends Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 141

Fig. 8. O2 release and uptake by Mb mutants during flow through a 25 mm diameter capillary. The residence time is given in ( ) seconds. In the release experiments a , MbO2 flowed through the lumen while the capillary wafer was continuously flushed with ( ) N2 . In the uptake experiments b , deoxyMb was sent through the same capillary wafer that was being flushed with humidified air. ( ) ( ) In c , the fractional saturation at a residence time of 0.5 s is plotted vs. log P50 . Note that for this physiologically relevant resident time, there would be little O2 delivery from the L29F mutant and almost 100% delivery from the L29W myoglobin. The capillary ( ) transport experiments were carried out at 37 8C; the P50 values listed in a are for measurements at 20 8C to be consistent with w x Table 1. For these mutants and wild-type Mb, the P50 values increase by a factor of 3–4 going from 20 to 37 8C 77 .

Table 2

Observed vs. Predicted P50 and k9 NO,ox values for multiple mutants designed as blood substitute prototypes

( ) ( ) Multiple mutant P50 P50 DGintP 50 k9NO,ox k9NO,ox DGintk9 NO,ox (mM)(mM)(kJy1 moly1) (mMsy1 y1) (mMsy1 y1) (kJy1 moly1) observed predicted observed predicted L29FyV68F 0.013 0.16 6.0 2.9 2.2 y0.63 H64FyV68F 130 300 2.0 1.8 17 5.4 H64QyV68F 4.0 12 2.8 7.5 17 1.9 H64QyV68W 4.6 8.9 1.6 2.0 7.2 3.1 L29FyH64Q 2.2 0.40 I4.2 17 14 y0.42 L29WyH64Q 5.6 200 8.7 3.0 5.6 1.5 L29FyH64QyV68I 15 1.9 I5.0 1.5 14 5.4 L29FyH64QyV68IL 0.23 0.13 y1.4 5.1 10 1.7 L29FyH64QyV68IF 0.26 0.92 3.0 13 4.0 y2.9 The predicted parameters were calculated from Eq. (4) and the interaction free energy term was computed from Eq. (5). Those N NG y1 parameters with dDGint 4.0 kJ mol are marked in bold. The results indicate that it is harder to predict P50 values than the rate constant for NO dioxygenation. 142 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

) ( ) ( ) directly on the P50 of the myoglobin sample. In V68I mutants . Both the Gln E7 and Ile E11 contrast, there is little or no dependence of the replacements have little effect on the rate of ligand rate of O250 uptake on P in capillary experiments capture in the distal pocket as judged by their NO ( ) Fig. 8b . dioxygenation rate constants, k9NO,ox in Table 1. These results reflect the diffusion gradients The Leu(B10) to Trp replacement has both across the capillary wall in the two types of effects. NO dioxygenation is reduced ;10-fold by experiments. The gradient is roughly proportional filling the distal pocket with the indole ring, and w x to the difference between O2 in the capillary wall P50 is increased ;30-fold due to direct steric w x w x and the free O2 in the lumen 42,85 .At50% hindrance with the bound ligand. The major prob- saturation of the Mb solution, the free concentra- lem with the Trp(B10) mutation is that the rate of f tion of O250 in the lumen P of the protein, which O2 association is reduced dramatically by this w x s ( ) is defined as the free O2 at Y 0.50. In uptake combination of effects. Adding the Gln E7 sub- ( ) experiments, the partial pressure of O2 gas in the stitution to Trp B10 Mb produces a prototype ( s capillary wall is nearly equal to that in the gas with optimized O2O2 binding properties k9 12 ( ) y1 y1 s y1) space ;150 mmHg in air and is much greater mMs,kO2 67 s but still a low value for ( ) y1 y1 than the P50 0.1–30 mmHg of the Mb sample. k9NO,ox,3mMs. As a result the gradient for uptake at ;50% The results in Table 1 demonstrate unambigu- ( y )f saturation is PwallP 50P wall . In release experi- ously that NO dioxygenation rates can be manip- w x ments, the external O2 is 0 due to flushing with ulated independently of O2 affinity. Different pure N2 gas, and the gradient for release at 50% structural features of the active site control these ( y )fy desaturation is PwallP 50P. 50 two properties. Ligand entry and the size of the The magnitude of these effects are shown in capture volume in the back of the distal pocket

Fig. 8c where the change in saturation after a govern the rate of NO dioxygenation. O2 affinity residence time of 0.5 s is plotted vs. the log of the is determined by steric and electrostatic interac-

P50 of the myoglobin sample. All of the samples tions on the distal side of the heme group and the ( ) are able to take up O2 effectively. Even the L29W geometry of the heme–His F8 complex on the mutant, which has an O2 association rate constant proximal side. Distal regulation can be achieved f y1 y1 F 0.2 mM s , was able to take up O2 in 0.5 by mutation of residues B10, E7, E11, and G8 ( ) s. These results show that the exact value of k9O2 Figs. 2 and 3, Table 1 . In the case of hemoglobin, is much less important than the P50 value of the proximal regulation of oxygen affinity can be protein. In the case of release, there is a roughly manipulated by mutations that affect the allosteric linear relationship between the amount of O2 transition. This strategy is particularly effective, w delivered as measured by DY and logP50 Fig. 7c, since replacements that affect the R to T transition (s)x. Clearly, any successful second generation are normally located along the subunit interfaces blood substitute candidate should have a high P50 and are far removed from the distal pocket. Thus, value as well as a low rate of NO dioxygenation. NO reactivity can be reduced by distal pocket

mutations, whereas P50 can be increased or 3.3.4. Raising P50 and lowering NO scavenging decreased by allosteric substitutions that are far Solutions to the problem of decoupling P50 and removed from the heme group but favor the low the rate of NO dioxygenation are relatively affinity quaternary conformation w27,28x. straightforward and based on the interactions In myoglobin, the most dramatic decoupling of shown in Figs. 2 and 3. Oxygen affinity can be NO dioxygenation from O2 affinity occurs for the decreased by weakening electrostatic interactions L29FyV68F and H64FyV68F double mutants. The with the polar FeO2 complex or sterically hinder- P50 values of these myoglobins are ;0.01 and ing the bound ligand. The strength of hydrogen 100 mM, respectively, whereas their rates of NO bonding can be reduced by replacing His(E7) with dioxygenation are virtually identical, ;2 mMy1 Gln and steric hindrance can be increased by sy1 . The Phe68(E11) side chain markedly decreas- replacing Val(E11) with Ile (Table 1, H64Q and es the space available for ligand capture, reducing Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 143

w x ( ) k9NO,ox in both mutants 91 . The Phe29 B10 side the effects of the individual replacements were chain displaces water in the distal pocket of additive on a free energy scale. If additivity deoxyMb and stabilizes bound O2 electrostatically applies, the free energy change for O2 binding to w x ( ) 52 . Both effects enhance O2 affinity significantly. a double mutant DG1,2 could be calculated from ( ) The Phe64 E7 mutation causes loss of the favor- the free energy for O2 binding to the wild-type ( ) able hydrogen bonding interaction that normally protein DGwt plus the sum of the differences in ( ) occurs between bound O2 and the native His E7 the free energy change for O2 binding to each ( ) side chain, resulting in a dramatic decrease in O2 mutant 1 and 2 compared with that of the wild- ( q w x affinity see H64F and H64FyV68F entries in type protein, dDG12dDG 93 : ) Table 1 . In combination, these strategies allow s q q ( ) DG1,2DG wtdDG 1dDG 2 3 selective variation of O2 affinity over a 10 000- fold range without altering NO scavenging and where: refute emphatically the criticism that the two proc- D sy ln D sy ln esses cannot be decoupled by protein engineering. G1,2RT K; 1,2G wtRT K wt sy dDG11wtRTlnŽ.K yK; 4. Discussion sy dDG22wtRTlnŽ.K yK 4.1. Applications to recombinant hemoglobin These logarithmic relationships should hold for The strategies and ideas developed using the kinetic as well as equilibrium parameters if an myoglobin prototypes listed in Table 1 have been Eyring or Arrhenius-type transition state model ( s y ‡ applied successfully to hemoglobin. The research applies to the rate process i.e. k Aexp{ DG y ) group at Baxter Hemoglobin Therapeutics (for- RT} . The predicted parameter can be calculated merly Somatogen) has constructed genetically from the parameter ratios if the free energy effects cross-linked recombinant hemoglobins molecules are additive: with 30-fold smaller NO dioxygenation rate con- K Ž.Ž.Ž.predict sKKyKKyK (4) f 1,2 wt 1 wt 2 wt stants and, at the same time, P50 values 65 mM at 37 8C. This was achieved by insertion of Gln(E7) and Trp(B10) into a subunits and If the effects are not additive on a free energy Trp(E11) in b subunits w27x. Brunori and cowork- scale, then an additional interaction term, DGint, ( ) ers w30x adopted a similar approach. They con- needs to be added to Eq. 3 . DGint represents the structed His(E7) to Gln and Leu(B10) to Tyr amount of interference or enhancement that occurs mutations in a and b subunits of human hemoglo- when both mutations are present. As pointed out w x bin, and the resultant mutant tetramer had a much by Wells 93 , this term is not easily predicted and reduced rate of NO scavenging. Ho and coworkers normally has to be determined empirically from w x observed data as: 92 have constructed a low O2 affinity hemoglo- ( ) bin mutant, which contains a Leu B10 to Phe DG syRTlnµK Ž.Ž.predicted yK observed ∂ mutation in a subunits to confer resistance to auto- int 1,2 1,2 (5) and chemical oxidation and an allosteric Asn108 to Gln mutation in b subunits to stabilize the low affinity T quaternary state. In the case of kinetic parameters, this term may reflect a change in the rate limiting step by the 4.2. Predicting multiple mutant properties combination of mutations. Comparisons of observed and predicted values

Multiple mutations are normally required to of P50 and k9 NO,ox for key double and triple myo- minimize NO scavenging and maximize O2 trans- globin mutants are presented in Table 2. In those ( ) port properties i.e. large P50 values . The design cases where the predictions differ from the of multiple mutants would be straightforward if observed values by a factor of G5 (i.e. 144 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

N NG y1) DGint 4.0 kJ mol , the parameters are marked heme iron atom and various helical segments in bold. Two conclusions can be reached from w54,95x. these data and hold for the larger set of data in our library w55x. First, it is easier to predict the 4.3. Optimizing stability as well as function rate constant for NO dioxygenation, which is governed only by ligand entry into the protein, Although most attention has been focused on than O2 affinity, which involves at least five reducing NO scavenging and optimizing O2 trans- distinct processes (Fig. 3A). Second, there are port, the resistance of recombinant hemoglobin to steric constraints when attempting to reduce NO oxidation, heme loss, and denaturation is equally scavenging by filling the interior portion of the important if the prototype is to be commercialized. distal pocket with larger amino acid side chains. Autooxidation and subsequent hemin loss are the Combining Phe(B10) with Ile(E11) or Leu(E11) principle causes of particulate formation in hemo- globin solutions. Free hemin is insoluble and gives a lower than expected value of k9NO,ox, suggesting favorable packing in the protein interi- apohemoglobin denatures quickly at temperatures or. A similar result is obtained for Phe(E7) and above 10 8C. Apoglobin stability is a key factor Phe(E11) replacements. However, when we con- in holoprotein production yield during cytoplasmic structed multiple Trp and Phe substitutions at the expression in bacteria. When induction is used, B10, E11, and G8 positions, the rates of NO protein synthesis is more rapid than both endoge- dioxygenation are much higher than expected; neous heme biosynthesis and the transport of complex heterogeneous kinetic behavior is often externally added heme. Newly synthesized apog- observed; and the proteins are generally unstable. lobin must be able resist denaturation long enough to pair up with its heme cofactor w15,18,20x. The data in Table 2 demonstrate the need for Often stability is sacrificed for function, and a crystal structures and molecular dynamics simula- dramatic example is the distal histidine found in tions to interpret the DG terms and, eventually, int almost all mammalian myoglobins. For example, to predict them de novo. At present, we have the apoprotein of the Phe(E7) Mb mutant is ;40 general strategies but specific implementation can times more resistant to unfolding than wild-type be difficult. For example, the rate of NO scaveng- apoMb; the H64F holoprotein is expressed at ing can be reduced by decreasing the capture higher levels in E. coli; and the corresponding volume in the distal pocket, but it is not clear Phe(E7) deoxyMb is more resistant to acid dena- from primary sequence alignments alone which turation w15,19x. In contrast, the H64F mutation residues should be changed to large amino acids. causes a dramatic 130-fold decrease in O2 affinity In human hemoglobin a chains, the rate of ligand (Table 1) and a 30-fold increase in the rate of entry is best slowed by substitutions at the B10 autooxidation w77x, properties that are highly det- but not the E11 position, whereas in b chains the rimental to the physiological function of myoglo- most dramatic decreases in the rate of ligand bin in myocytes. Keeping a His(E7) side chain in capture occur for Trp and Phe insertions at the the apolar heme pocket is unfavorable with respect E11 and G8 positions. Myoglobin shows interme- to apoprotein stability. However, the distal histidine diate behavior with positive effects occurring at is highly conserved because it is required for all three positions. Similar ambiguities occur when electrostatic stabilization of bound O2 and resis- trying to regulate steric hindrance of the bound tance to autooxidation. Clearly a compromise has ligand and hydrogen bonding by distal pocket evolved to maintain function at the expense of mutants w94x. This kind of problem is particularly resistance to unfolding. severe when trying to mimic the behavior of It is our goal to develop a database that will naturally occurring hemoglobins by small sets of contain enough information to allow the design of mutations in recombinant mammalian myoglobins. stable recombinant myoglobins and hemoglobins Often the two proteins have significantly different for a wide variety of pharmaceutical and industrial spatial relationships between the position of the uses. Despite the problems of predicting multiple Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 145 mutant behavior, this rational mutagenesis w4x R.S. Wade, C.E. Castro, Reactions of oxymyoglobin y with NO, NO22 , and NO under argon and in air, Chem. approach has produced recombinant hemoglobin ( ) molecules with little or no pressor effect but P Res. Toxicol. 9 1996 1382–1390. 50 w5x M. Brunori, Nitric oxide moves myoglobin centre stage, values capable of efficient transport and mainte- Trends Biochem. Sci. 26 (2001) 209–210. nance of oxygen consumption in rats after com- w6x M. Brunori, Nitric oxide, cytochrome-c oxidase and plete isovolemic exchange w86x. One of these myoglobin, Trends Biochem. Sci. 26 (2001) 21–23. molecules is being developed by Baxter Hemoglo- w7x P. Ascenzi, M. Brunori, Myoglobin: a pseudo-enzymatic bin Therapeutics as a safer, second generation scavenger of nitric oxide, Biochem. Mol. Biol. Educ. ( ) blood substitute. Perhaps even more important, 29 2001 183–185. w8x M.T. Gladwin, F.P. Ognibene, L.K. Pannell, et al., this database has led and will continue to lead to Relative role of heme nitrosylation and beta-cysteine 93 the development and testing of key biochemical nitrosation in the transport and metabolism of nitric and biophysical mechanisms of heme protein oxide by hemoglobin in the human circulation, Proc. function. Natl. Acad. Sci. USA 97 (2000) 9943–9948. w9x K. Nagai, M.F. Perutz, C. Poyart, Oxygen binding properties of human mutant hemoglobins synthesized in Acknowledgments Escherichia coli, Proc. Natl. Acad. Sci. USA 82 (1985) 7252–7255. We would like to dedicate this paper to Maurizio w10x R. Varadarajan, A. Szabo, S.G. Boxer, Cloning, expres- Brunori, whose friendship and scientific collabo- sion in Escherichia coli, and reconstitution of human ( ) rations have added both pleasure and insight into myoglobin, Proc. Natl. Acad. Sci. USA 82 1985 5681–5684. our work over the past 30 years. There has been w11x B.A. Springer, S.G. Sligar, High-level expression of a competition between the Rome group and our sperm whale myoglobin in Escherichia coli, Proc. Natl. and Quentin Gibson’s laboratories, which has been Acad. Sci. USA 84 (1987) 8961–8965. fun, honorable, and highly productive in terms of w12x G. Dodson, R.E. Hubbard, T.J. Oldfield, S.J. Smerdon, the exchange of students and ideas. Maurizio’s wit A.J. Wilkinson, Apomyoglobin as a molecular recogni- has added warmth and humor to what can some- tion surface: expression, reconstitution and crystalliza- tion of recombinant porcine myoglobin in Escherichia times be a cold and abrasive profession, and he is coli, Protein Eng. 2 (1988) 233–237. always able to bring out the best in his colleagues w13x E. Lloyd, A.G. Mauk, Formation of sulphmyoglobin and competitors. during expression of horse heart myoglobin in Escher- Supported by Robert A. Welch Foundation ichia coli, FEBS Lett 340 (1994) 281–286. Grant C-612 (JSO); National Institutes of Health w14x B.A. Springer, S.G. Sligar, J.S. Olson, G.N. Phillips, ( ) ( ) Mechanisms of ligand recognition in myoglobin, Chem. Grants HL47020 JSO , and GM35649 JSO , and ( ) the W.M. Keck Center for Computational Biology. Rev. 94 1994 699–714. w15x M.S. Hargrove, S. Krzywda, A.J. Wilkinson, Y. Dou, David H. Maillett and Raymund F. Eich were M. Ikeda-Saito, J.S. Olson, Stability of myoglobin: a recipients of predoctoral fellowships from the NIH model for the folding of heme proteins, Biochemistry training grants GM08280 and GM08362, respec- 33 (1994) 11767–11775. tively. Yi Dou was the recipient of a Welch w16x T.L. Whitaker, M.B. Berry, E.L. Ho, et al., The D-helix Foundation Postdoctoral Fellowship. in myoglobin and in the beta subunit of hemoglobin is required for the retention of heme wpublished erratum appears in Biochemistry 1997;36(38):15542x, Biochem- References istry 34 (1995) 8221–8226. w17x M.S. Hargrove, A.J. Wilkinson, J.S. Olson, Structural w1x A. Verma, D.J. Hirsch, C.E. Glatt, G.V. Ronnett, S.H. factors governing hemin dissociation from metmyoglo- Snyder, Carbon monoxide: a putative neural messenger, bin, Biochemistry 35 (1996) 11300–11309. Science 259 (1993) 381–384. w18x J.S. Olson, R.F. Eich, L.P. Smith, J.J. Warren, B.C. w2x M.P. Doyle, J.W. Hoekstra, Oxidation of nitrogen oxides Knowles, Protein engineering strategies for designing by bound dioxygen in hemoproteins, J. Inorg. Biochem. more stable hemoglobin-based blood substitutes, Artif. 14 (1981) 351–358. Cells Blood Substit. Immobil. Biotechnol. 25 (1997) w3x R.F. Eich, T. Li, D.D. Lemon, et al., Mechanism of NO- 227–241. induced oxidation of myoglobin and hemoglobin, Bio- w19x Q. Tang, W.A. Kalsbeck, J.S. Olson, D.F. Bocian, chemistry 35 (1996) 6976–6983. Disruption of the heme iron-proximal histidine bond 146 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

requires unfolding of deoxymyoglobin, Biochemistry 37 and oxyhemoglobin, Biochemistry 40 (2001) (1998) 7047–7056. 3385–3395. w20x E.E. Scott, E.V. Paster, J.S. Olson, The stabilities of w35x J.S. Olson, Stopped-flow, rapid mixing measurements mammalian apomyoglobins vary over a 600-fold range of ligand binding to hemoglobin and red cells, Methods and can be enhanced by comparative mutagenesis, J. Enzymol. 76 (1981) 631–651. Biol. Chem. 275 (2000) 27129–27136. w36x R.J. Rohlfs, E. Bruner, A. Chiu, et al., Arterial blood w x 21 M.S. Hargrove, E.W. Singleton, M.L. Quillin, et al., pressure responses to cell-free hemoglobin solutions and ( )™ His64 E7 Tyr apomyoglobin as a reagent for meas- the reaction with nitric oxide, J. Biol. Chem. 273 (1998) uring rates of hemin dissociation, J. Biol. Chem. 269 12128–12134. (1994) 4207–4214. w x w x 37 H. Sakai, H. Hara, M. Yuasa, et al., Molecular dimen- 22 C.L. Hunter, A.G. Mauk, D.J. Douglas, Dissociation of sions of Hb-based O(2) carriers determine constriction heme from myoglobin and cytochrome b5: comparison of resistance arteries and hypertension, Am. J. Physiol. of behavior in solution and the gas phase, Biochemistry Heart Circ. Physiol. 279 (2000) H908–H915. 36 (1997) 1018–1025. w38x T.C. Page, W.R. Light, J.D. Hellums, Prediction of w23x R.M. Winslow, New transfusion strategies: red cell microcirculatory oxygen transport by erythrocyteyhemo- substitutes, Annu. Rev. Med. 50 (1999) 337–353. globin solution mixtures, Microvasc. Res. 56 (1998) w24x A.I. Alayash, Hemoglobin-based blood substitutes and ( ) 113–126. the hazards of blood radicals, Free Rad. Res. 2000 in w x press. 39 T.C. Page, W.R. Light, C.B. McKay, J.D. Hellums, Oxygen transport by erythrocyte hemoglobin solution w25x J.R. Hess, Blood substitutes for surgery and trauma: y efficacy and toxicity issues, BioDrugs 12 (1999) 81–91. mixtures in an in vitro capillary as a model of hemoglo- bin-based oxygen carrier performance, Microvasc. Res. w26x A. Gulati, A. Barve, A.P. Sen, Pharmacology of hemo- ( ) 55 (1998) 54–64. globin therapeutics, J. Lab. Clin. Med. 133 1999 w x 112–119. 40 T.C. Page, Oxygen transport by erythrocyteyhemoglobin w27x J.C. Hartman, G. Argoudelis, D. Doherty, D. Lemon, R. solution mixtures in an in vitro capillary, Department Gorczynski, Reduced nitric oxide reactivity of a new of Chemical Engineering, Rice University, Houston, recombinant human hemoglobin attenuates gastric dys- 1997. w x motility, Eur. J. Pharmacol. 363 (1998) 175–178. 41 E.J. Boland, P.K. Nair, D.D. Lemon, J.S. Olson, J.D. w28x D.H. Doherty, M.P. Doyle, S.R. Curry, et al., Rate of Hellums, An in vitro capillary system for studies on reaction with nitric oxide determines the hypertensive microcirculatory O2 transport, J. Appl. Physiol. 62 effect of cell-free hemoglobin, Nat. Biotechnol. 16 (1987) 791–797. (1998) 672–676. w42x D.D. Lemon, P.K. Nair, E.J. Boland, J.S. Olson, J.D. w x 29 M.P. Doyle, I. Apostol, B.A. Kerwin, Glutaraldehyde Hellums, Physiological factors affecting O2 transport by modification of recombinant human hemoglobin alters hemoglobin in an in vitro capillary system, J. Appl. its hemodynamic properties, J. Biol. Chem. 274 (1999) Physiol. 62 (1987) 798–806. 2583–2591. w43x M.F. Perutz, Stereochemistry of cooperative effects in w x 30 A.E. Miele, S. Santanch, C. Travaglini-Allocatelli, B. haemoglobin, Nature 228 (1970) 726–739. Vallone, M. Brunori, A. Bellelli, Modulation of ligand w44x M.F. Perutz, Mechanisms regulating the reactions of binding in engineered human hemoglobin distal pocket, human hemoglobin with oxygen and carbon monoxide, ( ) J. Mol. Biol. 290 1999 515–524. Annu. Rev. Physiol. 52 (1990) 1–25. w31x Y. Dou, S.J. Admiraal, M. Ikeda-Saito, et al., Alteration w45x J.S. Olson, G.N. Phillips, Myoglobin discriminates of axial coordination by protein engineering in myoglo- between O , NO, and CO by electrostatic interactions bin. Bisimidazole ligation in the His64™ValyVal68™ 2 with the bound ligand, J. Biol. Inorg. Chem. 2 (1997) His double mutant, J. Biol. Chem. 270 (1995) 544–552. 15993–16001. w x w32x R.J. Rohlfs, A.J. Mathews, T.E. Carver, et al., The 46 S. Unzai, R. Eich, N. Shibayama, J.S. Olson, H. Mori- effects of amino acid substitution at position E7 (residue moto, Rate constants for O2 and CO binding to the 64) on the kinetics of ligand binding to sperm whale alpha and beta subunits within the R and T states of myoglobin, J. Biol. Chem. 265 (1990) 3168–3176. human hemoglobin, J. Biol. Chem. 273 (1998) w33x S. Herold, Kinetic and spectroscopic characterization of 23150–23159. an intermediate peroxynitrite complex in the nitrogen w47x G.N. Phillips, M. Teodoro, T. Li, B. Smith, M.M. Gilson, monoxide induced oxidation of oxyhemoglobin wcor- J.S. Olson, Bound CO is a molecular probe of electro- rected and republished article originally printed in FEBS static potential in the distal pocket of myoglobin, J. Lett. 1998;439(1y2):85–88x, FEBS Lett. 443 (1999) Phys. Chem. B. 103 (1999) 8817–8829. 81–84. w48x E.E. Scott, Q.H. Gibson, J.S. Olson, Mapping the w x 34 S. Herold, M. Exner, T. Nauser, Kinetic and mechanistic pathways for O2 entry into and exit from myoglobin, J. studies of the NO*-mediated oxidation of oxymyoglobin Biol. Chem. 276 (2001) 5177–5188. Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148 147 w49x E.E. Scott, Q.H. Gibson, Ligand migration in sperm w64x E. Clementi, G.C. Brown, N. Foxwell, S. Moncada, On whale myoglobin, Biochemistry 36 (1997) the mechanism by which vascular endothelial cells 11909–11917. regulate their oxygen consumption, Proc Natl Acad Sci w50x R.F. Tilton, I.D. Kuntz, G.A. Petsko, Cavities in pro- USA 96 (1999) 1559–1562. teins: structure of a metmyoglobin–xenon complex w65x E. Clementi, G.C. Brown, M. Feelisch, S. Moncada, solved to 1.9 A,˚ Biochemistry 23 (1984) 2849–2857. Persistent inhibition of cell respiration by nitric oxide: w51x M.L. Quillin, Crystal structures of distal pocket mutants crucial role of S-nitrosylation of mitochondrial complex of myoglobin, Biochemistry & Cell Biology, Rice Uni- I and protective action of glutathione, Proc Natl Acad versity, Houston, 1995. Sci USA 95 (1998) 7631–7636. w52x T.E. Carver, R.E. Brantley, E.W. Singleton, et al., A w66x X. Liu, M.J. Miller, M.S. Joshi, H. Sadowska-Krowicka, novel site-directed mutant of myoglobin with an unu- D.A. Clark, J.R. Lancaster, Diffusion-limited reaction

sually high O2 affinity and low autooxidation rate, J. of free nitric oxide with erythrocytes, J. Biol. Chem. Biol. Chem. 267 (1992) 14443–14450. 273 (1998) 18709–18713. w53x M. Brunori, F. Cutruzzola, C. Savino, C. Travaglini- w67x D.D. Thomas, X. Liu, S.P. Kantrow, J.R. Lancaster, The Allocatelli, B. Vallone, Q.H. Gibson, Structural dynam- biological lifetime of nitric oxide: implications for the

ics of ligand diffusion in the protein matrix: a study on perivascular dynamics of NO and O2 , Proc. Natl. Acad. a new myoglobin mutant Y(B10) Q(E7) R(E10), Bio- Sci. USA 98 (2001) 355–360. phys. J. 76 (1999) 1259–1269. w68x J.C. Liao, M.W. Vaughn, K.T. Huang, L. Kuo, Intravas- w x 54 F. Draghi, A.F. Miele, C. Travaglini-Allocatelli, et al., cular flow decreases erythrocyte consumption of nitric Controlling ligand binding in myoglobin by mutagene- oxide, Proc. Natl. Acad. Sci. USA 96 (1999) sis, J. Biol. Chem. 277 (2001) 7509–7519. 8757–8761. w x 55 R.F. Eich, Reactions of nitric oxide with myoglobin, w69x L. Jia, C. Bonaventura, J. Bonaventura, J.S. Stamler, S- Biochemistry & Cell Biology, Rice University, Houston, Nitrosohaemoglobin: a dynamic activity of blood TX, 1997. involved in vascular control wsee commentsx, Nature w x 56 S. Herold, T. Matsui, Y. Watanabe, Peroxynitrite isom- 380 (1996) 221–226. erization catalyzed by his64 myoglobin mutants, J. Am. w70x J.S. Stamler, L. Jia, J.P. Eu, et al., Blood flow regulation Chem. Soc. 123 (2001) 4085–4086. w x by S-nitrosohemoglobin in the physiological oxygen 57 P.R. Gardner, A.M. Gardner, L.A. Martin, A.L. Salzman, gradient, Science 276 (1997) 2034–2037. Nitric oxide dioxygenase: an enzymic function for w71x J.R. Pawloski, D.T. Hess, J.S. Stamler, Export by red flavohemoglobin, Proc. Natl. Acad. Sci. USA 95 (1998) blood cells of nitric oxide bioactivity, Nature 409 10378–10383. (2001) 622–626. w58x A.M. Gardner, P.R. Gardner, W. Brashear, K.D.R. Setch- w72x T.M.S. Chang, Blood Substitutes: Principles, Methods, ell, J.S. Olson, Fidelity of O-atom transfer in the Products and Clinical Trials, 1,, Karger-Landes, Basel, mechanism of nitric oxide dioxygenation by oxy-hemo- 1997. globins wAbstr.x (2001) submitted. w73x E.R. Sloan, M. Koenigsberg, D. Gens, et al., Disapirin w59x P.R. Gardner, A.M. Gardner, L.A. Martin, et al., Nitric- cross-linked hemoglobin (DCLHb) in the treatment of oxide dioxygenase activity and function of flavohemo- severe traumatic hemorrhagic shock: a randomized con- globins. Sensitivity to nitric oxide and carbon monoxide ( ) inhibition, J. Biol. Chem. 275 (2000) 31581–31587. trolled efficacy trial, J. Am. Med. Assoc. 282 1999 w x 1857–1864. 60 A.M. Gardner, L.A. Martin, P.R. Gardner, Y. Dou, J.S. w x Olson, Steady-state and transient kinetics of Escherichia 74 R. Saxena, A.D. Wijnhous, H. Carton, et al., Controlled coli nitric-oxide dioxygenase (flavohemoglobin). The safety study of a hemoglobin-based oxygen carrier, ( ) B10 tyrosine hydroxyl is essential for dioxygen binding DCLHb, in acute ishemic stroke, Stroke 30 1999 and catalysis, J. Biol. Chem. 275 (2000) 12581–12589. 993–996. w x w61x D.M. Minning, A.J. Gow, J. Bonaventura, R. Braun, M. 75 A.I. Alayash, R.E. Cashon, Hemoglobin and free radi- Dewhirst, D.E. Goldberg, J.S. Stamler, Ascaris haemo- cals: implications for the development of a safe blood ( ) globin is a nitric oxide-activated ’deoxygenase’ wsee substitute, Mol. Med. Today 1 1995 122–127. w x commentsx, Nature 401 (1999) 497–502. 76 M.S. Hargrove, T. Whitaker, J.S. Olson, R.J. Vali, A.J. w62x P.R. Gardner, G. Costantino, A.L. Salzman, Constitutive Mathews, Quaternary structure regulates hemin disso- w and adaptive detoxification of nitric oxide in Escheri- ciation from human hemoglobin published erratum x chia coli. Role of nitric-oxide dioxygenase in the pro- appears in J Biol Chem 1997;272(38):24096 , J. Biol. tection of aconitase, J. Biol. Chem. 273 (1998) Chem. 272 (1997) 17385–17389. 26528–26533. w77x R.E. Brantley, S.J. Smerdon, A.J. Wilkinson, E.W. Sin- w63x P. Sarti, A. Giuffre, E. Forte, D. Mastronicola, M.C. gleton, J.S. Olson, The mechanism of autooxidation of Barone, M. Brunori, Nitric oxide and cytochrome c myoglobin, J. Biol. Chem. 268 (1993) 6995–7010. oxidase: mechanisms of inhibition and NO degradation, w78x F. D’Agnillo, A.I. Alayash, Redox cycling of diaspirin Biochem. Biophys. Res. Commun. 274 (2000) 183–187. cross-linked hemoglobin induces G2yM arrest and apop- 148 Y. Dou et al. / Biophysical Chemistry 98 (2002) 127–148

tosis in cultured endothelial cells, Blood 98 (2001) w89x A.G. Tsai, H. Kerger, M. Intaglietta, Microvascular 3315–3323. oxygen distribution: effects due to free hemoglobin in w79x G.S. Hughes, V.L. DeSmith, P.K. Locker, S.F. Francom, plasma, in: R.D. Winslow, K.D. Vandegriff, M. Intag- Phlebotomy of 500 or 750 milliliters of whole blood lietta (Eds.), Blood Substitutes: New Challenges, Birk- followed by isovolemic hemodilution or autologous hauser, Boston, 1996, pp. 124–131. transfusion yields similar hemodynamic, hematologic, w90x R.M. Winslow, A. Gonzales, M.L. Gonzales, et al., ( ) and biochemical effects, J. Lab. Clin. Med. 123 1994 Vascular resistance and the efficacy of red cell substi- 290–298. tutes in a rat hemorrhage model, J. Appl. Physiol. 85 w80x G.S. Hughes, E.J. Antal, P.K. Locker, S.F. Francom, (1998) 993–1003. W.J. Adams, E.E. Jacobs, Physiology and pharmacoki- w x netics of a novel hemoglobin-based oxygen carrier in 91 M.L. Quillin, T. Li, J.S. Olson, et al., Structural and humans, Crit. Care Med. 24 (1996) 756–764. functional effects of apolar mutations of the distal valine w81x S. Takeoka, T.S.H. Ohgushi, T. Kose, H. Nishde, E. in myoglobin, J. Mol. Biol. 245 (1995) 416–436. Tsuchida, Construction of artificial methemoglobin w92x C.H. Tsai, T.Y. Fang, N.T. Ho, C. Ho, Novel recombi- reduction systems in Hb vesicles, Art. Cells Blood Subs. nant hemoglobin, rHb (beta N108Q), with low oxygen Immob. Biotech. 25 (1997) 31–41. affinity, high cooperativity, and stability against autoxi- w82x F. D’Agnillo, T.M. Chang, Polyhemoglobin-superoxide dation, Biochemistry 39 (2000) 13719–13729. dismutase-catalase as a blood substitute with antioxidant w93x J.A. Wells, Additivity of mutational effects in proteins, w x ( ) properties see comments , Nat. Biotechnol. 16 1998 Biochemistry 29 (1990) 8509–8517. 667–671. w94x A.J. Mathews, R.J. Rohlfs, J.S. Olson, J. Tame, J.P. w83x R.M. Winslow, Artificial blood: ancient dream, modern Renaud, K. Nagai, The effects of E7 and E11 mutations enigma wnews; commentx, Nat. Biotechnol. 16 (1998) on the kinetics of ligand binding to R state human 621–622. ( ) w84x M.R. McCarthy, K.D. Vandegriff, R.M. Winslow, The hemoglobin, J. Biol. Chem. 264 1989 16573–16583. w x role of facilitated diffusion in oxygen transport by cell- 95 S.J. Smerdon, S. Krzywda, A.M. Brzozowski, et al., free hemoglobins: implications for the design of hemo- Interactions among residues CD3, E7, E10, and E11 in globin-based oxygen carriers, Biophys. Chem. 92 myoglobins: Attempts to simulate the ligand binding (2001) 103–117. properties of Aplysia myoglobin, Biochemistry 34 w85x P.K. Nair, J.D. Hellums, J.S. Olson, Prediction of oxygen (1995) 8715–8725. transport rates in blood flowing in large capillaries, w96x X. Zhao, K. Vyas, B.D. Nguyen, et al., A double mutant Microvasc. Res. 38 (1989) 269–285. of sperm whale myoglobin mimics the structure and w x 86 M.P. Doyle, M.D. Cassens, D.H. Doherty, D.D. Lemon, function of elephant myoglobin, J. Biol. Chem. 270 Lowered nitric oxide reactivity decreases vasoconstric- (1995) 20763–20774. tion induced by recombinant human hemoglobin wAbstr. w97x B.D. Nguyen, X. Zhao, K. Vyas, et al., Solution and presentationx, Artif. Cells Blood Sub. Immobil. Biotech. crystal structures of a sperm whale myoglobin triple 29 (2001) 100. w x mutant that mimics the sulfide-binding hemoglobin from 87 M. Wolzt, R.J. MacAllister, D. Davis, et al., Biochemical ( ) characterization of S-nitrosohemoglobin. Mechanisms Lucina pectinata, J. Biol. Chem. 273 1998 underlying synthesis, no release, and biological activity, 9517–9526. w x J. Biol. Chem. 274 (1999) 28983–28990. 98 M.L. Quillin, R.M. Arduini, J.S. Olson, G.N. Phillips, w88x A.L. Baldwin, Modified hemoglobins produce venular High-resolution crystal structures of distal histidine interendothelial gaps and albumin leakage in the rat mutants of sperm whale myoglobin, J. Mol. Biol. 234 mesentery, Am J Physiol 277 (1999) H650–H659. (1993) 140–155. WCRF 2007 – Meat and Health Roundtable

INFORMATIONAL MEMORANDUM

ISSUE: NUTRITION AND HEALTH

IMPORTANCE: The number of overweight and obese Americans has increased dramatically during the last twenty years, to the point that the U.S. Centers for Disease Control and Prevention (CDC) refers to the obesity problem as an epidemic. Research indicates that the situation is worsening, not improving. Obesity has been implicated as a factor in the development of numerous health problems including diabetes, heart disease, asthma, and some forms of cancer. In 2000, over 300,000 deaths in the U.S. were associated with obesity and overweight, approaching the number (400,000) linked to cigarette smoking. The Surgeon General estimated that in 2000 the total direct and indirect costs attributed to overweight and obesity was $117 billion.

