ift expert report on emerging microbiological food safety issues

Microbiological food safety is a complex, fundamental issue

of continuing concern. Contributing to this complexity and

the emergence of food safety issues are ongoing changes in

demographics, geographic origin of food, food production

and processing, food consumption patterns, and microorgan-

isms themselves. These host, environmental, and pathogen

changes challenge our food safety policies and our ability to

manage food safety throughout the food system.

Recognizing this, the Institute of Food Technologists (IFT),

the 28,000-member nonprofit society for food science and tech-

nology, convened a panel of internationally renowned experts to

review the science related to emerging microbiological food safety

issues and implications for their control and to produce a com-

prehensive, scientific report. IFT’s objective for this Expert Report

is to increase the understanding—among IFT members, senior

policy officials, and other interested groups—of the scientific in-formation on emerging foodborne pathogens (from a broad eco-logical perspective) relative to public policy issues and strategiesfor preventing foodborne illness.

This report is the second Expert Report produced by IFTsince the establishment of its Office of Science, Communica-tions, and Government Relations, which led the production ofthis report and the IFT Expert Report on Biotechnology andFoods. In the seven sections of this report, the expert panel fo-cuses on the complexity of emerging foodborne pathogens andfactors influencing emergence; manifestation of clinical food-borne disease; human susceptibility; ecology of pathogens inpre-harvest and post-harvest environments; microbial viru-lence, evolution, selection, adaptation, stress, and driving forces;risk analysis, the Hazard Analysis and Critical Control Pointsystem, Food Safety Objectives, microbiological performancecriteria, microbial testing, and surveillance; and steps for man-aging food safety in the future.

IFT Expert Report onEmerging Microbiological Food Safety IssuesImplications for Control in the 21st Century

Founded in 1939, the Institute of Food Technologists is a nonprofit scientific society with 28,000 membersworking in food science, technology, and related professions in the food industry, academia, and government.As the society for food science and technology, IFT brings sounds science to the public discussion of food issues.

2 INSTITUTE OF FOOD TECHNOLOGISTS

IFT Expert Report Panelists

IFT is deeply grateful to the expert report panelists for the time and effort that each of them expended on this project, bringingtheir expertise and insight into the state-of-the-science on the numerous topics addressed in the report. Panelists traveled to Chicagoto participate in full-day meetings and devoted considerable additional time to drafting the report, participating in conference calls todiscuss drafts, and reviewing the drafts. IFT sincerely appreciates these experts’ invaluable dedication to furthering the understandingof emerging microbiological food safety issues and food safety management.

Morris Potter, D.V.M., Panel ChairLead Scientist for EpidemiologyCenter for Food Safety and AppliedNutritionU.S. Food and Drug Administration,Atlanta, GA

Douglas Archer, Ph.D.Professor, Food Science and HumanNutritionUniversity of Florida, Gainesville

Andrew Benson, Ph.D.Assistant Professor, Food MicrobiologyUniversity of Nebraska, Lincoln

Frank Busta, Ph.D.Emeritus Professor, Food Scienceand NutritionUniversity of Minnesota, St. Paul

James S. Dickson, Ph.D.Dept. Chair and Associate Professor,Dept. of MicrobiologyIowa State University, Ames

Michael Doyle, Ph.D.Director, Center for Food SafetyUniversity of Georgia, Griffin

Jeffrey Farber, Ph.D.Director, Bureau of Microbial HazardsHealth CanadaOttawa, Ontario, Canada

B. Brett Finlay, Ph.D.Professor, Biotechnology LaboratoryUniversity of British Columbia,Vancouver

Michael Goldblatt, Ph.D.Director, Defense Advanced ResearchProjects AgencyDefense Sciences Office, Arlington, VA

Craig Hedberg, Ph.D.Associate Professor, Division of Environ-mental and Occupational HealthUniversity of Minnesota, Minneapolis

Dallas Hoover, Ph.D.Professor, Dept. of Animal andFood SciencesUniversity of Delaware, Newark

Michael Jahncke, Ph.D.Associate Professor and Director,Dept. of Food Science and TechnologyVirginia Seafood Agricultural Researchand Extension Center at HamptonVirginia Polytechnic Institute and StateUniversity, Hampton

Lee-Ann Jaykus, Ph.D.Associate Professor, Food MicrobiologyDept. of Food ScienceNorth Carolina State University, Raleigh

Charles Kaspar, Ph.D.Associate Professor, Food ResearchInstitute and Environmental ToxicologyCenterUniversity of Wisconsin, Madison

Arthur P. Liang, M.D., M.P.H.Director, Food Safety Initiative ActivityDivision of Bacterial and Mycotic DiseasesNational Center for Infectious DiseasesCenters for Disease Control and Preven-tion, Atlanta, GA

James Lindsay, Ph.D.National Program Leader, Food SafetyAgricultural Research ServiceU.S. Department of Agriculture,Beltsville, MD

James Pestka, Ph.D.Professor, Food Science and HumanNutritionMichigan State University, East Lansing

Merle Pierson, Ph.D.Professor, Dept. of Food Science andTechnologyVirginia Polytechnic Institute and StateUniversity, Blacksburg

Peter Slade, Ph.D.Director of Research & Technical ServicesNational Center for Food Safety andTechnology, Summit-Argo, IL

R. Bruce Tompkin, Ph.D.Vice President of Food SafetyConAgra Refrigerated Prepared Foods,Downers Grove, IL

Mary Lou Tortorello, Ph.D.Research MicrobiologistNational Center for Food Safetyand TechnologyU.S. Food and Drug Administration,Summit-Argo, IL

IFT Staff

Mary Helen Arthur, M.T.S.C.Lead Editor, Expert ReportDept. of Science and Communications,Chicago, IL

Rosetta Newsome, Ph.D.Director, Dept. of Science and Commu-nicationsChicago, IL

Fred Shank, Ph.D.Vice President, Office of Science, Com-munications, and Government RelationsWashington, D.C.

The participants on the expert panel were chosen based on theirscientific expertise. Their contributions represent their individualscientific perspective and do not represent the perspective of theiremployer.

3EXPERT REPORT

Introduction ...................................................................................... 4Trinity of Factors ..................................................................... 4Evolution of Controls ............................................................. 4Fig. 1a-Foodborne Illness ....................................................... 4Fig. 1b-Reducing One Factor .................................................. 4Fig. 1c-Reducing Multiple Factors .......................................... 4Table 1. Evolution of Food Processing .................................... 5Evolution of Food Safety Policies ........................................... 5Microbiology 101 .................................................................... 6Incidence and Prevalence of Foodborne Illness ..................... 8Emergence of Pathogens ......................................................... 8Fig. 2. Foodborne Illness Identification .................................. 8Table 2. Foodborne Disease in the United States ................... 9Complex Drivers of Change ................................................. 11Framework for Food Safety Management ............................ 12

Science of Pathogenicity ................................................................ 13Nomenclature ........................................................................ 13Table 3. Classic Microbial Nomenclature ............................. 13Nomenclature of Salmonella and Fig. 3 ................................ 14Virulence ............................................................................... 15Fig. 4. Virulence and Foodborne Illness ............................... 15Quorum Sensing ................................................................... 16Virulence of Salmonella ........................................................ 17Pathogens Are More Than Just Bacteria ............................... 18Evolution ............................................................................... 19Fig. 5. Contrasting Views of Pathogen Evolution ................. 20Fig. 6. Genetic Material in E. coli ........................................... 20Evolution of Salmonella ........................................................ 21Selection ................................................................................. 21Stress ...................................................................................... 22F38 Regulated Proteins and Table 4 ....................................... 24Driving Forces in Pathogenicity ............................................ 25Emergence of Viruses, Parasitic Protozoa and Marine Biotoxins as Foodborne Pathogens ................................... 25Pathogenicity of E. coli O157:H7 .......................................... 27