BACKGROUND: Results from the 1999 National Health and Nutrition Examination Survey (NHANES) indicate that an estimated 61 percent of U.S. adults are either overweight or obese, defined as having a body mass index (BMI) of greater than 25. Among U.S. adults aged 20 – 74 years, the rate of overweight (defined as a BMI of 25 – 29.9) increased 2 percent from 1980 to 1999. In the same group of adults, the rate of obesity (defined as BMI greater than or equal to 30) nearly doubled from 1980 to 1999, from 15 percent to 27 percent. CDC estimates that more than 15 percent of children and adolescents are overweight, a rate that has at least doubled since the 1970’s.

In December 2001, Surgeon General David Satcher released a report entitled “The Surgeon General’s Call to Action to Prevent and Decrease Overweight and Obesity”. The report outlined strategies that U.S. communities should take to assist in combating the obesity problem. These strategies included: requiring physical education at all school grades; providing more healthy food options on school campuses; and providing safe and accessible recreational facilities for residents of all ages.

Research has revealed that the obesity problem is complex and no single factor may be implicated as the cause for rising trends in obesity. However, the issue may be simplified as an imbalance of energy expended and energy consumed. Body weight is a function of a combination of genetic, metabolic, behavioral, environmental, cultural and socioeconomic influences. The Surgeon General states that behavioral and environmental influences are large contributors to overweight and obesity and therefore provide the greatest opportunity for intervention designed to prevent the problem.

Informational Memorandum Nutrition and Health Page 2 of 3

Federal guidance on food nutrition, information and education programs is primarily driven by the Dietary Guidelines for Americans, developed jointly between HHS and USDA. This guidance provides science based information on the influence of physical activity and healthy dietary choices on major chronic diseases. Federal law requires that the guidelines be published at least every 5 years, and on January 12, 2005 the 6th edition of the Dietary Guidelines for Americans, was released. The primary nutrition education tool is the Food Guide Pyramid. The Pyramid is currently undergoing a scientific review and will likely be updated in conjunction with the release of the 2005 Dietary Guidelines.

AMI has become a member in the American Council for Fitness and Nutrition (ACFN), http://www.acfn.org/ which is a non-profit organization comprised of food, beverage, and consumer products companies, not-for-profit organizations and trade associations that have come together to combat the obesity problem through promoting healthy nutrition and wellness. ACFN’s mission is to advocate comprehensive long-term strategies and constructive public policies for improving health and wellness of all Americans, particularly youth, by promoting science- and behavior-based solutions focused on the critical balance between fitness and nutrition.

In February 2004, CDC published the National Health and Nutrition Examination Surveys which states that the prevalence of obesity in the United States increased from 14.5% to 30.9% from 1971-- 2000. In March 2004, it was announced that obesity is the second leading cause of preventable deaths in the United States, and is poised to become number one next year if the trend stays the same.

The Codex Alimentarius Commission (CAC) has undertaken a new initiative with the development of a discussion paper titled Implementation of the WHO Global Strategy on Diet, Physical Activity and Health: Action that could be taken by Codex. The CAC will review this discussion paper during its upcoming 29th session to be held July 3 – 7, 2006. The discussion paper describes the process that codex will take to gather input from member states and specifically describes the relevance of the issue to two of the existing codex committees, Committee on Food Labeling (CCFL) and the committee on Nutrition and Foods for Special Dietary Uses (CCNFSDU). The actions described in the discussion paper will be a circular letter that will be sent to member states for comment, sometime after the CAC meeting.

While obesity is a major issue, the relationship between diet and cancer continues to be studied by researchers and headlined in the media. AMI anticipates two major developments in the coming two years related to this topic. First, in September 2006 it is expected that the American Cancer Society (ACS), at their annual meeting in Hawaii, will release dietary advice on both cancer prevention through dietary modification and dietary advice to cancer survivors. It is expected that the ACS recommendations may include an “upper safe limit” for red meat consumption. AMI is working closely with National Cattlemen’s Beef Association (NCBA) and The National Pork Board (NPB) to monitor these developments and are currently planning a face to face meeting with officials at ACS to gain a better understanding of what they intend to release in September.

Secondly, in late 2007 it is expected that the World Cancer Research Fund (WCRF) and the American Institute for Cancer Research (AICR) will release a follow up report to the 1997 WCRF – AICR report. It is anticipated that the release of this report could have significant negative implications related to meat consumption, with specific emphasis in the following areas: Informational Memorandum Nutrition and Health Page 3 of 3 epidemiology showing association between red meat and processed meats and colo-rectal cancers (CRC), the growing concern related to the formation of heterocyclic amines during high-heat cookery and their relationship to CRC, and the role of N-nitroso compounds and heme iron in the development of CRC. AMI will again be coordinating our efforts very closely with NCBA and NPB to manage the issues surrounding this upcoming report.

STATUS: The AMI will continue to monitor the developments of nutrition-related guidance at the federal level including the reviews of the Dietary Guidelines and the Food Guide Pyramid. AMI will maintain a relationship with ACFN and will seek opportunities to support and advocate the policies and objectives of this initiative. AMI will continue to work closely with NCBA and NPB to monitor the release of the upcoming dietary advice from ACS. AMI staff will continue to actively participate in the International Meat Secretariat Human Health and Nutrition Committee and will work to develop a coordinated national and international strategy in response to the anticipated 2007 release of the WCRF/AICR report.

AMI POSITION: AMI supports the adoption of public policy aimed at curbing the obesity epidemic in the United States as long as these policies stress the importance of, and interdependent nature of, exercise and nutrition in achieving optimum long term health of consumers. AMI will work with other food industry organizations, including the ACFN, USDA, and HHS to ensure that regulatory nutrition initiatives are based in the best available science related to maintaining wellness and in making healthy food choices. AMI strongly advocates that a diet based upon a variety of foods, eaten in moderation, along with vigorous daily exercise, is the most appropriate and science based advice for consumers.

AMI CONTACT:

RANDALL HUFFMAN LYNN MORRISSETTE VICE PRESIDENT SENIOR DIRECTOR SCIENTIFIC AFFAIRS REGULATORY AFFAIRS AMI FOUNDATION

SUSAN BACKUS DIRECTOR OF RESEARCH AMI FOUNDATION

OCTOBER 2006 Downloaded from caonline.amcancersoc.org

MORE DETAILS ON RED MEAT, COLON CANCER LINK CA Cancer J Clin 2005;55;143-144 by guest on September 29, 2006 (©American Cancer Society, Inc.) This information is current as of September 29, 2006

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://caonline.amcancersoc.org/cgi/content/full/55/3/143

To subscribe to the print issue of CA: A Cancer Journal for Clinicians, go to (US individuals only): http://caonline.amcancersoc.org/subscriptions/

CA: A Cancer Journal for Clinicians is published six times per year for the American Cancer Society by Lippincott Williams & Wilkins. A bimonthly publication, it has been published continuously since November 1950. CA is owned, published, and trademarked by the American Cancer Society, 1599 Clifton Road, NE, Atlanta, Georgia 30329. (©American Cancer Society, Inc.) All rights reserved. Print ISSN: 0007-9235. Online ISSN: 1542-4863. CA Cancer J Clin 2005;55:139–144

mograms every two years, Smith said. Medi- cessed meat, poultry, or fish the people had care pays for women 40 and older in the eaten. program to have a mammogram every year. People who ate the most red meat in both time points (1982 and 1992/1993) were 30% to 40% more likely to develop cancers of MORE DETAILS ON RED MEAT, COLON the lower colon or rectum than those who ate CANCER LINK

the least red meat. Those who ate the most Downloaded from processed meats were 50% more likely to de- High consumption of red or processed meats velop cancers of the lower colon (descending has been linked to colorectal cancer in numerous and sigmoid) compared with those who ate the studies. Recent work by ACS epidemiologists least. These analyses compared risk among peo- helps put the relationship into perspective. ple in the top third (highest tertile) of meat

“Our study was better able to separate the caonline.amcancersoc.org consumption with those in the bottom third risk associated with meat consumption from (lowest tertile). that associated with other factors that affect For red meat (beef, lamb, pork), the cutoff colorectal cancer risk, especially obesity and for the highest tertile was three or more ounces physical inactivity,” said coauthor Michael per day for men—approximately the amount of Thun, MD, MS, Vice President for Epidemi- meat in a large fast-food hamburger. For ology and Surveillance Research at the ACS. women, the top tertile started at two or more The verdict: eating large amounts of red or ounces per day. For processed meat (bacon, processed meat over a long period of time can sausage, hot dogs, ham, cold cuts), the highest by guest on September 29, 2006 (©American Cancer Society, Inc.) indeed raise colorectal cancer risk. But the risks tertile started at one ounce eaten five or six days from such a diet are smaller than those from per week for men, and two or three days per obesity and lack of exercise, both for colon week for women. A slice of bologna weighs cancer and for overall health. about one ounce; two slices of cooked bacon ͓ ͔ “While these risks from meat consumption weigh a little more than half an ounce. to overall health are not in the same league as Eating poultry and fish did not raise the risk the risk from smoking, obesity, and physical of colon cancer. To the contrary, people in the inactivity,” said Thun, “these findings are im- highest tertile of poultry and fish consumption portant because red and processed meat are were 30% less likely to develop descending and major components of the diet of many Amer- sigmoid colon cancers than were people in the icans, and because there is now substantial ev- lowest tertile. The ratio of red meat to poultry idence that long-term high consumption and fish consumption was also predictive; com- increases the risk of colon cancer.” pared with those in the lowest tertile, people in The findings, published in JAMA (2005; the highest tertile were about 50% more likely 293:172–182), are based on a long-term study to be diagnosed with cancers of the descending of nearly 149,000 men and women aged 50 to and sigmoid colon. 74 in the Cancer Prevention Study (CPS) II The study findings add weight to current Nutrition Cohort, a subset of the CPS II Mor- ACS dietary guidelines, which recommend tality Cohort. The participants filled out a limiting red meat in favor of other sources of questionnaire about their eating habits in 1982 protein like poultry, fish, or beans, Thun said. and again in 1992/1993. Thun and his col- “This is not a condemnation of red meat, leagues looked at how many people had devel- but it is part of a growing body of evidence that oped colon cancer by 2001, then analyzed the red meat shouldn’t be the mainstay of your risk according to how much red meat, pro- diet,” he explained.

Volume 55 Y Number 3 Y May/June 2005 143 News & Views

This study did not address what it is about smaller portions and lean cuts, and bake, red meat that might influence cancer risk. The broil, or poach the meat rather than frying or iron and fat it contains may be culprits. Or it charbroiling it. may have to do with the way the meat is “The overall message is that less is better,” prepared; cooking at high temperatures can Thun said, “but it depends where you’re create carcinogenic heterocyclic amines. For starting from. If you’re eating red meat three

processed meat, the salt, smoke residue, and times a day, it’s reasonable to scale back to Downloaded from nitrates and nitrites used as preservatives may once a day. If you’re eating it once a day, it’s play a role. reasonable to scale back to a few times per For these reasons, the ACS guidelines ad- week.” vise people who do eat red meat to choose © American Cancer Society, Inc., 2005. caonline.amcancersoc.org by guest on September 29, 2006 (©American Cancer Society, Inc.)

144 CA A Cancer Journal for Clinicians AICR: Problems with Processed Meats

Home Search Donate Now Contact Us Research | Diet & Cancer | Publications | Cancer Survivors | How Can I Help? | Press | About AICR Publications AICR Newsletter e.Newsletter Search: Brochures September 2006 Diet and Cancer News New American Plate Problems with Processed Meats Shop AICR good-news-letter Americans love their processed meat—from sizzling bacon in the mornings to hot dogs on the grill and deli meats in between. According to a 2005/2006 business trend report, sales of sliced bacon alone are projected to hit almost 1.8 billion pounds this year. Subscribe to e.Newsletter Yet a mounting body of recent evidence supports a longstanding concern Home that processed meats are associated with increased cancer risk. The studies point to a likely connection between eating high quantities of processed meats and certain cancers, including colon, stomach, gastric and rectal.

A Preponderance of Processed Findings

In general, studies have found that the more processed meats you eat, the higher the cancer risk. One of the largest studies has linked high consumption of processed meats with an increased risk of both colorectal and one type of stomach cancer. This European study began in 1992 to track the diets and diseases of 521,483 people throughout ten countries. The colorectal cancer study, for example, found that after an average of almost five years Europeans who ate the highest amounts of processed meat had a 42 percent increased colorectal cancer risk than those who ate the lowest amounts.

In America, a similar-type of study published last year compared the amount of red meat, processed meat, or fish eaten by 149,000 Americans to the incidence of one type of colon cancer. Individuals were asked about their eating habits in 1982 and again ten years later. Researchers found that people who ate the most processed meats over the long term were 50 percent more likely to develop lower colon cancer compared with people who ate the least.

Several meta-analyses, studies that evaluate multiple related studies, have reiterated the hazards of processed meats. A recent example of one meta- analysis looked at 15 studies for a possible association between processed meat and stomach cancer. The Swedish researchers concluded that the risk of developing stomach cancer increases by between 15 and 38 percent when consumption of processed meats increases by about an ounce per day.

How’s It Doing That?

Exactly how processed meats may increase cancer risk is still unknown, yet fingers generally point to the preservatives nitrates and nitrites. Manufactures add these preservatives primarily to prolong the meat’s red color and prevent botulism, a potentially deadly disease caused by bacteria. (Many of these packaged meats can stay tucked inside the refrigerator for weeks or even years think SPAM without concerns of spoilage.)

http://www.aicr.org/site/News2?abbr=pub_&page=NewsArticle&id=10437&JServSessionIdr006=4fgpt7slb3.app1a (1 of 2)9/14/2006 12:09:06 PM AICR: Problems with Processed Meats

Nitrates can convert into nitrites, and when nitrites combine with certain compounds in the stomach they can form nitrosamines. These molecules were found to cause cancer in laboratory animal studies. To prevent the formation of nitrosamines, food manufacturers lowered the levels of the two preservatives in the 1970s and added a form of vitamin C, which helps prevent the formation of nitrosamines.

Genetics may cause some people to more readily form or be affected by nitrosamines. There are several other factors connected with processed meats’ association with increased cancer, including its high salt content.

And while the research continues…

In order to conclusively establish that processed meats are associated with certain cancers, scientists say more research is needed; several studies even found little or no association between the two. Yet the strength and relatively large number of studies that have found a link with high consumption and cancer has produced a turning point in the debate.

So for those who enjoy eating processed meats—what to do? Limiting your consumption of these meats is a start. AICR recommends eating no more than 3 ounces a day of red meat in general and arranging your plate with two- thirds (or more) of dishes from plant-based foods and the remaining one- third (or less) from animal protein. The “high” and “low” amounts of processed meats expressed in studies means different amounts in different studies. In the U.S. study, for example, the researchers described a high consumption of processed meat as at least 1.2 ounces a day for men and slightly over half an ounce a day for women. In the European study of colon cancer, a high consumption was 2.8 ounces a day. An ounce of processed meat is about one slice of bologna or half a hot dog.

Check the ingredients to see if the meats contain nitrites or nitrates. Some manufacturers, restaurants, and organic meats do not add these preservatives. Also, white processed meats may not contain them because there is no need to preserve a red color.

Another healthy safeguard is to eat more vitamins C- and E-rich fruits and vegetables. Along with other health benefits, research suggest these nutrients can mitigate the formation of nitrosamines. Avocado, spinach and dark green leafy vegetables contain high amounts of Vitamin E; citrus fruits, tomatoes, and broccoli are a few good sources of Vitamin C.

For recipes that feature foods high in these vitamins visit our Recipe Corner online.

Home | Your Privacy Rights | Site Map | Contact Us | Printer-Friendly Version

Copyright © 2006 American Institute for Cancer Research 1759 R Street NW, Washington, DC 20009 1-800-843-8114 (202-328-7744 in D.C.)

http://www.aicr.org/site/News2?abbr=pub_&page=NewsArticle&id=10437&JServSessionIdr006=4fgpt7slb3.app1a (2 of 2)9/14/2006 12:09:06 PM CAST Report on Use of GMO Feed for Livestock

Antibiotic Resistance and NPB On-Farm GMP Program

Take Care – Use Antibiotics Responsibly™

A Producer’s Guide to Responsible Antibiotic Use

1 Your Investment. Your Future. Antibiotic Use Concerns n Public health, medical, and consumer groups have concerns that antibiotic use in agriculture may contribute to the problem of antibiotic resistance in humans n Producers have a responsibility to protect public health while preserving animal health and welfare

2 Your Investment. Your Future. Pork Board Initiatives

n 1989 – Pork Quality Assurance n 1990-2005 – Research Funding n 1999 – Judicious Use of Therapeutic Antibiotic Guidelines (over 70,000 distributed with the help of FDA) n 2002, 2004 – Antibiotic Resistance and Alternatives Summits n 2000 – 2004 – Alternative Production Enhancers Programming n 2002 – Pork Board Statement on Antibiotic Use n 2005 - Take Care Release

3 Your Investment. Your Future. Pork Quality Assurance

QualityQuality Assurance® Assurance® A Program for America’s Pork Producers A Program for America’s Pork Producers

4 Your Investment. Your Future. Judicious Use Guidelines

5 Your Investment. Your Future. Pork Board Statement Antimicrobial Use in Pork Production “The National Pork Board supports the use of antimicrobials only when they provide demonstrable benefits and urges producers to: · take appropriate steps to decrease the need for their application; · adhere to judicious use guidelines; · assess the benefits and costs of all uses of antimicrobials; and · complete the Pork Quality Assurance Program and fully implement into their daily operations the management practices described for responsible use of animal health products.” July, 2002 6 Your Investment. Your Future. Responsible Use Program n Logical “Next Step” n Build on current programs n Raise producer awareness and education n Give producers a defensible position n Involve numerous pork chain stakeholders n Proactive public relations n Animal health AND public health focus

7 Your Investment. Your Future. Collaborative Effort n American Association of Swine Veterinarians n NPPC n Animal Health Institute n Sponsor Companies – Meeting Sponsors – Other Partners n American Feed Industry Association n Centers for Disease Control and Prevention n Food and Drug Administration

8 Your Investment. Your Future. Program Elements

n Program elements – Manual with principles and guidelines – Video, brochures, etc. n Antibiotic uses addressed – Nutritional efficiency – Disease control and prevention – Disease treatment – Injectable, water and/or feed administered antibiotics

9 Your Investment. Your Future. Responsible Use Principles

1. Take appropriate steps to decrease the need for the application of antibiotics 2. Assess the advantages and disadvantages of all uses of antibiotics 3. Use antibiotics only when they provide measurable benefits 4. Complete the Pork Quality Assurance Program and fully implement into daily operations the Good Production Practices described for responsible use of animal health products 5. Follow the Take Care Responsible Use Guidelines

10 Your Investment. Your Future. Responsible Use Guidelines

A. Use professional veterinary input as the basis for all medication decision-making B. Antibiotics should be used for treatment only when there is an appropriate clinical diagnosis C. Limit antibiotic treatment to ill or at-risk animals, treating the fewest animals indicated D. Antibiotics that are important in treating antibiotic resistant infections in human or veterinary medicine should be used in animals only after careful review and reasonable justification E. Mixing together injectable or water medications, including antibiotics, by producers is illegal F. Minimize environmental exposure through proper handling and disposal of all animal health products, including antibiotics

11 Your Investment. Your Future. Program Summary This program entails: n optimizing management of your pigs n involving a veterinarian in your decision making n evaluating every use of antibiotics n re-evaluating your antibiotic uses n and, doing so in accordance with regulation

12 Your Investment. Your Future. Producer Endorsement n Over 45,000,000 pigs marketed in the 12 months following the introduction of the program will come from producers who have committed to preserving animal health and welfare while protecting public health by endorsing the Take Care – Use Antibiotics Responsibly program

13 Your Investment. Your Future. 14 Your Investment. Your Future.

Mycobacterium avium – NACMCF Update

Developments in the Crohn’s and Johne’s Debate

Mark Klassen Beef Industry Safety Summit April 20th/2005 The Key Question

„ Does Mycobacterium Avium Subspecies Paratuberculosis (MAP) the causative organism of Johne’s disease in cattle cause Crohn’s disease in humans? Mycobacterium Avium sbsp Paratuberculosis Photos courtesy of M.Collins – U of Wisconsin Presentation Overview

„ Overview of the Johne’s / Crohn’s debate „ Recent Scientific Developments „ Science Based Strategy for MAP

„ Presentation NOT intended to provide a summary of evidence for/against MAP in CD JOHNES/CROHN’S DEBATE

„ Mycobacterium „ Johne’s Disease „ Crohn’s Disease „ The MAP Hypothesis for Crohn’s

Heinrich Albert Johne Photo Courtesty of M.Collins – U of Wisconsin What is Mycobacterium Avium subspecies paratuberculosis ? „ Comes from the genus mycobacterium

Diagram courtesy of M.Collins – U of Wisconsin „ Causes Johne’s and some say involved with Crohn’s Genus Mycobacterium

„ Some important characteristics of Mycobacterial Pathogens „ Intracellular infection - M. Avium Intracellulare „ Granulomas - complexes of immune cells „ Latent Infection - Mycobacterium TB Complex 1/3 of the world infected (15 million in U.S.) „ Most grow very slowly – leprosy 15 days generation vs 20 minutes for E.coli

„ As a result difficult infections to treat „ Mycobacteria responsible for more human illness and death than all other bacteria combined

MAP and Johne’s Disease

„ Two forms of Johne’s found „ Multi-bacillary – many visible MAP organisms and organism easily cultured e.g. cattle „ Pauci-bacillary – few if any visible MAP organisms and very difficult to culture e.g. sheep and goats „ MAP can produce intestinal inflammation primarily in ruminants but also other species „ ruminants (cattle, sheep, goats, bison, deer, elk etc..) „ non-ruminants (rabbits, monkeys, fox, badger, pigs, chickens ) „ sub-human primates Crohn’s Disease

„ Chronic intestinal inflammation - diarrhea, weight loss, cramps Healthy - granulomas often present „ Inflammation (mouth, throat, lungs, liver, joints, eyes, skin) „ Fistula approximately 30% „ Treatment (immuno-supression drugs and surgery in 90%+) „ Youthful Onset 15-21 yrs „ Life long disease which Crohn’s shortens lifespan slightly „ Mortality 2-4%

Photos from www.shafran.net Crohn’s Disease

„ 500,000 ill at present in Canada/U.S.A. „ Currently new case every 30 minutes and many studies report increasing incidence „ Genetic and Environmental Factors „ Genetic factors (NOD 2 / CARD 15) mutations „ Environmental factors – increasing incidence „ May be Crohn’s disease(s) „ Cause(s) Unknown The MAP Hypothesis for Crohn’s

„ 1905 – UK Surgeon Dalziel „ 1913 BMJ - Johne’s and Crohn’s pathology was so similar "as to justify a proposition that the diseases may be the same“ „ 1930’s – Dr. Crohn reclassified as auto-immune disease because he could not see billions of MAP so clearly visible in cattle intestine „ 1980s Dr. Chiodini – cultured MAP in CD in 3 patients after 2 years incubation

Photos courtesy of M.Collins – U of Wisconsin The MAP Hypothesis for Crohn’s

„ Humans with CD may have pauci-bacillary Johne’s „ organism hard to culture (e.g. goats and sheep) „ MAP in CD is does not usually have cell wall which promotes survival as immune system is best Immunostained CWD able to recognize cell wall MAP in Human CD proteins Photo courtesy S. Naser - UCF The MAP Hypothesis for Crohn’s

„ Advocates believe consumption of contaminated food/water with cell wall deficient form causes CD in susceptible people „ Most individuals infected have no symptoms just as in 90% of individuals with M.TB „ Exposure to MAP organism with cell wall (typically form present in manure) may permit a protective immune response e.g. farmers SCIENTIFIC DEVELOPMENTS

„ MAP in Food „ Dairy (Milk etc..) „ Beef „ Water

„ Medical MAP Culture Slants Photos courtesy of M.Collins – U of Wisconsin Dairy Products – The Issue

„ Do Pasteurization Standards which were originally created for Mycobacterium Tuberculosis Complex destroy Mycobacterium Paratuberculosis?

„ Why the concern? „ Milk is an important food for very young children who may be more susceptible to MAP just as young cattle are more susceptible. U.S. Retail Milk Study

„ Laboratory studies are split on whether current standards are sufficient „ Retail studies have been performed and have shown viable MAP in milk in several countries „ U.S. Retail study to be published in several weeks (Journal of Food Protection – May Issue) „ Retail milk from Wisconsin, California, Minnesota „ 2.8% positive on solid culture (gold standard) and 6% on liquid culture which is more sensitive „ Approximately 30% of cartons had MAP DNA „ May be cross contamination / pasteurization failure Other Dairy Products

„ Unpasteurized cheese „ Approximately 33% of cheese in U.S. „ Some data indicates MAP can survive cheese manufacturing process „ Aging process reduces MAP „ Yogurt – lab studies shows MAP may survive „ Ice Cream - unstudied „ Baby Formula – retail study in Europe Dairy Products - Conclusions

„ Human exposure to viable MAP through milk is known to occur in the U.S.A. „ Presently, the presence of viable MAP in milk is not considered a food safety issue by regulatory agencies

„ When a dairy animal is culled with Johne’s it typically becomes beef Beef Products– The Issue

„ Is MAP in beef and can it be killed by cooking? „ MAP could potentially enter beef three ways „ Fecal contamination „ Blood (white blood cells) inside meat „ Lymph nodes „ May be easier to kill MAP in a liquid and with lower fat content „ Some beef cuts cooked to lower temperatures „ About 10% use thermometer (Canada) „ Raw product contamination potential in beef which is not a concern in milk MAP in Beef (Cornell Study 2001)

„ Isolation of Mycobacterium paratuberculosis (M.ptb) from thin market cows at slaughter. (Journal of Animal Science vol. 79) „ Is MAP Detectable in 539 thin dairy and beef cows from 3 large US plants…..

„ Johne’s – 34% D and 3% B cows sampled ƒ ( detectable M.ptb in IC/FC)

„ Systemic Johne’s– 8% D and .3% B sampled ƒ detectable M.ptb in liver ƒ 3% D and .3% B cows M.ptb in P/SC LN associated with g. beef MAP in Beef (Cornell Study 2001)

„ If we extrapolate to the total population of market cows assuming data from study „ Fecal Contamination (Intestinal Johne’s) ƒ 12% of Dairy Cattle and .3% of Beef „ MAP in tissues (Systemic Johne’s)

„ 3% of Dairy Cattle and .03% of Beef

„ MAP presence is much lower in beef cattle „ Fecal contamination is most likely source of MAP vs edible tissues MAP in Beef (Swiss Study 2004)

„ MAP in Skeletal Muscle „ Correlation between MAP in lymph nodes vs. muscle from thin animals used for food „ If MAP in lymph node (mesenterial or Ileocecal lymph nodes) 75% PCR + in skeletal muscle (diaphragm) „ PCR does not necessarily indicate live MAP but authors state that a 99% correlation existed between PCR positive lymph nodes and culture „ Culture of muscle was attempted from 2 very sick animals and was positive MAP in Beef - Conclusions

„ MAP from fecal contamination is the most likely route of exposure but the organism in manure may not be a concern according to advocates of MAP hypothesis (has cell wall)

„ MAP in edible tissues could occur in systemically infected animals

„ Presence of MAP in raw product is not a potential concern unless it survives cooking Water – The Issue

„ Could MAP infection also occur through water as occurs with its close relative M Avium Avium? „ M. Avium sbsp Avium (MAA) infection in humans known to occur through water consumption and through aerosols in hot tubs and showers „ MAA responsible for 95% of infections (intestinal and lung disease) in AIDS and often reason for death „ MAA increasingly recognized as lung pathogen in humans and lymph node infection in children „ Individuals with children who may be more susceptible to MAP often turn down hot water heater perhaps promoting bio-films/ survival MAP in Water

„ MAP much more resistant to chlorination than other organisms e.g. E. coli „ Water trough – 1.5 years „ Lagoon – 9 months „ Pasture - >1 year „ MAP can live in Amoeba which can make some bacteria more virulent e.g. listeria, M. avium „ Presence of Bovine MAP in Taff River and potential relationship to Crohn’s cluster in Cardiff in UK Applied and Environmental Microbiology – April Issue „ lung involvement not uncommon common in CD and may precede intestinal symptoms (Am J Surg Pathol. 2003 Feb;27(2):213-9. ) Medical Developments

„ Presence of MAP in Humans with Crohn’s „ Antibiotic Trials in Crohn’s „ Developments to Watch

„ Quick Overview Only of Complex Area Presence of MAP in Crohn’s Disease

„ Supporters of MAP hypothesis believe that humans with CD will have MAP organisms in similar locations to cattle with Johne’s „ Dr. Crohn did not classify as mycobacterial disease because he did not have sensitive enough techniques for pauci-bacillary form of JD

Cattle w Johne’s Humans w Crohn’s Intestine MAP + ? Milk MAP + ? Blood MAP + ? MAP in Intestinal Tissue

„ PCR evidence does indicate presence of MAP DNA in CD tissues in many but not all studies „ Modern PCR methods increased percentage positive „ MAP in CD tissues 50-90% vs 5-30% non-IBD controls „ Labs must be capable of finding DNA in pauci-bacillary JD in sheep/goats vs cow multi-bacillary disease „ Most Recent Study „ Feb 2005 – Journal of Inflammatory Bowel Disease.

„ 83% Crohn’s Disease positive for MAP DNA

„ 17% positives in Controls „ Two additional studies to be published shortly „ American Journal of Gastroenterology 83% CD „ Journal Gut PCR study CD 52% Controls 5% MAP in Intestinal Tissues

„ PCR evidence does not directly indicate whether MAP was alive „ If dairy products have MAP DNA in significant percentage of cartons could be from foods and not evidence of infection „ Since 2000 new more rapid culture methods from Beckton-Dickson (MGIT) have been developed „ Paper in press indicating majority of CD patients have live MAP in intestinal tissue MAP in Human Breast Milk

„ Has not been widely studied „ Difficult to get CD patients to participate „ Am J Gastroenterol. 2000 Apr;95(4):1094-5. „ 2/2 CD patients MAP in Breast Milk „ 0/5 Controls „ Since then a third CD patient positive and one CD negative MAP in Human Blood

„ In a similar manner as advanced cases of Johne’s, advocates of MAP hypothesis believe the organism should be in human blood – Dr. Saleh Naser (UCF) „ Lancet. 2004 Sep 18;364(9439):1039-44. „ 14/28 (50%) Crohn’s culture positive „ 2/9 (22%) Ulcerative Colitis „ 0/15 (0%)Healthy people

„ Study must be replicated by other labs „ Changes the nature of CD research and raised profile of the issue significantly in Europe due to media Antibiotic Trials

„ Challenges with Antibiotic trials involving MAP „ Very difficult to determine drug susceptibilities in very slow growing organisms „ Slow growth and lack of cell wall makes organism resistant to most antibiotics useful in TB „ New macrolide antibiotics thought to be active based on test tube studies but not certain „ Four uncontrolled trials with multiple agents/upto three year treatment reported clinical improvement and in some healing of intestinal ulceration „ Crohn’s disease relapses and remits spontaneously so need controlled trial Australian Controlled Trial

„ Started in 1999 and finished in Sept. 2004 „ Compared adding antibiotics potentially effective against MAP to standard treatment „ Placebo group was conventional treatment „ Results not yet published but some data presented at British Society of Gastroenterology Meeting in March 2005 Australian Controlled Trial

Australian RCT of R/C/C l 120 4mth,1yr,2yr 3yr (off antibiotics) not significant 100 statistically significant using ITT analysis using ITT analysis 80 Treatment 60 Placebo 40

20

Percenage of patientsin the tria 0 0 50 100 150 Weeks Australian Controlled Trial

„ Appears Antibiotics of benefit but why? „ To interpret results conclusively need 4 things „ 1. Test patients for MAP at the start of trial „ 2. Verify strains sensitive to antibiotics „ 3. Test patients for MAP during trial and relate presence/absence/reduction to clinical improvement „ 4. Need to ensure patients not reinfected following removal of antibiotics Australian Controlled Trial

„ These four types of data are not available „ In fairness it was not possible when trial started in 1999 and would still be very difficult „ Results could be attributed to activity on other organisms „ Trial results not clear enough to produce public health issue „ It is very possible that MAP selectively colonizes CD patients and makes treatment more difficult without being a causal factor. Developments to Watch

„ Studies of MAP involvement in other diseases „ Ongoing studies on Irritable Bowel Syndrome for potential MAP involvement. This is much less severe form of intestinal disorder which affects approximately 10%-15% of the U.S. population „ Blau Syndrome „ arthritis (joints), uveitis (eyes), skin rash (face) and granulomatous inflammation (organs) „ Very rare inherited disease (12 pedigrees) „ Like Crohn’s - mutation in CARD 15/NOD 2 region Blau Syndome

„ Data to be presented at IAP in August „ 6 tissue samples from five individuals from U.S. and Australia „ IS900 sequence (almost always) from MAP in 6/6 samples (kidney, skin, synovial tissue from joints) „ preserved specimens on slides with little tissue „ coincidental presence from food less likely vs intestine „ hspX - absolute confirmation of MAP in at least 3 of 5 individuals „ Other diseases with CARD 15 mutations and granulomas to be studied for MAP presence „ CARD 15 may influence ability to clear intracellular pathogens Medical Developments - Conclusions

„ New molecular and microbiological culture techniques have increasingly supported the presence of viable MAP in CD patients but it is uncertain if this presence; „ Coincidental - MAP exposure with no consequence „ Complicating – infection making existing CD worse „ Causal – necessary for the disease to occur.

„ Associations representing doctors which specialize in treating intestinal disease and Crohn’s Disease patient groups do not generally support the MAP hypothesis Science Based Strategy for MAP

„ Technical Advisory group of 13 scientists including experts in medical and food aspects to support Science Based Strategy „ Product based interventions are likely more effective, cheaper, and much quicker than farm based control programs given MAP world-wide geographic dist. „ First priority is to address MAP inside edible tissue as fecal contamination is already being addressed by multiple interventions „ Determine level/presence of MAP and then determine interventions required to effectively control and prevent viable organisms in beef „ Need to find solutions and not just identify problems MAP in Edible Tissue – Pilot Trial

„ Utilized 5 clinical JD animals acceptable for consumption „ A) Test detection methods (4 labs test identical tissues) „ B) Compare PCR and culture results „ C) Attempt to determine if MAP present in edible tissue

„ 8 skeletal muscle samples from different regions within the carcass

„ offal (liver and kidney)

„ lymph nodes beyond the gut including those potentially associated with hamburger „ D) Relate MAP presence back to standard diagnostic tests for Johnes MAP in Edible Tissues

„ Preliminary data appear to support previous studies

„ skeletal muscles

„ offal

„ lymph nodes „ Likely low levels in muscle tissue if present which are to be tested in cooking trials MAP in Edible Tissues

„ Ground Beef Sampling „ Sampled 600 packages of ground beef „ Could not detect MAP in any sample „ Sensitivity of test was 10 CFU/gram at best „ Irradiation and Cooking research „ Irradiation approximately as effective as cooking „ 7 log reduction (as required for E.coli) likely not possible and perhaps not necessary Future Research on MAP

„ More work on cooking methods and irradiation and ability to reduce MAP „ Carcass decontamination treatments „ Rapid methods development rtPCR and use of indicator organisms „ Cell Wall Deficient (ZN-) form „ HACCP based systems for MAP

„ Closely monitor medical research Research Challenges

„ Working with MAP is very time consuming (E. coli 24-48 hrs vs MAP 3-4 months) „ Working with MAP is very expensive „ Only a few properly qualified scientists „ Methods for meat must be developed as almost no published literature exists „ Need at least 2 years and probably 5 years to develop solid answers to basic questions Medical and Regulatory Perspective

„ MAP in food is NOT considered a food safety issue by regulatory bodies in Canada or the U.S. „ Almost all doctors specializing in gastrointestinal disease do NOT support the MAP hypothesis „ Regulators in Canada/U.S.A./UK/EU all state there is “not enough evidence to dismiss or confirm” this hypothesis Thank You AMIF/CDC Collaboration on O157 Publication

E. coli Trends WG call – Minutes for July 26, 2006

Present: CA (Duc Vugia), CT (Sharon Hurd), GA (Stepy Thomas), MD (Stephanie Mickelson), NM (Karen Johnson), TN (John Dunn, Tim Jones), CDC (Fred Angulo, Linda Demma).

Topics discussed:

1. Welcome and review of minutes from June 28 WG call – John Dunn/Fred Angulo ƒ Discu s sed action items from last call o Potential for incorporation of figure from FSIS regarding prevalence of E. coli O157:H7 in ground beef—very important component of the paper but inclusion of a figure might be difficult due to space, needs additional followup with FSIS o AMI survey and potential data tables or figures- AMIF tables / figures included in packet for discussion ƒ Manuscript will highlight the trend in FoodNet incidence of E. coli O157:H7 (and HUS) and AMI survey describing industry initiatives designed to make ground beef safer, will highlight regulatory guidance and FSIS testing data as well

2. Discussed idea (from Duc Vugia) generated since last call to present data showing that c o nsumption of ground beef has been stable ƒ Figure from AMI showing per capita “disappearance” of beef in the US by year, 1979-200 4 at the end of the packet o This data is published by AMI each year, Randy Huffman has the appropriate reference o All agreed that the data (figure or text?) adds credibility and should be included but limited to period of interest for FoodNet data, 1996-2005 • Olga Henao will continue to check the FoodNet pop surveys to see if ground beef consu m ption has changed in the FoodNet sites between the surveys o Two relevant questions in the FoodNet pop surveys to explore: 1) Consumed ground beef in the last 7 days?, 2) Any undercooked?

3. Went through the selected tables and figures from the previous call • Table 1- no comments, will reorganize order of states (currently alphabetical) to highlight year of entry in FoodNet, for consideration on next call • Table 2- discussed clarification of title to indicate d+ HUS and children <15 yo, discussed format / footnotes to clarify variation in the obtaining HDD by year and site (CT) • Explore combining Tables 1 and 2 in a single table (FJA) • Figure 1- Suggestion to list original sites in legend • Figure 2- discussed format and presentation o suggestion to present bars in side-by-side format o need to include stacked bar format highlighting HDD and active surveillance components of HUS data o footnote with appropriate information for HUS surveillance (eg CT – no HDD, HDD started in 2000)

4. Dis c ussed new AMIF data tables • Table 1- consensus that demographics from AMI survey can be included as text 1 of 12

o other demographic information that would be relevant includes 1) % of ground beef production that the surveyed processors represents in US?, % of this ground beef consumed domestically? • Tables 2 and 3- discussed conversations between RH, JD, OH held since last call a bo ut creation of figures from these data o planning to look at presentation of data by year for comparison to FNet incidence, interesting ecological relationships may become evident (eg in 2002, 82% of establishments began “test and hold”—corresponds to large decline in FNet incidence) o discussed subjective “effectiveness score” and difficulties in presentation, “effectiveness score” is good marker for selection of interventions to depict graphically and corresponds to the perceived important interventions in industry o consensus on call that the “effectiveness score” could be used to help select top 5 or 6 interventions to describe in more detail o we should also explore presentation of these data using % of ground beef produced by the surveyed processors that has been subjected to the various interventions (top 5 or 6) and testing strategies by year

5. Ot he r topics • Consider inclusion of data showing decline in ground beef recalls from FSIS, suggestion to explore including with testing data from FSIS in a single figure • Consider inclusion of outbreak data from CDC outbreak unit / eFORS, figure depicting outbreaks (and ground beef associated ob’s) by year

6. Ac ti on items • continue to develop draft manuscript (JD, OH) • correspond with FSIS (K Holt) regarding inclusion of recall and testing data (JD) • FoodNet pop survey consumption variables (OH) • work with AMIF to explore graphic presentation of data by % of US ground beef and by year (JD, RH, OH) • explore interest in presentation of outbreak data in manuscript with FDDB (FA)

7. Next call is August 23, 2006 - 3-4pm EST ƒ CALL INFORMATION: TBA

2 of 12

Title: Decline in laboratory-confirmed E. coli O157:H7 infections temporally associated with implementation of industry control measures targeting reduction in ground beef contamination, FoodNet, 1996-2005

Table 1. Crude incidence per 100,000 population of laboratory-confirmed E. coli O157:H7 by site, FoodNet, 1996-2005 Year FoodNet site All Sites CA CO CT GA MD MN NM NY OR TN 1996 0.96 - 2.28 0.22 - 5.16 - - 2.22 - 2.62 1997 0.80 - 1.39 0.19 - 4.25 - - 2.47 - 2.09 1998 1.63 - 1.71 1.20 0.94 4.42 - 1.99 3.08 - 2.37 1999 1.06 - 2.86 0.50 0.65 3.64 - 4.51 1.87 - 1.94 2000 1.45 - 2.46 0.51 0.64 4.38 - 3.51 3.35 1.06 2.03 2001 1.11 1.72 1.11 0.51 0.49 4.41 - 1.47 2.19 1.40 1.55 2002 1.39 2.31 1.30 0.53 0.55 3.19 - 1.68 5.03 0.87 1.69 2003 0.90 1.46 1.06 0.26 0.29 2.63 - 1.23 2.42 0.58 1.06 2004 0.78 0.82 0.88 0.26 0.40 2.16 0.37 1.25 1.70 0.81 0.90 2005 0.87 1.02 1.20 0.37 0.49 2.35 0.53 1.71 1.84 0.64 1.04

Table 2. Crude incidence per 100,000 population of Hemolytic Uremic Syndrome by site, FoodNet, 1996-2004 Year FoodNet site All Sites CA CO CT GA MD MN NM NY OR TN 1996 1997 1998 1999 2000 2001 2002 2003 2004

3 of 12

Figure 1. Relative rate compared with 1996-1998 baseline period of laboratory-confirmed E. coli O157 infection in all FoodNet sites and original five sites, 1996-2005

2

All sites Original sites

Relative 1 rate

0.8 0.7 0.6

0.5

1996-98 1999 2000 2001 2002 2003 2004 2005

Year

Figure 2. Incidence of E. coli O157:H7 infection and Hemolytic Uremic Syndrome (HUS) per 100,000 children aged < 15 in FoodNet, 1997-2004

1.2 9

E. coli O157 8 1 HUS 7 0.8 6

5 0.6 4 HUS incidence HUS 0.4

E. coliE. O157incidence 3

2 0.2 1

0 0 1997 1998 1999 2000 2001 2002 2003 2004 Year

4 of 12

AMI Tables -- First draft….

Table 1. Establishment Demographics

Establishment functions Number of establishments Slaughter, Fabrication, and Grinding 23 Slaughter and Fabrication only 6 Grinding only 4 Slaughter only 1 Total Establishments 34

Table 2. Intervention implementation

Intervention Year % using Effectiveness intervention score, avg* Slaughter Practices Hide-on carcass wash ’03-’04 53 3.0 Sanitary hide removal and evisceration ’94 100 3.7 Spot Cleaning Knife trimming ’94 100 3.3 Steam vacuum ’94-’96 97 3.3 Hock steam blow off ’97-’04 94 2.3 Emerge fecal detection system ’03 56 2.3 Physical barriers Paper covering of critical areas ’94-’02 82 3.6 Pre-evisceration treatment Pre-evisceration carcass rinse, hot water ’94-’98 24 4.0 Pre-evisceration carcass rinse, organic ’94-’01 97 4.0 acid Final Carcass Treatment Final wash ‘94 100 3.7 Organic acid rinse ’94-’00 100 3.6 Thermal pasteurization ’94-’01 100 4.8 Chilled Carcass Treatment Chlorine spray chill ’94-’03 79 2.2 Organic acid treatment ‘03 26 4.5

5 of 12

Table 3. E. coli O157:H7 Testing Practices – Sampling Sites

Sampling Site Initial Date initiated Current

n/N % n/N %

Hide 20/34 59% 98 – 04 8/34 24%

Carcass 20/34 59% 97 – 03 20/34 59%

Trim 34/34 100% 94 – 03* 34/34 100%

Primal cut 6/ 30 20% 00 – 03 6/30 20%

Ground beef 25/34 74% 02 – 03 14/34 41%

*100% of establishments survey answered Yes, they have tested or are testing for E. coli O157:H7

Per capita consumption* of beef, United States, 1996-2004

Hamburger 100 Processed beef 90 Beef cuts 80 Total beef l 70 60 50 40

Pounds retai 30 20 10 0

03 998 1996 1997 1 1999 2000 2001 2002 20 2004 Ye ar

*Consumption or “disappearance” data derived from US Department of Agriculture by AMI. Meat and Poultry Facts, 2005.

6 of 12 All Confirmed O157 Outbreaks Range OB Outbreaks Ill Avg OB Size Median OB Size Size 1996** 13 260 20 7 2 to 61

1997** 10 212 21 10 3 to 108 1998 27 574 21.3 11 2 to 143

1999 25 710 28.4 10 2 to 321 2000 24 1199 49.9 14.5 2 to 736 2001 16 233 14.6 10.5 2 to 55

2002 14 204 14.6 5 2 to 74 2003 16 187 11.7 10.5 2 to 51

2004 10 94 9.4 4 2 to 59 **Pre-eFORS reports

All O157 outbreaks where beef was the confirmed vehicle of transmission* (confirmed with reasons 1 and 2≠) Outbreaks Ill Avg OB Size Median OB Size Range OB Size 1996** 1 3 3 3 3 to 3 1997** 1 15 15 15 15 to 15 1998 1 34 34 34 34 to 34 1999 5 358 71.6 9 5 to 321 2000 4 70 17.5 9 2 to 50 2001 0 0 0 0 -- 2002 1 74 74 74 74 to 74 2003 0 0 0 0 -- 2004 3 68 22.7 5 4 to 59 **Pre-eFORS reports * Beef outbreaks are defined as outbreaks where either 1) the contaminated ingredient identified was beef or 2) no other ingredient was identified as contaminated and one of the food items implicated was an ‘essential’ beef item and not a complex food dish (i.e.- beef stroganoff without any ingredients listed would not be captured in this analysis, this would be a ‘Contains’ beef item, but we are not certain that beef was the causative commodity) ≠ Reason 1: statistical evidence from epidemiologic investigation; Reason 2: Laboratory evidence (e.g. identification of agent in food)

All O157 outbreaks where beef was the suspected vehicle of transmission (includes confirmed and non-confirmed beef vehicles)* Median OB Year Outbreaks Ill Avg OB Size Size Range OB Size 1996** NA NA NA NA NA 1997** NA NA NA NA NA 1998 1 34 34 34 34 to 34 1999 12 408 34 9 2 to 321 2000 10 171 17.1 14 2 to 50 2001 0 0 0 0 -- 2002 4 87 21.8 5.5 2 to 74 2003 2 15 7.5 7.5 2 to 13 2004 6 77 12.8 4 2 to 59 * Beef outbreaks are defined as outbreaks where either 1) the contaminated ingredient identified was beef or 2) no other ingredient was identified as contaminated and one of the food items implicated was an ‘essential’ beef item and not a complex food dish (i.e.- beef stroganoff without any ingredients listed would not be captured in this analysis, this would be a ‘Contains’ beef item, but we are not certain that beef was the causative commodity)

7 of 12 Confirmed Escherichia coli O157 outbreaks, 1996-2004

30 80 70 25 All O157 60 outbreaks 20 50 O157 outbreaks 15 40 where beef is 30 confirmed vehicle 10 Median number ill 20 (all O157)

5 10 ill number Median Number of outbreaks Median number ill 0 0 (confirmed beef)

9 0 1 2 3 0 996 997 998 99 00 00 00 1 1 1 1 2 2 2 20 2004 Year

8 of 12 Percent of ground beef samples positive for E. coli O157:H7

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3

Percent Positives 0.2 0.1 0 94 95 96 97 98 99 00 01 02 03 04 05 Year

9 of 12 Percent of Population Reporting Consumption of Ground Beef FoodNet Population Survey, 1996-2003 100

90 Undercooked ground beef Fully cooked ground beef 80 70

60 50

Percent* 40

30 20 10 0 1996-97 1998-99 2000-01 2002-03 Year * Based on weighted estimates 10 of 12 Percent of Population Reporting Consumption of Ground Beef FoodNet Population Survey, 1996-2003

80

70

60

50

40

Percent* 30

20

10

0 1996-97 1998-99 2000-01 2002-03 Year * Based on weighted estimates 11 of 12 Percent of Population Reporting Consumption of Undercooked (Pink) Ground Beef FoodNet Population Survey, 1996-2003

20

15

10 Percent*

5

0 1996-97 1998-99 2000-01 2002-03 Year * Based on weighted estimates 12 of 12 AMIF Research Update

DRAFT

White Paper on Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources

M. Ellin Doyle1*, John Archer2, Charles W. Kaspar1, and Ronald Weiss1

1Food Research Institute, University of Wisconsin–Madison, Madison, WI 53706 2Wisconsin Division of Public Health, Bureau of Communicable Diseases and Preparedness, Communicable Disease Epidemiology Section, Madison, WI 53702

Contents Page Introduction ...... 2 Epidemiology of E. coli O157:H7...... 3 Outbreak Data...... 3 Reservoirs of E. coli O157:H7 ...... 3 Cattle—the primary reservoir ...... 4 Other ruminants ...... 4 Other animals...... 5 Transport Hosts...... 5 Routes of Human Infection ...... 6 Direct contact...... 6 Contaminated food...... 6 Contaminated water...... 7 Responses to E. coli O157:H7 Outbreaks ...... 7 Surveillance...... 7 Animals...... 7 Foods...... 8 Human illness...... 8 Regulations ...... 9 Beef...... 10 Juice ...... 11 Fresh produce...... 11 Drinking water...... 11 Swimming pools and beaches...... 12 Fairs and petting zoos...... 12 Industry Initiatives: Intervention Strategies ...... 12 Discussion and Summary...... 13 Epidemiology...... 13 Evaluation of current practices ...... 13 Recommendations ...... 14 Interventions—Recent Improvements...... 14 Further Research ...... 15 Figures ...... 17 Tables ...... 19 Reference List ...... 23 Appendix...... 37

* Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 2 White Paper:

INTRODUCTION E. coli O157:H7 was first identified as a possible human pathogen in 1975 in a California patient with bloody diarrhea and was first associated with a foodborne (ground beef) outbreak of disease in 1982 (299;353). This serotype (defined by its O and H surface antigens) and some non-O157 serotypes of E. coli produce verocytotoxins, also called Shiga-like toxins because of their similarity to toxins produced by Shigella dysenteriae. These E. coli are called VTEC (verocytotoxin-producing E. coli), STEC (Shiga- toxin producing E. coli), and also EHEC (enterohemorrhagic E. coli) because of the symptoms they produce. Serotypes of VTEC bacteria may include different strains that differ in some virulence factors or other characteristics such as motility and sorbitol fermentation. It has been estimated that E. coli O157:H7 causes about 73,500 cases of illness and 60 deaths annually in the U.S. while non-O157 VTEC serotypes cause about 37,000 cases annually (73;254). Epidemiological studies of foodborne outbreaks have indicated that fewer than 40 cells of E. coli O157:H7 can cause illness in some people (162;328;335). Waterborne outbreaks, involving both drinking water and pools and lakes, also indicate a very low infectious dose because of the great dilution factor in large quantities of water. Depending on their age, immune status, general health, dose of bacteria, and virulence factors in the bacteria, persons exposed to VTEC may experience mild diarrhea, severe bloody diarrhea, hemorrhagic colitis, or hemolytic uremic syndrome (HUS) with kidney failure. Some cases, usually among young children and older people, are fatal. Children <5.5 years of age are more likely to develop HUS and some studies have indicated that antibiotics or antimotility agents increase risk for HUS (53;101). Other recent studies indicated that correlations of HUS with antibiotic use were not very strong (304;339). Patients with higher temperatures, higher white blood cell counts, and a longer periods of diarrhea appear to be more likely to develop HUS (181;339). Although this literature review is focused on epidemiology of E. coli O157:H7, which is the primary serotype causing outbreaks of VTEC in the U.S., more than 200 different serotypes of E. coli produce shiga-like toxins and many have been implicated in outbreaks and cases of HUS in the U.S. and other countries (73;196;251). Verotoxin-producing E. coli can be sorted into 4 clonal groups according to different virulence factors and other characteristics that are encoded by genes on their chromosomes, plasmids, and phages. Many of these genes, including those encoding Shiga toxins, were apparently acquired from other organisms while some other functions, such as motility and sorbitol fermentation, were lost during the evolution of E. coli O157:H7 from avirulent ancestors. [Some recent European isolates of E. coli O157:H7 have been found to ferment sorbitol (43).] Approximately 25% of the E. coli O157:H7 genome is derived from bacteriophages indicating that these viruses were very important in the horizontal transfer of genes (80;213;344;355). VTEC bacteria continue to evolve even during the course of infections as strains isolated from stool samples early in infection sometimes differ from those isolated later (195;240;355). Analyses by CDC indicate that 10.8% of E. coli O157:H7 tested in 2003 have become resistant to one or more antibiotics (85). Non-O157:H7 strains vary in their ability to cause severe human illness and outbreaks but there is evidence that many are associated with cattle and other ruminants as is E. coli O157:H7 (295). During 2004, 110 non-O157 VTEC infections were identified in the ten states surveyed by FoodNet with serogroups O111, O103, and O26 most frequently detected (84). Outbreaks in the U.S. include: a 1994 outbreak in MT associated with milk (serogroup O104) (245), a 1999 outbreak in CT associated with lake water (serogroup O121) (235), a 1999 outbreak in TX associated with a salad (serogroup O111) (72), a 2000 outbreak in WA associated with punch (serogroup O103), a 2001 outbreak in SD in a day care (serogroup O111), and a 2001 outbreak in MN associated with lake water (serogroup O26) (243). In Australia, E. coli O157:H7 is not as commonly isolated in cases of bloody diarrhea and HUS as are other serotypes. Serogroup O111 was identified in a 1995 outbreak associated with sausage (271) and a 2003 outbreak thought to result from person to person transfer (106). Other Australian outbreaks most likely resulting from contact with cattle were caused by serotype O86:H27 (247) and O48:H21 (154). Non- O157 VTEC have also caused outbreaks in several European countries and Argentina (10;155;192;257;310;354).

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 3

EPIDEMIOLOGY OF E. coli O157:H7

Outbreak Data E. coli O157:H7 has been isolated from ill people around the world. It tends to be reported more often from more developed countries but this may be an artifact caused by the paucity of sophisticated diagnostic laboratories in developing countries. Appendix 1 lists, chronologically, published details of 204 reported outbreaks of E. coli O157:H7. Undoubtedly, other outbreaks have occurred but details have not been published in accessible journals. A review of data on outbreaks of infectious intestinal disease in the UK from 1992 to 2003 pointed out that there were 1763 outbreaks reported to authorities but only 55 of these were presented in peer-reviewed journals. Of 45 VTEC outbreaks reported to National Surveillance Program, only 7 were described in the literature (261). FoodNet data indicate that E. coli O157:H7 causes significantly more cases of sporadic infections than cases linked to an outbreak (84). For example, in 2004, only 9% of 402 confirmed cases of infection with E. coli O157:H7 were associated with outbreaks (www.cdc.gov/foodnet/annual/2004/report.pdf). Sporadic infections appear to be associated with some of the same factors that cause outbreaks: undercooked hamburgers and exposure to farms and cattle. Some sporadic infections are also associated with use of immunosuppressive medications and dining at table service restaurants (197). Vehicles of infection, suspected or confirmed, have been identified for most outbreaks listed in the appendix. Table 1 lists outbreaks with >100 cases. Of the eight largest outbreaks, four were associated with meat, three with drinking water and the largest outbreak to date, in Japan, was associated with radish sprouts shipped from one farm. Importance of different vehicles of infection has changed somewhat over time. Table 2 lists the first recognized outbreaks for different vehicles of infection starting with the first outbreak in 1982 linked to ground beef. Figure 1 shows this data on a timeline. During the 1980s most outbreaks of E. coli O157:H7 were associated with inadequately cooked hamburgers and unpasteurized milk (80;176;178;305). Some later outbreaks have been traced to other dairy products such as cheese and yogurt (24;172;248). In most cases, dairy products were made from unpasteurized milk (121); in others, there was a problem with post- pasteurization contamination (342). Increasingly, contaminated water has been reported as a source of human infection. This includes drinking water sources contaminated with animal feces and also contaminated lake and pool water used for swimming and playing (112). Fruits and vegetables have also been cited as vehicles for human infection with E. coli O157:H7. They have presumably been exposed to untreated manure in the environment or else foods have been washed or irrigated with contaminated water. In a 2005 letter, FDA stated that it was aware of 18 outbreaks of E. coli O157:H7 associated with lettuce and one associated with spinach in the U.S. since 1995. (www.cfsan.fda.gov/~dms/prodltr2.html) A large outbreak in Japan, affecting more than 12,000 persons, was associated with contaminated radish sprouts (145;241) and other outbreaks have been associated with contaminated fruit juices, melon, and salad greens (3;59;105;162;294). More recently a number of outbreaks have occurred among children visiting farms and petting zoos where they come into direct contact with animals carrying E. coli O157:H7 and their environment (158). Outbreaks at county fairs may also result from airborne dispersion of bacteria in buildings used to show animals (57;345). Finally, direct person-to-person infection occurs particularly among children and their caregivers, such as in day care facilities and also within families (205;357).

Reservoirs of E. coli O157:H7 Understanding the epidemiology of this organism requires a knowledge of where these bacteria live and grow in nature (their reservoir) and of how humans come into contact them. Ruminants have been identified as the major reservoir of E. coli O157:H7 with cattle as the most important source of human infections. Other ruminants known to harbor these bacteria include sheep, goats, and deer. STEC bacteria are occasionally isolated from other animals but it is believed that the bacteria are present as transients

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 4 White Paper:

and that the animals acquired these bacteria from meat, foods, or water contaminated by fecal material from ruminants (80). STEC bacteria usually do not cause illness in animals—with a few exceptions such as diarrhea in calves (194). Cattle—the primary reservoir Cattle are probably the most important ultimate source of infections for humans. Of the outbreaks listed in Appendix 1, 64 appear to be associated directly with cattle. These include 32 associated with beef, 10 with “meat,” and 22 with dairy products from cows. In addition there were 15 outbreaks associated with contact with animals at farms or petting zoos and 5 outbreaks linked to exposure to mud, dust or other environmental sources around farm fields and buildings where farm animals are shown. (Some of these animal associated outbreaks may have been due to infected sheep or goats.) Many other outbreaks associated with contaminated water and fresh produce may be indirectly associated with cattle. VTEC have been detected in calves, dairy cows and beef cattle worldwide. Prevalence of these bacteria in cattle and their excreta appears to vary seasonally as well as with the age of the animals and other factors but is generally <10%. Examination of naturally and experimentally infected calves and cattle demonstrated that most E. coli O157:H7 adhere to mucosal epithelium in a short 5 cm long region just proximal from the recto-anal junction. As a result, E. coli O157:H7 is present predominantly on the surface of the cow pats (223;255). Evidence has accumulated indicating that some infected animals may shed E. coli at much higher levels than others. Analyses of 440 E. coli-positive fecal pats in Scotland revealed that E. coli O157:H7 levels ranged from <100 cfu/g (in about 75% of pats) to a high of about 9 × 105 cfu/g in the most contaminated pat. Data on cattle from more than 900 farms suggested that the 20% of animals that are most infectious are responsible for approximately 80% of transmission of this pathogen on the farms. It is not presently known if these “super-shedders” are genetically more hospitable to E. coli O157:H7 (232;233). Generally, shedding of E. coli O157:H7 is higher during warm months and higher in calves just after weaning than earlier or later in life (80;149). Other ruminants Sheep are another significant source of E. coli O157:H7 for human infection and these bacteria have been detected in meat (48;93;94) and in animals from several countries (62– 64;90;166;167;179;199;236;259). Monitoring of a flock of sheep for 16 months revealed that the animals shed E. coli O157:H7 only during the summer and several different strains were shed by sheep in a flock at single time and the strains shed by different sheep changed over time (209). Diet may affect shedding of E. coli (208). Sheep may shed as many as 104 cfu of E. coli O157:H7 per gram feces (265) and these bacteria were found to survive for 15 weeks in manure in a field associated with an outbreak of E. coli O157:H7 among scouts at a camp (264). Infected lambs may have been associated with outbreaks involving visits to farms or petting zoos. Goats are another reservoir of E. coli O157:H7 and other verotoxin producing E. coli (63;124;199) and outbreaks of human disease have been linked to cheese made from unpasteurized goats’ milk (135) and to petting zoos with goats (116;166). Following an oral dose of E. coli O157:H7, attaching and effacing lesions, similar to those seen in cattle, develop in the colon and recto-anal junction of goats (210). Deer are present in significant numbers in some environments also used by cattle, sheep and goats. Since they are also ruminants, deer may serve as a reservoir for E. coli O157:H7 and their droppings may contaminate some fresh fruits and vegetables such as apples (105;139). Numerous studies in the U.S. and abroad have documented the presence of E. coli O157:H7 and other verotoxin-producing E. coli in wild deer (46;129;139;211;296;298;307;349). E. coli O157:H7 has also been detected in some farmed deer (91) and in raw deer meat (253;336). Several cases of human infection with E. coli O157:H7 have been traced to contaminated deer meat (202;288).

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 5

Other animals E. coli O157:H7 has been detected in numerous other animals but none of them are considered a significant source of human infection. In one case, a farmer handling a horse infected with E. coli O157:H7 subsequently developed an infection with the same bacterial strain (89). VTEC bacteria have been detected in several domestic and wild animals including horses (50;161;280), dogs (62;63;161;198), rats (102;103), an opossum (297) and cats (62) and in a few zoo animals including monkeys and lemurs (50) and an orangutan (64). Swine have been found, in several studies, to be infected with E. coli O157:H7 but usually only a small percentage (0.4–14%) of animals test positive (62;63;68;95;168;191;198;198;199;203;269;313;369). Experimentally infected pigs shed these bacteria in their feces for at least two months (108). Rabbits, both wild individuals and animals being raised commercially on some farms, were reported to harbor enterohemorrhagic E. coli (150;214;286). Fecal pellets collected in the summer from wild rabbits on four of six UK farms that harbored VTEC-shedding cattle tested positive for VTEC, including E. coli O157:H7. Rabbit fecal pellets collected in winter tested negative (309). Poultry meat sometimes has E. coli O157:H7 on its surface and these bacteria do persist in the ceca of experimentally infected chicks for as long as 11 months (80;312;326). It was recently reported that 26 of 720 cloacal swab samples from living layer hens in Italian intensive management layer hen farms tested positive for E. coli O157:H7 (123). There are a few reports of the isolation of E. coli O157:H7 in chicken feces (281;313) and turkey feces (168). Shellfish in contaminated waters are known to concentrate some pathogens such as Cryptosporidium. E. coli can be detected in sewage and the possibility exists that pathogenic strains such as E. coli O157:H7 could be present in water contaminated by sanitary sewer overflow or runoff from farm fields. Although there has been one recent report of several strains of STEC, including E. coli O157:H7, isolated from shellfish collected from coastal areas of France (156), it appears that E. coli O157:H7 does not significantly contaminate shellfish as yet (227).

Transport Hosts Birds are thought to be possible transport hosts for E. coli O157:H7. Some wild birds harbor these bacteria and pigeons, for example, might spread these bacteria around a farm environment. E. coli O157:H7 has been isolated from gulls (347), a rook (relative of crows) (133), and pigeons (159;246;311;316). Experimentally infected pigeons continue to shed these bacteria for about two weeks (103). However, a recent study in Colorado suggests that pigeons may not be a major route of transmission of E. coli O157:H7. None of the E. coli isolated from 406 pigeon samples collected at dairies produced shiga-like toxins (276). Flies and beetles, including houseflies and filth flies of several species (7;161;199;249;289;316;332) and dung beetles (362), collected on farms with animals shedding E. coli O157:H7, contain detectable levels of these bacteria. These insects frequent fecal deposits and may be able to transfer these bacteria to foods, feed and water. In experiments with houseflies, E. coli O157:H7 survived and replicated in the mouthparts and crop of the flies for up to 4 days (204;308). Fruit flies, collected from a compost pile of decaying apples and peaches contaminated with E. coli, contained these bacteria both internally and externally and were able to transfer them to wounds in uncontaminated apples. E. coli O157:H7 can grow rapidly in apple wounds. Fruit flies could contribute to widespread contamination of wounded apples that may be processed into cider (186). Slugs are known vegetarian pests that frequently traverse leafy vegetables and may be present on these foods when harvested. Slugs ingest bacteria from the environment and also accumulate bacteria in the mucus surrounding their bodies. Some common gray field slugs collected on a farm in Scotland were found to carry the same pathogenic strain of E. coli as detected in feces from sheep grazing there. Slugs may travel 12 m or more per night so there is a potential for slugs to carry E. coli O157:H7 from manure to vegetables (325).

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 6 White Paper:

Routes of Human Infection Various routes for human infection with E. coli O157:H7 were reviewed in a recent article on the epidemiology of outbreaks of this bacterium in the U.S. (1982–2002) (291). Studies in Canada and France demonstrated that the incidence of HUS and VTEC infection in humans is correlated with indicators of cattle density (165;343). E. coli O157:H7 in ruminant feces may be directly ingested by persons interacting or working with animals. Fecal material may contaminate meat during slaughter, may enter lakes or drinking water sources by action of rain or wind, and may be deposited on fruits and vegetables inadvertently or by use of manure for fertilization. In addition, some animals may transport these bacteria from a fecal source to drinking water or foods. All of these routes are variations of a pattern “from turd to tongue” (278). Direct contact E. coli O157:H7 shed by infected animals may be spread to many surfaces in enclosures where ruminants are kept including the hides of other animals (98). Depending on moisture and humidity, these bacteria may persist on gates, stiles and other farm surfaces for more than four weeks (356). E. coli O157:H7 survives in cattle feces for up to 18 weeks at 15ºC (146). This helps explain why a substantial number of people residing on dairy farms have evidence of current (stool cultures) or past (serologic status) infection with VTEC (361). Several outbreaks among children who visited farms or petting zoos resulted from direct exposure to these bacteria followed by inadequate handwashing. Person-to-person spread of E. coli O157:H7 has been the primary mode of infection in many outbreaks in day cares, schools and hospitals, particularly where there have been lapses in hygiene (55;278;293). In many other outbreaks, some of the cases who consumed contaminated food or water, passed the infection directly to others. Although a majority of children infected with E. coli O157:H7 shed these bacteria in their feces for only a few days, in more seriously ill children, cells of E. coli O157:H7 may be shed for 20–30 days or longer. VTEC bacteria may be present in stool samples even after children become asymptomatic (195). Contaminated food Beef, lamb, and mutton can be contaminated during slaughter and processing by exposure to feces or hides containing E. coli O157:H7. In a recent study in the Midwest, more than 45% of over 330 carcasses tested during July–August contained detectable levels of E. coli O157:H7 (134). Prevalence of these bacteria on carcasses was 43% at pre-evisceration (immediately after hide removal), 18% at postevisceration (after evisceration, splitting and trimming), and 2% after postprocessing (after antimicrobial treatments including hot water and organic acid washes and steam pasteurization). The initial high level of contamination was greatly reduced during processing suggesting that sanitary procedures within these plants were effective. However, it was also true that some carcass samples, from lots in which no preslaughter hide or fecal samples contained E. coli O157:H7, were found to test positive for these bacteria. This suggests that cross-contamination can occur in processing plants. Milk from dairy cows, sheep, and goats may be contaminated with E. coli and other bacteria from the environment. Proper pasteurization will kill these bacteria. Outbreaks of E. coli O157:H7 due to contaminated dairy products are usually associated with unpasteurized milk but there have been some cases of post-pasteurization contamination. Manure is a valuable fertilizer for crops but manure containing E. coli O157:H7 may be a source of contamination for vegetables or fruits that are not normally cooked before eating. In one study, these bacteria were able to survive for 42 days in manure heaps that were turned and for 90 days in unturned heaps (141) while another study found that E. coli O157:H7 was undetectable after 4 weeks in biowaste compost piles (218). E. coli O157:H7 did not grow or survive in dairy wastewater lagoons (292) but did survive for more than two months in garden soil treated with contaminated manure (183;251). Field and greenhouse experiments have demonstrated that both E. coli O157:H7-contaminated manure and irrigation water may cause contamination of vegetables. Onions and carrots grown in soils treated with contaminated manure or irrigated with contaminated water had detectable levels of E. coli

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 7

O157:H7 on their subterranean parts for 2.5 to 5.5 months (183). Lettuce grown in soil amended with contaminated manure did not contain E. coli O157:H7 in leaves (189;190) but spray irrigation of lettuce with contaminated water deposited E. coli O157:H7 on lettuce leaves and these bacteria persisted for up to 30 days (324). Experiments with shredded lettuce, carrots, and cucumbers demonstrated that E. coli O157:H7 could survive and grow on these vegetables even under modified atmospheres used in commercial packaging (2). Foods can also be contaminated with E. coli O157:H7 by cross-contamination during food preparation and by infected workers who don’t practice good hygiene. Several restaurant outbreaks in OR and WA in 1993 were associated with a variety of items from the salad bar but not with steak. All the restaurants obtained their beef from the same source and it was the practice to trim, macerate, and marinate the beef in the same kitchens used for preparation of fruits and vegetables for the salad bar. It appeared that the beef itself was cooked well enough to destroy E. coli O157:H7 but that some raw beef was the source of contamination for the fresh produce (185). Contaminated water Water used for drinking or recreation has been reported as the vehicle of infection for 49 outbreaks: 6 outbreaks associated with water parks and pools, 18 with lakes, springs, canals, and streams, 10 with well water, 12 with “drinking water,” and 3 with tap water. Fecal material from ruminant animals, domestic and/or wild, is the probable source of E. coli O157:H7 in lakes, streams, and wells and for some “drinking water” outbreaks. Drinking water from an unchlorinated source was implicated in a large Missouri outbreak (331). Infected persons are likely the source of bacteria in the pools and water parks and possibly for some other waterborne outbreaks.

RESPONSES TO E. coli O157:H7 OUTBREAKS Today, outbreaks of foodborne disease are featured in almost instant and broad media coverage. This causes anxiety among the general population, lawsuits, and requests for action to protect the health and safety of consumers. The Federal Government through the Centers for Disease Control and Prevention (CDC), United States Department of Agriculture (USDA), Food and Drug Administration (FDA), and the Environmental Protection Agency (EPA) is under pressure and scrutiny and this in turn results in new guidelines, directives, and regulations. State and local public health departments may also be pressured to increase surveillance activities and respond more rapidly to outbreaks. In addition, reports of foodborne disease prompt formation of organizations such as STOP (Safe Tables Our Priority) dedicated to the support of survivors and families of outbreak victims and to dissemination of information on methods for preventing foodborne disease.

Surveillance Animals There is no regular surveillance program to monitor prevalence of infection with E. coli O157:H7 among ruminant animals in the U.S. Some other countries, including Sweden (8), Denmark (256), and The Netherlands (313), have established national monitoring programs to detect VTEC in herds or cattle at slaughter. Some U.S. surveys indicated that prevalence of E. coli O157:H7 was 13% in feces of one group of summer feedlot cattle (217) and 28% in feedlot cattle presented for slaughter at Midwestern plants (134). A recent review reported that the prevalence of E. coli O157:H7 in beef cattle ranged from 0.2 to 27.8% in published data. In addition, a number of non-O157 VTEC serotypes have been reported from beef cattle (177). A review of numerous papers on dairy cattle in the U.S., Canada, Europe, Japan, and Brazil indicated that reported prevalence of E. coli O157:H7 varied widely. (Sampling and analytical methods also differed among these studies.) While there were many reports of prevalences of <20%, there were several very high rates of 40–70%. Ranges of prevalence rates in the U.S. for cows, heifers, and calves were,

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 8 White Paper:

respectively, 0.2–8.4%, 1.6–3.0%, and 0.4–40% (178). Recent surveys of dairy farms in Minnesota (99) indicated that 4.5–5.2% of cows harbored E. coli O157:H7 while 6.5% of dairy cattle in Louisiana tested positive for these bacteria (130). Downer dairy cattle in WI were found to be three times as likely as healthy cattle to harbor E. coli O157:H7 (77). Many surveys may underestimate the actual carriage of E. coli O157:H7 because they are based on analyses of single samples from fecal pats. Rectoanal mucosal swab cultures were found to detect a higher prevalence of infection with E. coli O157:H7 (9.5%) than fecal cultures (4.7%) (157). A higher prevalence of infection was also observed when fecal pats were sampled in several locations rather than at just one site (132). A survey of fecal samples from livestock at 32 county and state fairs found that some of the animals at 31 fairs were shedding E. coli O157:H7. Prevalence among cattle was 11.4% and among sheep and goats was 3.6%. VTEC were also isolated from flies and from environmental samples even after the livestock left the fair (199). Nine of 12 county fairs in Minnesota had cattle infected with E. coli O157:H7 with 11% of manure samples testing positive (99). The relatively high prevalence of E. coli O157:H7 in fair animals was surprising because these animals are usually raised individually or in small groups and are washed frequently. Foods FSIS began a microbiological testing program for ground beef in October 1994. A subset of plants and retail outlets that grind beef are selected randomly each month for testing. This sampling and testing program is based on data on outbreaks of foodborne illness and on information from the Office of Public Health Science. Testing results are available on the FSIS website (www.fsis.usda.gov/Science/Ground_Beef_E.Coli_Testing_Results/index.asp). Figure 2 shows the number of ground beef samples tested from 2001–2005. Figure 3 shows a dramatic decrease in the number of samples positive for E. coli O157:H7 between 2002 and 2003. This is believed to be a result of a 2002 order by FSIS that beef plants reassess their food safety plans. Most plants made major changes to their operations by installing and validating new technologies to control and test for E. coli O157:H7. Results from 2006 testing so far indicate that E. coli O157:H7 contamination is continuing to decrease with 14 positive samples detected among 7,295 samples tested. However, a further recall of approximately 900 lb. ground beef reported on August 18 demonstrates that there may still be some issues to be addressed at some plants. Human illness FoodNet (Foodborne Diseases Active Surveillance Network) is a collaborative active surveillance project to track foodborne illness and involves CDC, USDA, FDA, and ten states in the Emerging Infections Program (CA, CO, CT, GA, MD, MN, NM, NY, OR, and TN). FoodNet began collecting information from five sites in 1996 and has now expanded to monitor about 15% of the U.S. population. In the ten states, public health officials frequently contact directors of over 650 laboratories testing stool samples to find new cases of foodborne disease and HUS and report these to CDC. Goals of FoodNet include: determining the burden of foodborne illness in the U. S., monitoring trends in specific foodborne illness, determining specific foods and settings associated with foodborne illness, and developing and assessing interventions to reduce foodborne illness. Reports of foodborne illness from clinical laboratories are reported by all state health departments to CDC under The National Notifiable Diseases Surveillance System (NNDSS). However, there is some variation among states in the priority and funding given to investigation of foodborne illness and notifiable illnesses and their aggressiveness in tracking down causes of outbreaks and sporadic cases. Many persons with foodborne illness are not seriously ill and do not seek medical care and it is likely that only a fraction of cases is reported to CDC by passive surveillance systems. Incidence of E. coli O157:H7 infections in the U.S. has declined in recent years according to available surveillance data although there was a slight increase in 2005 (Table 3). According to the latest data from FoodNet (2005), incidence of several other foodborne infections has also decreased, including

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 9

Listeria, Campylobacter, and Salmonella. However, incidence of infections with Vibrio has increased (86). The most dramatic decline in cases, as noted by both FoodNet data and cases included in Notifiable Diseases, was between 2002 and 2003 when there was also a dramatic decrease in positive samples of ground beef detected by FSIS sampling (Figure 4). This trend was also detected in Wisconsin (Figure 5). Enter-Net reports that infections with E. coli O157:H7 decreased by 6% between 2000 and 2004 but infections with non-O157 VTEC serotypes has increased. These serogroups are likely underdiagnosed and underreported in most countries, including the U.S. (140). In response to the 1992–1993 Jack in the Box outbreak, scientists at CDC subjected bacterial isolates from several western states to DNA fingerprinting by pulsed-field gel electrophoresis (PFGE) thereby aiding in identifying the vehicle of infection and the extent of the outbreak. Prompt recognition of the outbreak probably prevented hundreds of cases of illness. CDC then developed standardized PFGE methods for several foodborne pathogens and created a network of national and state public health and regulatory laboratories that can submit PFGE patterns of bacterial isolates to a central database. In 1998, the inauguration of PulseNet was announced by Vice President Gore and it has since become instrumental in facilitating early recognition of common source outbreaks (www.cdc.gov/pulsenet/). Database administrators analyze PFGE data submitted by participating laboratories to identify clusters of infection and can electronically alert participants to outbreaks. Epidemiologists throughout the country can also rapidly compare PFGE patterns of local foodborne bacterial pathogens and determine whether there are similar strains causing illness elsewhere (151;160). Regional and national meetings, such as the First Foodborne Epidemiologist Meeting held in 2004, bring together foodborne epidemiologists from throughout the country. These meetings are being held more frequently and serve to educate public health workers on the latest methods and research and allow participants to share information on state programs. State and federal laboratory personnel present PulseNet data on sporadic and outbreak cases of foodborne illness.