Humans as Hosts of Foodborne Disease ........................................ 28Manifestations of Clinical Disease ....................................... 28Table 5. Causes of Foodborne Illness ................................... 29Pfiesteria piscicida and Pfiesteria-like Microbes as Potential Foodborne Pathogens ........................................ 30Resistance to Microbial Foodborne Disease ........................ 31Susceptibility to Microbial Foodborne Disease ................... 34Table 6. Factors That Increase Host Susceptibility ............... 34Cryptosporidiosis .................................................................. 35Individual Choices that Affect Disease Risk ......................... 36Table 7. Factors That Increase Risk of Foodborne Disease .. 37Modification of Susceptibility ............................................... 39

Microbial Ecology and Foodborne Disease ..................................... 40PRE-HARVEST ENVIRONMENT ...................................... 40Overarching Issues ................................................................ 40Table 8. Sources of Imported Fresh and Frozen Produce .... 41Table 9. Percentage of Total U.S. Consumption Provided By Imports ......................................................................... 41Typical Pre-Harvest Environment for Foods of Plant Origin ... 42Production Practices and Mycotoxins .................................. 42Typical Pre-Harvest Environment for Foods of Animal Origin ...................................................................... 43

Development and Dissemination of Resistant Organisms .. 44Wild-Caught Shellfish and Fish ............................................ 45The Role of Microbiological Indicators in Assuring Food Safety ........................................................................ 46Specific Production Methods ................................................ 47HARVEST ENVIRONMENT ............................................... 48Produce .................................................................................. 48Food Animals ........................................................................ 48Aquaculture and Wild-Caught Fish and Shellfish ............... 49POST-HARVEST ENVIRONMENT .................................... 49Food Animal Slaughter and Meat Processing ...................... 49Post-Harvest Processing of Other Commodities ................. 52Water ...................................................................................... 52Alternative Processing Technologies ..................................... 53Table 10. Limitations to Alternative Processing Technologies Currently Under Development ................... 54Validation of Treatment Effectiveness Using Microbiological Surrogates ............................................... 55Transportation and Storage .................................................. 60Retail and Food Service ......................................................... 61Outbreaks of Shigella sonnei Infection Associated with Fresh Parsley ...................................................................... 61Microbial Stress Responses to Processing ............................ 62Table 11. Conditions That Can Produce Sublethally Injured Cells ...................................................................... 64New Tools for Pathogen Research ......................................... 64Ability of Pathogens To Survive in the Environment ........... 66

Application of Science to Food Safety Management ...................... 67Risk Assessment .................................................................... 67Risk Management Using Food Safety Objectives ................. 69Fig. 7. Framework for Food Safety Management ................. 70Hazard Control and Monitoring .......................................... 71Table 12. FSOs in the Food Safety Management Framework ....72Fig. 8. Establishing Performance Criteria ............................. 74Fig. 9. Unequal Levels of Food Safety ................................... 75Table 13. Probability of Acceptance (P

a) of Defective

Product Using a 2-class Sampling Plan ............................ 76Table 14. USDA Monitoring Program for Salmonella ......... 76Value of Test Results and Fig. 10 ........................................... 77Testing for Mycotoxins .......................................................... 79Surveillance for Foodborne Hazards and Illness ................. 79Outbreak Investigations and New Foodborne Pathogens ... 80Animal Surveillance for E. coli O157:H7 .............................. 82

Next Steps in Food Safety Management ......................................... 85Strategic Prioritization to Reduce Foodborne Disease ........ 85Strategies for the Future ........................................................ 87A Cooperative Approach to the Safety of Sprouts ................ 89Anticipating the Future: Food Safety Issues on the Horizon.....92

Conclusions .................................................................................... 93

References ..................................................................................... 97

Table of Contents

On the CoverThe image on the cover is a scanning electron micrograph ofListeria monocytogenes, an emerging foodborne pathogen.

© 2000 S. Lowry/Univ. Ulster/Stone


The continued occurrence of

foodborne illness is not evidence of the

failure of our food safety system. In

fact, many of our prevention and

control efforts have been—and

continue to be—highly effective. The

U.S. food supply is arguably among the

safest in the world, but significant

foodborne illness continues to occur.

Despite great strides in the area of

microbiological food safety, much

remains to be done.Foodborne illness is not a simple

problem in need of a solution; it is a com-plex combination of factors that must bemanaged on a continual basis. Aside fromthe inherent ability of pathogens them-selves to evolve, pathogens’ victims andthe microbial environment play a role inthe changing nature of foodborne illness,opening new niches and creating new vul-nerabilities. No matter how sophisticatedand complex a system is developed, foodsafety management is never finished orcomplete, because change is constant.

Recognizing that food safety is a fun-damental and continuing issue, the Insti-tute of Food Technologists commis-sioned an expert panel to review theavailable scientific literature related toemerging microbiological food safety is-sues. The experts were charged withidentifying the factors that make a mi-croorganism “emerge” as an importantfoodborne pathogen and identifyingmechanisms that use this knowledge toimprove the safety of our food supply.The objective of this report is to increaseunderstanding, among IFT membersand other interested parties, of the scien-

tific information on emerging foodbornepathogens relative to public policy issuesand strategies for preventing foodborneillness.

Trinity of Factors

At the simplest level, foodborne ill-ness can be reduced to three factors: thepathogen, the host, and the environmentin which they exist and interact (see Fig.1a). Complex relationships exist amongthese factors, and all three factors arenecessary for foodborne illness to occur.For example, a susceptible host may con-sume food that contains a significantamount of a microorganism, but if themicroorganism does not possess thetraits necessary to cause illness, food-borne disease does not occur. Similarly,adequately cooking a food to kill thepathogenic microbes can eliminate theexposure factor and render the food safe.

When one or more of the three fac-tors changes, new foodborne pathogens“emerge.” For example, host susceptibil-ity can increase so much that existingmicroorganisms that do not cause ill-ness in the general population achievepathogen status in the newly immuno-compromised individuals. The changealso can be increased virulence, e.g.,when a microorganism acquires charac-teristics that help it invade the humanbody. Or it can be new exposure, e.g.,when fruit from one region carries apathogenic microorganism to popula-tions in a different geographic regionthat have never before been exposed.

This trinity is also the key to reduc-ing foodborne illness. Prevention andcontrol efforts often focus on the contri-bution of one of these factors, such aswashing vegetables to remove surface

contamination (Fig. 1b). In many cases,however, the most effective approach ad-dresses more than one factor. Currenttechnologies and production methodscannot provide a food supply that iscompletely free of all pathogenic micro-organisms. Fortunately, even small re-ductions in several factors can have asubstantial combined effect (Fig. 1c).

Evolution of Controls

New technologies, production prac-tices and food manufacturing processesare developed to meet the needs of achanging society. Early food preservationmethods—such as canning, cheese mak-ing, bread making, brewing, fermentation,pickling, salting, and drying—were usedto provide sufficient year-round foodavailability. Later developments reflected anew focus on food safety, variety, conve-nience, and nutritive and sensory quality.

At the beginning of the 20th century,contaminated milk, meat, and otherfoods caused large outbreaks and manysporadic cases of foodborne disease, of-ten with fatal consequences. The revolu-tion in sanitation and hygiene related tofood and water and the almost universaladoption of thermal pasteurization formilk produced tremendous improve-ments in food safety. New technologieswith increasing sophistication have yield-ed continued improvements in microbi-ological food safety while delivering bet-ter quality foods with greater nutritionalvalue and superior sensory characteris-tics (see Table 1).

Innovations in packaging have beenintegral to the developments in food pro-cessing and product development. Pack-ages contain and protect their food con-tents and inform consumers; they also

Introduction

Fig. 1a- Foodborne Illness Fig. 1b-Reducing One Factor Fig. 1c-Reducing Multiple Factors

5EXPERT REPORT

preserve, perform, communicate, and sell(Downes, 1989). From the hermeticallysealed containers for shelf-stable foodsdeveloped in the early part of the 19th

century, to the metal cans for heat-pro-cessed foods, folding cartons, and milkbottles of the latter part of the 19th centu-ry, to the boil-in-bags, plastic tubs, highdensity polyethylene gallon milk jugs,aseptic cartons, and microwavable poly-mers of the 20th century, packaging hasplayed a key role in the development ofthe food industry in the Western worldin the 20th century (Downes, 1989).