Regulations Federal regulations for various aspects of food handling and processing are promulgated by FDA and USDA while EPA is in charge of clean water regulations. In addition, many state and local agencies have regulations that impact food, drinking water, swimming beaches and pools and agricultural fairs and petting zoos. Many of these regulations were established after significant human disease outbreaks and deaths and/or broad media coverage highlighted shortcomings in handling of food, water, or animals. Figure 6 depicts a timeline showing the association between significant outbreaks and new regulations. Public outrage, aroused by the book The Jungle, lead to the passage and implementation of food inspection laws, including the Federal Meat Inspection Act (FMIA) and the Pure Food and Drug Act for non-meat products in 1906. The Poultry Products Inspection Act (PPIA) was enacted in 1956. FMIA and PPIA require mandatory inspection of livestock before slaughter and mandatory post-mortem inspection of all carcasses, establish sanitary standards for slaughterhouses and meat processing establishments, and authorize USDA to inspect meat processing and slaughtering operations. Only unadulterated carcasses are approved for further distribution to customers. Meat is considered adulterated “if it bears or contains any poisonous or deleterious substance which may render it injurious to health” or “if it consists in whole or in part of any filthy, putrid, or decomposed substance or is for any other reason unsound, unhealthful, unwholesome, or otherwise unfit for human food” (www.fda.gov/opacom/laws/meat.htm#SUBCHAPTER_1). USDA regulations implementing FMIA and PPIA are found in Title 9 of the Code of Federal Regulations and FSIS periodically issues Directives that provide instructions for inspectors. In 1967, the Wholesome Meat Act updated FMIA to require inspection of all meat processed and sold within the same state. The Wholesome Poultry Act of 1968 instituted similar requirements for intrastate processing of poultry. In addition, at least twenty-eight states have their own meat and/or poultry inspection programs covering small and very small establishments. These programs are run cooperatively with FSIS.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 10 White Paper:

Beef E. coli O157:H7 was first recognized as a foodborne pathogen after a 1982 outbreak affecting 20 people who had consumed undercooked hamburger. During the next 9 years several other outbreaks associated with beef caused illness in 30–70 persons each. Then in 1991, there was a major ground meat-related outbreak in Canada and in 1992–1993, an outbreak associated with a fast food restaurant, affected over 700 people who had eaten undercooked hamburger in the western U.S. with 37 children developing HUS and 4 deaths. This prompted the passage of a several regulations by FSIS to improve meat safety (254). • 1994: Under the authority of the Federal Meat Inspection Act, Michael Taylor of FSIS, in a speech in September, declared E. coli O157:H7 to be an adulterant in ground beef. This was challenged in a Texas court but the court decided that USDA had good reason to consider these bacteria as adulterants. • 1994: FSIS Directive 7235.1 required the placement of safe handling labels on packages of raw meat and poultry. These labels address storage, cooking, and holding practices to minimize or prevent growth of pathogenic bacteria (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/7235.1.pdf). • 1994: FSIS Notice 50-94 instituted testing of raw ground beef for E. coli O157:H7. This notice was later replaced in 1998 by Directive10,010.1. • 1996: HACCP (Hazard Analysis and Critical Control Point) systems were mandated as a systematic procedure for determining critical points during processing when meat could be contaminated and instituting appropriate controls to prevent contamination. This rule establishes a testing program for the pathogen, requires slaughter plants to routinely test carcasses for generic E. coli, and requires all plants to incorporate an antimicrobial process and have in place sanitation standard operating procedures (SSOPs) (www.fsis.usda.gov/OPPDE/rdad/FRPubs/93- 016F.pdf ) (Fed. Register 61(144):38806-38989). • 1998: FSIS Directives 6150.1, rev 1 and 6420.1 told inspectors to enforce zero tolerance for visible fecal, ingesta and milk contamination of poultry and livestock carcasses at slaughter. This directive was revised in 2004 to include head, cheek and wesand meats because these may be included in ground beef (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/6420.2.pdf). • 1998: FSIS Directive10,010.1 revised policy for sample collection and testing to emphasize establishments perceived to be a greater risk. Establishments using validated pathogen reduction interventions on beef carcasses and that had not identified a positive sample within the previous six months would not need to be tested by FSIS (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/10.010.1.pdf). • 1998: Performance standards for lethality and stabilization for cooking of meat were updated (9CFR 318.17). • 1999: USDA issued rules allowing irradiation of refrigerated or frozen/uncooked red meat and meat products to destroy pathogenic bacteria including E. coli O157:H7 (64 FR 72150) (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/7700.1Rev1.pdf). • 1999: Raw ground beef products, including trimmings, would also be considered adulterated if they contain E. coli O157:H7 (Fed. Register 64:2803-2805). • 2002: FSIS directed producers of ground beef to reassess HACCP plans in light of new epidemiological data on E. coli O157:H7, including improvements in analytical tests, data on conditions in feedlot pens as related to shedding of VTEC, and data on contamination of carcasses as related to prevalence of VTEC on hides and in feces. Implementation of critical controls was required (Fed. Register 67(194):62325). • 2003: USDA banned all downer cattle from the human food chain. This was intended to prevent possible transmission of BSE but may also have decreased the prevalence of E. coli O157:H7 in

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 11

meat from culled animals since downer dairy cattle have been shown to have a higher prevalence of E. coli O157:H7 (77) (www.usda.gov/news/releases/2004/01/0457.htm). • 2004: Directive 10.010.1 was revised to require all federally inspected plants be subject to testing for E. coli O157:H7. This directive also provided instructions for follow up actions if a sample tested positive and for verifying control of products that are presumptive or proven positive for E. coli O157:H7 (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/10.010.1.pdf). • 2005: Directive 7700.1, FSIS revised and updated instructions regarding irradiation of meat and poultry products in official establishments including off-site irradiation of product (www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/7700.1Rev1.pdf). • 2006: Directive 7120.1, Amend 8 provided updated information on safe and suitable ingredients used in the production of meat and poultry products. This includes substances that could be used to destroy bacteria or inhibit their growth. Juice Following a multistate outbreak of disease linked to E. coli O157:H7 in unpasteurized apple juice in 1996 (105), FDA published a final rule (Federal Register 63(130):37029-37056) requiring a warning label on fruit juices that have not been processed to prevent, reduce, or eliminate pathogenic microorganisms. Following another juice-borne outbreak of Salmonella in orange juice in 2000, another final rule published by FDA in 2001 required all juice manufacturers to develop and implement a HACCP plan to prevent contamination of juices with dangerous pathogens. Treatments should reduce pathogen levels by 5 logs (Federal Register 66(13):6138-6201). Fresh produce FDA has published regulations requiring current good manufacturing practices (GMPs) for food processors under its jurisdiction. However, these regulations do not include “establishments engaged solely in the harvesting, storage, or distribution of one or more raw agricultural commodities. …FDA will, however, issue special regulations if it is necessary to cover these excluded operations” (21U.S.C. §321(r)). Despite this exclusion, FDA can still regulate fresh produce as “food” subject to the adulteration provisions of the Food, Drug, and Cosmetic Act. A food shall be deemed to be adulterated “if it has been prepared, packed, or held under insanitary conditions whereby it may have become contaminated with filth, or whereby it may have been rendered injurious to health.” Therefore, FDA can take enforcement action against an agricultural producer if it determines that food is being produced under insanitary conditions. In 1998, FDA issued a “Guide to Minimize Microbial Food Safety Hazards for Fruits and Vegetables,” that recommended good agricultural practices and GMPs (www.fda.gov/ohrms/dockets/98fr/97n0451a.pdf). However, foodborne outbreaks related to fresh produce continued. In a 2005 letter, FDA noted that 8 outbreaks of E. coli O157:H7 associated with lettuce and spinach could be traced to Salinas, California. Creeks and rivers in this area test positive for E. coli O157:H7 periodically and some areas are highly susceptible to localized flooding. FDA reminded producers that it considers ready-to-eat crops (such as lettuce) to be adulterated if they have been in contact with flood waters due to potential exposure to sewage, animal waste, pathogens, and other contaminants (www.cfsan.fda.gov/~dms/prodltr2.html). Drinking water According to EPA regulations, a system that operates at least 60 days per year, and serves 25 people or more or has 15 or more service connections, is regulated as a public water system under the Safe Drinking Water Act of 1974. If a system is not a public water system as defined by EPA’s regulations, it is not regulated under the Safe Drinking Water Act, although it may be regulated by state or local authorities. Revised National Primary Drinking Water regulations are published in the Code of Federal Regulations (40CFR141.1).

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 12 White Paper:

Under the Safe Drinking Water Act, EPA requires public water systems to monitor for coliform bacteria. Systems analyze first for total coliforms, because this test is fast. Any time a sample is positive for total coliforms, the same sample must be analyzed for either fecal coliforms or E. coli. Both are indicators of contamination with animal waste or human sewage (www.epa.gov/safewater/sdwa/index.html). Systems serving 25 to 1,000 people typically take one sample per month. Some states reduce this frequency to quarterly for ground water systems if a recent sanitary survey shows that the system is free of sanitary defects. Systems using surface water, rather than ground water, are required to take extra steps to protect against bacterial contamination because surface water sources are more vulnerable to such contamination. At a minimum, all systems using surface waters must disinfect. Disinfection, including chlorination, will kill E. coli O157:H7. Bottled water is regulated by FDA as a food. Standards of quality are listed in the Code of Federal Regulations (21CFR165.110); these include allowable limits for turbidity, color, odor, coliforms, radioactivity and for 70 chemicals. Coliform levels should not exceed 4 cfu/100 mL in any single sample. State and local governments also regulate bottled water. Swimming pools and beaches These are regulated by state and local authorities. Pools and water parks are usually tested regularly for fecal contamination that may result from accidents with young children. Lakes and rivers can also be contaminated by human visitors but are perhaps more at risk of contamination during heavy rains when manure may wash into nearby bodies of water. Water at swimming beaches may not be tested as frequently for coliforms but should be tested when there is likely to be a problem with run off. Fairs and petting zoos These are regulated by state and county agriculture or public health departments. During the past several years, regulations requiring more hand-washing stations, warning signs, disinfecting of handrails, etc., and cleanliness in animal enclosures have been enacted in a number of states, including North Carolina and Washington, that experienced outbreaks of E. coli O157:H7 at county fairs. The CDC has published a compendium of standardized recommendations to prevent disease associated with animals in public settings. The single most important recommendation was proper washing of hands (57).

Industry Initiatives: Intervention Strategies The emergence of E. coli O157:H7 as an important foodborne pathogen has prompted industry initiatives to reduce contamination and improve food safety. These included the use of better equipment and testing procedures as well as improved management systems (HACCP). From 1996–2000, the Economic Research Service of USDA estimates that the meat and poultry industry spent about $180 million/year on improvements in food safety. More than $7 million has been spent since 1994 on applied research related to control of E. coli O157:H7. This research led to commercialization of several processes to reduce pathogens on carcasses during processing: (a) vacuuming of carcasses with steam or hot water (126); (b) thermal pasteurization in which carcasses are rinsed with water at 180ºF; and (c) rinsing with mild organic acids. Steam pasteurization was found to be particularly effective (279). By 1997, the two largest beef packing companies ordered such equipment for all their plants. Following the large 1992–1993 outbreak in the western U.S. linked to hamburgers, the beef industry created a blue-ribbon task force to develop plans to aggressively address the problem of E. coli O157:H7 in cattle and beef. Among the accomplishments of this task force were development and implementation strategies for HACCP programs, safe handling labels on packages of beef, and steam vacuuming technology. The Beef Industry Food Safety Council was formed in 1997 to address the problem of foodborne pathogens in beef. This council facilitates research activities and develops and implements education programs for both consumers and the industry. Because it is not readily apparent to consumers whether meat or produce is contaminated with E. coli or other bacteria, if a significant outbreak of foodborne disease is associated with a particular company, results can be economically devastating. One of the most well known cases was Hudson Foods which

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 13

went out of business after its hamburger was implicated in an outbreak and 25 million pounds of product had to be recalled. The industry has instituted a 100% test and hold program for any product that will be ground. Positive product was diverted to cooked products. It should be noted that this testing, because of sampling limitations, will not detect all lots that contain E. coli O157:H7. However, improvements in cleaning carcasses and testing programs have greatly reduced positive samples of ground beef detected by FSIS in recent years (Figure 3). Demands by large meat and poultry buyers and many foreign buyers have driven some improvements in food safety practices. After the 1992–1993 outbreak, Jack in the Box canceled all its current contracts with hamburger suppliers and required superior food safety controls including more stringent testing and strict temperature control from its future suppliers. Safer hamburgers were produced using this Bacterial Pathogen Sampling and Testing Program. Many other major meat buyers including McDonald’s, Burger King, Kroger and other fast food and grocery chains have also mandated their own safety standards in contracts with their suppliers. Ground beef manufacturers are using test and hold procedures to ensure the safety of ground that is shipped out to restaurants and consumers. Fresh salad vegetables have also been implicated as a source of foodborne illness. Following the 1998 Guidance from the FDA (www.fda.gov/ohrms/dockets/98fr/97n0451.pdf) to producers of fresh fruits and vegetables, a majority of the lettuce/leafy green industry adopted the suggested good agricultural practices (GAPs). In April 2006, representatives of trade associations for growers and marketers of fresh produce published a document to aid their members in implementing safeguards during growing, harvesting, value-added operations, distribution, retail, and food service operations: Commodity Specific Food Safety Guidelines for the Lettuce and Leafy Greens Supply Chain (www.cfsan.fda.gov/~acrobat/lettsup.pdf). Restaurants have also established their own procedures for ensuring that meat, particularly hamburger, is well-cooked to destroy bacterial pathogens and to prevent cross-contamination during food preparation. The 2005 the FDA Model Food Code also was directed toward a more proactive approach through active managerial control in reducing foodborne illness rather than reacting to reported illnesses and outbreaks. This along with the more strenuous recommendation related to no bare hand contact of foods, single use of gloves, and more enhanced reporting of employee health from the employee and restrictions by employers may also have contributed to improved public health as witness by the yearly reduction in outbreaks by enteric bacterial pathogens in the U. S. (CDC – personal communication).An important part of this process is education of employees on the importance of cleanliness and proper cooking. Menus contain a warning that the restaurant will not be responsible for illness if patrons request that meat not be fully cooked.

DISCUSSION AND SUMMARY

Epidemiology Evaluation of current practices and recommendations Active surveillance programs, such as FoodNet, identify a greater proportion of cases of E. coli O157:H7 than passive programs. It has been estimated that as many as 4–8 times as many symptomatic cases occur for each case identified by active surveillance (56). Sporadic cases of E. coli O157:H7, confirmed in laboratories, have remained stable over several years but incidence varies for different states. Many factors may affect whether cases are ascertained: • It is not always possible to obtain samples of implicated foods after an outbreak is recognized but greater efforts could probably be made to identify pathogens and food vehicles if more resources were put into epidemiological investigations. In Japan, it is the practice to reserve a portion of foods served and store it in the freezer for later analysis, if necessary (162). • Successful investigation of foodborne disease outbreaks requires sufficient resources for collecting and analyzing food and clinical samples. Of 336 foodborne outbreaks reported to

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 14 White Paper:

FoodNet in 1998–1999, 71% had no identified etiology. Reports on these 237 outbreaks indicated that no stool samples were collected in 156 outbreaks and neither food nor stool samples were analyzed in 130 outbreaks. Causative agents cannot be determined if food and/or clinical samples are not examined. • Patients may not seek medical attention or doctors may not order analyses of stool samples. For example, the percentage of stool samples with bloody diarrhea that were cultured for E. coli O157:H7 was 58% in Georgia and 96% in Connecticut (56). • Not all laboratories routinely culture stool samples for E. coli O157:H7 and results are not always reported (56). Stool samples from suspected VTEC infections may be cultured for E. coli O157:H7 but test negative if a non-O157 serotype is responsible. Extrapolating from data provided by FoodNet and from data in peer-reviewed articles in the literature can be problematic on two counts. (a) FoodNet sites were not chosen to be representative of the U.S. population as a whole A demographic comparison of the population in the FoodNet sites with the total U.S. population reveals a similar composition by age. However, Hispanics are underrepresented in FoodNet populations as are those in lower socio-economic groups (163). (b) Only a small fraction of outbreaks reported to national surveillance programs are subsequently described in peer-reviewed literature. A review of data on outbreaks of infectious intestinal disease in the UK from 1992 to 2003 pointed out that there were 1763 outbreaks reported to authorities but only 55 were presented in peer- reviewed journals. Of 45 VTEC outbreaks reported to the national surveillance program, only 7 were described in the literature (261). Outbreaks reported in the literature tend to be those for which microbiological and epidemiological evidence is strongest or more out of the ordinary. This may skew perceptions on the frequency of different vehicles of infection as the UK study found that reports in the literature overemphasized the importance of milk/milk products, deserts, and miscellaneous foods and underestimated importance of meat, fish, poultry and eggs. In addition, if incidence reports are not adjusted for outbreaks, then data may also be misleading. For example, in one day care outbreak in Minnesota, 43 cases were confirmed but most of these were identified because all children and staff at the day care were cultured and many would probably not have been identified if the outbreak hadn’t occurred (56). Recommendations: • Greater uniformity is needed in statewide investigations and reporting of foodborne illness. Public health systems in some states, such as Minnesota, are very well organized and all specimens are sent to state labs for analysis. Other states with different priorities are less aggressive and rigorous in testing samples and conducting investigations. They should be encouraged to improve funding for public health and to adopt and incorporate better epidemiological procedures and improved laboratory methods. • Federal grants could be targeted to improving laboratory facilities and training for state epidemiologists, as needed. Greater participation by all states will aid in the rapid identification of multistate outbreaks. (Improving laboratory facilities and expertise at state levels can be considered part of Homeland Security preparations to detect possible outbreaks associated with biological or chemical weapons.)

Interventions—Recent Improvements Interventions undertaken during recent years by numerous organization concerned with food production and food safety have led to a decline in outbreaks and cases of E. coli O157:H7 (Figures 4 and 5). These include: • Increased emphasis on HACCP and pathogen control by industry and government. This includes improvements at slaughter and processing plants (hide washing, steam vacuuming, hot water and antimicrobial rinses, hold and test programs), at restaurants (proper cooking of hamburger, more

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 15

care about cross-contamination, notes on menus), and improvements at packing plants and retail stores. • Increased educational efforts targeted at doctors (1999, American Medical Association Program), workers, and consumers. Numerous web sites and publications maintained by academic institutions, industry trade groups, state and local health departments, and the federal government (e.g. www.fightbac.org/, originally launched in 1996) have made the public more aware of the potential hazard of undercooked hamburger and the importance of hand washing. Articles in the popular press also reinforce these concepts. Increasingly, agricultural fairs and petting zoos have been posting more signs and providing more hand washing facilities. • Increased cooperation between federal and state agencies with CDC and FDA sharing more PFGE patterns through PulseNet and federal regulators attending more state and regional meetings on epidemiology and foodborne disease. • Improved training to make APHIS inspectors more knowledgeable about risk-based inspection. • Litigation surrounding some high profile outbreaks of foodborne disease. While not primarily intended to be an educational tool, publicity surrounding some outbreaks has undoubtedly alerted consumers to important food safety issues.

Further Research Much of the research on controlling E. coli O157:H7 has been devoted to ensuring the safety of meat after slaughter. Yet outbreaks of illness associated with meat and fresh produce still occur, indicating that further efforts need to address preslaughter conditions of cattle (78;344;371). Several studies have documented that some individual animals shed very high levels of E. coli O157:H7 while fecal material from other animals in the same herd contains much lower concentrations of these bacteria (223;232;233). Carcass contamination with E. coli O157:H7 is more likely when feces contain higher concentrations of this pathogen (136;267). Research to determine why some cattle are “supershedders” and how to identify them would help reduce transmission in herds and carcass contamination. If super-shedding cattle can be identified, then some interventions noted below could be targeted at those animals. Although it is impossible to completely eliminate pathogenic bacteria from all cattle, some products and procedures are being tested for efficacy in reducing levels of E. coli O157:H7 in cattle. Further research is needed to establish parameters that will make these interventions more consistently effective and commercially useful. These include: • probiotics and competitive exclusion: Probiotic bacteria are harmless or beneficial and compete with pathogens to reduce or prevent their colonization of the gut. Results have not been consistent in “real world” studies but research should be directed at finding more effective combinations of probiotic microbes (337;365;367). • antibiotics: These compounds are used to treat illness in animals and to improve feed efficiency and growth. Neomycin treatment has been shown to significantly decrease fecal shedding of E. coli O157:H7 in cattle and might be useful under some conditions (78). However, there is a trend to reduce antibiotic use in animals to minimize the potential for inducing drug-resistant pathogenic bacteria. This should be considered in assessing recommendations for future use of antibiotics. • bacteriophages: Phages are viruses that attack bacteria. Some are very specific in regard to the bacteria they destroy. As with probiotics, results in real world studies are not consistent but research should continue (263;315). FDA has recently approved the use of certain phages on foods to reduce or prevent growth of pathogens. • vaccines: Virulence factors secreted by the type III system of E. coli O157:H7 were found to be effective components of a vaccine that significantly reduced shedding of these bacteria by cattle in a feedlot setting (284). Further field tests of different vaccine types should determine whether this is a cost effective way to reduce E. coli O157:H7 levels in cattle.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 16 White Paper:

• chlorate: Chlorate can be metabolized by E. coli O157:H7 to form chlorite but as chlorite accumulates the bacteria die. Some experiments indicate this might be a useful way to reduce carriage of E. coli O157:H7 shortly before slaughter (13;321). • ruminant diet: Diet may affect shedding of E. coli O157:H7 as several studies have demonstrated that inclusion of barley increases shedding (76;114). Other studies have indicated that an abrupt change from finishing rations to forage decreases E. coli O157:H7 (78). But feeding trials with different compositions of dietary grains do not always yield consistent results and further research is needed to determine if this approach can be an effective control. • water trough hygiene: E. coli O157:H7 can be present and persist in water troughs on farms and in feedlots (316;317). These bacteria can survive in sediments in troughs for as long as eight months and serve as a continuing source of reinfection for a herd (216). More data on the effects of frequency of cleaning troughs and on antimicrobials that might be used for cleaning or added to water could suggest methods for reducing transmission among cattle on farms and in feedlots. In addition to reducing levels of E. coli O157:H7 in cattle, further efforts are needed to ensure safety of fresh produce. It is very difficult to sanitize raw fruits and vegetables after harvest because pathogens may reside in inaccessible parts of the food. Therefore, strategies for prevention of preharvest contamination from manure or runoff from livestock operations is essential and should be further investigated (60;61). Outbreaks at petting zoos and fairs can be decreased by hand washing but animal owners and caretakers should be aware that E. coli O157:H7 can survive for weeks in feces if conditions are right (58) and can also persist for days on surfaces (356). E. coli O157:H7 can also survive in bedding, such as wood chips and sawdust, and can even grow in bedding contaminated with urine (120). There is some controversy as to whether bedding for animals should be cleaned out daily (with the attendant risk of creating aerosols of the pathogens that may be inhaled or deposited on other surfaces) or whether new clean bedding should be placed on top of the old bedding for several days to cover up and keep the pathogens in place during the fair. Further data on effects of different methods of handling bedding and wastes and on sanitizing surfaces may reduce outbreaks at fairs and petting zoos. More effective educational efforts may also be useful in minimizing outbreaks at swimming pools, water parks, and beaches. Disease caused by E. coli O157:H7 in humans is still not well understood. Reported data on the use of antibiotics for treatment of infections with E. coli O157:H7 is conflicting and should be resolved. As yet there is no good animal model for HUS to test various hypotheses about mechanisms for dispersal of shiga toxins through the body, in particular to the kidney. Mechanisms of action of the toxins, interactions of different virulence factors, and potential for transfer of virulence to other bacteria need to be further delineated.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 17

82 95 hamburger Drinking water 98 95 Wat er par k Deer jerky 86 04 95 98 Unpasteurized Petting zoo lettuce Tap water milk 96 90 apple juice 02 Wel l wat er 91 96 Alfalfa sprouts Lake water 98 91 Pool water Cheese curds Apple cider 93 96 Salad bar 94 00 Person to person salami Farm animals

83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03

Figure 1. Timeline of appearance of different vehicles for human infection with E. coli O157:H7.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 18 White Paper:

60 12,000

50 10,000

40 8,000

30 6,000

20 4,000

10 2,000

0 0 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005

Figure 2. FSIS ground beef samples analyzed by year 2001-2005 Figure 3. E. coli O157:H7 isolates from FSIS sampled ground beef 2001-2005

4,000 300

3,500 250 3,000 200 2,500

2,000 150

1,500 100 1,000 50 500

0 0 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005

Figure 4. Nationwide E.coli O157:H7 cases by year , 2001–2005 Figure 5. Wisconsin E. coli O157:H7 cases by year, 2001–2005

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 19

Vehicle (Reference) New Regulations

2006

FDA warning letter to lettuce producers lettuce (252) 2005 All plants to be tested for E. coli O157:H7 2004

blade tenderized beef (212) 2003 Downer cattle not allowed in food

beef (370) 2002 HACCP and pathogen reduction required for juices beef (319;346) 2001 beef (370) 2000 beef (370) Irradiation for ground beef beef (182;187) 1999 Revised testing requirements

beef (370) 1998 Zero tolerance for feces on carcasses Guide to minimize beef (318) 1997 HACCP required for meat processors apple cider (170) 1996 Warning labels required on unpasteurized juices apple juice (14;105) 1995 beef (79) Testing required for E. coli O157:H7 meat (9;144) 1994 Safe handling labels for raw meat E. coli declared an adulterant beef (300;340) 1993

ground beef (341) 1992

apple cider (59) 1991

1990 ground beef (54;238;270;273;302;303) to 1983 ground beef (299;353) 1982 First E. coli O157:H7 outbreak

Figure 6. Timeline of important U.S. outbreaks and new regulations to control E. coli O157:H7.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 20 White Paper:

Table 1. Large outbreaks of E. coli O157:H7 (>100 cases).

Year # Cases Location; Vehicle Reference

1996 12,680 Sakai, Japan; radish sprouts (145;241;363) 2000 2,300 Canada; drinking water (Walkerton) (175) 1999 >1,000 US; well water (New York) (97) 2000 788 US; raw beef, cross contamination of other foods (Sizzler) (370) 1992–93 >700 US; hamburger at fast food restaurants (Jack in the Box) (52;69;119;341) 1995 633 Fife, Scotland; sewage contamination of drinking water (193) 1991 521 Canada; minced beef and caribou (268) 1996–67 512 Scotland; meat from one shop (109) 1996 503 Scotland; lunch foods (277) 1997 332 UK; restaurant food (17) 1999 329 US: IL, KY, MO; beef (182) 1990 243 US: MO; well water (331) 1999 159 Canada; petting zoo (283) 2005 157 UK; sliced cooked meat (305) 1998 157 US: WY; tap water (266) 1999 143 Canada; salami (226) 1999 127 US: NY; well water, total of 781 cases, some C. jejuni (258;260) 2005 120 Sweden; lettuce (109) 1999 114 UK; milk, pasteurized (153) 1995–96 110 Sweden; unknown (368) 2002 109 Canada; salads (67) 1997 108 US; alfalfa sprouts (107) 2004 108 US: NC; petting zoo (116) 2000 102 US: UT; irrigation water used for drinking (215) 1994 >100 Scotland; pasteurized milk (342)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 21

Table 2. E. coli O157:H7: First reports of vehicles/modes of infection.

Year Vehicle Location Reference

1982 beef, ground (1st reported outbreak) US (299;353) 1985 meat, cooked Canada (81;206) 1986 milk cows, unpasteurized US; Canada (178;230) 1988 beef, roast US: WI (302) 1989 water, well Canada (294) 1990 food, restaurant UK (229) 1990 person-person transfer Scotland and Israel (205;219) 1990 water, drinking UK (122) 1990 water, tap Japan (5) 1991 yogurt UK (248) 1991 water, lake US: OR (201) 1991 apple cider US: MA (59) 1992–93 cheese, cows’, unpasteurized France (121) 1992 water, pool UK: Scotland (71) 1993 salads US: OR, WA (184) 1994 milk, cows pasteurized UK: Scotland (301;342) 1994 salami US: WA, CA (9) 1994 sandwiches US: WI (370) 1994 animals, farm UK: England (320) 1995 milk, goats, unpasteurized Czech Republic (65) 1995 ham, cooked UK: England (359) 1995 lettuce US and Canada (3;285) 1995 deer jerky, meat US: OR (202;288) 1995 potatoes UK (96) 1995–96 sausage Germany (12) 1996 apple juice (Odwalla) US and Canada (105) 1996 food handler Australia (234) 1996 meat, butcher shop UK: Scotland (109) 1996 sprouts, radish Japan (145;241) 1997 sprouts, alfalfa US: MI, VA (70;107) 1997 environment (farm field) UK (110) 1997 cakes, cream filled UK (262) 1998 cheese curds US: WI (131) 1998 cream, unpasteurized UK (24) 1998 water park US: GA (49) 1999 beach UK (31;164) 1999 animals, petting zoo Canada (283) 2000 foods, deli UK (36) 2000 water, stream US: CA (215) 2003 beef, steak US: MN (212) 2003 spinach US: CA (294) 2004 cheese, goats’ France (135)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 22 White Paper:

Table 3. Data from surveillance summaries of foodborne and notifiable diseases from CDC.

Notifiable cases Foodborne Outbreaks FoodNet Data

(84;86) (E. coli) (82;83) (E. coli O157:H7) (87;88) Year E. coli O157:H7 non-O157 Outbreaks Cases Deaths cases/100,000 relative rates

2006 *1,391 2005 2,621 908 1.06 2004 2,544 626 0.9 0.6 2003 2,671 408 1.06 0.7 2002 3,840 254 1.68 1.05 2001 3,287 191 1.60 0.95 2000 4,528 2.0 1.2 1999 4,513 2.0 1.25 1998 3,161 2.4 1.0 1997 2,555 2.1 1.0 1996 2,741 11 325 1 3.0 1.0 1995 2,139 25 393 1 1994 1,420 25 902 0 1993 15 1,340 5 1992 3 19 0 1991 3 33 0 1990 2 80 0 1989 1 3 0 1988 2 109 0

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 23

Reference List 1. Abbas Z, Balram C, MacDonald BW, Giffin CS, Armini J, and Panaro L. 2005. An investigation of two simultaneous E. coli O157:H7 outbreaks in health region 3, New Brunswick, August to September 2003. Can Commun Dis Rep 31(22):229–235. 2. Abdul-Raouf UM, Beuchat LR, and Ammar MS. 1993. Survival and growth of Escherichia coli O157:H7 on salad vegetables. Appl Environ Microbiol 59:1999–2006. 3. Ackers ML, Mahon BE, Leahy E, Goode B, Damrow T, Hayes PS, Bibb WF, Rice DH, Barrett TJ, Hutwagner L, Griffin PM, and Slutsker L. 1998. An outbreak of Escherichia coli O157:H7 infections associated with leaf lettuce consumption. J Infect Dis 177:1588–1593. 4. Ackman D, Marks S, Mack P, Caldwell M, Root T, and Birkhead G. 1997. Swimming-associated haemorrhagic colitis due to Escherichia coli O157:H7 infection—evidence of prolonged contamination of a fresh water lake. Epidemiol Infect 119:1–8. 5. Akashi S, Joh K, Tsuji A, Ito H, Hoshi H, Hayakawa T, Ihara J, Abe T, Hatori M, Mori T, and Nakamura T. 1994. A severe outbreak of haemorrhagic colitis and haemolytic uraemic syndrome associated with Escherichia coli O157:H7 in Japan. Eur J Pediatr 153:650–655. 6. Al-Jader L, Salmon RL, Walker AM, Williams HM, Willshaw GA, and Cheasty T. 1999. Outbreak of Escherichia coli O157 in a nursery: lessons for prevention. Arch Dis Child 81:60–63. 7. Alam MJ and Zurek L. 2004. Association of Escherichia coli O157:H7 with houseflies on a cattle farm. Appl Environ Microbiol 70:7578–7580. 8. Albihn A, Eriksson E, Wallen C, and Aspan A. 2003. Verotoxinogenic Escherichia coli (VTEC) O157:H7 a nationwide Swedish survey of bovine faeces. Act Vet Scan 44:43–52. 9. Alexander ER, Boase J, Davis M, Kirchner L, Osaki C, Tanino T, Samadpour M, Tarr P, Goldoft M, Lankford S, Kobyashi J, Stehrgreen P, Bradley P, Hinton B, Tighe P, Pearson B, Flores GR, Abbott S, Bryant R, Werner SB, and Vugia DJ. 1995. Escherichia coli O157:H7 outbreak linked to commercially distributed dry-cured salami—Washington and California, 1994. Morbid Mortal Weekly Rep 44:157–160. 10. Allerberger F, Rossboth D, Dierich MP, Aleksic S, Schmidt H, and Karch H. 1996. Prevalance and clinical manifestations of shiga-toxin producing Escherichia coli infections in Austrian children. Eur J Clin Microbiol Infect Dis 15:545–550. 11. Allerberger F, Wagner M, Schweiger P, Rammer HP, Resch A, Dierich MP, Friedrich AW, and Karch H. 2001. Escherichia coli O157 infections and unpasteurised milk. Euro Surveill 6(10):147–151. 12. Ammon A, Petersen LR, and Karch H. 1999. A large outbreak of hemolytic uremic syndrome caused by an unusual sorbitol-fermenting strain of Escherichia coli O157:H–. J Infect Dis 179:1274–1277. 13. Anderson RC, Carr MA, Miller RK, King DA, Carstens GE, Genovese KJ, Callaway TR, Edrington TS, Jung YS, and McReynolds JL. 2005. Effects of experimental chlorate preparations as feed and water supplements on Escherichia coli colonization and contamination of beef cattle and carcasses. Food Microbiol 22:439–447. 14. Anonymous. 1996. Outbreak of Escherichia coli O157:H7 infections associated with drinking unpasteurized commercial apple juice—British Columbia, California, Colorado, and Washington, October 1996 Morbid Mortal Weekly Rep 45(44)::975. 15. Anonymous. 1996. VTEC O157 infection in West Yorkshire associated with the consumption of raw milk. CDR Weekly 6:181. 16. Anonymous. 1997. Escherichia coli O157 family outbreak in the Highlands. SCIEH Weekly Rep 31:77. 17. Anonymous. 1997. Escherichia coli O157 outbreak in Lincolnshire. CDR Weekly 7:101. 18. Anonymous. 1997. European collaboration identifies an outbreak of Escherichia coli O157 infection in visitors to Fuerteventura, Canary Islands. CDR Weekly 7(15):127. 19. Anonymous. 1997. Outbreak of E. coli infection in a nursing home in Arboath. SCIEH Weekly Rep 31:29. 20. Anonymous. 1997. Outbreak of E. coli O157 infection associated with a butchers shop in Hawick. SCIEH Weekly Rep 31:51. 21. Anonymous. 1997. Two outbreaks of Vero cytotoxin producing Escherichia coli O157 infection associated with farms. CDR Weekly 7(30):263, 266. 22. Anonymous. 1997. Update on the nosocomial outbreak of E. coli O157 at Falkirk and District Royal Infirmary. SCIEH Weekly Rep 31:117. 23. Anonymous. 1997. Vero cytotoxin producing Escherichia coli O157. CDR Weekly 7:409. 24. Anonymous. 1998. Cases of Escherichia coli O157 infection associated with unpasteurized cream. CDR Weekly 8:389, 392. 25. Anonymous. 1998. E. coli O157 contamination of unpasteurised cheese in England. SCIEH Weekly Rep 32:108. 26. Anonymous. 1998. Outbreak of Vero cytotoxin producing Escherichia coli O157 infection in Dorset. CDR Weekly 8:183, 186. 27. Anonymous. 1999. Escherichia coli O157 associated with eating unpasteurized cheese—update. CDR Weekly 9:131, 134. 28. Anonymous. 1999. Outbreak of VTEC O157 infection at a prison in the Midlands. CDR Weekly 9:281, 284. 29. Anonymous. 1999. Sporadic cases of VTEC O157 infection associated with travel to southern Turkey. CDR Weekly 9:443, 446. 30. Anonymous. 1999. VTEC O157 infection in three nurseries in Preston. CDR Weekly 9:167, 170. 31. Anonymous. 1999. VTEC outbreak linked to beach holidays. CDR Weekly 9:327, 330.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 24 White Paper:

32. Anonymous. 2000. E. coli O157 infection in Grampian: update 13 June 2000. SCIEH Weekly Rep 34:137. 33. Anonymous. 2000. Outbreak of VTEC O157 in South Yorkshire. CDR Weekly 10:359. 34. Anonymous. 2000. Outbreaks of Escherichia coli O157 infection in two prisons. CDR Weekly 10:375. 35. Anonymous. 2000. Outbreaks of VTEC O157 infection linked to consumption of unpasteurised milk. CDR Weekly 10:203, 206. 36. Anonymous. 2000. Two outbreaks of VTEC O157 infection in northern England. CDR Weekly 10:229. 37. Anonymous. 2001. E. coli O157, Inverclyde. SCIEH Weekly Rep 35:153. 38. Anonymous. 2001. Outbreak of vero cytotoxin producing Escherichia coli O157 infection in a children’s nursery in Suffolk. CDR Weekly 11(26) (28 June 2001). 39. Anonymous. 2002. E. coli O157 outbreak. SCIEH Weekly Rep 36:206, 208. 40. Anonymous. 2002. Vero cytotoxic-producing E. coli VTEC O157 PT21/28 outbreak associated with a nursery. CDR Weekly 12(50) (12 Dec 2002). 41. Anonymous. 2003. Outbreak of verocytotoxin-producing Escherichia coli O157 (VTEC O157) and Campylobacter spp associated with a campsite in North Wales. CDR Weekly 13(26) (26 June 2003). 42. Anonymous. 2004. E. coli O157 outbreak in Highland. SCIEH Weekly Rep 38:169. 43. Anonymous. 2006. E. coli O157 infections in the UK. Euro Surveill 11(6):E060601.2 www.eurosurveillance.org/ew/2006/060601.asp#2. 44. Anonymous. 2006. National increase of vero-cytotoxin producing E. coli O157 phage type 8- case-control study. CDR Weekly 16(17) (27 April 2006). 45. Anonymous. 2006. Outbreak of vero cytotoxin-producing E. coli (O157 VTEC) associated with a producing butcher in Leeds. CDR Weekly 16(28) (13 July 2006). 46. Asakura H, Makino S, Shirahata T, Tsukamoto T, Kurazono H, Ikeda T, and Takeshi K. 1998. Detection and genetical characterization of shiga toxin-producing Escherichia coli from wild deer. Microbial Immunol 42:815–822. 47. Banatvala N, Magnano AR, Cartter ML, Barrett TJ, Bibb WF, Vasile LL, Mshar P, Lambertfair MA, Green JH, Bean NH, and Tauxe RV. 1996. Meat grinders and molecular epidemiology—two supermarket outbreaks of Escherichia coli O157:H7 infection. J Infect Dis 173:480–483. 48. Barlow RS, Gobius KS, and Desmarchelier PM. 2006. Shiga toxin-producing Escherichia coli in ground beef and lamb cuts: Results of a one-year study. Int J Food Microbiol 111(1):1–5. 49. Barwick RS, Levy DA, Craun GF, Beach MJ, and Calderon RL. 2000. Surveillance for waterborne disease outbreaks— United States, 1997–1998. Morbid Mortal Weekly Rep 49:SS04:1–35. 50. Bauwens L, De Meurichy W, and Vercammen F. 2000. Isolation of Escherichia coli O157 from zoo animals. Vlaams Diergeneeskundig Tijdschr 69:76–79. 51. Beecher C. 2005. Shareholder dairy’s milk tests positive for E. coli. Capital Press. www.ecolilitigation.com/deecreek008.htm 52. Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC, Tarr PI, Bartleson CA, Lewis JH, Barrett TJ, Wells JG, and et al. 1994. A multistate outbreak of Escherichia coli O157:H7 associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. The Washington experience. J Am Med Assoc 272:1349–1353. 53. Bell BP, Griffin PM, Lozano P, Christie DL, Kobayashi JM, and Tarr PI. 1997. Predictors of hemolytic uremic syndrome in children during a large outbreak of Escherichia coli O157:H7 infections. Pediatrics 100:E121–E126. 54. Belongia EA, MacDonald KL, Parham GL, White KE, Korlath JA, Lobato MN, Strand SM, K. Casale, and Osterholm MT. 1991. An outbreak of Escherichia coli O157:H7 colitis associated with consumption of precooked meat patties. J Infect Dis 164:338–343. 55. Belongia EA, Osterholm MT, Soler JT, Ammend DA, Braun JE, and MacDonald KL. 1993. Transmission of Escherichia coli O157:H7 infection in Minnesota child day-care facilities. J Am Med Assoc 269:883–888. 56. Bender JB, Smith KE, McNees AA, Rabatsky-Ehr TR, Segler SD, Hawkins MA, Spina NL, Keene WE, Kennedy MH, Van Gilder TJ, and Hedberg CW. 2004. Factors affecting surveillance data on Escherichia coli O157 infections collected from FoodNet sites, 1996–1999. Clin Infect Dis 38:S157–S164. 57. Bender JB, N. 2005. Compendium of measures to prevent disease associated with animals in public settings, 2005. Morbid Mortal Weekly Rep 54:1–13. 58. Berry ED and Miller DN. 2005. Cattle feedlot soil moisture and manure content: II. Impact on Escherichia coli 0157. J Environ Qual 34:656–663. 59. Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ, Wells JG, and Griffin PM. 1993. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh-pressed apple cider. J Am Med Assoc 269:2217– 2220. 60. Beuchat LR. 2002. Ecological factors influencing survival and growth of human pathogens on raw fruits and vegetables. Microb Infect 4:413–423. 61. Beuchat LR. 2006. Vectors and conditions for preharvest contamination of fruits and vegetables with pathogens capable of causing enteric diseases. Brit Food J 108:38–53. 62. Beutin L, Geier D, Steinruck H, Zimmermann S, and Scheutz F. 1993. Prevalence and some properties of verocytotoxin (shiga-like toxin)-producing Escherichia coli in seven different species of healthy domestic animals. J Clin Microbiol 31:2483–2488. 63. Beutin L, Geier D, Zimmermann S, and Karch H. 1995. Virulence markers of shiga-like toxin-producing Escherichia coli

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 25

strains from healthy domestic animals of different species. J Clin Microbiol 33:631–635. 64. Beutin L, Knollmann-Schanbacher G, Rietschel W, and Seeger H. 1996. Animal reservoirs of Escherichia coli O157:H7. Vet Rec 139:70–71. 65. Bielaszewska M, Janda J, Blahova K, Minarikova H, Jikova E, Karmali MA, Laubova J, Sikulova J, Preston MA, Khakhria R, Karch H, Klazarova H, and Nyc O. 1997. Human Escherichia coli O157:H7 infection associated with the consumption of unpasteurized goats milk. Epidemiol Infect 119:299–305. 66. Blackburn BG, Craun GF, Yoder JS, Hill V, Calderon RL, Chen N, Lee SH, Levy DA, and Beach MJ. 2004. Surveillance for waterborne disease outbreaks associated with drinking water—United States, 2001–2002. Morbid Mortal Weekly Rep 53(SS08):23–45. 67. Bolduc D, Srour LR, Sweet L, Neatby A, Issacs A, and Lim G. 2004. Severe outbreak of Escherichia coli O157:H7 in health care institutions in Charlottetown, Prince Edward Island, Fall, 2002. Can Commun Dis Rep 30(9):81–88. 68. Borie C, Monreal Z, Guerrero P, Sanchez ML, Martinez J, Arellano C, and Prado V. 1997. Prevalence and characterization of enterohaemorrhagic Escherichia coli isolated from healthy cattle and pigs slaughtered in Santiago, Chile. Arch Med Vet 29:205–212. 69. Brandt JR, Fouser LS, Watkins SL, Zelikovic I, Tarr PI, Nazar-Stewart V, and Avner ED. 1994. Escherichia coli O157:H7-associated hemolytic-uremic syndrome after ingestion of contaminated hamburgers. J Pediatr 125:519–526. 70. Breuer T, Benkel DH, Shapiro RL, Hall WN, Winnett MM, Linn MJ, Neimann J, Barrett TJ, Dietrich S, Downes FP, Toney DM, Pearson JL, Rolka H, Slutsker L, and Griffin PM. 2001. A multistate outbreak of Escherichia coli O157:H7 infections linked to alfalfa sprouts grown from contaminated seeds. Emerg Infect Dis 7:977–982. 71. Brewster DH, Brown MI, Robertson D, Houghton GL, Bimson J, and Sharp JC. 1994. An outbreak of Escherichia coli O157 associated with a children’s paddling pool. Epidemiol Infect 112:441–447. 72. Brooks JT, Bergmire-Sweat D, Kennedy M, Hendricks K, Garcia M, Marengo L, Wells J, Ying M, Bibb W, Griffin PM, Hoekstra RM, and Friedman CR. 2004. Outbreak of Shiga toxin-producing Escherichia coli O111:H8 infections among attendees of a high school cheerleading camp. Clin Infect Dis 38:190–198. 73. Brooks JT, Sowers EG, Wells JG, Greene KD, Griffin PM, Hoekstra RM, and Strockbine NA. 2005. Non-O157 shiga toxin-producing Escherichia coli infections in the United States, 1983–2002. J Infect Dis 192:1422–1429. 74. Bruce MG, Curtis MB, Payne MM, Gautom RK, Thompson EC, Bennett AL, and Kobayashi JI. 2003. Lake-associated outbreak of Escherichia coli O157:H7 in Clark County, Washington, August 1999. Arch Pediatr Adoles Med 157:1016– 1021. 75. Bruneau A, Rodrigue H, Ismäel J, Dion R, and Allard R. 2004. Outbreak of E. coli O157:H7 associated with bathing at a public beach in the Montreal-Centre region. Can Commun Dis Rep 30(15)133–136. 76. Buchko SJ, Holley RA, Olson WO, Gannon VPJ, and Veira DM. 2000. The effect of different grain diets on fecal shedding of Escherichia coli O157:H7 by steers. J Food Prot 63:1467–1474. 77. Byrne CM, Erol I, Call JE, Kaspar CW, Buege DR, Hiemke CJ, Fedorka-Cray PJ, Benson AK, Wallace FM, and Luchansky JB. 2003. Characterization of Escherichia coli O157:H7 from downer and healthy dairy cattle in the upper Midwest region of the United States. Appl Environ Microbiol 69:4683–4688. 78. Callaway TR, Anderson RC, Edrington TS, Genovese KJ, Bischoff KM, Poole TL, Jung YS, Harvey RB, and Nisbet DJ. 2004. What are we doing about Escherichia coli O157:H7 in cattle? J. Anim Sci 82:E93–E99. 79. Cannon M, Thomas H, Sellers W, Bates M, Blake P, Stetler H, Toomey K, Fowler J, Halford S, Young G, Hall S, Erwin P, Boaz V, and Swinger G. 1996. Outbreak of Escherichia coli O157:H7 infection—Georgia and Tennessee, June 1995. Morbid Mortal Weekly Rep 45:249–251. 80. Caprioli A, Morabito S, Brugere H, and Oswald E. 2005. Enterohaemorrhagic Escherichia coli: emerging issues on virulence and modes of transmission. Vet Res 36:289–311. 81. Carter AO, Borczyk AA, Carlson JA, Harvey B, Hockin JC, Karmali MA, Krishnan C, Korn DA, and Lior H. 1987. A severe outbreak of Escherichia coli O157:H7—associated hemorrhagic colitis in a nursing home. N Engl J Med 317:1496–1500. 82. Centers for Disease Control. 1996. Surveillance for Foodborne Disease Outbreaks—United States, 1988–1992. Morbid Mortal Weekly Rep 45:1–67. 83. Centers for Disease Control. 2000. Surveillance for foodborne-disease outbreaks—United States, 1993–1997. Morbid Mortal Weekly Rep 49:1–62. 84. Centers for Disease Control. 2006. FoodNet Surveillance Report for 2004 (Final Report). 203 p. www.cdc.gov/foodnet/annual/2004/Report.pdf 85. Centers for Disease Control. 2006. National antimicrobial resistance monitoring system for enteric bacteria (NARMS): 2003 human isolates final report. 67 p. www.cdc.gov/narms/annual/2003/NARMS2003AnnualReport.pdf 86. Centers for Disease Control. 2006. Preliminary FoodNet data on the incidence of infection with pathogens commonly transmitted through food—10 states, United States, 2005. Morbid Mortal Weekly Rep 55:392–395. 87. Centers for Disease Control. 2006. Provisional cases of selected notifiable diseases. Morbid Mortal Weekly Rep 55:927. 88. Centers for Disease Control. 2006. Reported cases of notifiable diseases by geographic division and area — United States. Morbid Mortal Weekly Rep 55:886. 89. Chalmers RM, Salmon RL, Willshaw GA, Cheasty T, Looker N, Davies I, and Wray C. 1997. Vero-cytotoxin-producing Escherichia coli O157 in a farmer handling horses. Lancet 349:1816. 90. Chapman PA. 2000. Sources of Escherichia coli O157 and experiences over the past 15 years in Sheffield, UK. J Appl

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 26 White Paper:

Microbiol 88:51S–60S. 91. Chapman PA and Ackroyd HJ. 1997. Farmed deer as a potential source of verocytotoxin-producing Escherichia coli O157. Vet Rec 141:314–315. 92. Chapman PA, Cornell J, and Green C. 2000. Infection with verocytotoxin-producing Escherichia coli O157 during a visit to an inner city open farm. Epidemiol Infect 125:531–536. 93. Chapman PA, Malo ATC, Ellin M, Ashton R, and Harkin MA. 2001. Escherichia coli O157 in cattle and sheep at slaughter, on beef and lamb carcasses and in raw beef and lamb products in South Yorkshire, UK. Int J Food Microbiol 64:139–150. 94. Chapman PA, Siddons CA, Malo ATC, and Harkin MA. 1996. Lamb products as a potential source of E. coli O157. Vet Rec 139:427–428. 95. Chapman PA, Siddons CA, Malo ATC, and Harkin MA. 1997. A 1-year study of Escherichia coli O157 in cattle, sheep, pigs and poultry. Epidemiol Infect 119:245–250. 96. Chapman PA, Siddons CA, Manning J, and Cheetham C. 1997. An outbreak of infection due to verocytotoxin-producing Escherichia coli O157 in four families—the influence of laboratory methods on the outcome of the investigation. Epidemiol Infect 119:113–119. 97. Charatan F. 1999. New York outbreak of E. coli poisoning affects 1000 and kills two. Brit Med J 319:873. 98. Childs KD, Simpson CA, Warren-Serna W, Bellenger G, Centrella B, Bowling RA, Ruby J, Stefanek J, Vote DJ, Choat T, Scanga JA, Sofos JN, Smith GC, and Belk KE. 2006. Molecular characterization of Escherichia coli O157:H7 hide contamination routes: Feedlot to harvest. J Food Prot 69:1240–1247. 99. Cho S, Bender JB, Diez-Gonzalez F, Fossler CP, Hedberg CW, Kaneene JB, Ruegg PL, Warnick LD, and Wells SJ. 2006. Prevalence and characterization of Escherichia coli O157 isolates from Minnesota dairy farms and county fairs. J Food Prot 69:252–259. 100. Cieslak PR, Noble SJ, Maxson DJ, Empey LC, Ravenholt O, Legarza G, Tuttle J, Doyle MP, Barrett TJ, Wells JG, Mcnamara AM, and Griffin PM. 1997. Hamburger-associated Escherichia coli O157:H7 infection in Las Vegas—a hidden epidemic. Am J Publ Health 87:176–180. 101. Cimolai N, Basalyga S, Mah DG, Morrison BJ, and Carter JE. 1994. A continuing assessment of risk factors for the development of Escherichia coli O157:H7-associated hemolytic uremic syndrome. Clin Nephrol 42:85–89. 102. Cizek A, Alexa P, Literak I, Hamrik J, Novak P, and Smola J. 1999. Shiga toxin-producing Escherichia coli O157 in feedlot cattle and Norwegian rats from a large-scale farm. Lett Appl Microbiol 28:435–439. 103. Cizek A, Literak I, and Scheer P. 2000. Survival of Escherichia coli O157 in faeces of experimentally infected rats and domestic pigeons. Lett Appl Microbiol 31:349–352. 104. Clark A, Morton S, Wright P, Corkish J, Bolton FJ, and Russell J. 1997. A community outbreak of Vero cytotoxin producing Escherichia coli O157 infection linked to a small farm dairy. Commun Dis Rep Rev 7:R206–R211. 105. Cody SH, Glynn MK, Farrar JA, Cairns KL, Griffin PM, Kobayashi J, Fyfe M, Hoffman R, King AS, Lewis JH, Swaminathan B, Bryant RG, and Vugia DJ. 1999. An outbreak of Escherichia coli O157:H7 infection from unpasteurized commercial apple juice. Ann Int Med 130:202–209. 106. Combs BG, Wise RP, Tribe IG, Mwanri L, and Raupach JCA. 2003. Investigation of two clusters if shiga toxin-producing Escherichia coli cases in South Australia. Commun Dis Intelligence 27:517–519. 107. Como-Sabetti K, Reagan S, Allaire S, Parrott K, Simonds CM, Hrabowy S, Ritter B, W. Hall, Altamirano J, Martin R, Downes F, Jennings G, Barrie R, Dorman MF, Keon N, Kucab M, Alshab A, Robinsondunn B, Dietrich S, Moshur L, Reese L, Smith J, Wilcox K, Tilden J, Wojtala G, and et al. 1997. Outbreaks of Escherichia coli O157:H7 infection associated with eating alfalfa sprouts—Michigan and Virginia, June–July 1997. Morbid Mortal Weekly Rep 46:741–744. 108. Cornick NA and Helgerson AF. 2004. Transmission and infectious dose of Escherichia coli O157:H7 in swine. Appl Environ Microbiol 70:5331–5335. 109. Cowden JM, Ahmed S, Donaghy M, and Riley A. 2001. Epidemiological investigation of the Central Scotland outbreak of Escherichia coli O157 infection, November to December 1996. Epidemiol Infect 126:335–341. 110. Crampin M, Willshaw G, Hancock R, Djuretic T, Elstob C, Rouse A, Cheasty T, and Stuart J. 1999. Outbreak of Escherichia coli O157 infection associated with a music festival. Eur J Clin Microbiol Infect Dis 18:286–288. 111. Cransberg K, Vandenkerkhof JHCT, Banffer JRJ, Stijnen C, Wernars K, Vandekar NCAJ, Nauta J, and Wolff ED. 1996. Four cases of hemolytic uremic syndrome—source contaminated swimming water. Clin Nephrol 46:45–49. 112. Craun GF, Calderon RL, and Craun MF. 2005. Outbreaks associated with recreational water in the United States. Int J Environ Health Res 15:243–262. 113. Crump JA, Sulka AC, Langer AJ, Schaben C, Crielly AS, Gage R, Baysinger M, Moll M, Withers G, Toney DM, Hunter SB, Hoekstra RM, Wong SK, Griffin PM, and Van Gilder TJ. 2002. An outbreak of Escherichia coli O157:H7 infections among visitors to a dairy farm. N Engl J Med 347:555–560. 114. Dargatz D.A., S. J. Wells, L. A. Thomas, D. D. Hancock, and L. P. Garber. 1997. Factors associated with the presence of Escherichia coli O157 in feces of feedlot cattle. J Food Prot 60:466–470. 115. David ST, MacDougall L, Louie K, McIntyre L, Paccagnella AM, Schleicher S, and Hamade A. 2004. Petting zoo- associated Escherichia coli O157:H7- secondary transmission, asymptomatic infection, and prolonged shedding in the classroom. Can Commun Dis Rep 30(20):173–180. 116. Davies M, Engel J, Griffin D, Ginzl D, Hopkins R, Blackmore C, Lawaczec E, Nathan L, Levy C, Briggs G, Kioski C, Kreis S, Keen J, Durso L, Schulte J, Fullerton K, Long C, Smith S, Barton C, Gleit C, Joyner M, Montgomery S, Braden

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 27

C, Goode B, and Chertow D. 2005. Outbreaks of Escherichia coli O157:H7 associated with petting zoos—North Carolina, Florida, and Arizona, 2004 and 2005. Morbid Mortal Weekly Rep 54:1277–1280. 117. Davis BS and Brogan RT. 1994. Butchers, burgers, and E. coli O157. Commun Dis Environ Health Scotland 28:3–6. 118. Davis BS and Brogan RT. 1995. A widespread community outbreak of E. coli O157 infection in Scotland. Public Health 109:381–388. 119. Davis M. 1994. Update—Multistate outbreak of Escherichia coli O157:H7 infections from hamburgers—western United States, 1992–1993 Morbid Mortal Weekly Rep 42(14):258–264. 120. Davis MA, Cloud-Hansen KA, Carpenter J, and Hovde CJ. 2005. Escherichia coli O157:H7 in environments of culture- positive cattle. Appl Environ Microbiol 71:6816–6822. 121. Deschenes G, Casenave C, Grimont F, Desenclos JC, Benoit S, Collin M, Baron S, Mariani P, Grimont PAD, and Nivet H. 1996. Cluster of cases of haemolytic uraemic syndrome due to unpasteurised cheese. Pediatr Nephrol 10(2):203–205. 122. Dev VJ, Main M, and Gould I. 1991. Waterborne outbreak of Escherichia coli O157. Lancet 337:1412. 123. Dipineto L, Santaniello A, Fontanella M, Lagos K, Fioretti A, and Menna LF. 2006. Presence of Shiga toxin-producing Escherichia coli O157:H7 in living layer hens. Lett Appl Microbiol 43:293–295. 124. Dontorou A, Papadopolou C, Filioussis G, Apostolou I, Economou V, Kansouzidou A, and Levidiotou S. 2004. Isolation of a rare Escherichia coli O157:H7 strain from farm animals in Greece. Comp Immunol Microbiol Infect Dis 27:201–207. 125. Doorduyn Y, de Jager CM, van der Zwaluw, Friesma IHM, Heuvelink AE, de Boer E, Wannet WJB, and van Duynhoven YTHP. 2006. Shiga toxin-producing Escherichia coli (STEC) O157 outbreak, The Netherlands, September–October 2005. Euro Surveill Monthly 11(7–8) www.eurosurveillance.org/em/v11n07/1107-223.asp. 126. Dorsa WJ, Cutter CN, and Siragusa GR. 1997. Effects of steam-vacuuming and hot water spray wash on the microflora of refrigerated beef carcass surface tissue inoculated with Escherichia coli O157:H7, Listeria innocua, and Clostridium sporogenes. J Food Prot 60:114–119. 127. Duffell E, Espié E, Nichols T, Adak GK, De Valk H, Anderson K, and Stuart JM. 2003. Investigation of an outbreak of E. coli O157 infections associated with a trip to France of schoolchildren from Somerset, England. Euro Surveill 8(4):81–86. 128. Dundas S, Todd WTA, Stewart AI, Murdoch PS, Chaudhuri AKR, and Hutchinson SJ. 2001. The central Scotland Escherichia coli O157:H7 outbreak: Risk factors for the hemolytic uremic syndrome and death among hospitalized patients. Clin Infect Dis 33:923–931. 129. Dunn JR, Keen JE, Moreland D, and Thompson RA. 2004. Prevalence of Escherichia coli O157:H7 in white-tailed deer from Louisiana. J Wildlife Dis 40:361–365. 130. Dunn JR, Keen JE, and Thompson RA. 2004. Prevalence of shiga-toxigenic Escherichia coli O157:H7 in adult dairy cattle. J Am Vet Med Assoc 224:1151–1158. 131. Durch J, Ringhand T, Manner K, Barnett M, Proctor M, Davis J, and Boxrud D. 2000. Outbreak of Escherichia coli O157:H7 infection associated with eating fresh cheese curds—Wisconsin, June 1998. Morbid Mortal Weekly Rep 49:911–913. 132. Echeverry A, Loneragan GH, Wagner BA, and Brashears MM. 2005. Effect of intensity of fecal pat sampling on estimates of Escherichia coli O157 prevalence. Am J Vet Res 66:2023–2027. 133. Ejidokun OO, Walsh A, Barnett J, Hope Y, Ellis S, Sharp MW, Paiba GA, Logan M, Willshaw GA, and Cheasty T. 2006. Human Vero cytotoxigenic Escherichia coli (VTEC) O157 infection linked to birds. Epidemiol Infect 134:421–423. 134. Elder RO, Keen JE, Siragusa GR, Barkocy-Gallagher GA, Koohmaraie M, and Laegreid WW. 2000. Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing. Proc Nat Acad Sci USA 97:2999–3003. 135. Espié E, Vaillant V, Mariani-Kurkdjian P, Grimont F, Martin-Schaller R, De Valk H, and Vernozy-Rozand C. 2006. Escherichia coli O157 outbreak associated with fresh unpasteurized goats’ cheese. Epidemiol Infect 134:143–146. 136. Fegan N, Higgs G, Vanderlinde P, and Desmarchelier P. 2005. An investigation of Escherichia coli O157 contamination of cattle during slaughter at an abattoir. J Food Prot 68:451–457. 137. Feldman KA, Mohle-Boetani JC, Ward J, Furst K, Abbott SL, Ferrero DV, Olsen A, and Werner SB. 2002. A cluster of Escherichia coli O157: Nonmotile infections associated with recreational exposure to lake water. Pub Health Rep 117:380–385. 138. Ferguson DD, Scheftel J, Cronquist A, Smith K, Woo-Ming A, Anderson E, Knutsen J, De AK, and Gershman K. 2005. Temporally distinct Escherichia coli O157 outbreaks associated with alfalfa sprouts linked to a common seed source— Colorado and Minnesota, 2003. Epidemiol Infect 133:439–447. 139. Fischer JR, Zhao T, Doyle MP, Goldberg MR, Brown CA, Sewell CT, Kavanaugh DM, and Bauman CD. 2001. Experimental and field studies of Escherichia coli O157:H7 in white-tailed deer. Appl Environ Microbiol 67:1218–1224. 140. Fisher I, Meakins S, and Enter-net participants. 2006. Surveillance of enteric pathogens in Europe and beyond: Enter-net annual report for 2004. Eurosurveillance 11. 141. Fremaux B, Delignette-Muller ML, Prigent-Combaret C, Gleizal A, and Vernozy-Rozand C. 2006. Growth and survival of non-O157:H7 shiga-toxin-producing Escherichia coli in cow manure. J Appl Microbiol (in press). 142. French mult-agency outbreak investigation team. 2005. Outbreak of E. coli O157:H7 infections associated with a brand of beefburgers in France. Euro Surveill 10(11):E051103.1 www.eurosurveillance.org/ew/2005/051103.asp#1 143. Friedman MS, Roels T, Koehler JE, Feldman L, Bibb WF, and Blake P. 1999. Escherichia coli O157:H7 outbreak associated with an improperly chlorinated swimming pool. Clin Infect Dis 29:298–303. 144. Frost B, Chaos C, Ladaga L, Day W, Tenney M, Mcwilliams D, Barrett E, Branch L, Jenkins S, Linn M, Turf E, Woolard

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 28 White Paper:

D, Miller GB, Henderson S, Campbell B, Mismas M, Dvorak J, Patel D, Peery D, Morano J, and Campbell K. 1995. Escherichia coli O157:H7 outbreak at a summer camp—Virginia, 1994. Morbid Mortal Weekly Rep 44:419–421. 145. Fukushima H, Hashizume T, Morita Y, Tanaka J, Azuma K, Mizumoto Y, Kaneno M, Matsu-ura MO, Konma K, and Kitani T. 1999. Clinical experiences in Sakai City Hospital during the massive outbreak of enterohemorrhagic Escherichia coli O157 infections in Sakai City, 1996. Pediatr Int 41:213–217. 146. Fukushima H, Hoshina K, and Gomyoda M. 1999. Long-term survival of shiga toxin-producing Escherichia coli O26, O111, and O157 in cattle feces. Appl Environ Microbiol 65:5177–5181. 147. Gage R, Crielly A, Baysinger M, Chernak E, Herbert G, Johnson-Entsuah A, Fraser G, Rinehardt C, Solomon M, Withers G, Berman R, Moll M, Rankin J, Carroll J, Ettinger M, Henderson S, Mismas M, Patel D, Reed T, Smith E, Wozniak J, Toney D, Pearson J, Hofmann J, Grendon J, and Kobayashi J. 2001. Outbreaks of Escherichia coli O157:H7 infections among children associated with farm visits—Pennsylvania and Washington, 2000. Morbid Mortal Weekly Rep 50:293– 297. 148. Galanis E, Longmore K, Hasselback P, Swann D, Ellis A, and Panaro L. 2003. Investigation of an E. coli O157:H7 outbreak in Brooks, Alberta, June–July 2002. Can Commun Dis Rep 29(3)2–28. 149. Garber LP, Wells SJ, Hancock DD, Doyle MP, Tuttle J, Shere JA, and Zhao T. 1995. Risk factors for fecal shedding of Escherichia coli O157:H7 in dairy calves. J Am Vet Med Assoc 207:46–49. 150. Garcia A and Fox JG. 2003. The rabbit as a new reservoir host of enterohemorrhagic Escherichia coli. Emerg Infect Dis 9:1592–1597. 151. Gerner-Smidt P, Kincaid J, Kubota K, Hise K, Hunter SB, Fair MA, Norton D, Woo-Ming A, Kurzynski T, Sotir MJ, Head M, Holt K, and Swaminathan B. 2005. Molecular surveillance of Shiga toxigenic Escherichia coli 0157 by PulseNet USA. J Food Prot 68:1926–1931. 152. Gilbert M, Srour L, Paccagnella A, MacDougall L, Fung J, Nelson E, and Fyfe M. 2005. An outbreak of Escherichia coli O157:H7 associated withg a children’s water spray park and identified by two rounds of pulsed-field gel electrophoresis testing. Can Commun Dis Rep 31(12):133–140. 153. Goh S, Newman C, Knowles M, Bolton FJ, Hollyoak V, Richards S, Daley P, Counter D, Smith HR, and Keppie N. 2002. E. coli O157 phage type 21/28 outbreak in North Cumbria associated with pasteurized milk. Epidemiol Infect 129:451– 457. 154. Goldwater PN and Bettelheim KA. 1995. Hemolytic uremic syndrome due to shiga-like toxin producing Escherichia coli O48:H21 in South Australia. Emerg Inf Dis 1(4):132–133. 155. Goldwater PN and Bettelheim KA. 1998. New perspectives on the role of Escherichia coli O157:H7 and other enterohemorrhagic E. coli serotypes in human disease. J Med Microbiol 47:1039–1045. 156. Gourmelon M, Montet MP, Lozach S, Le Mennec C, Pommepuy M, Beutin L, and Vernozy-Rozand C. 2006. First isolation of Shiga toxin 1d producing Escherichia coli variant strains in shellfish from coastal areas in France. J Appl Microbiol 100:85–97. 157. Greenquist MA, Drouillard JS, Sargeant JM, Depenbusch BE, Shi XR, Lechtenberg KF, and Nagaraja TG. 2005. Comparison of rectoanal mucosal swab cultures and fecal cultures for determining prevalence of Escherichia coli O157:H7 in feedlot cattle. Appl Environ Microbiol 71:6431–6433. 158. Grif K, Orth D, Lederer I, Berghold C, Roedl S, Mache CJ, Dierich MP, and Wurzner R. 2005. Importance of environmental transmission in cases of EHEC O157 causing hemolytic uremic syndrome. Eur J Clin Microbiol Infect Dis 24:268–271. 159. Grossmann K, Weniger B, Baljer G, Brenig B, and Wieler LH. 2005. Racing, ornamental and city pigeons carry Shiga toxin producing Escherichia coli (STEC) with different Shiga toxin subtypes, urging further analysis of their epidemiological role in the spread of STEC. Berlin Munch Tierarzt Wochenschr 118:456–463. 160. Gupta A, Hunter SB, Bidol SA, Dietrich S, Kincaid J, Salehi E, Nicholson L, Genese CA, Todd-Weinstein S, Marengo L, Kimura AC, and Brooks JT. 2004. Escherichia coli O157 cluster evaluation. Emerg Infect Dis 10:1856–1858. 161. Hancock DD, Besser TE, Rice DH, Ebel ED, Herriott DE, and Carpenter LV. 1998. Multiple sources of Escherichia coli O157 in feedlots and dairy farms in the northwestern USA. Prev Vet Med 35:11–19. 162. Harakudo Y, Konuma H, Iwaki M, Kasuga F, Sugitakonishi Y, Ito Y, and Kumagai S. 1997. Potential hazard of radish sprouts as a vehicle of Escherichia coli O157:H7. J Food Prot 60:1125–1127. 163. Hardnett FP, Hoekstra RM, Kennedy M, Charles L, and Angulo FJ. 2004. Epidemiologic issues in study design and data analysis related to FoodNet activities. Clin Infect Dis 38:S121–S126. 164. Harrison S and Kinra S. 2004. Outbreak of Escherichia coli O157 associated with a busy bathing beach. Commun Dis Public Health 7:47–50. 165. Haus-Cheymol R, Espie E, Che D, Vaillant V, DeValk H, and Desenclos JC. 2006. Association between indicators of cattle density and incidence of paediatric haemolytic-uraemic syndrome (HUS) in children under 15 years of age in France between 1996 and 2001: an ecological study. Epidemiol Infect 134:712–718. 166. Heuvelink AE, Van Heerwaarden C, Zwartkruis-Nahuis JTM, Van Oosterom R, Edink K, Van Duynhoven YTHP, and De Boer E. 2002. Escherichia coli O157 infection associated with a petting zoo. Epidemiol Infect 129:295–302. 167. Heuvelink AE, Vandenbiggelaar FLAM, Deboer E, Herbes RG, Melchers WJG, Huisintveld JHJ, and Monnens LAH. 1998. Isolation and characterization of verocytotoxin-producing Escherichia coli O157 strains from Dutch cattle and sheep. J Clin Microbiol 36:878–882. 168. Heuvelink AE, Zwartkruis-Nahuis JTM, van den Biggelaar FLAM, van Leeuwen WJ, and de Boer E. 1999. Isolation and

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 29

characterization of verocytotoxin-producing Escherichia coli O157 from slaughter pigs and poultry. Int J Food Microbiol 52:67–75. 169. Hilborn ED, Mermin JH, Mshar PA, Hadler JL, Voetsch A, Wojtkunski C, Swartz M, Mshar R, Lambert-Fair MA, Farrar JA, Glynn MK, and Slutsker L. 1999. A multistate outbreak of Escherichia coli O157:H7 infections associated with consumption of mesclun lettuce. Arch Intern Med 159:1758–1764. 170. Hilborn ED, Mshar PA, Fiorentino TR, Dembek ZF, Barrett TJ, Howard RT, and Cartter ML. 2000. An outbreak of Escherichia coli O157:H7 infections and haemolytic uraemic syndrome associated with consumption of unpasteurized apple cider. Epidemiol Infect 124:31–36. 171. Hildebrand JM, Maguire HC, Holliman RE, and Kangesu E. 1996. An outbreak of Escherichia coli O157 infection linked to paddling pools. Commun Dis Rep Rev 6:R33–R36. 172. Honish L, Predy G, Hislop N, Chui L, Kowalewska-Grochowska K, Trottier L, Kreplin C, and Zazulak I. 2005. An outbreak of E. coli O157:H7 hemorrhagic colitis associated with unpasteurized gouda cheese. Can J Pub Health 96:182– 184. 173. Howatt G. 2006. Origin of beef linked to Longville death is still mystery. Star Tribune. www.startribune.com/462/v- print/story/632284.html. 174. Howie H, Mukerjee A, Cowden J, Leith J, and Reid T. 2003. Investigation of an outbreak of Escherichia coli O157 infection caused by environmental exposure at a scout camp. Epidemiol Infect 131:1063–1069. 175. Hrudey SE, Payment P, Huck PM, Gillham RW, and Hrudey EJ. 2003. A fatal waterborne disease epidemic in Walkerton, Ontario: comparison with other waterborne outbreaks in the developed world. Water Sci Technol 47:7–14. 176. Hussein HS and Bollinger LM. 2005. Prevalence of Shiga toxin-producing Escherichia coli in beef. Meat Sci 71:676–689. 177. Hussein HS and Bollinger LM. 2005. Prevalence of Shiga toxin-producing Escherichia coli in beef cattle. J Food Prot 68:2224–2241. 178. Hussein HS and Sakuma T. 2005. Invited review: Prevalence of Shiga toxin-producing Escherichia coli in dairy cattle and their products. J Dairy Sci 88:450–465. 179. Hussein HS, Thran BH, and Glimp HA. 2003. Verotoxin-producing Escherichia coli in sheep grazing an irrigated pasture or arid rangeland forages. Exp Biol Med 228:358–364. 180. Ihekweazu c, Barlow M, Roberts S, Christensen H, Guttridge B, Lewis D, and Paynter S. 2006. Outbreak of E. coli O157 infection in the south west of the UK: risks from streams crossing seaside beaches. Euro Surveill 11(4):128–130. 181. Ikeda K, Ida O, Kimoto K, Takatorige T, Nakanishi N, and Tatara K. 2000. Predictors for the development of haemolytic uraemic syndrome with Escherichia coli O157:H7 infections: with focus on the day of illness. Epidemiol Infect 124:343– 349. 182. Illinois Dept. of Health. 1999. Petersburg E. coli outbreak traced to beef. www.idph.state.il.us/public/press99/ecolinews.htm 183. Islam M, Doyle MP, Phatak SC, Millner P, and Jiang XP. 2005. Survival of Escherichia coli O157:H7 in soil and on carrots and onions grown in fields treated with contaminated manure composts or irrigation water. Food Microbiol 22:63– 70. 184. Jackson LA, Keene WE, McAnulty JM, Alexander ER, Diermayer M, Davis MA, Hedberg K, Boase J, Barrett TJ, Samadpour M, and Fleming DW. 2000. Where’s the beef? The role of cross-contamination in 4 chain restaurant- associated outbreaks of Escherichia coli O157:H7 in the Pacific northwest. Arch Intern Med 160:2380–2385. 185. Jackson SG, Goodbrand RB, Johnson RP, Odorico VG, Alves D, Rahn K, Wilson JB, Welch MK, and Khakhria R. 1998. Escherichia coli O157:H7 diarrhoea associated with well water and infected cattle on an Ontario farm. Epidemiol Infect 120:17–20. 186. Janisiewicz WJ, Conway WS, Brown MW, Sapers GM, Fratamico P, and Buchanan RL. 1999. Fate of Escherichia coli O157:H7 on fresh-cut apple tissue and its potential for transmission by fruit flies. Appl Environ Microbiol 65:1–5. 187. Jay MT, Garrett V, Mohle-Boetani JC, Barros M, Farrar JA, Rios R, Abbott S, Sowadsky R, Komatsu K, Mandrell R, Sobel J, and Werner SB. 2004. A multistate outbreak of Escherichia coli O157:H7 infection linked to consumption of beef tacos at a fast-food restaurant chain. Clin Infect Dis 39:1–7. 188. Jensen C, Ethelberg S, Gervelmeyer A, Nielsen EM, Olsen KEP, and Molbak K. 2006. First general outbreak of verocytotoxin-producing Escherichia coli O157 in Denmark. Euro Surveill 11(2):55–58. 189. Johannessen GS, Bengtsson GB, Heier BT, Bredholt S, Wasteson Y, and Rorvik LM. 2005. Potential uptake of Escherichia coli O157:H7 from organic manure into crisphead lettuce. Appl Environ Microbiol 71:2221–2225. 190. Johannessen GS, Froseth RB, Solemdal L, Jarp J, Wasteson Y, and Rorvik LM. 2004. Influence of bovine manure as fertilizer on the bacteriological quality of organic Iceberg lettuce. J Appl Microbiol 96:787–794. 191. Johnsen G, Wasteson Y, Heir E, Berget OI, and Herikstad H. 2001. Escherichia coli O157 H7 in faeces from cattle, sheep and pigs in the southwest part of Norway during 1998 and 1999. Int J Food Microbiol 65:193–200. 192. Johnson RP, Clarke RC, Wilson JB, Read SC, Rahn K, Renwick SA, Sandhu KA, Alves D, Karmali MA, Lior H, McEwen SA, Spika JS, and Gyles CL. 1996. Growing concerns and recent outbreaks involving non-O157:H7 serotypes of verotoxigenic Escherichia coli. J Food Prot 59:1112–1122. 193. Jones IG and Roworth M. 1996. An outbreak of Escherichia coli O157 and campylobacteriosis associated with contamination of a drinking water supply. Public Health 110:277–282. 194. Kang SJ, Ryu SJ, Chae JS, Eo SK, Woo GJ, and Lee JH. 2004. Occurrence and characteristics of enterohemorrhagic Escherichia coli O157 in calves associated with diarrhoea. Vet Microbiol 98:323–328.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 30 White Paper:

195. Karch H, Russmann H, Schmidt H, Schwarzkopf A, and Heesemann J. 1995. Long-term shedding and clonal turnover of enterohemorrhagic Escherichia coli O157 in diarrheal diseases. J Clin Microbiol 33:1602–1605. 196. Karmali MA. 2005. Use of comparative genomics as a tool to assess the clinical and public health significance of emerging Shiga toxin-producing Escherichia coli serotypes. Meat Sci 71:62–71. 197. Kassenborg HD, Hedberg CW, Hoekstra M, Evans MC, Chin AE, Marcus R, Vugia DJ, Smith K, Ahuja SD, Slutsker L, and Griffin PM. 2004. Farm visits and undercooked hamburgers as major risk factors for sporadic Escherichia coli O157:H7 infection: Data from a case-control study in 5 FoodNet sites. Clin Infect Dis 38:S271–S278. 198. Kaufmann M, Zweifel C, Blanco M, Blanco JE, Blanco J, Beutin L, and Stephan R. 2006. Escherichia coli O157 and non- O157 Shiga toxin-producing Escherichia coli in fecal samples of finished pigs at slaughter in Switzerland. J Food Prot 69:260–266. 199. Keen JE, Wittum JE, Dunn JR, Bono JL, and Durso LM. 2006. Shiga-toxigenic Escherichia coli O157 in agricultural fair livestock, United States. Emerg Infect Dis 12:780–786. 200. Keene WE, Hedberg K, Herriott DE./Hancock DD, McKay RW, Barrett TJ, and Fleming DW. 1997. Prolonged outbreak of Escherichia coli O157:H7 infections caused by commercially distributed raw milk. J Infect Dis 176:815–818. 201. Keene WE, McAnulty JM, Hoesly FC, Williams LP Jr, Hedberg K, Oxman GL, Barrett TJ, Pfaller MA, and Fleming DW. 1994. A swimming-associated outbreak of hemorrhagic colitis caused by Escherichia coli O157:H7 and Shigella sonnei. N Engl J Med 331:579–584. 202. Keene WE, Sazie E, Kok J, Rice DH, Hancock DD, Balan VK, Zhao T, and Doyle MP. 1997. An outbreak of Escherichia coli O157:H7 infections traced to jerky made from deer meat. J Am Med Assoc 277:1229–1231. 203. Kijima-Tanaka M, Ishihara K, Kojima A, Morioka A, Nagata R, Kawanishi M, Nakazawa M, Tamura Y, and Takahashi T. 2005. A national surveillance of Shiga toxin-producing Escherichia coli in food-producing animals in Japan. J Vet Med Ser B 52:230–237. 204. Kobayashi M, Sasaki T, Saito N, Tamura K, Suzuki K, Watanabe H, and Agui N. 1999. Houseflies: Not simple mechanical vectors of enterohemorrhagic Escherichia coli O157:H7. Am J Trop Med Hyg 61:625–629. 205. Kohli HS, Chaudhuri AK, Todd WT, Mitchell AA, and Liddell KG. 1994. A severe outbreak of E. coli O157 in two psychogeriatric wards. J Pub Health Med 16:11–15. 206. Krishnan C, Fitzgerald VA, Dakin SJ, and Behme RJ. 1987. Laboratory investigation of outbreak of hemorrhagic colitis caused by Escherichia coli O157:H7. J Clin Microbiol 25:1043–1047. 207. Kudaka J, Asato R, Itokazu K, Nakamura M, Taira K, Kuniyosi H, Kinjo Y, Terajima J, Watanabe H, Swaminathan B, Braden CR, and Dunn JR. 2005. Escherichia coli O157:H7 infections associated with ground beef from a U.S. military installation—Okinawa, Japan, February 2004. Morbid Mortal Weekly Rep 54:40–42. 208. Kudva IT, Hatfield PG, and Hovde CJ. 1995. Effect of diet on the shedding of Escherichia coli O157:H7 in a sheep model. Appl Environ Microbiol 61:1363–1370. 209. Kudva IT, Hatfield PG, and Hovde CJ. 1997. Characterization of Escherichia coli O157:H7 and other shiga toxin- producing E. coli serotypes isolated from sheep. J Clin Microbiol 35:892–899. 210. La Ragione RM, Ahmed NMY, Best A, Clifford D, Weyer U, Cooley WA, Johnson L, Pearson GR, and Woodward MJ. 2005. Colonization of 8-week-old conventionally reared goats by Escherichia coli O157:H7 after oral inoculation. J Med Microbiol 54:485–492. 211. Lahti E, Hirvela-Koski V, and Honkanen-Buzalski T. 2001. Occurrence of Escherichia coli O157 in reindeer (Rangifer tarandus). Vet Rec 148:633–634. 212. Laine ES, Scheftel JM, Boxrud DJ, Vought KJ, Danila RN, Elfering KM, and Smith KE. 2005. Outbreak of Escherichia coli 0157:H7 infections associated with nonintact blade-tenderized frozen steaks sold by door-to-door vendors. J Food Prot 68:1198–1202. 213. Law D. 2000. The history and evolution of Escherichia coli O157 and other Shiga toxin-producing E. coli. World J Microbiol Biotechnol 16:701–709. 214. Leclercq A and Mahillon J. 2003. Farmed rabbits and ducks as vectors for VTEC O157:H7. Vet Rec 152:723–724. 215. Lee SH, Levy DA, Craun GF, Beach MJ, and Calderon RL. 2002. Surveillance for waterborne disease outbreaks—United States, 1999–2000. Morbid Mortal Weekly Rep 519SS08):1–28. 216. LeJeune JT, Besser TE, and Hancock DD. 2001. Cattle water troughs as reservoirs of Escherichia coli O157. Appl Environ Microbiol 67:3053–3057. 217. LeJeune JT, Besser TE, Rice DH, Berg JL, Stilborn RP, and Hancock DD. 2004. Longitudinal study of fecal shedding of Escherichia coli O157:H7 in feedlot cattle: Predominance and persistence of specific clonal types despite massive cattle population turnover. Appl Environ Microbiol 70:377–384. 218. Lemunier M, Francou C, Rousseaux S, Houot S, Dantigny P, Piveteau P, and Guzzo J. 2005. Long-term survival of pathogenic and sanitation indicator bacteria in experimental biowaste composts. Appl Environ Microbiol 71:5779–5786. 219. Lerman Y, Cohen D, Gluck A, Ohad E, and Sechter I. 1992. A cluster of cases of Escherichia coli O157 infection in a day-care center in a communal settlement (kibbutz) in Israel. J Clin Microbiol 30:520–521. 220. Levy DA, Bens MS, Craun GF, Calderon RL, and Herwaldt BL. 1998. Surveillance for waterborne disease outbreaks— United States, 1995–1996. Morbid Mortal Weekly Rep 47(SS05). 80 p. 221. Licence K, Oates KR, Synge BA, and Reid TMS. 2001. An outbreak of E. coli 0157 infection with evidence of spread from animals to man through contamination of a private water supply. Epidemiol Infect 126:135–138. 222. Liptakova A, Siegfried L, Rosocha J, Podracka L, Bogyiova E, and Kotulova D. 2004. A family outbreak of haemolytic

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 31

uraemic syndrome and haemorrhagic colitis caused by verocytotoxigenic Escherichia coli O157 from unpasteurised cow’s milk in Slovakia. Clin Microbiol Infect 10:576–578. 223. Low JC, McKendrick IJ, McKechnie C, Fenlon D, Naylor SW, Currie C, Smith DGE, Allison L, and Gally DL. 2005. Rectal carriage of enterohaemorrhagic Escherichia coli O157 in slaughtered cattle. Appl Environ Microbiol 71:93–97. 224. Ludwig K, Ruder H, Bitzan M, Zimmermann S, and Karch H. 1997. Outbreak of Escherichia coli 0157:H7 infection in a large family. Eur J Clin Microbiol Infect Dis 16:238–241. 225. MacDonald C, Drew J, Carlson R, Dzogan S, Tataryn S, Macdonald A, Ali A, Amhed R, Easy R, Clark C, and Rodgers F. 2000. Outbreak of Escherichia coli O157:H7 leading to the recall of retail ground beef—Winnipeg, Manitoba, May 1999. Can Commun Dis Rep 26(13)109–111. 226. MacDonald DM, Fyfe M, Paccagnella A, Trinidad A, Louie K, and Patrick D. 2004. Escherichia coli O157:H7 outbreak linked to salami, British Columbia, Canada, 1999. Epidemiol Infect 132:283–289. 227. MacRae M, Hamilton C, Strachan NJC, Wright S, and Ogden ID. 2005. The detection of Cryptosporidium parvum and Escherichia coli O157 in UK bivalve shellfish. J Microbiol Meth 60:395–401. 228. Mannix M. 2005. Large E. coli O157 outbreak in Ireland, October–November 2005. Euro Surveill 10(12):E051222.3 www.eurosurveillance.org/ew/2005/051222.asp#3 229. Marsh J, MacLeod AF, Hanson M F, Emmanuel FX, Frost JA, and Thomas A. 1992. A restaurant-associated outbreak of E. coli O157 infection. J Pub Health Med 14:78–83. 230. Martin ML, Shipman LD, Wells JG, Potter ME, Hedberg K, Wachsmuth IK, Tauxe RV, Davis JP, Arnoldi J, and Tilleli J. 1986. Isolation of Escherichia coli O157:H7 from dairy cattle associated with two cases of haemolytic uraemic syndrome. Lancet 2(8514):1043. 231. Maruzumi M, Morita M, Matsuoka Y, Uekawa A, Nakamura T, and Fuji K. 2005. Mass food poisoning caused by beef offal contaminated by Escherichia coli O157. Jap J Infect Dis 58:397. 232. Matthews L, Low JC, Gally DL, Pearce MC, Mellor DJ, Heesterbeek JAP, Chase-Topping M, Naylor SW, Shaw DJ, Reid SWJ, Gunn GJ, and Woolhouse MEJ. 2006. Heterogeneous shedding of Escherichia coli O157 in cattle and its implications for control. Proc Nat Acad Sci USA 103:547–552. 233. Matthews L, McKendrick IJ, Ternent H, Gunn GJ, Synge B, and Woolhouse MEJ. 2006. Super-shedding cattle and the transmission dynamics of Escherichia coli O157. Epidemiol Infect 134:131–142. 234. McCall B, Strain D, Hills S, Heymer M, Bates J, Murphy D, Kelly R, and Price D. 1996. An outbreak of Escherichia coli O157 infection on the Gold Coast. Commun Dis Intelligence 20:236–239. 235. McCarthy TA, Barrett NL, Hadler JL, Salsbury B, Howard RT, Dingman DW, Brinkman CD, Bibb WF, and Cartter ML. 2001. Hemolytic-uremic syndrome and Escherichia coli O121 at a Lake in Connecticut, 1999. Pediatrics 108: NIL_9- NIL_15. 236. McCluskey BJ, Rice DH, Hancock DD, Hovde CJ, Besser TE, Gray S, and Johnson RP. 1999. Prevalence of Escherichia coli O157 and other Shiga-toxin-producing E. coli in lambs at slaughter. J Vet Diagn Invest 11:563–565. 237. McDonnell RJ, Rampling A, Crook S, Cockcroft PM, Wilshaw GA, Cheasty T, and Stuart J. 1997. An outbreak of Vero cytotoxin producing Escherichia coli O157 infection associated with takeaway sandwiches. Commun Dis Rep Rev 7:R201–R205. 238. McDonough S, Heer F, and Shireley L. 1991. Foodborne outbreak of gastroenteritis caused by Escherichia coli O157:H7—North Dakota, 1990. Morbid Mortal Weekly Rep 40:265–267. 239. McIntyre L, Fung J, Paccagnella A, Issac-Renton J, Rockwell F, Emerson B, and Preston T. 2002. Escherichia coli O157 outbreak associated with the ingestion of unpasteurized goat’s milk in British Columbia, 2001. Can Commun Dis Rep 28(1):6–8. 240. Mellmann A, Bielaszewska M, Zimmerhackl LB, Prager R, Harmsen D, Tschape H, and Karch H. 2005. Enterohemorrhagic Escherichia coli in human infection: In vivo evolution of a bacterial pathogen. Clin Infect Dis 41:785– 792. 241. Michino H, Araki K, Minami S, Takaya S, Sakai N, Miyazaki M, Ono A, and Yanagawa H. 1999. Massive outbreak of Escherichia coli O157:H7 infection in schoolchildren in Sakai City, Japan, associated with consumption of white radish sprouts. Am J Epidemiol 150:787–796. 242. Milne LM, Plom A, Strudley I, Pritchard GC, Crooks R, Hall M, Duckworth G, Seng C, Susman MD, Kearney J, Wiggins RJ, Moulsdale M, Cheasty T, and Willshaw GA. 1999. Escherichia coli O157 incident associated with a farm open to members of the public. Commun Dis Public Health 2:22–26. 243. Misselwitz J, Karch H, Bielazewska M, John U, Ringelmann F, Ronnefarth G, and Patzer L. 2003. Cluster of hemolytic- uremic syndrome caused by Shiga toxin-producing Escherichia coli O26:H11. Pediatr Infect Dis J 22:349–354. 244. Mohle-Boetani JC, Farrar JA, Werner SB, Minassian D, Bryant R, Abbott S, Slutsker L, and Vugia DJ. 2001. Escherichia coli O157 and Salmonella infections associated with sprouts in California, 1996–1998. Ann Int Med 135:239–247. 245. Moore K, Damrow T, Abbott DO, and Jankowski S. 1995. Outbreak of acute gastroenteritis attributable to Escherichia coli serotype O104:H21—Helena, Montana, 1994. Morbid Mortal Weekly Rep 44:501–503. 246. Morabito S, Dell’Omo G, Agrimi U, Schmidt H, Karch H, Cheasty T, and Caprioli A. 2001. Detection and characterization of Shiga toxin-producing Escherichia coli in feral pigeons. Vet Microbiol 82:275–283. 247. Morgan AK, Piispanen, Humphreys, and Murphy D. 2005. A cluster of cases of haemolytic uraemic syndrome in north Queensland associated with a novel Shiga-like toxin-producing Escherichia coli. Commun Dis Intelligence 29:191–193. 248. Morgan D, Newman CP, Hutchinson DN, Walker AM, Rowe B, and Majid F. 1993. Verotoxin producing Escherichia coli

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 32 White Paper:

O157 infections associated with consumption of yoghurt. Epidemiol Infect 111:181–187. 249. Moriya K, Fujibayashi T, Yoshihara T, Matsuda A, Sumi N, Umezaki N, Kurahashi H, Agui N, Wada A, and Watanabe H. 1999. Verotoxin-producing Escherichia coli 0157:H7 carried by the housefly in Japan. Med Vet Entomol 13:214–216. 250. Mudgett CC, Ruden R, and Austin CC. 1998. A beach-associated outbreak of Escherichia coli O157:H7. J Environ Health 60:7–13. 251. Mukherjee A, Cho S, Scheftel J, Jawahir S, Smith K, and Diez-Gonzalez. 2006. Soil survival of Escherichia coli O157:H7 acquired by a child from garden soil recently fertilized with cattle manure. J Appl Microbiol 101:429–436. 252. Murphy J. 2005. Minnesota finds E. coli in lettuce bags. www.producenews.com/storydetail.cfm?ID=5449. 253. Nagano H, Hirochi T, Fujita K, Wakamori Y, Takeshi K, and Yano S. 2004. Phenotypic and genotypic characterization of beta-D-glucuronidase-positive Shiga toxin-producing Escherichia coli O157:H7 isolates from deer. J Med Microbiol 53:1037–1043. 254. Naugle AL, Holt KG, Levine P, and Eckel R. 2005. Food safety and inspection service regulatory testing program for Escherichia coli O157:H17 in raw ground beef. J Food Prot 68:462–468. 255. Naylor SW, Low JC, Besser TE, Mahajan A, Gunn GJ, Pearce MC, McKendrick IJ, Smith DG, and Gally DL. 2003. Lymphoid follicle-dense mucosa at the terminal rectum is the principal site of colonization of enterohemorrhagic Escherichia coli O157:H7 in the bovine host. Infect Immun 71:1505–1512. 256. Nielsen EM and Scheutz F. 2002. Characterisation of Escherichia coli O157 isolates from Danish cattle and human patients by genotyping and presence and variants of virulence genes. Vet Microbiol 88:259–273. 257. Notario R, Fain JC, Prado V, Rios M, Borda N, and Gambande T. 2000. Animal reservoir and genotipic characterization of enterohemorrhagic Escherichia coli (EHEC) in Argentina. Rev Med Chile 128:1335–1341. 258. Novello A. 1999. Outbreak of Escherichia coli O157:H7 and Campylobacter among attendees of the Washington County Fair-New York, 1999. Morbid Mortal Weekly Rep 48:803–805. 259. Novotna R, Alexa P, Hamrik J, Madanat A, Smola J, and Cizek A. 2005. Isolation and characterization Shiga toxin- producing Escherichia coli from sheep and goats in Jordan with evidence of multiresistant serotype O157:H7. Vet Medicina 50:111–118. 260. NY State Dept Health. 2000. Health Commissioner releases E. coli outbreak report. www.health.state.ny.us/press/releases/2000/ecoli.htm 261. O’Brien SJ, Gillespie IA, Sivanesan MA, Elson R, Hughes C, and Adak GK. 2006. Publication bias in foodborne outbreaks of infectious intestinal disease and its implications foe evidence-based food policy. England and Wales 1992– 2003. Epidemiol Infect 134:667–674. 262. O’Brien SJ, Murdoch PS, Riley AH, King I, Barr M, Murdoch S, Greig A, Main R, Reilly WJ, and Thomson-Carter FM. 2001. A foodborne outbreak of Vero cytotoxin-producing Escherichia coli O157:H-phage type 8 in hospital. J Hosp Infect 49:167–172. 263. O’Flynn G, Ross RP, Fitzgerald GF, and Coffey A. 2004. Evaluation of a cocktail of three bacteriophages for biocontrol of Escherichia coli O157:H7. Appl Environ Microbiol 70:3417–3424. 264. Ogden ID, Hepburn NF, MacRae M, Strachan NJC, Fenlon DR, Rusbridge SM, and Pennington TH. 2002. Long-term survival of Escherichia coli O157 on pasture following an outbreak associated with sheep at a scout camp. Lett Appl Microbiol 34:100–104. 265. Ogden ID, MacRae M, and Strachan NJC. 2005. Concentration and prevalence of Escherichia coli O157 in sheep faeces at pasture in Scotland. J Appl Microbiol 98:646–651. 266. Olsen SJ, Miller G, Breuer T, Kennedy M, Higgins C, Walford J, McKee G, Fox K, Bibb W, and Mead P. 2002. A waterborne outbreak of Escherichia coli O157:H7 infections and hemolytic uremic syndrome: Implications for rural water systems. Emerg Infect Dis 8:370–375. 267. Omisakin F, MacRae M, Ogden ID, and Strachan NJC. 2003. Concentration and prevalence of Escherichia coli 0157 in cattle feces at slaughter. Appl Environ Microbiol 69:2444–2447. 268. Orr P, Lorencz B, Brown R, Kielly R, Tan B, Holton D, Clugstone H, Lugtig L, Pim C, Macdonald S, Hammond G, Moffatt M, Spika J, Manuel D, Winther W, Milley D, Lior H, and Sinuff N. 1994. An outbreak of diarrhea due to verotoxin-producing Escherichia coli in the Canadian Northwest Territories. Scan J Infect Dis 26:675–684. 269. Osek J. 2004. Phenotypic and genotypic characterization of Escherichia coli O157 strains isolated from humans, cattle and pigs. Vet Medicina 49:317–326. 270. Ostroff SM, Griffin PM, Tauxe RV, Shipman LD, Greene KD, Wells JG, Lewis JH, Blake PA, and Kobayashi JM. 1990. A statewide outbreak of Escherichia coli O157:H7 infections in Washington State. Am J Epidemiol 132:239–247. 271. Paton AW, Ratcliff RM, Doyle RM, Seymour-Murray J, Davos D, Lanser JA, and Paton JC. 1996. Molecular microbiological investigation of an outbreak of hemolytic-uremic syndrome caused by dry fermented sausage contaminated with shiga-like toxin-producing Escherichia coli. J Clin Microbiol 34:1622–1627. 272. Paunio M, Pebody R, Keskimaki M, Kokki M, Ruutu P, Oinonen S, Vuotari V, Siitonen A, Lahti E, and Leinikki P. 1999. Swimming-associated outbreak of Escherichia coli O157: H7. Epidemiol Infect 122:1–5. 273. Pavia AT, Nichols CR, Green DP, Tauxe RV, Mottice S, Greene KD, Wells JG, Siegler RL, Brewer ED, Hannon D, and Blake PA. 1990. Hemolytic-uremic syndrome during an outbreak of Escherichia coli O157:H7 infections in institutions for mentally retarded persons: clinical and epidemiologic observations. J Pediatr 116:544–551. 274. Payne CJI, Petrovic M, Roberts RJ, Paul A, Linnane E, Walker M, Kirby D, Burgess A, Smith RMM, Cheasty T, Willshaw G, and Salmon RL. 2003. Vero cytotoxin-producing Escherichia coli O157 gastroenteritis in farm visitors,

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 33

North Wales. Emerg Infect Dis 9:526–530. 275. Pebody RG, Furtado C, Rojas A, McCarthy N, Nylen G, Ruutu P, Leino T, Chalmers R, de Jong B, Donnelly M, Fisher I, Gilham C, Graverson L, Cheasty T, Willshaw G, Navarro M, Salmon R, Leinikki P, Wall P, and Bartlett C. 1999. An international outbreak of Vero cytotoxin-producing Escherichia coli O157 infection amongst tourists; a challenge for the European infectious disease surveillance network. Epidemiol Infect 123:217–223. 276. Pedersen K, Clark L, Andelt WF, and Salman MD. 2006. Prevalence of shiga toxin-producing Escherichia coli and Salmonella enterica in rock pigeons captured in fort Collins, Colorado. J Wildlife Dis 42:46–55. 277. Pennington H. 1998. Factors involved in recent outbreaks of Escherichia coli O157:H7 in Scotland and recommendations for its control. J Food Safety 18:383–391. 278. Pennington TH. 2000. VTEC: lessons learned from British outbreaks. J Appl Microbiol 88:90S–98S. 279. Phebus RK, Nutsch AL, Schafer DE, Wilson RC, Riemann MJ, Leising JD, Kastner CL, Wolf JR, and Prasai RK. 1997. Comparison of steam pasteurization and other methods for reduction of pathogens on surfaces of freshly slaughtered beef. J Food Prot 60:476–484. 280. Pichner R, Sander A, Steinruck H, and Gareis M. 2005. Occurrence of Salmonella spp. and shigatoxin-producing Escherichia coli (STEC) in horse faeces and horse meat products. Berlin Munch Tierarzt Wochenschr 118:321–325. 281. Pilipcinec E, Tkacikova L, Naas HT, Cabadaj R, and Mikula I. 1999. Isolation of verotoxigenic Escherichia coli O157 from poultry. Folia Microbiol 44:455–456. 282. Pinto C. 2006. E. coli sends four toddlers to hospital. http://www.hendersonvillestarnews.com/apps/pbcs.dll/article?AID=/20060628/NEWS07/60628016 June 29, 2006. 283. Pollett GL and Warshawsky B. 1999. E. coli O157:H7 investigation complete. Middlesex-London Health Unit. www.healthunit.com/index.asp?mode=article&lang=english&articleID=11329&rank=313 284. Potter AA, Klashinsky S, Li Y, Frey E, Townsend H, Rogan D, Erickson G, Hinkley S, Klopfenstein T, Moxley RA, Smith DR, and Finlay BB. 2004. Decreased shedding of Escherichia coli O157:H7 by cattle following vaccination with type III secreted proteins. Vaccine 22:362–369. 285. Preston M, Borczyk A, Davidson R, McGeer A, Bertoli J, Harris S, Thususka J, Goldman C, Green K, Low D, Proctor P, Johnson W, and Khakhria R. 1997. Hospital outbreak of Escherichia coli O157:H7 associated with a rare phage type— Ontario. Can Commun Dis Rep 23(5):71–73. 286. Pritchard GC, Williamson S, Carson T, Bailey JR, Warner L, Willshaw G, and Cheasty T. 2001. Wild rabbits—novel vector for verocytotoxigenic Escherichia coli O157. Vet Rec 149:567. 287. Pritchard GC, Willshaw GA, Bailey JR, Carson T, and Cheasty T. 2000. Verocytotoxin-producing Escherichia coli O157 on a farm open to the public: outbreak investigation and longitudinal bacteriological study. Vet Rec 147:259–264. 288. Rabatsky-Ehr T, Dingman D, Marcus R, Howard R, Kinney A, and Mshar P. 2002. Deer meat as the source for a sporadic case of Escherichia coli O157:H7 infection, Connecticut. Emerg Infect Dis 8:525–527. 289. Rahn K, Renwick SA, Johnson RP, Wilson JB, Clarke RC, Alves D, McEwen S, Lior H, and Spika J. 1997. Persistence of Escherichia coli O157:H7 in dairy cattle and the dairy farm environment. Epidemiol Infect 119:251–259. 290. Rajpura A, Lamden K, Forster S, Clarke S, Cheesbrough J, Gornall S, and Waterworth S. 2003. Large outbreak of infection with Escherichia coli O157 PT21/28 in Eccleston, Lancashire, due to cross contamination at a butcher’s counter. Commun Disease Public Health 6:279–284. 291. Rangel JM, Sparling PH, Crowe C, Griffin PM, and Swerdlow DL. 2005. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982–2002. Emerg Infect Dis 11:603–609. 292. Ravva SV, Sarreal CZ, Duffy B, and Stanker LH. 2006. Survival of Escherichia coli O157:H7 in wastewater from dairy lagoons. J Appl Microbiol (in press). 293. Reida P, Wolff M, Pohls HW, Kuhlmann W, Lehmacher A, Aleksic S, Karch H, and Bockemuhl J. 1994. An outbreak due to enterohaemorrhagic Escherichia coli O157:H7 in a children day care centre characterized by person-to-person transmission and environmental contamination. Int J Med Microbiol Virol Parasitol Infect Dis 281:534–43. 294. Reiss G, Kunz P, Koin D, and Keeffe EB. 2006. Escherichia coli O157:H7 infection in nursing homes: Review of literature and report of recent outbreak. J Am Geriatr Soc 54:680–684. 295. Renter DG, Morris JG, Sargeant JM, Hungerford LL, Berezowski J, Ngo T, Williams K, and Acheson DWK. 2005. Prevalence, risk factors, O serogroups, and virulence profiles of Shiga toxin-producing bacteria from cattle production environments. J Food Prot 68:1556–1565. 296. Renter DG, Sargeant JM, Hygnstorm SE, Hoffman JD, and Gillespie JR. 2001. Escherichia coli O157:H7 in free-ranging deer in Nebraska. J Wildlife Dis 37:755–760. 297. Renter DG, Sargeant JM, Oberst RD, and Samadpour M. 2003. Diversity, frequency, and persistence of Escherichia coli O157 strains from range cattle environments. Appl Environ Microbiol 69:542–547. 298. Rice DH, Hancock DD, and Besser IE. 2003. Faecal culture of wild animals for Escherichia coli O157:H7. Vet Rec 152:82–83. 299. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, Hebert RJ, Olcott ES, Johnson LM, Hargrett NT, Blake PA, and Cohen ML. 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N Engl J Med 308:681–685. 300. Roberts CL, Mshar PA, Cartter ML, Hadler JL, Sosin DM, Hayes PS, and Barrett TJ. 1995. The role of heightened surveillance in an outbreak of Escherichia coli O157:H7. Epidemiol Infect 115:447–454. 301. Roberts JA, Upton PA, and Azene G. 2000. Escherichia coli O157: H7; an economic assessment of an outbreak. J Pub

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 34 White Paper:

Health Med 22:99–107. 302. Rodrigue DC, Mast EE, Greene KD, Davis JP, Hutchinson MA, Wells JG, Barrett TJ, and Griffin PM. 1995. A university outbreak of Escherichia coli O157:H7 infections associated with roast beef and an unusually benign clinical course. J Infect Dis 172:1122–1125. 303. Ryan CA, Tauxe RV, Hosek GW, Wells JG, Stoesz PA, McFadden HW Jr, Smith PW, Wright GF, and Blake PA. 1986. Escherichia coli O157:H7 diarrhea in a nursing home: clinical, epidemiological, and pathological findings. J Infect Dis 154:631–638. 304. Safdar N, Said A, Gangnon RE, and Maki DG. 2002. Risk of hemolytic uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 enteritis—A meta-analysis. J Am Med Assoc 288:996–1001. 305. Salmon R. 2005. Outbreak of verotoxin producing E. coli O157 infections involving over forty schools in south Wales, September, 2005. Euro Surveill 10(10):E051006.1 www.eurosurveillance.org/ew/2005/051006.asp#1 306. Samadpour N, Stewart J, Steingart K, Addy C, Louderback J, McGinn M, Ellington J, and Newman T. 2002. Laboratory investigation of an E. coli 0157:H7 outbreak associated with swimming in Battle Ground Lake, Vancouver, Washington. J Environ Health 64:16–20. 307. Sargeant JM, Hafer DJ, Gillespie JR, Oberst RD, and Flood SJA. 1999. Prevalence of Escherichia coli O157:H7 in white- tailed deer sharing rangeland with cattle. J Am Vet Med Assoc 215:792–794. 308. Sasaki T, Kobayashi M, and Agui N. 2000. Epidemiological potential of excretion and regurgitation by Musca domestica (Diptera: Muscidae) in the dissemination of Escherichia coli O157:H7 to food. J Med Entomol 37:945–949. 309. Scaife HR, Cowan D, Finney J, Kinghorn-Perry SF, and Crook B. 2006. Wild rabbits (Oryctolagus cuniculus) as potential carriers of verocytotoxin-producing Escherichia coli. Vet Rec 159:175–178. 310. Schimmer B, Outbreak Investigation Team. 2006. Outbreak of haemolytic uraemic syndrome in Norway: update. Euro Surveill 11(4):E060406.2 www.eurosurveillance.org/ew/2006/060406.asp#2 311. Schmidt H, Scheef J, Morabito S, Caprioli A, Wieler LH, and Karch A. 2000. A new shiga toxin 2 variant (Stx2f) from Escherichia coli isolated from pigeons. Appl Environ Microbiol 66:1205–1208. 312. Schoeni JL and Doyle MP. 1994. Variable colonization of chickens perorally inoculated with Escherichia coli O157:H7 and subsequent contamination of eggs. Appl Environ Microbiol 60:2958–2962. 313. Schouten JM, van de Giessen AW, Frankena K, De Jong MCM, and Graat EAM. 2005. Escherichia coli O157 prevalence in Dutch poultry, pig finishing and veal herds and risk factors in Dutch veal herds. Prev Vet Med 70:1–15. 314. Shefer AM, Koo D, Werner SB, Mintz ED, Baron R, Wells JG, Barrett TJ, Ginsberg M, Bryant R, Abbott S, and Griffin PM. 1996. Cluster of Escherichia coli O157:H7 infections with the hemolytic-uremic syndrome and death in California— a mandate for improved surveillance. West J Med 165:15–19. 315. Sheng H, Knecht HJ, Kudva IT, and Hovde CJ. 2006. Application of bacteriophages to control intestinal Escherichia coli O157:H7 levels in ruminants. Appl Environ Microbiol 72:5359–5366. 316. Shere JA, Bartlett KJ, and Kaspar CW. 1998. Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Appl Environ Microbiol 64:1390–1399. 317. Shere JA, Kaspar CW, Bartlett KJ, Linden SE, Norell B, Francey S, and Schaefer DM. 2002. Shedding of Escherichia coli O157:H7 in dairy cattle housed in a confined environment following waterborne inoculation. Appl Environ Microbiol 68:1947–1954. 318. Shillam P, Heltzel D, Beebe J, and Hoffman R. 1997. Escherichia coli O157:H7 infections associated with eating a nationally distributed commercial brand of frozen ground beef patties and burgers—Colorado, 1997. Morbid Mortal Weekly Rep 46:777–778. 319. Shillam P, Woo-Ming A, Mascola L, Bagby R, Lohff C, Bidol S, Stobierski MG, Carlson C, Schaefer L, Kightlinger L, Seys S, Kubota K, Mead PS, and Kalluri P. 2002. Multistate outbreak of Escherichia coli O157:H7 infections associated with eating ground beef—United States, June–July 2002. Morbid Mortal Weekly Rep 51:637–639. 320. Shukla R, Slack R, George A, Cheasty T, Rowe B, and Scutter J. 1995. Escherichia coli O157 infection associated with a farm visitor centre. Commun Dis Rep Rev 5:R86–R90. 321. Smith DJ, Oliver CE, Caton JS, and Anderson RC. 2005. Effect of sodium [36Cl]chlorate dose on total radioactive residues and residues of parent chlorate in beef cattle. J Agr Food Chem 53(18):7352–7360. 322. Smith KE, Stenzel SA, Bender JB, Wagstrom E, Soderlund D, Leano FT, Taylor CM, Belle-Isle PA, and Danila R. 2004. Outbreaks of enteric infections caused by multiple pathogens associated with calves at a farm day camp. Pediatr Infect Dis J 23:1098–1104. 323. Smith-Palmer A, Locking M, Reilly B, and Fisher I. 2005. Cluster of E. coli O157 infections in Scottish tourists returning from southwest Turkey, July–August 2005. Euro Surveill 10(8):E050818.2 www.eurosurveillance.org/ew/2005/050818.asp#2 324. Solomon EB, Pang HJ, and Matthews KR. 2003. Persistence of Escherichia coli O157:H7 on lettuce plants following spray irrigation with contaminated water. J Food Prot 66:2198–2202. 325. Sproston EL, Macrae M, Ogden ID, Wilson MJ, and Strachan NJC. 2006. Slugs: Potential novel vectors of Escherichia coli O157. Appl Environ Microbiol 72:144–149. 326. Stavric S, Buchanan B, and Gleeson TM. 1993. Intestinal colonization of young chicks with Escherichia coli O157:H7 and other verotoxin-producing serotypes. J Appl Microbiol 74:557–563. 327. Stevenson J and Hanson S. 1996. Outbreak of Escherichia coli O157 phage type 2 infection associated with eating precooked meats. Commun Dis Rep Rev 6:R116–R118.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 35