However, changes in technology arenot without risk. Conventional wisdomof decades ago held that properly refrig-erated foods would remain safe becauseit was thought that pathogenic bacteriawould not grow at refrigeration tempera-tures, but this is not always the case(Marth, 1998). Innovations in food pro-cessing, such as modified atmospherepackaging, can offer the benefit of greatlyextending the shelf life of refrigeratedfoods but may present microbiological

safety challenges. For example, modifiedatmosphere packaging of fresh packed,sliced mushrooms may allow the growthof Clostridium botulinum and potentialtoxin production (Doyle, 1998). Alteringthe package, incorporating microscopicholes to allow oxygen to permeate the in-terior of the packaged product, was an-other factor critical to ensuring the safetyof modified atmosphere packaged freshpacked, sliced mushrooms. For extendedshelf life refrigerated foods, strict tem-perature control and acceptable productshelf life are critical factors to consider(Doyle, 1998). As new technologies areintroduced, they must be evaluated fortheir potential effect on microbiologicalfood safety.

Despite all of the significant advanc-es to date, our growing knowledge basecontinues to expose the role of variousfoods and technological innovations infoodborne hazards, and changes in thefood, the consuming public, and thepathogens themselves continue to makefoodborne disease an important and

ever-changing challenge both for the in-dustrialized and the developing world.

Evolution of Food Safety Policies

Current U.S. food safety policies arethe accretions of decades of relatively in-dependent efforts to address specificproblems. Most are rooted in the sani-tary revolution that occurred at the be-ginning of the 20th century, and theyhave characteristics that have served uswell during the transition from an agrar-ian to an industrialized society.

Generally, these regulatory policiesrespond in one of three ways to obvioushazards that pose clear risk to humanhealth. First, for hazards that havestraightforward technical fixes, regula-tions require the application of the ap-propriate technologies. Regulatory stan-dards frequently have been set at the per-formance limit of the technology or thedetection limit for the analytical test usedfor process verification. However, tech-nologies to mitigate hazards are not al-

Early times HeatingCooking food kills many foodborne pathogens

1770s-1800s Canning/Thermal ProcessingSignificant discoveries in response to industrialization forces and Napoleon’s armies’ need for less dependence on localprovisions

1890s PasteurizationThermal treatment of raw milk to prevent milk from transmitting pathogens

1920s - 1930s Safe Canning/Processing ParametersCalculation of product heat penetration curve and initial microbial contamination level to determine minimum time-temperaturecombination for commercial sterility

1940s FreezingMechanical quick-freezing methods to preserve while maintaining quality

1950s Controlled Atmosphere PackagingReduced oxygen levels, increased concentrations of carbon dioxide, or selective mixtures of atmospheric gases to limitrespiration and ethylene production, delay ripening and decay, and increase refrigerated product shelf life

1960s Freeze DryingRapid deep-freezing followed by sublimation of water by heating the frozen product in a vacuum chamber. Best known appliedto coffee, to preserve delicate aroma compounds and maintain flavor and odor

1940s-1980s Aseptic Processing and PackagingHigh-temperature, short-time sterilization of food product independent of the container, container sterilization, and filling ofproduct in sterile atmosphere, resulting in increased food quality and nutrient retention

IrradiationNon-thermal process to kill pathogens, insects, and larvae, inhibit sprouting, and delay ripening and food spoilage

1990s Carcass Spray Treatments (e.g., water, acid), Steam Vacuuming, Steam PasteurizationCarcass decontamination interventions to meet microbiological performance criteria

High Pressure ProcessingFoods subjected to specified pressures and temperatures to preserve food while maintaining quality

Table 1. Evolution of Food Processing (Goldblith, 1989; IFT, 2000a, 1989; Lund, 1989)

ways apparent. In these cases, the regula-tory response has been to either keep thehazardous food out of the marketplaceor to forgo regulatory action and rely onprudent people to protect themselves.

Numerous food safety concernshave been successfully addressed by thisregulatory paradigm. The promulgationof regulations for low acid canned foodvirtually ended the historic associationof botulism with commercially cannedfood. Under the regulations, commer-cially canned foods undergo a mini-mum calculated destruction of 12D for

C. botulinum. Another very successfulexample is the water quality standardsfor shellfish growing waters. Thesestandards protected consumers fromshellfish-associated typhoid fever at atime when typhoid was fairly commonin coastal communities and contami-nated shellfish was an important sourceof infection. Historically, the safety offoods without a pathogen eliminationstep earlier in the line has dependedupon proper cooking.

The extraordinary complexity ofcontemporary food safety issues presents

major challenges to food safety policyformulation. Factors like the globalsourcing of products and ingredients,changes in land use, and evolution ofscience and technology have radicallychanged hazards associated with a par-ticular food and the control optionsavailable.

These challenges present themselvesin many ways depending on the particu-lar hazard. For example, the indicator or-ganisms used to predict the presence ofSalmonella Typhi in shellfish growingwaters poorly predict the presence of

Microbiology 101

The characteristics of the variousmicroorganisms are part of whatmakes microbiological food safety is-sues so complex. One type of micro-organism may thrive under condi-tions that are fatal to a different mi-crobe. Some microbial pathogenscause disease by infecting the humanhost, while others produce toxins thatcause illness. Some pathogens canmultiply in food during storage whileothers cannot. Because most micro-organisms can reproduce within amatter of minutes, these pathogenscan evolve quickly when environ-mental stresses select for strains withunique survival characteristics.

Types of Microorganisms

Microorganisms are divided intothree distinct categories: prokaryotes,eukaryotes, and viruses. Bothprokaryotes and eukaryotes are highlyregulated cells that possess elaborate“sensing” systems, enabling them to beaware of and react to their environ-ment as it changes.

Prokaryotes—Bacteria

Prokaryotes are single-celled livingmicroorganisms that have no nuclearmembrane separating their geneticmaterial from the cytoplasm withinthe cell. They are microscopic, in thatthey cannot be seen with the unassist-ed human eye. Bacteria are generallyfree-living in the environment, al-though some have complex nutrientrequirements and can grow only inspecial niches.

Bacteria are arguably the world’s

most successful life form. Their diversityand “complexity through simplicity”have and will continue to assure theirsurvival. Although the vast majority ofbacteria are harmless or helpful to hu-mans, some are pathogenic.

Eukaryotes—Parasites and Fungi

Eukaryotes are multicellular livingorganisms that possess a nuclear mem-brane that separates their nucleic acidfrom the cytoplasm. They are larger thanbacteria and are sometimes able to beseen by the human eye. Eukaryotes aregenerally free-living in the environment.

Some protozoa can be foodbornepathogens (e.g., Cyclospora, Giardia).They usually exist in multiple forms,some of which are environmentally sta-ble, but they seldom multiply in or onhuman food. Fungi such as molds andyeasts can multiply in or on human foodand also can be pathogenic.

Viruses

Foodborne pathogenic viruses arecomprised of a single type of nucleicacid surrounded by a protein coat. Vi-ruses are not free-living. In fact, theyare not living beings at all, but are obli-gate intracellular parasites. They aresmaller than bacteria (10-350nm), andsome viruses prey on bacteria (bacte-riophages).

Types of Bacteria

Bacteria come in various shapes,such as rods, spheres (cocci), and spirals.As a response to certain adverse environ-mental conditions, some bacteria canform spores. The spores start as denseregions within the cell, but as the cell de-

teriorates, the spore is released into theenvironment. Spores are extremely im-pervious to physical and chemical harm,making them difficult to inactivate in thefood processing environment.

In general, the bacterial kingdom canbe divided into gram-positive and gram-negative cells. These designations aregiven based on the results of a stainingprocedure that separates the two divi-sions by color, which is reflective of thecomposition of their cell wall.

Growth and Life SpanBacteria—Bacteria are the most

adaptable life form on Earth. Bacteriahave “optimal” (preferred) growth condi-tions, but some can grow and/or surviveat extremes of temperature, pH, osmoticpressure, and barometric pressure. Bacte-ria are genetically programmed for maxi-mum survival.

At optimal growth conditions, a bac-terial cell may divide every 10-20 min-utes. Assuming no death, a single cellcould thus give rise to a bacterial massequal to the Earth’s mass in one or twodays. Obviously death occurs, because offactors such as nutrient limitations orend product toxicity.