328. Strachan NJC, Doyle MP, Kasuga F, Rotariu O, and Ogden ID. 2005. Dose response modelling of Escherichia coli O157 incorporating data from foodborne and environmental outbreaks. Int J Food Microbiol 103:35–47. 329. Sugiyama A, Iwade Y, Akachi S, Nakano Y, Matsuno Y, Yano T, Yamauchi A, Nakayama O, Sakai H, Yamamoto K, Nagasaka Y, Nakano T, Ihara T, and Kamiya H. 2005. An outbreak of Shigatoxin-producing Eshcherichia coli O157:H7 in a nursery school in Mie Prefecture. Jap J Infect Dis 58:398–400. 330. Sutcliffe P, Picard L, Fortin BA, Malaviarachchi D, Hohenadel J, and O’Donnell B. 2004. Escherichia coli O157:H7 outbreak at a summer hockey camp, Sudbury, 2004. Can Commun Dis Rep 30(22):189–194. 331. Swerdlow DL, Woodruff BA, Brady RC, Griffin PM, Tippen S, Donnell HD Jr, Geldreich E, Payne BJ, Meyer A Jr, Wells JG, and et al. 1992. A waterborne outbreak in Missouri of Escherichia coli O157:H7 associated with bloody diarrhea and death. Ann Int Med 117:812–819. 332. Szalanski AL, Owens CB, Mckay T, and Steelman CD. 2004. Detection of Campylobacter and Escherichia coli O157:H7 from filth flies by polymerase chain reaction. Med Vet Entomol 18:241–246. 333. Söderström A, Lindberg A, and Andersson Y. 2005. EHEC O157 outbreak in Sweden from locally produced lettuce, August–September 2005. Euro Surveill 10(9):E050922.1 www.eurosurveillance.org/ew/2005/050922.asp#1 334. Tamblyn S, deGrosbois BA, Taylor D, and Stratton J. 1999. An outbreak of Escherichia coli O157:H7 infection associated with unpasteurized non-commercial, custom-pressed apple cider—Ontario, 1998. Can Commun Dis Rep 25(13):113–120. 335. Teunis P, Takumi K, and Shinagawa K. 2004. Dose response for infection by Escherichia coli O157:H7 from outbreak data. Risk Anal 24:401–407. 336. Thoms B. 1999. Detection of verotoxin-producing Escherichia coli in deer’s meat. Arch Lebensmittelhyg 50:52–54. 337. Tkalcic S, Zhao T, Harmon BG, Doyle MP, Brown CA, and Zhao P. 2003. Fecal shedding of enterohemorrhagic Escherichia coli in weaned calves following treatment with probiotic Escherichia coli. J Food Prot 66:1184–1189. 338. Tozzi AE, Niccolini A, Caprioli A, Luzzi I, Montini G, Zacchello G, Gianviti A, Principato F, and Rizzoni G. 1994. A community outbreak of haemolytic-uraemic syndrome in children occurring in a large area of northern italy over a period of several months. Epidemiol Infect 113:209–219. 339. Tserenpuntsag B, Chang HG, Smith PF, and Morse DL. 2005. Hemolytic uremic syndrome risk and Escherichia coli O157:H7. Emerg Infect Dis 11:1955–1957. 340. Turney C, Greensmith M, Shipp M, Mordhorst C, Whittingslow C, Brawley L, Koppel D, Bridges E, Davis G, Voss J, Lee R, Jay M, Abbott S, Bryant R, Reilly K, Werner SB, Barrett L, Jackson RJ, Rutherford GW, and Lior H. 1994. Escherichia coli O157:H7 outbreak linked to home-cooked hamburger—California, July 1993. Morbid Mortal Weekly Rep 43:213–216. 341. Tuttle J, Gomez T, Doyle MP, Wells JG, Zhao T, Tauxe RV, and Griffin PM. 1999. Lessons from a large outbreak of Escherichia coli O157:H7 infections: insights into the infectious dose and method of widespread contamination of hamburger patties. Epidemiol Infect 122:185–192. 342. Upton P and Coia JE. 1994. Outbreak of Escherichia coli O157 infection associated with pasteurised milk supply. Lancet 344:1015. 343. Valcour JE, Michel P, McEwen SA, and Wilson JB. 2002. Associations between indicators of livestock farming intensity and incidence of human shiga toxin-producing Escherichia coli infection. Emerg Infect Dis 8:252–257. 344. Vanselow BA, Krause DO, and McSweeney CS. 2005. The Shiga toxin-producing Escherichia coli, their ruminant hosts, and potential on-farm interventions: a review. Austral J Agr Res 56:219–244. 345. Varma JK, Greene KD, Reller ME, DeLong SM, Trottier J, Nowicki SF, DiOrio M, Koch EM, Bannerman TL, York ST, Lambert-Fair MA, Wells JG, and Mead PS. 2003. An outbreak of Escherichia coli O157 infection following exposure to a contaminated building. J Am Med Assoc 290:2709–2712. 346. Vogt RL and Dippold L. 2005. Escherichia coli O157:H7 outbreak associated with consumption of ground beef, June– July 2002. Pub Health Rep 120:174–178. 347. Wallace JS, Cheasty T, and Jones K. 1997. Isolation of vero cytotoxin-producing Escherichia coli O157 from wild birds. J Appl Microbiol 82:399–404. 348. Warrner M, Kuo K, Williams L, Adam B, Langkop C, Ruden R, Francis B, and Haupt T. 1996. Lake-associated outbreak of Escherichia coli O157:H7—Illinois, 1995. Morbid Mortal Weekly Rep 45:437–439. 349. Wasteson Y, Arnemo JM, Johansen BK, Vold L, Mathiesen SD, Olsen MA, Wiig O, and Derocher AE. 1999. Analysis of faecal samples from wild animals for verocytotoxin producing Escherichia coli and E. coli O157. Vet Rec 144:646–647. 350. Watanabe H, Wada A, Inagaki Y, Itoh K, and Tamura K. 1996. Outbreaks of enterohaemorrhagic Escherichia coli O157:H7 infection by two different genotype strains in Japan, 1996. Lancet 348:831–832. 351. Watanabe Y, Ozasa K, Mermin JH, Griffin PM, Masuda K, Imashuku S, and Sawada T. 1999. Factory outbreak of Escherichia coli O157:H7 infection in Japan. Emerg Infect Dis 5:424–428. 352. Welinder-Olsson C, Stenqvist K, Badenfors M, Brandberg A, Florén K, Holm M, Holmberg L, Kjellin E, Marild S, Studahl A, and Kaijser B. 2004. EHEC outbreak among staff at a children’s hospital—use of PCR for verocytotoxin detection and PFGE for epidemiological investigation. Epidemiol Infect 132:43–49. 353. Wells JG, Davis B R, Wachsmuth IK, Riley LW, Remis RS, Sokolow R, and Morris GK. 1983. Laboratory investigation of hemorrhagic colitis outbreaks associated with a rare Escherichia coli serotype. J Clin Microbiol 18:512–520. 354. Werber D, Fruth A, Liesegang A, Littmann M, Buchholz U, Prager R, Karch H, Breuer T, Tschape H, and Ammon A. 2002. A multistate outbreak of Shiga toxin-producing Escherichia coli O26:H11 infections in Germany, detected by

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 36 White Paper:

molecular subtyping surveillance. J Infect Dis 186:419–422. 355. Wick LM, Qi WH, Lacher DW, and Whittam TS. 2005. Evolution of genomic content in the stepwise emergence of Escherichia coli O157:H7. J Bacteriol 187:1783–1791. 356. Williams AP, Avery LM, Killham K, and Jones DL. 2005. Persistence of Escherichia coli O157 on farm surfaces under different environmental conditions. J Appl Microbiol 98:1075–1083. 357. Williams LD, Hamilton PS, Wilson BW, and Estock MD. 1997. An outbreak of Escherichia coli O157:H7 - involving long term shedding and person-to-person transmission in a child care center. J Environ Health 59:9–14. 358. Williams RC, Isaacs S, Decou ML, Richardson EA, Buffett MC, Slinger RW, Brodsky MH, Ciebin BW, Ellis A, and Hockin A. 2000. Illness outbreak associated with Escherichia coli O157:H7 in Genoa salami. Can Med Assoc J 162:1409–1413. 359. Willshaw GA, Smith HR, Cheasty T, Wall PG, and Rowe B. 1997. Vero cytotoxin-producing Escherichia coli O157 outbreaks in England and Wales, 1995: phenotypic methods and genotypic subtyping. Emerg Infect Dis 3:561–565. 360. Willshaw GA, Thirlwell J, Jones AP, Parry S, Salmon RL, and Hickey M. 1994. Vero cytotoxin-producing Escherichia coli O157 in beefburgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome in Britain. Lett Appl Microbiol 19:304–307. 361. Wilson J, Spika J, Clarke R, McEwen S, Johnson R, Rahn K, Renwick S, Karmali M, Lior H, Alves D, Gyles C, and Sandhu K. 1998. Verocytotoxigenic Escherichia coli infection in dairy farm families. Can Commun Dis Rep 24(3). 362. Xu JG, Liu QY, Jing HQ, Pang B, Yang JC, Zhao GF, and Li HW. 2003. Isolation of Escherichia coli O157:H7 from dung beetles Catharsius molossus. Microbial Immunol 47:45–49. 363. Yamasaki S and Takeda Y. 1997. Enterohemorrhagic Escherichia coli O157:H7 episode in Japan with a perspective on vero toxins (shiga-like toxins). J Toxicol Toxin Rev 16:229–240. 364. Yoder JS, Blackburn BG, Craun GF, Hill V, Levy DA, Chen N, Lee SH, and Calderon RL Beach MJ. 2004. Surveillance for waterborne disease outbreaks associated with recreational water—United States, 2001–2002. Morbid Mortal Weekly Rep 53(SS08):1–22. 365. Younts-Dahl SP, Osborn GD, Galyean ML, Rivera JD, Loneragan GH, and Brashears MM. 2005. Reduction of Escherichia coli O157 in finishing beef cattle by various doses of Lactobacillus acidophilus in direct-fed microbials. J Food Protection 68:6–10. 366. Zhang J, Xia S, Shen G, Chen Z, Huang P, Fu B, and Tu G. 2002. A study on acute renal failure after an outbreak of diarrhea in Suixian county, Henan province [Chinese]. Chung-Hua Liu Hsing Ping Hsueh Tsa Chih Chinese J Epidemiol 23:105–107. 367. Zhao T, Doyle MP, Harmon BG, Brown CA, Mueller POE, and Parks AH. 1998. Reduction of carriage of enterohemorrhagic Escherichia coli O157:H7 in cattle by inoculation with probiotic bacteria. J Clin Microbiol 36:641– 647. 368. Ziese T, Anderson Y, de Jong B, Löfdahl S, and Ramberg M. 1996. Outbreak of Escherichia coli O157 in Sweden. Euro Surveill 1(1):2–3. 369. Zweifel C, S. Schumacher, Beutin L, Blanco J, and Stephan R. 2006. Virulence profiles of Shiga toxin 2e-producing Escherichia coli isolated from healthy pig at slaughter. Vet Microbiol. (in press). 370. Archer, John (personal communication). Epidemiologist, Wisconsin Division of Public Health, Bureau of Communicable Diseases and Preparedness, Communicable Disease Epidemiology Section. http://dhfs.wisconsin.gov/communicable/Communicable/Contacts.htm 371. Loneragan GH, and Brashears MM. 2005. Pre-harvest interventions to reduce carriage of E. coli O157 by harvst-ready feedlot cattle. Meat Sci 71:72–78.

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 37

Appendix 1

Chronological List of Outbreaks of E. coli O157:H7

Date Location Cases Deaths HUS Vehicle Ref # 1982 (a) US: OR, MI 20 beef, ground (299;353) 1982 (b) Canada 31 1 0 beef, ground (294) 1984 US: NE 34 4 1 beef, ground (nursing home) (303) 1985 Canada (nursing home) 73 17 12 meat;person-person (81;206) 1986 (a) US: WA 40 2 1 beef, ground (270) 1986 (b) US: WI 2 0 2 milk, unpasteurized (230) 1986 (c) Canada 74 1 milk, unpasteurized (178) 1988 (a) US: MN 32 0 0 meat patties (54) 1988 (b) US: WI 61 0 0 beef, roast (302) 1988 (c) US: UT 51 4 8 beef, ground (273) 1989 Canada 7 1 water, well (nursing home) (294) 1990 (a) Japan 14 2 14 water, tap (5) 1990 (b) US: MO 243 4 2 water, well (331) 1990 (c) UK: Scotland 11 4 person-person;geriatric hospital (205) 1990 (d) UK 16 4 restaurant food (229) 1990 (e) UK 4 0 4 water, drinking (122) 1990 (f) Israel 4 0 0 person-person (219) 1990 (g) US: ND 70 2 beef, roast (238) 1991 (a) Canada 521 2 22 beef, ground?;person-person (268) 1991 (b) US: MA 18 0 0 apple cider (59) 1991 (c) UK 16 0 5 yogurt (248) 1991 (d) US: OR 21 0 3 water, lake (201) 1992 (a) Germany 45 1 3 person-person (293) 1992 (b) Scotland 6 1 water, pool (71) 1992-3 (c) US: WA, ID, CA, NV >700 3 37 beef, ground (52;69;100;119; (Jack in Box) 314;341) 1992-3 (f) France 4 0 4 cheese, cows’ unpasteurized (121) 1992-4 US: OR 14 0 0 milk, unpasteurized (200) 1993 (a) UK: Wales 8 1 beef, ground (360) 1993 (b) US: OR, WA, 93 0 0 salad bar ? (184) 1993 (c) Netherlands 11 4 water, lake (111) 1993 (d) US: CA 15 0 0 beef, ground (340) 1993 (e) UK 6 1 3 water, paddling pool (171) 1993 (f) US: CT 20 0 0 beef, ground (300) 1993 (g) Germany 6 1 person-person (224) 1994 (a) US: VA 20 0 1 beef, ground (144) 1994 (b) US: WA, CA 23 0 2 salami (9) 1994 (c) Italy 15 1 15 poultry ? (338) 1994 (d) UK: Scotland >100 1 9 milk, pasteurized (342) 1994 (e) US: NY 12 0 2 water, lake (4) 1994 (f) UK: Scotland 22 beef, ground (118;278) 1994 (g) UK: Scotland 71 2 10 milk, pasteurized (301) 1994 (h) UK: Scotland 26 0 1 beef, ground (117)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 38 White Paper:

1994 (i) US: CT 21 supermarket, food? (47) 1994 (j) UK: England 7 4 animals, farm (320) 1994 (k) US: WI 26 0 0 sandwiches (370) 1994 (l) UK: Scotland 22 1 cheese (178) 1995 (a) US: IL 12 3 water, lake (348) 1995 (b) US: Mont. 92 1 lettuce, leaf (3) 1995 (c) UK: Wales 31 2 person-person (6) 1995 (d) US: GA, TN 10 1 beef, ground (79) 1995 (e) US: IL 12 water, lake (250) 1995 (f) US: OR 11 0 0 deer jerky (202) 1995 (g) UK 8 1 potatoes, person-person (96) 1995 (h) Canada 21 0 0 lettuce (285) 1995 (j) UK: Scotland 633 0 2 water, tap (193) 1995 (k) UK 14 0 2 meats, precooked (327) 1995 (l) UK 26 0 2 sandwiches, takeaway (237) 1995 (m) Czech Republic 9 4 milk, goat unpasteurized (65) 1995 (n) US: WI 27 0 0 water (370) 1995 (o) UK: England 11 0 4 ham, cooked (359) 1995 (p) US: MN 33 water, drinking (220) 1995-6 (a) Sweden 110 0 29 unknown (368) 1995-6 (b) Germany 28 3 28 sausage (12) 1996 (a) US and Canada (Odwalla) 70 1 14 apple juice (14;105) 1996 (b) US: CT, IL 61 3 lettuce, mesclun (169) 1996 (c) US: GA 18 1 water, pool (143) 1996 (d) Japan, Sakai 12,680 12 121 sprouts, radish (schools) (145;241;350;363) 1996 (e) US: CT 14 0 4 apple cider (170) 1996 (f) Japan 47 1 3 sprouts,radish (factory) (351) 1996 (g) Scotland (Wishaw) 503 20 34 meats, cooked (128;277;278) 1996 (h) US: CO 24 0 1 person-person (357) 1996 (i) Australia 6 0 0 food handler (234) 1996 (j) US: FL 2 water, well (220) 1996 (k) UK 3 0 0 milk, unpasteurized (15) 1996 (l) UK 12 0 1 milk, raw and "pasteurized" (104) 1996-7 Scotland 512 17 meat from a shop (109) 1997 Europe 15 0 3 water, well (275) 1997 (a) US: CO (Hudson) 15 beef, ground (318) 1997 (c) US: MI, VA 108 0 4 sprouts, alfalfa (70;107) 1997 (d) UK: England 5 2 animals, farm (287) 1997 (e) UK 8 0 1 environment (dairy farm field) (21;110) 1997 (f) UK 45 0 0 cakes, cream-filled (262) 1997 (g) Finland 14 water, lake (272) 1997 (h) UK 332 0 1 restaurant food (17) 1997 (i) UK 2 0 0 cheese, cows’ unpasteurized (23) 1997 (j) UK 3 0 2 animals, farm (92;242) 1997 (k) Europe (travel overseas) 10 0 3 water, drinking ? (18) 1997 (l) UK: Scotland 6 unknown (nursing home) (19) 1997 (m) UK: Scotland 10 meat, cooked, butcher shop (20)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 39

1997 (n) UK: Scotland 5 1 unknown:farm animals? (16) 1997 (o) UK: Scotland 34 unknown; hospital (22) 1997 (p) US: WA 4 water, well (49) 1997 (q) US: MO 8 water, lake (49) 1998 (a) US: WI 69 0 1 cheese curds (131; 370) 1998 (b) US: WY 157 4 water, tap (266) 1998 (c) US: CA, NV 8 0 sprouts, alfalfa (244) 1998 (d) Canada 39 0 2 salami (358) 1998 (e) Canada 11 0 0 apple cider (334) 1998 (f) UK (Dorset) 10 water, drinking (26) 1998 (g) UK 7 0 0 cream, unpasteurized (24) 1998 (h) UK 1 1 cheese, cows’ unpasteurized (25) 1998 (i) US: WI 47 0 3 fruit salad (370) 1998 (j) US: WI 13 0 0 beef, ground (370) 1998 (k) US: GA 26 water park (49) 1998 (l) US: MN 5 water, lake (49) 1998 (m) US: IL 3 water, well (49) 1999 (a) US: NY (Wash. Co. fair) 127 2 14 water, well; 781cases E. coli and/or (258;260) C. jejuni 1999 (b) UK 114 3 milk, pasteurized (153) 1999 (c) Canada 143 0 6 salami (226) 1999 (d) US: CA, NV, AZ 13 0 3 beef, tacos (187) 1999 (e) US: CT 1 deer meat (288) 1999 (f) US: WA 37 0 3 water, lake (74;306) 1999 (g) Scotland 6 water, drinking (221) 1999 (h) UK: Wales 24 3 animals, farm (274) 1999 (i) Canada 7 0 0 beef, ground (225) 1999 (j) US: NY >1,000 2 11 water, well (97) 1999 (k) UK 3 0 0 cheese, cows’ unpasteurized (27) 1999 (l) UK 9 3 milk?; person-person (30) 1999 (m) UK 27 0 0 unknown (28) 1999 (n) UK 5 0 2 unknown-visit to Turkey (29) 1999 (o) UK 14 1 3 beach (31;164) 1999 (p) US: CA 7 0 0 water, lake (137) 1999 (q) Canada 159 animals, petting zoo (283) 1999 (r) Sweden 11 0 0 lettuce (352) 1999 (s) US: TX 22 water, drinking (215) 1999 (t) US: IL, KY, MO 329 beef (182) 2000 (a) US: WA 5 0 animals, farm (147) 2000 (b) Canada (Walkerton) 2,300 7 27 water,drinking (175) 2000 (c) UK: Scotland 20 environment (agr. showground) (174) 2000 (d) UK 6 0 1 milk, unpasteurized (2 outbrk) (35) 2000 (e) UK 9 0 1 foods, deli (36) 2000 (f) UK 15 0 3 unknown (36) 2000 (g) UK 7 person-person (33) 2000 (h) UK 45 0 0 foods? (prison) (34) 2000 (i) UK: Scotland 18 animals, farm ?; camp (32) 2000 (j) US: WI (Sizzler) 788 1 4 beef, unknown foods (370)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation 40 White Paper:

2000 (k) US: OH 48 0 2 water, drinking (294) 2000 (l) Netherlands 2 1 animals, petting zoo (166) 2000 (m) US: PA 51 0 8 animals, farm (113) 2000 (n) US: WI 15 0 3 unknown food (370) 2000 (o) US: WI 9 0 1 beef, ground (370) 2000 (p) US: ID 4 water, canal (215) 2000 (q) US: CA 5 water, stream (215) 2000 (r) US: UT 102 water, irrigation; also C. jejuni (215) 2000 (s) US: ID 15 water, spring (215) 2000 (t) US: CT 11 water, lake (215) 2001 (a) US: OH 23 1 environment (fair building) (345) 2001 (b) Canada 5 0 2 milk, goat, unpasteurized (239) 2001 (c) Canada 4 0 0 water, lake (75) 2001 (d) Austria 2 0 1 milk, unpasteurized (11) 2001 (e) UK 20 1 person-person (38) 2001 (f) UK: Scotland 15 water private well (37) 2001 (g) China 39 32 28 water, animals?( 91%>60yrs) (366) 2001 (h) UK 30 0 2 meat (butcher shop) (290) 2001 (i) US: MN 4 0 0 animals, farm (322) 2001 (j) US: WI 55 1 animals, fair (370) 2001 (k) US: WI 34 0 1 environment (stock pavilion) (370) 2001 (l) US: MN 20 water, lake (364) 2001 (m) US: SC 45 water, lake (364) 2001 (n) US: ME 9 water, wading pool (364) 2002 (a) US: CO, etc. (ConAgra) 28 5 beef, ground (319;346) 2002 (c) Canada 13 0 2 cheese, cows’ unpasteurized (172) 2002 (d) Canada 17 0 2 person-person;daycare (148) 2002 (e) Canada 109 2 0 salads (67) 2002 (f) UK, France 10 0 0 salad, cucumber (127) 2002 (g) UK: Wales 16 0 0 water, drinking (41) 2002 (h) UK 16 person-person (40) 2002 (i) UK 15 water, drinking (39) 2002 (j) US: WI 54 0 0 beef, ground (370) 2002 (k) US: WA 76 lettuce, romaine Ecoliblog 2002 (l) US: KY 2 water, well (66) 2002 (m) US: OR 60 12 environment, animal dust (fair) (57) 2003 (a) US: MN, CO 20 sprouts, alfalfa (138) 2003 (b) US 12 1 beef, steak, blade tenderized (212) 2003 (c) Canada 45 0 0 animals, petting zoo (115) 2003 (d) Canada 8 0 3 person-person (1) 2003 (e) Australia 3 0 0 foods? (106) 2003 (f) US: CA 32 2 3 spinach (nursing home) (294) 2003? Slovakia 9 3 milk, unpasteurized (222) 2003-4 Denmark 25 0 0 milk, pasteurized (188) 2004 (a) Japan: Okinawa 6 0 0 beef, ground (207) 2004 (b) France 3 2 cheese, goats’ (135) 2004 (c) Canada 34 0 2 beef, ground (330) 2004 (d) Canada 13 0 1 water park (152)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation Human Illness Caused by E. coli O157:H7 from Food and Non-food Sources 41

2004 (e) UK: Scotland 5 0 2 water, drinking (42) 2004 (f) Japan: Mie 23 1 person-person (329) 2004 (g) UK: England 7 0 0 water, stream (180) 2004 (h) US: NC 108 15 animals, petting zoo (116) 2005 (a) France 13 13 beef, ground (142) 2005 (b) UK: Wales 157 1 2 meat, cooked (305) 2005 (c) UK: Scotland; Turkey 15 0 0 unknown (323) 2005 (d) Sweden 120 0 7 lettuce (333) 2005 (e) Ireland 18 0 2 water, drinking; person-person (228) 2005 (f) UK 79 0 beef (44) 2005 (g) Japan: Kumamoto 11 0 0 beef, offal (231) 2005 (h) US: FL 73 0 12 animals, petting zoo (116) 2005 (i) US: AZ 2 animals, petting zoo (116) 2005 (j) US: WA 18 0 2 milk, unpasteurized (51) 2005 (k) Netherlands 32 0 0 beef, steak tartare (125) 2005 (l) US: MN, WI, OR 25 1 lettuce, packaged (Dole) (252) 2006 (a) UK: Scotland (Fife) 23 0 4 person-person? (nursery) (43) 2006 (b) UK: Scotland 3 0 0 meat (butcher shop,Lanarkshire) (43) 2006 (c) UK: England 4 1 4 unknown (nursery) (43) 2006 (d) UK: England 25 meat, cooked, butcher shop (45) 2006 (e) US: TN 10 3 person-person (daycare) (282) 2006 (f) US: MN 17 2 1 ground beef (173)

Corresponding author: M. Ellin Doyle, Ph.D., [email protected] August 2006 Food Research Institute, UW–Madison, www.wisc.edu/fri/ Funded in part by the American Meat Institute Foundation OCTOBER 2006 Volume 8 Issue 4 BOARD Nutrient Dense Lean Meats Key to a Healthy Rich Armstrong Sara Lee Food & Beverage Diet Says USDA Nutrition Policy Expert Van Ayvazian Boar’s Head Provision Co., Inc. Balancing Choices and Caloric Needs Essential To Fixing Nation’s Diet Richard Bond “Lean meats” fi t well into U.S. Department guidance business for over 100 years beginning in Tyson Foods, Inc. of Agriculture’s (USDA) recommendation to 1894 with food composition and dietary standards J. Patrick Boyle consume nutrient dense foods, said Eric Hentges, for the U.S. population. The current process of American Meat Institute Ph.D., executive director of USDA’s Center for federal nutrition guidance started with the 1980 Rod Brenneman Nutrition Policy and Promotion. Dietary Guidelines for Americans and the latest Seaboard Foods revision was issued January 2005. This current William Buckner Hentges, who has over 20 years of nutrition process is jointly administered by USDA and the Cargill Meat Solutions Corporation research and education experience, explains that Department of Health and Human Services with Philip Clemens lean meats are considered nutritionally dense – review and revision occurring every fi ve years. Clemens Family Corporation and thus key in dealing with issues of overweight Henry Davis Greater Omaha Packing Co., Inc. and obesity. In answering questions posed by The process by which the Guidelines are the AMI Foundation, Dr. Hentges talks about developed is designed not only to ensure scientifi c Tim Day Bar-S Foods Co. the USDA/Department of Health and Human accuracy, but also to give the public confi dence Services Dietary Guidelines, the nutritional and reassurance in the nutrition messages that Jeffrey Ettinger Hormel Foods Corporation guidance tool “MyPyramid,” and the importance are developed. The current scientifi c literature is of including meat in one’s diet. reviewed by an independent advisory committee Neil Genshaft Fresh Mark, Inc. of academics from leading universities and Q: Can you describe the role that USDA research institutions across the country. Members John Griffi n Griffi n Industries, Inc. plays in ensuring that U.S. consumers get of the Dietary Guidelines Advisory Committee sound nutritional guidance? are not paid by the government, but contribute James Hodges American Meat Institute their time and talent free of charge. All committee A: The USDA has been in the nutrition meetings, usually four or Carl Kuehne (continued on page 8) American Foods Group

Joseph Luter, III Smithfi eld Foods, Inc. Advanced Listeria Intervention and Control

Robert Manly Smithfi eld Foods, Inc. Conference Slated for Fall Michael McCain New breakout session topics, discussions on sanitation at Kraft Foods North America, Kraft Maple Leaf Foods Inc. advanced intervention techniques, analyzing a Foods, Inc. and John Butts, Ph.D., vice president, John Miller case study and a talk on successfully completing research and development for Land O’ Frost National Beef Packing Company, LLC a Food Safety and Inspection Service (FSIS) Inc., who will discuss building a fi rm foundation David Miniat food safety assessment are all scheduled for for Listeria control through sanitary equipment Ed Miniat, Inc. the revised Advanced Listeria Intervention and and facility design techniques. The “Back to Ed Nelson Control Conference, scheduled for Nov. 14-15, the Basics” track will conclude with talks on Indiana Packers Corporation 2006, at the Hyatt Regency Denver at Colorado sanitation best practices by Rory Redemann, food Larry Odom Odom’s Tennessee Pride Sausage, Inc. Convention Center in Denver, Colo. safety sanitarian at Kraft Foods North America, and Dale Fredell, manager for education services Stuart Prosser Cryovac Division, Sealed Air Conferees will be benefi t from a plenary at Ecolab, Inc. Corporation session led by Joe Meyer of Kraft Foods, Inc. on Robert Rebholtz the public health and regulatory implications of The second track on advanced intervention AB Foods, LLC Listeria control while producing safe ready-to-eat techniques is new to the 2006 conference. Several Rick Searer meat and poultry products universities have been invited to speak on recent Kraft Foods research in the areas of DNA tracking techniques, Dennis Vignieri Attendees can then choose between two new ingredient technologies for preventing Kenosha Beef International, Ltd. concurrent educational sessions. The fi rst track growth and alternative intervention techniques. Norbert Woodhams will focus on the basics for process control Pierre Foods, Inc. and will include talks by Joe Stout, director of (continued on page 5) AMI Foundation News Science Soundbites: A Review of Recent Research Hot Water Proves Better Than differences between the control samples invade the placenta, where – protected Lactic Acid Wash and the treated samples suggesting that from the body’s immune system the antimicrobial treatments did not – the bacteria is free to replicate before The application of hot water to reduce cause serious adverse sensory changes. spreading to infect other organs such as the levels of aerobic bacteria and Researchers believe these antimicrobial the liver and spleen. The study claims Enterobacteriaceae and the prevalence treatments can be a promising intervention to be the fi rst to trace the pathway of of Escherichia coli O157:H7 on available to ground beef processors who listeriosis infection. The Centers for preevisceration beef carcasses proved to currently have few interventions in their Disease Control and Prevention estimates be more successful than the use of, or in process. Journal of Food Protection, Vol. that 2,500 people fall seriously ill because addition to, lactic acid. Researchers with 69, No. 8, Pages 1802-1807. of listeriosis in the United States. Of those, the U.S. Department of Agriculture’s about one in three occurs in pregnant Roman L. Hruska U.S. Meat Animal CNS Tissue Detection on women. The study was published online Research Center found that a commercial Production Tools Successful in in the journal PLoS Pathogens in June. Researchers are now looking at how hot water carcass wash cabinet Study applying 74 C water for 5.5 seconds Listeria moves from the digestive tract reduced both aerobic plate counts and to the placenta in order to understand the The detection of central nervous system mechanisms of growth and create potential Enterobacteriaceae over one and a half (CNS) tissue on meat and carcass- times the reduction of a commercial prevention and treatment methods for splitting band saw blade surfaces was pregnant women. lactic acid spray cabinet that applied 2 successful with a glial fi brillary acidic percent lactic acid at approximately 42 protein fl uorescent enzyme-linked C. Additionally, the hot water treatments immunosorbent assay (GFAP F-ELISA) Study Shows Lm Count at Deli reduced E. coli O157:H7 prevalence in a study performed by researchers Counter Declining by 81 percent, while the lactic acid at Colorado State University. GFAP is treatment only reduced E. coli O157:H7 restricted to CNS tissue and has been used Less than one percent of all retail ready- prevalence by 35 percent. A combination as a marker for CNS tissue presence in to-eat meat and poultry luncheon meats of the two treatments produced only a meat products. The sample preparation, tested positive for Listeria monocytogenes 79 percent reduction. The results suggest extraction procedure and extraction in a study recently presented at the Annual that hot water would be more benefi cial temperature of the GFAP F-ELISA were Meeting of the International Association than lactic acid for decontamination of modifi ed to detect CNS tissue on meat for Food Protection in Calgary, Alberta. preevisceration beef carcasses. Journal surfaces and on carcass-splitting band Researchers from four universities tested of Food Protection, Vol. 69, No. 8, Pages saw blades. The process developed 8000 samples in four states (Georgia, 1808-1813. consistently provided detection of GFAP California, Minnesota and Tennessee) on meat surfaces and on carcass-splitting and found that, overall, 0.69 percent of Organic Acids, Acidifi ed Sodium band saw blades and recovery was not products tested had a positive result for Chlorite Control Bacteria On affected by storing samples overnight. Listeria monocytogenes. Deli products that Beef Trim, Ground Beef The study demonstrates the effectiveness were pre-sliced had a prevalence rate of of modifi ed sampling procedures and 0.15 percent, while those deli items that were sliced in the retail store were slightly Acidifi ed sodium chlorite and acetic and preparations, sample extraction buffer pH higher at 1.23 percent. This was reported lactic acids were effective in reducing and extraction temperatures. Additionally, to be very encouraging news when foodborne pathogens in beef trim prior the modifi cations introduced to the compared to a similar study conducted in to grinding in a simulated processing original F-ELISA sampling protocol 2003. When these data from 2006 were environment, a recent study showed. resulted in a sensitive and repeatable assay compared with the previous study, the Researchers at Texas Tech University for detection of CNS tissue. Journal of prevalence rates were reduced by more treated beef trim and ground beef with Food Protection, Vol. 69, No. 8, Pages than 50 percent for these two deli product a high and low inoculation and found 1966-1970. categories. that all treatments reduced the pathogens Salmonella Typhimurium and Escherichia Placenta May Be Culprit in coli O157:H7 to non-detectable numbers. Listeriosis Illnesses Materials treated with the high inoculation showed signifi cant reductions. Organic Scientists at the University of California, acids prove to be more effective in Berkeley, have challenged some long-held ground beef immediately after grinding, assumptions regarding listeriosis illness but after one day of storage, there were in pregnant women with a recent study. no differences between the treatments. The research, conducted in guinea pigs, A consumer panel found few sensory showed that Listeria monocytogenes can

Page 2 | October 2006 AMI Foundation News ‘Super-Shedders’ Responsible for Increasing E. coli Prevalence Cattle with high levels of E. coli O157:H7 play a key role in the course of sampling. increasing the prevalence of the bacteria throughout the feedlot, according to a recently released study, jointly funded by the AMI Researchers believe that interventions to identify super- Foundation and the National Cattlemen’s Beef Association. shedders and prevent their entry into pens or dissemination among other cattle are necessary; however, they did not identify The study, “The Role of Super-Shedders in Determining where this is best to occur, i.e. on entry into the feedlot, at Feedlot Pen Prevalence of E. coli O157:H7,” analyzed saleyards, during transport, the removal and addition of “super-shedders” and the during lairage, or at some transmission of E. coli O157:H7 between cattle. It also data derived from this combination of these points. looked at the association between the presence of a super- “study suggests a casual shedder in a pen and the prevalence of E. coli O157:H7 in the Moreover, the study pen environment. “Super-shedders” are cattle within feedlots link between the presence indicates that mixing cattle

that are colonized by E. coli O157:H7 more frequently, of a super-shedder and from different pens or herds persistently and in greater numbers than other cattle. has the potential to allow the prevalence of E. coli “ cohabitation of super-shedders The project, completed by Washington State University, O157:H7 in a cattle pen and low-shedding cattle, confi rmed the role of the recto-anal junction in the general which could result in the latter cattle population and in super-shedders as the main becoming more likely to carry colonization point for E. coli O157:H7. and shed E. coli O157:H7 and thus making them a greater potential as a source of the bacteria. In addition, data derived from this study suggests a casual link between the presence of a super-shedder and the prevalence An improved knowledge of how E. coli O157:H7 reside at of E. coli O157:H7 in a cattle pen and among cattle that reside the recto-anal junction and become excreted in feces will allow near by. development of interventions designed to reduce E. coli O157:H7 carriage and transmission to other cattle and the food chain. Removal of super-shedders from hot pens – pens where super-shedders reside – resulted in reductions in E. coli Rowland Cobbold, Ph.D., the principle investigator, now O157:H7 parameters for cattle colonization and environmental with the University of Queensland, will present these fi ndings contamination; however, these reductions were no greater than at the 6th International Symposium on Shiga-Toxin Producing those seen in control pens, where general reductions in E. coli Escherchia coli Infections in Melbourne, Australia. To read the O157:H7 prevalence, count and persistence also occurred over entire report online, visit AMIF.org and navigate to “Research.”

Listeria Intervention Research Shows Promise

The combination of pre- and post-package pasteurization pasteurization of large RTE processed meats. demonstrated reduced levels of Listeria monocytogenes (Lm) on ready-to-eat (RTE) meat and poultry products in a recent study by The hot dog pasteurization efforts included three Oklahoma State University researchers. components: fi rst, to determine the effectiveness of a hot water deluge system; and second, organic acid sprays and pre- and Lead investigators Peter Muriana and J. Roy Escoubas tested post-package for reduction of Lm to determine if equivalent pre- and post-package pasteurization methods of RTE meats reductions can be obtained with less processing time or together with anti-listerial ingredients to validate the utility of temperatures. Finally, researches aimed to determine the impact surface pasteurization as a tool for reducing Lm on hot dogs and of antimicrobials and organic acids on outgrowth of Lm during deli turkey. the product’s shelf life.

Results demonstrated that liquid smoke-treated samples had The large RTE processed meats pre- and post-package a slight reduction of Lm during outgrowth. Liquid smoke or lactic pasteurization efforts included two components - the fi rst to acid treatments, when combined with heat, suppressed outgrowth examine the effectiveness of the combination of pre- and post- for up to four weeks on hot dogs. The combination of pre- and package pasteurization of RTE deli turkey and the second to post-package processing proved to be very effi cient; however, examine the combination of pre- or post-package pasteurization none of the treatments were shown to be effective on uncured of RTE deli meats in combination with antimicrobials and impact turkey products. on outgrowth of Lm.

The two-part project focused fi rst on pre-package To read the entire report online, visit AMIF.org and click on pasteurization of hot dogs and later on pre- and post-package “Research.”