Viruses—In general, virusesreproduce more rapidly than bacteria,but they can only grow in an infectedhost cell, not in food. A single infectedhost cell may give rise to hundreds orthousands of new viruses within a fewhours, each of which may infect a newhost cell.

Roles of FoodborneMicroorganisms

Not all microorganisms in foods areharmful. In fact, only a small proportionof foodborne organisms have been asso-


7EXPERT REPORT

cooking will kill this parasite, the foodsassociated with human infection—rasp-berries, basil, and other fresh produce—are generally not cooked before con-sumption.

When microorganisms such as Shigel-la or the hepatitis A virus are found onfresh produce, it can be difficult to deter-mine whether the contamination oc-curred during food preparation, duringdistribution, during processing and pack-ing, or in the growing fields. Under thesecircumstances, the appropriate point ofintervention is difficult to identify.

Modifications to thermal processinghave made a wide array of food typespossible, including a proliferation of“pasteurized” food products distributedin flexible plastic packages that requirerefrigeration. Although thermal process-ing inactivates a large majority of thevegetative spoilage organisms, in con-trolled atmosphere packaging, it alsodrives off oxygen and fails to inactivatesporeformers like C. botulinum. If thecold chain is not properly maintained,botulinum toxin can be produced beforefood spoilage is detectable.

Vibrio, hepatitis A, or Norwalk-like vi-ruses. Typhoid fever is now extremelyuncommon in U.S. coastal waters, and itis now these other shellfish-associatedpathogens from which we need to protectourselves.

Pathogens on foods that are oftenconsumed without cooking present asignificant challenge. The pathogen Cy-clospora cayetanensis is very likely ofhuman origin, but our limited knowl-edge about its natural ecology does notenable us to assure its absence fromfresh produce. Although adequate

ciated with disease in normal, healthyanimals and humans.

Commensal Microbes

Virtually all raw food contains mi-croorganisms. The source of these is theproduction environment, where a widevariety of organisms are environmentallyubiquitous. Processing and handling offoods can also contribute to the typesand numbers of commensal organismson foods. Most of these commensal or-ganisms are harmless to animals and hu-mans; in fact, they may actually be bene-ficial in that they provide high levels of“competitive” microflora that usuallygrow faster than contaminating patho-gens. Although the purpose of manycommon food processing methods suchas pasteurization and canning is the de-struction of pathogens, commensal mi-croorganisms are often destroyed in theprocess as well.

Spoilage Microbes

Spoilage may be defined as a condi-tion in which food becomes inedible be-cause of undesirable changes in color,flavor, odor, appearance, and texture.This condition occurs primarily be-cause the organisms grow to high levels,producing enzymes that break downfood components such as fats, proteins,and sugars. In most instances, spoilageis caused by commensal organisms thathave been allowed to reach populationsin the range of 105 to 107 CFU/g offood. Different classifications of foods(such as red meats, vegetables, fish, etc.)have different spoilage profiles becausethe food environment will dictate whichorganisms will grow, dominate, and

cause spoilage. For instance, spoilage ofraw meats is almost always associatedwith gram-negative psychrotrophs, theso-called cold-thriving organisms, be-cause they grow at refrigeration temper-atures. Fresh fruits are frequentlyspoiled by yeasts and molds that areable to thrive in acidic conditions. Mostspoiled foods do not cause foodbornedisease; in reality, the high levels ofspoilage organisms have frequently“out-competed” the pathogens, keepingpathogen growth in check.

Beneficial Microbes

Perhaps the most widely recognizedgroup of beneficial foodborne microor-ganisms are the members of the lactic acidbacteria (LAB) group. This is a functionalname used to classify fermentative organ-isms that produce lactic acid as the prima-ry by-product of metabolic activity, al-though other products, such as alcoholand carbon dioxide may be produced aswell. These metabolic products are re-sponsible for the characteristic flavor,odor and texture of fermented food prod-ucts. The lactic acid bacteria are com-monly used in dairy, vegetable and meatfermentations. Notable members of thisgroup belong to the Lactococcus, Lactoba-cillus, Pediococcus, and Leuconostoc genera.Some species of yeasts and molds can alsobe used in the commercial production offermented foods, including Saccharomycescerevisiae, used frequently for producingbread, beer and wines, and Aspergillusoryzae, used in the fermentation of orien-tal foods such as soy sauce.

Foodborne Pathogens

Foodborne pathogens encompass arelatively small group of foodborne micro-

organisms that are associated with dis-ease in humans and/or animals. Patho-genic microbes are capable of causing ill-ness, either by infecting the host or byproducing toxins that cause the host tobecome ill. Some microorganisms arepathogenic in one host species but not inanother. For example, Escherichia coliO157:H7 causes illness in humans butnot in cattle, its primary host.

Microbiological Indicators

An index or indicator refers to asingle or group of microorganisms, oralternatively, a metabolic product,whose presence in a food or the envi-ronment at a given level is indicativeof a potential quality or safety prob-lem. Microbiological indicators areused in place of direct testing for apathogen, largely because they areeasier to work with. Indicators maybe a specific microorganism (e.g., E.coli), a metabolite (e.g., lactic acid ti-tration), or some other indirect mea-sure (e.g., ATP bioluminescence as ameasure of sanitation efficacy). Usinga specific microorganism as an indi-cator is difficult, because appropriateindicator organisms are difficult toidentify. An “ideal” indicator organ-ism: (1) has a history of presence infoods at any time that the targetpathogen or toxin might be present;(2) is present at concentrations di-rectly related to that of the targetpathogen or toxin; (3) is absent fromfood when the target is not present;(4) has growth rates equivalent to, orslightly greater than, the pathogen;(5) has rapid, simple, and inexpensivequantitative assays available; (6) hassimilar resistance profiles to the tar-get; and (7) is nonpathogenic.


Non-thermal processes also havebeen modified over time. Consumerswant foods with fewer preservatives, lesssalt, fewer calories, and better texture; thefood industry has responded with manynew formulations. However, substitut-ing ingredients with gums or other fat re-placers and reducing salt or sugar can re-quire a reevaluation of food safety con-trol measures. For example, the replace-ment of sugar with an alternative sweet-ener in hazelnut yogurt and failure toevaluate the impact of this change onfood safety resulted in an outbreak ofbotulism (O’Mahony et al., 1990).

In addition to the impact of changesin processing, scientists have discoveredthat some foodborne pathogens survivetraditional processes better than expect-ed. For example, Salmonella have beenfound in 60-day aged cheese and on rawalmonds, and newly recognized patho-gens such as E. coli O157:H7 are moretolerant of conditions of low pH andother traditional barriers than antici-pated. The resistance of pathogens totraditional treatments affects the safetyof our drinking and processing water aswell. We have relied on chlorination torid drinking water of pathogens for de-cades, but recent waterborne outbreakshave been caused by parasites, such asGiardia and Cryptosporidium, that arenot controlled by chlorine.

Handling during preparation in thehome or foodservice establishment mayaffect the pathogens present in the food.For foods in which preparation shouldkill the pathogens, recurring outbreaks offoodborne illness highlight our limitedability to tightly control food preparationbehaviors and practices. Certainly the in-cidence of foodborne disease would besignificantly reduced if we could eliminatepathogens earlier in the food chain. How-ever, the initial number of the pathogen inthe food is only one factor in the risk offoodborne illness. At some point, im-provements in sanitation will produceonly small incremental gains. As the levelof contamination becomes increasinglysmall, other food safety approaches willneed to be adopted.

Incidence and Prevalence ofFoodborne Illness

Food safety is a complex issue that de-pends on a number of interrelated envi-ronmental, cultural, and socioeconomicfactors. More than 200 known diseasesare transmitted through food, and more

than half of all recognized foodborne dis-ease outbreaks have unknown causes, in-dicating the real number of disease-caus-ing agents is likely much larger than 200.The recognized causes of foodborne ill-ness include viruses, bacteria, parasites,manmade chemicals, biotoxins, heavymetals, and prions. The symptoms offoodborne illnesses range from mild gas-troenteritis to life threatening neurologic,hepatic, and renal syndromes.

In the United States, foodborne dis-eases have been estimated to cause ap-proximately 76 million illnesses,325,000 hospitalizations, and 5,000deaths each year (Mead et al., 1999). Apathogen’s ability to cause illness can bevery different from the severity of theillness it causes. Some pathogens suchas Norwalk-like viruses cause a greatnumber of illnesses (9.2 million peryear) but the fatality rate is very small(0.001%) (see Table 2). Others such asVibrio vulnificus cause few illnesses (47per year), but many of those illnessesare fatal (38.3%). Salmonella, Listeriamonocytogenes, and Toxoplasma gondiiare responsible for more that 78% ofthe deaths but only approximately 11%of total cases of foodborne disease. Theissue is further complicated by patho-gens, such as L. monocytogenes, thatcause little or no illness in healthy indi-viduals but cause severe illness anddeath in sensitive populations, includ-ing the immunosuppressed, the elderlyand the developing fetus. Prioritizingfood safety resources can be difficult.

Scientists gather data about the inci-dence and severity of foodborne illnessthrough surveillance, both passive andactive. Mild cases of illness often are notcaptured by surveil-lance programs becausemedical intervention isnot required for recov-ery. Many steps are re-quired for a foodbornepathogen to be identi-fied as the cause of ill-ness and for data to begathered through sur-veillance programs (seeFig. 2). Furthermore,because many patho-gens transmittedthrough food also maybe spread by contami-nated water and per-son-to-person contact,the role of food can bedifficult to assess. Fi-

nally, some proportion of foodborne ill-ness is caused by pathogens that have notyet been identified and thus cannot bediagnosed. In fact, many of the patho-gens of concern today were not recog-nized as causes of foodborne illness just20 years ago (Mead et al., 1999).

New scientific advances make it pos-sible to approach foodborne illness froma different, broader perspective. Morepowerful diagnostic procedures and bet-ter communication technology allow im-proved tracking and surveillance forfoodborne illness. Genetic identificationmethods allow scientists to link geo-graphically distinct outbreaks of food-borne illness to a single source. Patho-gens can appear to emerge simply be-cause scientists develop methods to iden-tify the presence of certain microorgan-isms and link them to foodborne disease.

New technologies based on recentadvances in genomics also give scientistsgreater insight into pathogen virulenceand evolution, opening the door to bettercontrols and therapeutics. Future scien-tific advances will continue to enhanceefforts to identify and understand food-borne pathogens, and these insights willcontribute the data necessary for science-based risk assessment and food safetymanagement.

Emergence of Pathogens

The terminology “newly emergingpathogen” has become somewhat over-used, or perhaps it is merely ill defined.True pathogen “emergence” could be di-rectly linked to evolution, whether thatevolution occurs gradually or rapidly.

In a broader context, emergence can

Sick individual seeks medical attention.

Clinician considers cause to be foodborne,requests proper tests and collects appropriate specimen.

Proper test is performed correctly.

Results are reported to the health departmentand ultimately the CDC.

Regulatory agency investigates.

Identification of the source of the illnessprompts product recall or other action.

Fig. 2. Foodborne Illness Identification

9EXPERT REPORT

Microorganism

Bacteria

Bacillus cereus

Brucella spp.

Campylobacter spp. C. jejuni

Clostridium botulinum

Clostridium perfringens

Escherichia coli

Enterotoxigenic

Enteropathogenic

E. coli O157:H7

Enteroinvasive

Listeria monocytogenes

Salmonella spp.

S. Typhi and S. Paratyphi

Other Salmonella spp.

Shigella spp.

Staphylococcus aureus

Streptococcus spp.

Group A (S. pyogenes)

Principal Symptomsa

Diarrheal—Watery diarrhea, abdominal cramps and painEmetic—Nausea and vomiting

Sweating, headache, lack of appetite, fatigue, feverc

Watery diarrhea, fever, abdominal pain, nausea, headache,muscle pain

Weariness, weakness, vertigo, double vision, difficultyswallowing and speaking

Intense abdominal cramps, diarrhea

Watery diarrhea, abdominal cramps, fever, nausea, malaise

Watery or bloody diarrhea

Severe abdominal cramps, watery or bloody diarrhea,hemolytic uremic syndrome

Abdominal cramps, vomiting, fever, chills, generalizedmalaise, hemolytic uremic syndrome

Nausea, vomiting, diarrhea; influenza-like symptoms;meningitis, encephalitis; septicemia in pregnant women,their fetuses or newborns; intrauterine or cervical infectionthat may result in spontaneous abortion or stillbirth

Typhoid-like fever, malaise, headache, abdominal pain, bodyaches, diarrhea or constipation

Nausea, vomiting, abdominal cramps, fever, headache;chronic symptoms (e.g., arthritis)

Abdominal pain and cramps; diarrhea; fever; vomiting;blood, pus or mucus in stools; tenesmus

Nausea, vomiting, retching, abdominal cramps, prostration

Sore, red throat; pain on swallowing; tonsillitis; high fever;headache; nausea; vomiting; malaise; rhinorrhea

Onseta

6-15 hr0.5-6 hr

Days to weeksc

2-5 days

18-36 hr

8-22 hr

24 hr

n/a

1-2 days

12-72 hr

Few days-3 weeks

7-28 daysc

6-48 hr3-4 weeks

12-50 hr

1-7 hr

1-3 days

Potential Food Contaminationa

Meats, milk, vegetables and fishRice products, starchy foods (e.g., potato, pasta, and cheese products

Raw or unheated processed foods of animal origin (e.g., milk, milkproducts, cream, cheese, butter)c

Raw chicken, beef, pork, shellfish; raw milk

Improperly canned or fermented goods

Meat, meat products, gravies

Foods contaminated by human sewage or infected food handlers

Raw beef and chicken; food contaminated by feces or contami-nated water

Undercooked or raw hamburger, alfalfa sprouts, unpasteurized juices,dry-cured salami, lettuce, game meat, cheese curds, raw milk

Food contaminated by human feces or contaminated water,hamburger meat, unpasteurized milk

Raw milk, cheeses (particularly soft-ripened varieties), rawvegetables, raw meats, raw and smoked fish, fermented sausages

Raw meats, poultry, eggs, milk and dairy products, fish, shrimp,frog legs, yeast, coconut, sauces and salad dressings

Salads (potato, tuna, chicken, macaroni), raw vegetables, bakeryproducts (e.g., cream-filled pastries), sandwich fillings, milk anddairy products, poultry

Meat and meat products, poultry, egg products, salads (egg, tuna,chicken, potato, macaroni), cream-filled bakery products, milk anddairy products

Temperature-abused milk, ice cream, eggs, steamed lobster, groundham, potato salad, egg salad, custard, rice pudding, shrimp salad

Illnessesb

27,360n/an/a

777

1,963,141n/a

58

248,520

55,594

n/a

62,458

n/a

2,493

n/a

659

1,341,873

89,648

185,060

50,920

n/a

Deathsb

0n/an/a

6

99n/a

4

7

0

n/a

52

n/a

499

n/a

3

553

14

2

0

n/a

Table 2. Foodborne Disease in the United States, Including Estimated Annual Prevalence (FDA/CFSAN, 2002; Mead et al., 1999)

a n/a indicates FDA/CFSAN (2002) did not provide information.b n/a indicates Mead et al. (1999) did not provide an estimate for this pathogen.c As described by ICMSF (1996).