October 2006 | Page 3 AMI Foundation News White Paper Traces E. coli O157:H7 Roots A wide variety of government and scientifi c documents were priorities, are less aggressive and rigorous in testing reviewed as part of the recently released “White Paper on Human samples and conducting investigations. They should be Illness Caused by E. coli O157:H7 from Food and Non-Food encouraged to improve funding for public health and to Sources.” adopt and incorporate better epidemiological procedures and improved laboratory methods. Researchers at the Food Research Institute at the University • Federal grants could be targeted to improving laboratory of Wisconsin-Madison teamed up with the Wisconsin Division facilities and training for state epidemiologists, as needed. of Public Health’s Bureau of Communicable Diseases and Greater participation by all states will aid in the rapid Preparedness, Communicable Disease Epidemiological Section to identifi cation of multi-state outbreaks. examine scientifi c literature databases, industry publications and government publications and regulations from the United States, The report also points to improvements in recent years that Canada, Europe and Japan from the date that E. coli O157:H7 have led to a decline in outbreaks and cases of E. coli O157:H7. was fi rst recognized as a possible human pathogen in 1975 up These include: until July 2006. • Increased emphasis on HACCP and pathogen control by industry and government. Included in the search was information on the emergence • Increased educational efforts targeted at doctors, workers of the organism as a human pathogen, surveillance and outbreak and consumers. reports, epidemiological studies, important government • Increased cooperation and information sharing between regulations and industry initiatives to control E. coli federal and state agencies with the Centers for Disease O157:H7. From this, the authors of the paper generated an Control and Prevention and the Food and Drug historical timeline summarizing major events occurring from the Administration. fi rst recognized cases of illness cause by E. coli O157:H7 through • Improved training to make Food Safety and Inspection July 2006. Service inspectors more knowledgeable about risk-based inspection. Based on the information collected, researchers made the • Litigation surrounding some high profi le outbreaks of following two recommendations: foodborne disease alerting consumers to important food • Greater uniformity is needed in statewide investigations safety issues. and reporting of foodborne illness. Public health systems in some states are well organized – with all samples sent To read the entire report online, visit AMIF.org and navigate to state labs for analysis. Other states, with different to “Research.”

USDA Conference Focuses on Reaching At-Risk Populations Innovative information and research about reaching at-risk policy – including public health professionals, food safety and other populations with information on foodborne illness was educators, food industry professionals and scientifi c writers the theme of the 2006 Food Safety Education Conference. The and journalists - to work together to create positive behavior event, sponsored by the U.S. Department of Agriculture’s Food modifi cation during food preparation and consumption. Safety Inspection and Service (FSIS), was held Sept. 27-29 in Denver, Colo. The conference drew international, federal, state The conference focused on fi ve main themes, including: and local experts who discussed options and detailed ideas about • Foodborne Illness Surveillance and Epidemiological reaching targeted, at-risk populations with educational materials Insights; and preventative strategies. • Food Safety Behavioral and Attitudinal Research; • Social Marketing, Educational Interventions, and Expert presenters stressed that while foodborne illness Program Research; can strike anyone in the general population, some individuals • Role of Foodservice and Food Industries; and are more susceptible – including pregnant women, young • New Technologies. children, older adults and those with weakened immune systems – who make up the at-risk population. The primary goal of the In addition to FSIS, the conference was sponsored by NSF conference was to share current data on foodborne illness, present International, the Food and Drug Administration, the Centers strategies to enhance food safety knowledge and effect attitudinal for Disease Control and Prevention, USDA’s Cooperative State and behavioral modifi cation in general and at-risk populations Research, Education and Extension Service, USDA’s Food and communicate the latest science-based safe handling practices. and Nutrition Service and NSF/World Health Organization Collaborating Center for Food Safety. In addition, the conference emphasized the need for all To view presentations and information from the conference, groups with a stake in communicating and forming public health go to http://www.fsis.usda.gov/denver2006/.

Page 4 | October 2006 AMI Foundation News

CO2 System Useful in Reducing Bacteria in Ground Beef Bacteria counts in ground beef can be reduced with the use The highest lethality in challenged beef trimmings and of a controlled phase carbon dioxide (CPCO2) system, researchers ground beef occurred with 1500 psi CPCO2 for 15 minutes. Total at the Kansas State University Food Science Institute concluded Plate Count, generic E. coli, E. coli O157:H7 and Salmonella spp. in a recent AMI Foundation-funded study, “Elimination of reached 0.83, 0.96 , 1.00 and 1.06 log reductions, respectively. Escherichia coli O157:H7, Generic Escherichia coli, and Ground beef and beef trimmings saw similar bacterial reductions. Salmonella Spp. on Beef Trimmings Prior to Grinding Using a

Controlled Phase Carbon Dioxide (CPCO2) System.” Food processing applications of CPCO2 are becoming more popular as economically viable alternatives to heat treatments.

The experiment was designed to evaluate antimicrobial, The most potential for CPCO2 application is for products that are quality and shelf life effects of CPCO2 beef trimmings destined sensitive to heat and pressure, such as fresh produce, fruit juice for ground beef. The most signifi cant results demonstrated that and beverages, fresh and smoked fi sh and fresh meats. using CPCO2 on beef trimmings before grinding can reduce E. coli, coliforms, Salmonella spp. and aerobic bacteria in ground beef. The team of KSU researchers, including Curtis Kastner,

Additionally, the use of CPCO2 had minimal effects on color or Ph.D., James L. Marsden, Ph.D., Daniel Y.C. Fung, Ph.D. and odor characteristics of ground beef. Carlos Arturo Tanus, believe this technology should not only improve the microbial safety of ground beef, but also promote The investigation was divided into four separate studies. The extended ground beef color shelf life.

fi rst evaluated the effect of CPCO2 as an antimicrobial on sterile fi lter papers challenged with various foodborne pathogens. The “This research demonstrates a promising food safety applica- second study assessed the effects of CPCO2 on non-challenged tion for beef trimmings and ground beef,” said AMI Foundation and challenged beef trimmings. The third phase of the study President James H. Hodges. “Additional research will help to confronted concerns raised during the previous phase by further validate and refi ne the process.” analyzing the quality of ground beef treated with CPCO2. The

fi nal stage evaluated the effects of CPCO2 as an antimicrobial To read the entire report online, visit AMIF.org and navigate intervention process for beef trimmings destined for grinding. to “Research.”

Second Quarter Salmonella Risk-Based Sampling Closes Testing Data Released Out Listeria Conference

The Food Safety and Inspection Service’s (FSIS) released (continued from page 1) its second report for the revised Salmonella Verifi cation Testing Some highlights for the afternoon include a breakout session Program. on verifi cation and validation, led by Randy Huffman, Ph.D., vice president of scientifi c affairs at the AMI Foundation, a The new procedures focus inspection and laboratory discussion of data analysis by John Butts, Ph.D., and a session on resources on establishments that have had the most samples test developing a routine sampling plan for process control with Tim positive for Salmonella and the most samples with serotypes most Freier, Ph.D., corporate microbiologist with Cargill, Inc. frequently associated with human salmonellosis cases. The second day of the conference will kick off with a case The agency also began testing turkey carcasses using study on root cause identifi cation by John Weisgerber, director “baseline guidance results” that were originally published in the of corporate safety with Weisgerber Consulting, LLC. Bob Federal Register in February of 2005 and based on results from Reinhard, director of food safety at Sara Lee Foods, will discuss the turkey carcass sponge baseline study conducted from July successfully completing a routine Lm risk-based sampling 1997 through June 1998. and food safety assessment by FSIS. In between, sessions on investigation and corrective actions, lot and line segregation and Because of this restructuring, comparisons of results for this a breakout session on best practices are sure to keep attendees year to previous years will be inappropriate. Until a solid body interested and focused. of data is established, FSIS will use the results of an upcoming nationwide baseline study to provide valid estimates of the Registration can be completed online in the Events/ prevalence of certain pathogens of public health concern and Education section of MeatAMI.com. AMI members receive the permit valid statistical comparisons to be made over time. special conference rate of $595. Three or more member company attendees receive a discounted rate of $495. All other attendees A 12-month Young Chicken (Broiler) Baseline Study is may register for the regular rate of $695. scheduled to begin shortly and additional baseline studies are under development. For additional conference information, visit MeatAMI.com.

October 2006 | Page 5 AMI Foundation News Bacteria Approved For Use As Anti-Listerial Agent

A food additive petition to use bacteriophage as an The FDA evaluated the safety of the additive after reviewing antimicrobial agent against Listeria monocytogenes (Lm) on 1) the safety of the six phages constituting the food additive, 2) ready-to-eat meat (RTE) and poultry products has been approved the safety of potential residues from Lm used in the manufacture by the Food and Drug Administration (FDA). of the food additive and the need for limits related to their levels, 3) whether undesirable genes are potentially carried by the Intralytix, Inc. uses a mixture of equal proportions of six food additive, and 4) the need for additional identity and safety individually purifi ed phages that is sprayed directly on the surface specifi cations. of the RTE food product at a level of approximately 1 milliliter per 500 square centimeters of the food surface area just prior The petition to use phages was originally fi led in July 2002 to packaging. The formulation consists of multiple phages to for use on foods, including fresh meat, meat products, fresh minimize the possibility of Lm developing a resistance to the poultry and poultry products; however, the company amended treatment. the petition for use only on RTE products. The fi nal rule was published in the Aug. 18, 2006 Federal Register. The process was ruled safe by the FDA as the phages used infect only bacteria, and not mammalian or plant cells. The Food Safety and Inspection Service (FSIS) has also Additionally, humans are routinely and naturally exposed to reviewed the technology and has determined that it was effective phages at high levels through food, water and the environment in controlling Lm on meat and poultry and that it would not without adverse effect. Phages are also a part of the normal mislead consumers. FSIS will require products treated with the microbial population of the human digestive system. phage preparation to contain a label declaration on the package and in the ingredient statement.

Canadian Meat Council Symposium Explores Keys to Maintaining Meat Safety, Quality

The latest developments in food preservation and safety, for the specifi c pathogen that we are looking for. The problem including extending the shelf-life of meat while maintaining with these methods is that they’re slow because they require the optimum quality, were the focus the Canadian Meat Council’s organism to grow,” he said in his presentation. (CMC) fi rst-ever symposium on “Advanced and Rapid Methods in Quality Control of Meat and Meat Products.” The two-day Rick Holley, Ph.D., a professor in the department of food event, which attracted more than 90 participants and 12 supplier science at the University of Manitoba, outlined how scientists sponsors from throughout North America, was held Aug. 10-11, are currently investigating the various causes of meat and meat 2006 in Calgary, Alberta. product spoilage and methods of control. He noted that over the last 15 years, “we’ve seen shelf life extensions from 30 days The conference’s keynote speaker, Daniel Y.C. Fung, with pork to 55 days, largely as a result of improvements in Ph.D. of Kansas State University, discussed the last 25 years of plant sanitation, temperature control as well as improvements in developments in rapid microbiological methods and offered some packaging materials and also packaging approaches using some predictions for the future. Conferees discussed the remarkable of those materials.” advances the North American meat industry has made over the years. “The methods we have today have never been better,” said Holley stated that while excellent technology exists for CMC Technical Services Director Parthi Muthukumarasamy. keeping meat fresh over a lengthy period of time, temperature is “Over the years it has evolved and now we are talking a matter the key to success. “The best possible opportunity for attaining of 10 to 15 minutes [to detect and identify pathogens] compared shelf life is keeping the temperature as low as possible without to days for culture tests so we are really fast and most of the freezing the meat. What that means is using temperatures that are methods are accurate.” in the vicinity of about minus 1.5 C, plus or minus a half degree. That type of temperature control is expensive and requires Presentations covering food safety, laboratory methods for equipment that’s very precise so you don’t freeze the meat. As sampling and testing meat products, and the interpretation of soon as you freeze the meat you lose quality.” collected data were all examined in detail on the fi rst day of the conference. On the second day, Mansel Griffi ths, Ph.D., the “This conference was an excellent avenue for the exchange director of the Canadian Research Institute for Food Safety at of new ideas and different approaches, and benefi ted the entire the University of Guelph and senior industrial research chair in North American meat industry,” said Randall Huffman, Ph.D., dairy microbiology, outlined Molecular Methods for Detection of vice president of scientifi c affairs for the AMI Foundation, who Foodborne Pathogens. “Currently we use culture-based methods. attended the event. That is, we rely on growth of the organism on media that selects

Page 6 | October 2006 AMI Foundation News 2006 MIRC Focused on Industry Research Needs The annual Meat Industry Research Conference, co- by Bucky Gwartney, director of research and technical services sponsored by the American Meat Institute Foundation (AMIF) for the National Cattleman’s Beef Association. The session and the American Meat Science Association (AMSA), held Oct. also explored future trends in nutritional and quality traits that 4-5 at the Westin Diplomat Resort & Spa in Hollywood, Fla. was consumers value. driven by attendee contributions toward fi ve year meat science research goals. Additional details are available online now at General Session: Research Priorities 2007 and Beyond AMIF.org with a complete conference summary in the January Facilitators and attendees participated in an exciting session 2006 AMIF Newsletter. to develop long-term research goals.

Wednesday, October 4, 2006 Thursday, October 5, 2006

General Session: Hot Topic: Low Oxygen CO Packaging Update Keynote Address Randy Huffman, Ph.D., vice president of scientifi c affairs To welcome attendees and set the tone of the event, for the AMI Foundation and Janet Riley, senior vice president of Associate Dean at the University of Florida’s Institute of Food public affairs and professional development for AMI discussed and Agricultural Sciences, Douglas Archer, Ph.D., discussed the the benefi ts of modifi ed atmosphere packaging for processors, challenges academia has the opportunity to solve. retailers and consumers and the source of all the questions being raised by the press concerning the use of carbon monoxide in Development of a Long Range Plan for these systems. Meat Science Research – Setting the Stage While several key government, academia and industry Bacteriophage: Leveraging Nature To Treat Bacteria groups set research priorities for meat science, the discipline itself Developing one of the fi rst bactericides based on phage lacks a cohesive and coordinated long-term plan. Conference technology, Justin Reber, President & CEO of OmniLytics, Inc. Chairman and Pennsylvania State University Professor, Edward demonstrated the potential impact for phage technology as an Mills, Ph.D., initiated the day-long sessions to being the process application for various food processing operations. of prioritizing meat science research needs for the next fi ve years. Fight for the Cure: Sodium Nitrite Status Report Research Priority A: Product Quality New research continues to affi rm nitrite’s safety and University Distinguished Professor and Monfort Chair at benefi ts. Nevertheless, nitrite continues to face challenges. Colorado State University, Gary Smith, Ph.D., lead the review James Coughlin, Ph.D., president of Coughlin & Associates and of current research related to meat product quality. The session Andrew Milkowski, Ph.D., adjunct professor at the University included a discussion of practical genetic tools to predict quality of Wisconsin – Madison presented information on the true state and biochemical events to time and temperature interactions of science behind nitrites and the latest from the international during processing on muscle quality parameters. scientifi c arena.

Research Priority B: Food Safety Cloning: Vision for the Future Dane Bernard, vice president of food safety and quality Cloning holds promise as a means to maximize the potential assurance at Keystone Foods led the discussion on the current of superior genetics within livestock herds. Mark Walton, state of food safety in the U.S. meat supply and how meat safety president of ViaGen, discussed his perspective on cloning research has led to recent changes throughout the industry. Topics technology and the expected implications on relevant national included emerging pathogens, production unit-specifi c microbial and international policy issues. loads and ecology, pathogen intervention strategies and validation studies. Bernard also led the conversation on what tools the USDA Perspective on Natural Claims industry needs to meet the potential food safety challenges of the Unlike organic meat products, natural meats are not yet next fi ve years. offi cially defi ned. Martin O’Connor, branch chief for the U.S. Department of Agriculture’s Agricultural Marketing Service, Research Priority C: Processing and Packaging discussed the latest concerning the defi nition and the possible Iowa State University Professor of Animal Science, Food implications for product labeling. Science and Human Nutrition Joseph Sebranek, Ph.D. led this session on the current state of technology in processing and packaging and how new and innovative methods are being developed to meet changing consumer needs.

Research Priority D: Consumer Needs A discussion of how the meat science research agenda should be targeted based upon recent consumer research was led

October 2006 | Page 7 AMI Foundation News Hentges: ‘Balanced Consumption Across All Food Groups’ Key (continued from page 1) fi ve, are open to the public, including average for all government Web sites. When consumers were the press. All of the meetings are asked their reason for coming to the Web site, the number announced in the Federal Register and all of the proceedings are one reason was “change diet/eat healthier,” followed by “fi nd posted on government websites. The public has more than one nutrition information” and “lose weight.” Additionally we know opportunity to provide comments. We make every effort to make that depending on the time of year 18-36 percent of the visitors the development of the guidelines open and transparent. are students, which is a very important audience in terms of education for long-term behavior changes. Q: USDA recommends two to three servings of meat per person, per day. Can you expand on the role meat and poultry Q: Can you offer some insight into the steps the meat industry products play as a component of a healthy, balanced diet? has taken recently to improve both production effi ciency and respond to consumers’ health concerns about dietary fat. A: The new food patterns of the USDA Food Guide have What is the signifi cance of these changes for consumers? dropped the term “servings” and now use measurements such as cups and ounces (or ounce equivalent). The basic food group of A: Dietary fat and the type of fatty acids consumed was a meat and beans includes meat, poultry, fi sh, eggs, dry beans and growing health concern in the ‘70s. The most prominent of these nuts. Traditionally we think of this food group as providing high issues was the consumption of saturated fat and cholesterol and quality protein to the diet, but additionally the group is a good the relation to heart disease. Both the producer and the process source of iron, zinc, and B-vitamins. However, for good health segments of the industry took this very seriously and set out and weight maintenance it is important to more often choose lean, to reduce the fat content of their products. Through practice low fat or skinless products from this food group. in breeding, feeding and food technology, they were able to accomplish their fat reduction goal such that the USDA’s nutrient Q: Frequently we hear the concept that meat is considered composition databases were revised in the late 80’s and early 90’s a “nutrition dense” food source. Do you agree with this for beef, pork, lamb and veal. These efforts have continued and assertion and if so, what factors are used to make this more revision to these databases have been made in the past few determination? years. The end result is that consumers are seeing much less fat in the meat case today and have more lean or low-fat choices in A: One of the key recommendations of the Dietary their meat products. Guidelines is to “consume a variety of nutrient-dense foods and beverages within and among the basic food groups.” The Q: There are occasionally reports in the media about consumer brochure that accompanies the guidelines policy epidemiological studies that try to link meat consumption document advises, “Get the most nutrition out of your calories.” with adverse health effects. Can you put those studies in The nutrient density concept is very important as we seek to perspective? reverse the current trends of overweight and obesity in America. Lean meat fi ts into this concept in that three ounces may average A: Epidemiological studies are useful for identifying ten percent of your daily caloric needs, but provides more than potential trends between groups or among populations. They can ten percent of the daily requirement for protein, iron, zinc, B-12 also point to potential relationships between factors that were and several other nutrients. not previously observed. However, they do not determine cause and effect in the same way as human clinical trials. Due to the Q: What should people know about the 2005 Dietary expense of randomized human clinical trials, few trials have been Guidelines and MyPyramid guidance? Has USDA attempted conducted examining meat consumption and health outcome to measure the awareness and use of MyPyramid? and those trials have also been smaller in scale. However, these few trials generally have not detected adverse affects of meat A: The Dietary Guidelines are the science based federal consumed in appropriate portions as part of a balanced diet. The nutrition recommendations and MyPyramid interprets these Dietary Guidelines again point to the importance of balanced recommendations into actionable implementation for consumers. consumption across the food groups, while making choices that The original Food Guide Pyramid was released in 1992 and the limit the intake of saturated and trans-fat, added sugars, salt and new MyPyramid and its website, MyPyramid.gov, were released alcohol. in April 2005. To date, we have had over 2 billion hits to our Web site, which is enormous. We were told that if we received Q: Food in the U.S. is both abundant and inexpensive, yet 74 5 million hits on the day of the launch, it would be a red-letter percent of the U.S. population has a diet that USDA classifi es day. MyPyramid received 50 million hits browning out the entire as needing improvement. What is going wrong? system. The response to the interactive education tools has been overwhelming. Now we have stabilized at about 25 million hits a A: USDA has done a good job over the years of reaching week. its goal of a safe, abundant, and affordable food supply. The We have implemented an online customer satisfaction two biggest issues for us now are balancing our choices survey. Our overall satisfaction rating is 17 percent above the among all the food groups and doing so (continued on page 9)

Page 8 | October 2006 AMI Foundation News Ongoing AMIF Research – E. coli O157:H7 Investigator Institution Project Title John Scanga, Colorado State Use of Warm (55°C) 2.5% or 5.0% Lactic Acid for: (A) Reducing Microbial Counts on J.N. Sofos, University Beef Subprimal Cuts and Beef Trimmings Following Fabrication, and (B) Reducing K.E. Belk, Incidence of E. coli O157:H7 in Combo-Bins of Beef Trimmings and Inside (in the G.C. Smith interior) Beef Cuts Subjected to Blade/ Needle or Moisture-Enhancement Tenderization Ongoing AMIF Research – Listeria monocytogenes Investigator Institution Project Title Charles Carpenter, Utah State Anti-Listeria Action of Levulinate Jeff Broadbent University Mary Alice Smith, University Refi nement of Listeria monocytogenes (L. monocytogenes) Low Dose Data from Pregnant Joseph Frank of Georgia Guinea Pigs for Human Risk Assessment Kathy Glass, University of Minimum Nitrite Levels Required to Control Listeria monocytogenes on Ready-to-Eat James Claus Wisconsin Meat and Poultry Products

Ongoing AMIF Research – Targeted Research Investigator Institution Project Title Mindy Brashears, Mark Miller, Chance Texas Tech University, Risk Factors and Consequences Associated With Brooks, John Blanton, and Christine West Texas A&M Condensation in Fresh and Ready-to-Eat Processing Alvarado, Guy Loneragan University Facilities Bradley Marks, Michigan State Determine the Likelihood that Salmonella Develops Alicia Orta-Ramirez, Alden Booren, University Heat Resistance during Thermal Processing of Elliot Ryser Commercial, Whole-Muscle, Ready-to-Eat Meat Products Catherine Cutter, Pennsylvania State Determination of the Effi cacy of Chlorine Dioxide as an Ed Mills University anti-Listerial Agent in RTE Brine Chilling Solutions Charles Kaspar, University of White Paper on Human Illness Caused by E. coli Ellin Doyle, Wisconsin-Madison O157:H7 from Food and Non-Food Sources Ronald Weiss To view status reports for these projects, visit www.amif.org. Hentges: Importance of Physical Activity Cannot Be Forgotten (continued from page 8) without exceeding our caloric needs. research much the same as the medical community has employed MyPyramid education messages and for years. Use of an evidence-based system further assures the interactive tools seek to correct this imbalanced food selection the public, the policy makers, and scientists that the quality of by providing more personalized food patterns and allowing nutrition research is of the highest caliber. individuals to monitor and track their behavior. It has been said that as a nation we are overfed and undernourished. The concept Q: Do you have any fi nal thoughts concerning the nutritional of nutrient density discussed above is again very important in health of the U.S. population in relation to other countries? achieving this balance within our caloric budget. A: There are great commonalities between the health of Q: How does USDA evaluate the full breadth of both the U.S. population and other developed nations. This should negative and positive studies that are published addressing be expected as our global economy and communication make diet and health, and eventually reach a sound, science- our food supply and cultural behaviors more similar. I recently based conclusion on a given issue? participated in a European Union/United States conference looking at action on diet, physical activity and health. The A: As we all know, it seems that there is a new study or concept here is that we both are experiencing increases in research report featured in the press every day. Many of these overweight and obesity and that together we may identify featured studies are not ready for prime time reporting and it common “good practices” that can reverse this trend. We have is the body of work on a subject over time that is important mainly discussed diet and nutrition above, but we should not rather than an individual study. Obviously, not all research is of forget the importance of physical activity in obtaining energy equal quality or of equal strength for making nutrition policy. balance within our diets. New emphasis was placed on this issue The advisory committee for the 2005 Dietary Guidelines used a in the 2005 Dietary Guidelines and it was a new concept addition systematic, evidence-based review for analysis of the scientifi c to MyPyramid.

October 2006 | Page 9 AMI Foundation News Calendar of Events For additional information on any of these upcoming events, or to register, please visit our website at MeatAMI.com and navigate to Events/Education or contact Anne Nuttall at 202/587-4241 or [email protected].

Advanced Listeria Intervention & Control Annual Animal Care and Handling Conference Conference When: Mar. 28 – 30, 2007 When: Nov. 14 – 15, 2006 Where: Sheraton Overland Park, Overland Park, Kan. Where: Hyatt Regency Denver at Colorado Convention Center, What: Featuring the leading academic experts in the fi eld, this Denver, Colo. conference provides a wealth of information on the What: AMIF is pleased to present the new Advanced Listeria latest trends and ideas for implementing change and Intervention and Control Workshop. This highly rated improvement to animal care at the plant level. educational opportunity is designed to help manufacturers of ready-to-eat (RTE) meat and meat products examine Worker Safety, Health and Human Resources the issues surrounding testing and to provide experience Conference in developing appropriate sanitation standards and procedures for processing RTE products. In addition When: Apr. 3 – 4, 2007 to assuring optimal product safety, implementing best Where: Hyatt Regency Denver at Colorado Convention Center, practices for RTE processing offers a key benefi t: helping Denver, Colo. to assure compliance. Note: Registration is limited to 60 What: Leading experts provide attendees to this conference participants. authoritative and practical instruction for improving working conditions and employee relations. The 2007 Annual Meat Conference conference also features the AMI/National Safety Council Worker Safety Awards Program. When: Feb. 18 – 20, 2007 Where: Caribe Royale All-Suites Resort and Convention Center, Orlando, Fla. What: The Annual Meat Conference is the premier educational AMIF Request For event for retailers of meat and poultry products. Conference programming examines the hottest trends Proposals Closed from ethnic marketing to fl avor innovation, details pressing public policy issues in areas such as nutrition The AMI Foundation’s 2006 Request For Proposals closed and labeling and offers training in key areas such as on July 28, 2006. The pre-proposals are currently under crisis management and media relations. Attendees review by the AMIF Research Advisory Committee. Funding also sample hundreds of meat and poultry products recommendations will be made to the AMI Executive at the conference’s most popular event: the Product Committee at their January 2007 meeting. For additional Tasting Reception. A special Tech Fair Luncheon offers information about the process, visit amif.org or contact Susan exhibits of new technologies of interest to retailers Backus at [email protected]. and processors. The conference also provides ample networking opportunities to gather new ideas – and new customers. On the web: www.meatconference.com

AMI Foundation Contacts All AMI Foundation staff can be reached at: 1150 Connecticut Avenue, NW, 12th Floor, Washington, D.C. 20036 202/587-4200 or at the email address below. James Hodges Randall Huffman, Ph.D. president vice president, scientifi c affairs [email protected] [email protected]

Janet Riley David Ray senior vice president, public affairs & professional development vice president, public affairs [email protected] [email protected]

Susan Backus Patrick Wilson director of research manager, public affairs & online editor [email protected] [email protected]

Page 10 - October 2006 Fr: Tim Freier, Cargill T: Dane Bernard, Tracie Sheehan, Randy Huffman

I know that time is going to be very tight at the meeting, but I am very interested in getting in front of the following:

Deli industry gets wake up call from Salmonella lawsuit: Instant food poison test strips key in helping to preventing future cases 13.sep.06 PRWeb Press Release Orlando Florida–- In an open letter to the International Dairy Deli Bakery Association, Magna Medical General Manager Robert Greene addresses the importance of empowering deli managers with the necessary tools to “spot check” food product and machinery to prevent cross contamination. In a recent lawsuit filed last Thursday in Johnson Superior Court (Indiana), it was food from the deli department of a major national chain store that had been implicated as the alleged source of outbreak of Salmonella food poisoning. Salmonella bacteria, found in meat, and animal waste, particularly poultry, causes food poisoning in humans. Proper handling and cooking procedures prevent harmful infection, however, cross contamination can occur. The MMS Salmonella strip can detect 50 of the most common and deadliest strains. The strips are submerged in food samples, if the organism is present the strip will change color in less than 20 minutes. Greene feels that every produce, meat, fish, deli, and dairy manager should have one bottle of instant Salmonella and instant E.coli test strips to ensure food integrity for their departments. “Manager’s need to be empowered to go with their gut feeling” says Greene. “A $3.25 test strip could save them over $75,000 in lawsuits.” Escherichia coli (usually abbreviated to E. coli) is one of the many species of bacteria that is naturally found in the lower intestines of warm-blooded animals. “There are literally thousands of varieties of E.Coli, and the most common and deadliest strains have been included such as E.coli 103 and 0157:H7 for detection in the instant E.Coli strip. Lawsuit case number in reference is Johnson Superior Court (Indiana) Case number 41D030609CT00062. Magna Medical Services, Inc, a trusted provider of drug testing and screening products who recently expanded to include tests for safety, health, and sanitation, has been chosen to manage the sales and distribution worldwide.

Perhaps a brief discussion - has anyone evaluated these rapid strips - should we commission a quick lab study by an independent lab to verify these claims? I have a feeling these strips are going to lead to big problems - false positives, misinterpreted results - might be good to debunk them sooner than later. Update of MDR-Ground Beef Sample Processing

September 2006 Number of samples collected each month by region*

Region: 1 2 3 4 5 6 7 8 combined 2005 91 60 133 0 0 0 74 133 491 Jan-06 12 0 23 0 0 0 0 25 60 Feb-06 16 0 20 0 0 0 0 22 58 Mar-06 19 0 28 0 0 0 0 26 73 Apr-06 20 0 28 0 0 0 0 26 74 May-06 19 0 23 0 0 0 0 24 66 Jun-06 22 0 24 0 0 0 0 25 71 Jul-06 44 29 57 0 67 47 16 115 375 Aug-06 40 14 67 0 70 76 22 109 398 Sep-06 46 0 66 0 77 72 0 126 387 Oct-06 0 Nov-06 0 Dec-06 0

TOTAL 329 103 469 0 214 195 112 631 2053

*Preliminary Data Samples positive for Salmonella each month by region*

Region: 1 2 3 5 6 7 8 combined 2005 0/91 0/60 5/133 (4%) ns ns 1/74 (1%) 9/133 (7%) 15/491 (3%) Jan-06 0/12 ns 0/23 ns ns ns 0/25 0/60 (0%) Feb-06 1/16 (6%) ns 1/20 (5%) ns ns ns 1/22 (5%) 3/58 (5%) Mar-06 1/19 (5%) ns 0/28 ns ns ns 0/26 1/73 (1%) Apr-06 1/20 (5%) ns 0/28 ns ns ns 0/26 1/74 (1%) May-06 0/19 ns 2/23 (9%) ns ns ns 0/24 2/66 (3%) Jun-06 1/22 (5%) ns 1/24 (4%) ns ns ns 0/25 2/71 (3%) Jul-06 1/44 (2%) 2/29 (7%) 2/57 (4%) 2/67 (3%) 4/47 (9%) 0/16 1/115 (1%) 12/375 (3%) Aug-06 0/40 1/14 (7%) 10/67 (15%) 1/70 (1%) 10/76 (13%) 2/22 (9%) 2/109 (2%) 26/398 (7%) Sep-06 1/46 (2%) ns 11/66 (17%) 3/77 (4%) 10/72 (14%) ns 6/126 (5%) 31/387 (8%) Oct-06 Nov-06 Dec-06

Total 6/329 (2.0%)3/103 (3.0%)32/469 (6.8%) 6/214 (2.8%) 24/195 (12%) 3/112 (2.7%) 19/361 (5.3%) 93/2053 (4.5%)

ns = no samples collected for analysis *Preliminary Data Samples positive for MDR-Salmonella each month by region*

Region: 1 2 3 5 6 7 8 combined 2005 0/91 0/60 0/133 ns ns 0/74 0/133 0/491 Jan-06 0/12 ns 0/23 ns ns ns 0/25 0/60 Feb-06 0/16 ns 0/20 ns ns ns 0/22 0/58 Mar-06 1/19 (5%) ns 0/28 ns ns ns 0/26 1/73 (1.4%) Apr-06 0/20 ns 0/28 ns ns ns 0/26 0/74 May-06 0/19 ns 0/23 ns ns ns 0/24 0/66 Jun-06 0/22 ns 0/24 ns ns ns 0/25 0/71 Jul-06 0/44 1/29 (3%) 0/57 0/67 0/47 0/16 0/115 1/375 (0.3%) Aug-06 0/40 0/14 0/67 1/70 (1%) 1/76 (1.3%) 1/22 (5%) 0/109 3/398 (0.8%) Sep-06 0/46 ns 1/66 (2%) 0/77 1/72 (1.4%) ns 3/126 (2%) 5/387 (1.3%) Oct-06 Nov-06 Dec-06

Total 1/329 (0.3%)1/103 (1.0%)1/469 (0.2%) 1/214 (0.5%) 2/195 (1.0%) 1/112 (0.9%) 3/361 (0.8%) 10/2053 (0.5%)

ns = no samples collected for analysis *Preliminary Data Samples positive for Listeria species (including LM) each month by region*

Region: 1 2 3 5 6 7 8 combined 2005 15/91 (16%) 15/60 (25%) 39/133 (33%) ns ns 35/74 (47%) 94/133 (71%) 198/491 (40%) Jan-06 1/12 (8%) ns 12/23 (52%) ns ns ns 14/25 (56%) 27/60 (45%) Feb-06 3/16 (19%) ns 8/20 (40%) ns ns ns 17/22 (77%) 28/58 (48%) Mar-06 7/19 (37%) ns 10/28 (36%) ns ns ns 11/26(42%) 28/73 (38%) Apr-06 7/20 (35%) ns 10/28 (36%) ns ns ns 10/26 (38%) 27/74 (36%) May-06 5/19 (26%) ns 4/23 (17%) ns ns ns 19/24 (79%) 28/66 (42%) Jun-06 9/22 (41%) ns 3/24 (13%) ns ns ns 10/25 (40%) 22/71 (31%) Jul-06 10/44 (23%) 5/29 (17%) 7/57 (12%) 11/67 (16%) 12/47 (26%) 2/16 (13%) 25/115 (22%) 72/375 (19%) Aug-06 8/40 (20%) 3/14 (21%) 9/67 (13%) 14/70 (20%) 21/76 (28%) 1/22 (5%) 29/109 (27%) 85/398 (21%) Sep-06 7/47 (15%) ns 14/66 (21%) 6/77 (8%) 24/72 (33%) ns 31/126 (25%) 82/387 (21%) Oct-06 Nov-06 Dec-06

total 72/329 (22%)23/103 (22%) 116/469 (25%)31/214 (15%)57/195 (29%) 38/112 (34%)260/361 (72%) 597/2053 (29%)

ns = no samples collected for analysis. *Preliminary Data Samples positive for LM each month by region*

Region: 1 2 3 5 6 7 8 combined 2005 8/91 (9%) 14/60 (23%) 18/133 (14%) ns ns 9/74 (12%) 32/133 (24%) 81/491 (16%) Jan-06 1/12 (8%) ns 3/23 (13%) ns ns ns 7/25 (28%) 11/60 (18%) Feb-06 0/16 ns 4/20 (20%) ns ns ns 5/22 23%) 9/58 (16%) Mar-06 3/19 (16%) ns 4/28 (14%) ns ns ns 6/26 (23%) 13/73 (18%) Apr-06 2/20 (10%) ns 4/28 (14%) ns ns ns 6/26 (23%) 12/74 (16%) May-06 3/19 (16%) ns 3/23 (13%) ns ns ns 9/24 (38%) 15/66 (23%) Jun-06 5/22 (23%) ns 2/24 (8%) ns ns ns 7/25 (28%) 14/71 (20%) Jul-06 4/44 (9%) 2/29 (7%) 2/57 (4%) 11/67 (16%) 1/47 (2%) 1/16 (6%) 17/115 (15%) 38/375 (10%) Aug-06 5/40 (13%) 2/14 (14%) 4/67 (6%) 13/70 (19%) 10/76 (13%) 0/22 14/109 (13%) 48/398 (12%) Sep-06 2/47 (4%) ns 6/66 (9%) 5/77 (6%) 6/72 (8%) ns 22/126 (17%) 41/387 (11%) Oct-06 Nov-06 Dec-06

total 33/329 (10%)18/103 (17%) 50/469 (11%) 29/214 (14%) 17/195 (9%) 10/112 (9%) 125/361 (35%) 282/2053 (14%)

ns = no samples collected for analysis. *Preliminary Data Samples Positive for shiga toxin genes by PCR*

any stx gene Region n stx1 stx2 stx1+2 present 1 329 6 (2%) 55 (17%) 32 (10%) 93 (28%) 2 103 2 (2%) 10 (10%) 15 (15%) 27 (26%) 3 469 8 (2%) 60 (13%) 62 (13%) 130 (28%) 5 214 8 (4%) 18 (8%) 9 (4%) 35 (16%) 6 195 2 (1%) 19 (10%) 5 (3%) 26 (13%) 7 112 6 (5%) 28 (25%) 11 (10%) 45 (40%) 8 631 7 (1%) 119 (19%) 133 (21%) 259 (41%)

all 2053 39 (2%) 309 (15%) 267 (13%) 615 (30%)

*Preliminary Data Effects of antimicrobial interventions on MDR-Salmonella, non-MDR-Salmonella, and E. coli O157:H7*

reduction (log CFU/cm2) Treatment Strain Strain type (n=12) wash treatment chill total Salmonella MDR ozone 0.66 0.36 0.28 1.29 FX 0.54 1.41 0.08 2.03 hot water 0.62 1.82 -0.32 2.12 EO 0.30 0.66 0.18 1.14 average 0.54 1.08 0.02 1.65 non-MDR ozone 0.72 0.29 0.35 1.36 FX 0.57 1.08 0.27 1.92 hot water 0.59 1.42 -0.15 1.86 EO 0.37 0.69 0.17 1.23 average 0.56 0.87 0.16 1.59

E. coli non-virulent FX 0.76 1.93 -0.92 1.77 hot water 0.61 0.96 -0.06 1.51 EO 0.43 0.46 0.67 1.57 average 0.60 1.12 -0.10 1.62 virulent FX 0.82 1.68 -0.79 1.72 hot water 0.81 1.26 -0.50 1.57 EO 0.28 0.30 0.89 1.46 average 0.64 1.08 -0.13 1.58 *Preliminary Data Salmonella data – Plant B*

Trip 1 2 Feedlot Fecal Enumeration Feedlot Fecal Prev 0.0% 0.0% Feedlot Hide Enumeration Feedlot Hide Prev 0.0% 0.0%

Plant Hide after cabinet Enumeration 4 1 Plant Hide after cabinet Prev 100.0% 59.8% Plant Pre-evis Enumeration Plant Pre-evis Prev 0.0% 0.0%

*Preliminary Data