10INSTITUTE OF FOOD TECHNOLOGISTS

Principal Symptomsa

Diarrhea, abdominal cramps, nausea, vomiting, fever, chills, dizziness

Mild watery diarrhea, acute diarrhea, rice-water stools

Diarrhea, abdominal cramps, fever, vomiting, nausea; blood or mucus-containing stools

Diarrhea, abdominal cramps, nausea, vomiting, headache, fever, chills

Fever, chills, nausea, septicemia in individuals with some underlyingdiseases or taking immunosuppressive drugs or steroids

n/a

Fever, abdominal pain, diarrhea and/or vomiting

Tingling or tickling in the throat, vomiting or coughing up worm(s),abdominal pain, nausea

Exiting of roundworm, vague digestive tract discomfort, pneumonitis

Severe watery diarrhea (intestinal illness); coughing, fever and intestinaldistress (pulmonary and tracheal illness)

Watery diarrhea, explosive stools, loss of appetite, bloating, stomachcramps, vomiting, aching muscles

Diarrhea, abdominal cramps, bloating, weight loss, malabsorption

Discharge of proglottids, anal itching, abdominal pain, nausea, weakness,weight loss, intestinal disorder

Flu-like symptoms, swollen lymph glands, muscle aches and painsd

Severe gastrointestinal distress, nausea, vomiting, headaches, weakness,muscle pain, chills, difficulty breathing, body swelling, visual deficiencies,fever, night sweatingc

Fever, anorexia, malaise, nausea, abdominal discomfort; jaundice mayfollow

Nausea, vomiting, low-grade fever, diarrhea, abdominal pain

Vomiting, watery diarrhea, low-grade fever; severe in infants and youngchildren

Microorganism

Group D (other Streptococcusspp.)

Vibrio cholerae

V. cholerae serogroup O1

V. cholerae serogroup non-O1

Vibrio parahaemolyticus

Vibrio vulnificus

Vibrio, other

Yersinia enterocolitica

Parasites and Protozoa

Anisakis simplex

Ascaris lumbricoides

Cryptosporidium parvum

Cyclospora cayetanensis

Giardia lamblia

Taenia spp.

Toxoplasma gondii

Trichinella spiralis

Viruses

Hepatitis A

Norwalk-like viruses

Rotavirus

Onseta

2-36 hr

6hr-5 days

48 hr

4-96 hr

16 hr

n/a

24-48 hr

1hr-2 weeks

n/a

1-12 days

1 week

1 week

n/a

10-23 daysd

3-14 days

10-50 days

24-48 hr

1-3 days

Potential Food Contaminationa

Underprocessed or improperly prepared sausage, evaporatedmilk, cheese, meat croquettes, meat pie, pudding, raw milk,pasteurized milk

Raw, improperly cooked, or recontaminated shellfish

Raw, improperly cooked, or recontaminated shellfish

Raw, improperly cooked, or recontaminated shellfish and fish

Raw or recontaminated oysters, clams, crabs

n/a

Meats, oysters, fish, raw milk

Raw or undercooked seafood

Raw produce grown in soil contaminated by insufficientlytreated sewage

Foods contaminated via food handlers and manure

Water or food contaminated with infected stool

Food contaminated via infected food handlers

Raw or undercooked beef, pork

Raw or undercooked meats, especially pork, lamb,venisond

Raw or undercooked pork or wild game

Shellfish, salads, other foods contaminated via infectedfood handlers or water

Shellfish and salad ingredients contaminated by infectedfood handlers or fecally contaminated water

Foods contaminated via fecal contaminated food handlers

Illnessesb

n/a

49

n/a

n/a

n/a

47

5,122

86,731

n/a

n/a

30,000

14,638

200,000

n/a

112,500

52

4,170

9,200,000

39,000

Deathsb

n/a

0

n/a

n/a

n/a

18

13

2

n/a

n/a

7

0

1

n/a

375

0

4

124

0

Table 2. Foodborne Disease in the United States, Including Estimated Annual Prevalence, continued

a n/a indicates FDA/CFSAN (2002) did not provide information.b n/a indicates Mead et al. (1999) did not provide an estimate for this pathogen.c As described by ICMSF (1996).d From CDC (2002).

11EXPERT REPORT

be used to describe a recent significantchange. Using this interpretation, apathogen could be described as emergingwhen it is first linked to disease, when theillness it causes suddenly increases infrequency or severity, or even when apathogen recognized for a significantamount of time suddenly “reappears.” Interms of public perception, all these sce-narios may be considered emerging mi-crobiological food safety issues.

Public Perception of Emergence

C. cayetanensis, a recent arrival inthe United States, is an example of thekinds of surprises that can occur as aresult of increasing world trade inready-to-eat foodstuffs. As foodstuffsare transported ever greater distances,pathogens can be transported to newareas as well. The United States also hasexperienced illness outbreaks caused bySalmonella serotypes either rarely ornever before isolated here, apparentlyfor the same reason(s) that Cyclospora“emerged.”

Campylobacter jejuni and L. monocy-togenes were well accepted as foodbornepathogens in the 1970s and 1980s, re-spectively, and, as such, they are hardlynew. Prior to 1972, C. jejuni could not becultured from feces, so it was not recog-nized as a common human pathogen, norwas the frequency of its presence in food,particularly raw poultry, realized. Withimprovements in both the sensitivity andrapidity of the available methodology, C.jejuni may appear to be an emergingpathogen. However, scientists cannot de-termine if the frequency of C. jejuni isola-tion from human stool or food is truly in-creasing, or if laboratory-induced biasgives the appearance of an increase. Simi-larly, methods for isolation of L. monocy-togenes from foods and the environmenthave improved since its “emergence” as afoodborne pathogen in the mid-1980s. Asa result, it appears that the frequency of L.monocytogenes in foods and the environ-ment is increasing, but the influence oflaboratory-induced bias is difficult toweigh. For both of these pathogens, in-creased awareness of analytical laboratorypersonnel and physicians have affected thefrequency of isolation from foods and di-agnosis of human illness.

Changes in foods serving as vectorshave brought new attention to a long-recognized pathogen, C. botulinum, thecausative agent of botulism. Botulismfrom commercially canned foods has

been essentially eliminated; however, cas-es of botulism in the late 1980s and early1990s led scientists to discover that C.botulinum could grow and produce tox-ins in foods such as twice-baked pota-toes, grilled onions, and garlic-in-oil.When these new food associations werediscovered, even C. botulinum was de-scribed by some as an emerging patho-gen. Large outbreaks—although recur-rences of situations that have happenedin the past—can propel a pathogen to“emerging” status whether truly deservedor not. In many cases, it is the new foodvehicle that is the surprise, and not somuch the pathogen itself.

Emergence as a Function of Evolution

Evolution can produce pathogensthat are truly emerging, in that the mi-croorganisms have new characteristicsthat enable them to cause disease. Bacte-ria have evolved highly sophisticated sig-nal transduction systems that allow themicroorganisms to respond at a geneticlevel to environmental conditions in acoordinated manner. The many envi-ronmental stimuli that trigger such activ-ity are collectively referred to as stresses.

The genetic response to stress(es) canactivate certain virulence determinants, amicrobe’s “contingency plan” for surviv-ing in hostile environments. Beyond ac-tivating virulence determinants, externalstresses also accelerate the bacteria’s rateof evolution, meaning new pathogenscould “emerge” relatively quickly. Suchevents probably contributed substantial-ly to the evolution of the highly virulentO157:H7 strain of E. coli.

Bacteria are constantly mutating, andenvironmental forces may select a muta-tion that confers an advantage in the faceof the environmental challenge. Themany and varied environmental stressescommonly include starvation, high orlow pH, oxidation, heat, cold, and os-motic imbalance. The genetic responseto one stress may protect the microbefrom a different stress, a phenomenonknown as cross protection.

Bacteria possess specific genetic loci,particularly in “contingency genes,” thatare highly mutable when compared with“housekeeping genes” that are relativelystable. Contingency genes help a mi-crobe successfully interface with envi-ronmental change, while housekeepinggenes run the routine cellular machinery.Genetic variation can occur in manyways, including increased mutation.

Stresses can create microorganisms withgreatly enhanced mutation frequencies(1000-fold or more). The large numberof different mutations increases thechance of a mutation that will enablesurvival in the stressful environment.These hypermutable microorganismsalso may more readily share DNA withother microorganisms, even remotely re-lated species. Horizontal transmissionof genetic material from one microor-ganism to another can result in quantumjumps in evolution. Gene transfer be-tween separate lineages of a bacterialpathogen can lead to the emergence ofaltogether new pathogens. Recently se-quenced bacterial genomes reveal moreextensive exchange of genetic materialbetween species than had been expected.

Complex Drivers of Change

No matter how emergence is defined,it becomes clear that the interrelation-ship of pathogen, host and environmentplays a key role in microbiological foodsafety. A number of factors will drive theemergence of new food safety concerns,including changes in the characteristicsof the consuming public, changes in thefoods we manufacture and sell, changesin the hazards themselves, and changesin the ability of public health officials toidentify illnesses as foodborne and totrace the illnesses to their food source.

Host Factors

Changing demographic characteris-tics of consumers affect the number ofcases of foodborne illness. As theworld’s population continues to grow,constant rates of disease will increase thetotal number of cases. In addition, theproportion of the population that is athigh risk of foodborne infections, illness,and death is rising. Factors that increasethe impact of foodborne diseases includeage, chronic diseases, immunosuppres-sive conditions, and pregnancy. The im-mune system functions less effectively inthe elderly, putting them at greater risk. Agrowing proportion of our population isimmunocompromised due to HIV infec-tion, cancer chemotherapy, and drugsused to combat rejection of transplantedorgans. Larger numbers of people withchronic diseases, like diabetes, now livelonger and also are at increased risk offoodborne diseases.

Other consumers are at elevated riskof foodborne illness because of the in-


creased likelihood of exposure. This ele-vated risk is sometimes due to food pref-erences based on ethnicity, age, or gender.Young adult males, for example, aremore likely to eat inadequately cookedground beef.

Environmental Factors

Food industry economics and tech-nical factors continue to drive consolida-tion in primary agricultural productionand food processing. Although thishelps reduce costs and assure uniformquality, when a large lot of a contaminat-ed food enters distribution, the scope ofthe resulting outbreak is increased.

Global sourcing also carries the po-tential to move pathogens and toxinsfrom areas in which the pathogen is in-digenous to areas in which it has not pre-viously existed. Unfamiliarity compli-cates diagnosis and containment and canresult in outbreaks that become quitelarge before they are recognized. Haz-ards are truly mobile, and our food safe-ty programs must be very agile to reduceour risk.

Even slight increases in environmen-tal temperatures can significantly affectthe risk of foodborne illness. The growthof algae in surface waters, estuaries, andcoastal waters is sensitive to temperature.About 40 of the approximately 5,000known species of marine phytoplankton(algae) can produce biotoxins, whichmay reach human consumers throughshellfish. Warmer sea temperatures canencourage a shift in species compositionof algae toward the more toxic di-noflagellates. Upsurges of toxic phy-toplankton blooms in Asia are stronglycorrelated with El Niño, and in the Unit-ed States, paralytic shellfish poisoningand other marine biotoxin-induced dis-eases have been associated with shellstock harvested from beds traditionallyconsidered safe.

Consumer desires drive food prod-uct development. Food manufacturersrespond to desires for “fresher” food,low fat products, or ready-to-eat foodsby developing new processes or refor-mulating existing products. Changes inthe food processing environment orproduct formulations can create a newniche for pathogenic microorganisms.Producing familiar foods in nontradi-tional sites also may lead to introduc-tion of new food hazards; such was thecase with the first outbreaks of cy-closporiasis associated with raspberries

imported into the United States andCanada from Guatemala.

Pathogen Factors

Stable and accurate transfer of genet-ic information from parent to offspringis essential for the preservation of a spe-cies. However, keeping pace with anever-changing environment also requiresvariability. When naturally occurringbacteria, for example, divide, most of theoffspring look and act just like their par-ents, but a small proportion of the off-spring mutate, increasing the chance thatsome might survive in a new, hostile en-vironment. If the environment has notchanged, these new strains may not sur-vive, but this natural occurrence makes italmost certain that traditional food pro-cesses will fail to deliver their predictedlevel of safety at some point. This is partof nature and happens without humanintervention.

In addition to this unstimulated hy-permutability, the food production andprocessing environment can increase therate of change in foodborne pathogens.Bacterial stress responses may increasepathogen virulence, and other actionscan affect which microorganisms surviveand dominate in a particular environ-ment. For example, use of antimicrobialagents during livestock production mayselect resistant strains from a back-ground of susceptible microorganisms,increasing the likelihood that the micro-organisms in a food are resistant to thoseand related antimicrobials. Even if thesemicroorganisms are not pathogenic, theycan share the genetic material that en-ables them to resist antimicrobials withpathogenic microorganisms in the hu-man gut, producing pathogens that causeinfections that may be difficult to treat.

Through improved laboratory tech-niques, scientists are identifying adversehealth effects associated with ever-small-er levels of exposure to natural and an-thropogenic substances. New ELISA andradioimmunoassays for various myc-otoxins are pushing tolerances for com-mon mycotoxins down and are findingmore poorly characterized mycotoxinsin a broad array of commodities. Ourunderstanding of biology, however, isnot keeping up with our laboratoryskills, and judging the public health sig-nificance of “positive” laboratory resultsis becoming more difficult. Unlockingthe human genome and the genomes ofpathogenic microorganisms, however, is

beginning to clarify the very basis of theinteraction between humans and the mi-croorganisms that can make us sick.

Industrialized and developing nationshave improved their ability to conductsurveillance and investigate outbreaks ofdisease in humans during the last two de-cades of the 20th century, and this progressis continuing. In addition, the combina-tion of molecular biology and electronicinformation technology in centers aroundthe globe is refining the quality of the datathat links cases together around commonexposures. National and multinationalnetworks of collaborators are beingpulled together with help and guidancefrom the World Health Organization andthe Food and Agriculture Organizationof the United Nations, to facilitate therapid sharing of data.

The process of ensuring the safety offood is dynamic as well as complex.Changes in the types of food that are con-sumed, the geographic origins of foodproducts, and the ways in which differentfoods are processed affect both the risk forcontamination and the adequacy of safetymeasures. The processes used to controlfoodborne hazards to limit the potentialfor foodborne disease must be continu-ously reviewed and judged against newinformation and new hazards. Advancesin risk assessment methodologies andavailability of additional data make it pos-sible to integrate information from thevarious stages in the food productionprocess for those foodborne hazards weknow about. This capability can be usedto identify particular steps in the foodsupply system for targeted intervention tocontrol hazards and prevent disease. It ismore difficult to provide specific adviceon how to prevent foodborne hazards thathave not yet been identified.

Framework for Food SafetyManagement

Our existing approach to food safetymanagement has given the United Statesan extremely safe food supply. However,estimates of the incidence of foodborneillness clearly show that, in some cases,the existing approach to control is inade-quate. The complex, ever-changing na-ture of microbiological food safety guar-antees that new challenges will continueto emerge.

Microbiological food safety is not anissue only for microbiologists. Just as thefarm-to-table approach to food safetyhas provided an overall picture of food

13EXPERT REPORT

safety management, many scientific dis-ciplines contribute to our knowledgeabout food safety. The scientific commu-nity must pull together multidisciplinaryteams that combine microbiology, epide-miology, genetics, evolutionary biology,immunology and other areas of expertiseto enhance our understanding of the in-terrelated factors that drive emergingfood safety issues.

Just as the issues change over time, so

too must our management strategies andour regulatory framework. Regulatoryprograms must be flexible to address is-sues as they arise and to benefit from sci-entific advances. Continued research willimprove our understanding of the com-plex factors that cause foodborne illness,and surveillance programs will gatherdata to document the effectiveness of ourcontrols and identify new problems asthey emerge. A science-based food safety

management framework should use foodsafety objectives to translate data aboutrisk into achievable public policy goals.

Microbiological food safety issueswill continue to emerge. Although wecannot expect to accurately predict thedetails of complex changes such aspathogen evolution, scientific knowledgecan be used to identify the areas of great-est concern, so that we may be ready torespond as issues arise.

Science of PathogenicityPathogenicity is the ability to cause

illness. Because pathogens are living

organisms that rapidly adapt and

evolve, the methods they use to cause

illness are never static. Pathogen

evolution is continuous and is driven

by a variety of forces, only some of

which relate to human activities. The

continual evolution of foodborne

pathogens forces us to change food

production processes and products to

maintain and improve microbiological

food safety. Control strategies that

were once effective may not remain so

if the pathogens become tolerant.Fortunately, genomic and improved

molecular and imaging techniques havevastly expanded scientific understandingof the organisms that cause foodbornedisease. These tools also have enabledscientists to attribute foodborne diseaseto microorganisms that had not previ-ously been identified as pathogenic or asfoodborne.

However, researchers still have manyquestions to answer: What makes onestrain of a microbe pathogenic whenother microorganisms within the samespecies are not? How do microorganismsbecome pathogenic? Understandingpathogenicity is not just necessary fordeveloping methods to treat illness but isalso needed for pathogen control.

Pathogen control includes preven-tion of food contamination, eliminationfrom the food, reduction to an acceptablelevel, or prevention of multiplication andtoxin formation. In addition, when sci-

entists understand how a particularpathogen is able to cause illness, thenthey can look for ways to disrupt thisprocess and render the microorganismharmless or find treatments that mitigateillness. The very factors that createpathogenicity are opportunities for con-trol. Just as the pathogens adapt andevolve, so can our understanding andour response.

Nomenclature

Traditionally, the first step in under-standing foodborne pathogens has beento develop a system of nomenclature anddescriptions of microorganisms withinthis system. For the purposes of study,scientists try to classify microorganismsbased on a set of common characteristicsthat sometimes include presumed patho-genic attributes. However, as our scientif-ic understanding has improved, theinitial classifications often no longerpresent a full and accurate picture. Whennomenclature becomes outdated, ques-tions are raised about the scientific valid-ity of regulatory policies based on classi-fication schemes that predict pathogenic-

ity poorly or that cluster pathogenic andnonpathogenic microbes together underone name.

The names used to describe variousmicrobiological foodborne pathogensare based on systematic nomenclature. Itis common practice to identify an organ-ism based on its genus and species. Toprovide additional detail, classificationssuch as subspecies, strain, serotype,pathovar, and toxin type may be used(see Table 3).

In the past, the classification of mi-croorganisms has relied primarily onstructural (morphological) and func-tional (physiological) characteristics.For example, shape is a morphologicalcharacteristic, and the pattern of en-zymes produced is a physiological char-acteristic. The commonly used morpho-logical distinctions of gram-positive andgram-negative are based on differencesin cell wall composition. Morphologicalfeatures remain the primary means ofclassification for molds. Although mor-phological characteristics can classifybacteria into broad categories (e.g.,spherical, rod-shaped, or curved), bacte-ria generally have few morphological fea-

Nomenclature

Family

Genera/genus

Species

Subspecies

Serovar

Table 3. Classic Microbial Nomenclature

Example 1

Enterobacteriaceae

Escherichia

coli

O157:H7

Example 2

Enterobacteriaceae

Salmonella

enterica

enterica

Typhimurium

Example 3

Mycobacteriaceae

Mycobacterium

avium

paratuberculosis


Nomenclature ofSalmonella

Bacteria in the genus Salmonellaare important contaminants in foodand water. Recently, efforts have beenmade to simplify the nomenclatureof Salmonella. Instead of using sero-type designations (of which there aremore than 2,000) incorrectly as spe-cies designations, most Salmonellaspecies are now classified as Salmo-nella enterica and then further iden-tified by serovar (e.g., Salmonella ty-phimurium becomes S. enteri-ca serovar Typhimurium, seeFig. 3). For convenience, thespecies (enterica) designationis frequently eliminated, leav-ing Salmonella Typhimurium.

The diverse population oforganisms sharing the samename has complicated effortsto study Salmonella, but re-cent advances illustrate thatmost Salmonella are actuallyquite similar to each other, es-pecially at the virulence factorlevel.

Salmonella typically causethree diseases in humans:

gastroenteritis (caused by S. Typhimu-riuim, S. Enteritidis, and others); enter-ic fever (S. Typhi and S. Paratyphi); andan invasive systemic disease (S. Choler-aesuis). In the United States, nonty-phoidal Salmonella account for an esti-mated 1.3 million illnesses annually,with several hundred deaths (Mead etal., 1999). The U.S. incidence of typhoidfever is relatively low—approximately700 cases annually, mainly as a result ofinternational travel. Worldwide, the in-

cidence of typhoid fever is declining(due, in part, to better distribution ofsafe water and successful vaccines),while the incidence of nontyphoidalsalmonellosis is increasing rapidly.A portion of this increase correlatesto changes in the food productionenvironment that may have givenSalmonella the opportunity to spreadand to contaminate foods that aredistributed through large complexnetworks.

tures that are readily discernible by lightmicroscopy or that are stable under abroad range of environmental conditions.To create a system with more precision,taxonomists were forced to base classifica-tion schemes on both morphologicalcharacteristics and physiological charac-teristics that generally reflect the biochem-ical diversity among bacterial species.

As the available techniques and tech-nology advanced, scientists found newways to classify microorganisms. The ad-vent of ribosomal RNA (ribonucleicacid) sequencing began a new era of tax-onomy (Woese et al., 1990). rRNA ispresent in organisms in all kingdomsand performs the same essential func-tions in all organisms. rRNA evolvesslowly so it serves as the ideal evolution-ary clock. Scientists soon developed largedatabases of rRNA sequences used toclassify new species (Olsen et al., 1992).This era also produced many of the cur-rent DNA (deoxyribonucleic acid) hy-bridization strategies used to detect andidentify pathogenic microorganisms in

food and clinical samples (King et al.,1989). Although the use of rRNA se-quences as a chronometer to measure re-lationships among species has disruptedtraditional groupings based on pheno-typic characteristics, many of the taxa ofimportance to food microbiology remainintact, with some modifications amonggroupings of certain species. For exam-ple, Campylobacter pylori, initiallynamed pyloric campylobacter for itssimilarity to Campylobacter jejuni, wassubsequently renamed Helicobacter py-lori (Dubois, 1995).

The era of microbial genomicsreached full swing in the 1990s. Usingseveral available microbial genome se-quences, scientists have been able to com-pare the evolutionary histories of differentbacterial “races” (phylogenies) to the uni-versal phylogenetic tree predicted fromrRNA sequences. These efforts resulted inthe disappointing discovery that the ge-nome-based phylogenies were frequentlydiscordant with rRNA predictions (Brownand Doolittle, 1997).

Mining the sequences for the relative“time” that specific segments appearedwithin the genome revealed the reasonsfor the discordance: significant portionsof microbial genomes have been ac-quired through gene transfer from othermicroorganisms (Lawrence andOchman, 1998). Examining large sets ofvirulence genes on contiguous segmentsof DNA (known as pathogenicity is-lands) also demonstrated that genesconferring virulence characteristics areoften some of the most recent acquisi-tions among the genomes of pathogenicspecies (Hacker and Kaper, 2000). Asidefrom the impact on nomenclature, thisnew information has profoundlychanged scientific thinking about theevolution of virulence. The ability of apathogen to suddenly obtain a criticalvirulence factor through genetic ex-change is at odds with the idea of slow,gradual evolution of virulence.

Now that scientists know that micro-bial genomes change more rapidly thanpreviously believed, the concept of bac-

Genus Salmonella

enterica

entericasalamai arizonidiarizoni houtenae indica bongori

Enteritidis Typhi Paratyphi CholeraesuisTyphimurium

Species

Subspecies

Serovar

Fig. 3. Nomenclature of Salmonella

15EXPERT REPORT

terial “species” is in flux. In terms of di-versity, genome size may vary by 20%among subpopulations of a single spe-cies (for example, see Bergthorsen andOchman, 1998). This variability cancover as many as 1-2 megabases of DNAthat code for thousands of genes, whichcan confer unique characteristics—suchas virulence—on specific subpopula-tions. Still, a subset of genes that definesignature characteristics of the speciesmust be shared among all members ofthat species.

These contrasting views of what con-stitutes a bacterial species challenge theconcepts that underlie microbiologicalcriteria and testing for the control ofpathogens in food manufacturing. Insome cases, the potential to cause food-borne disease can be characteristic of anentire species, such as S. enterica, but inother cases it may be a consequence ofrecently evolved virulence characteristicsamong specific subpopulations of a spe-cies, such as Escherichia coli O157:H7.The sensitivity of genome-based finger-printing methods has even called intoquestion the equality of virulence in ge-netically distinct subpopulations of E.coli O157:H7 (Kim et al., 1999). As an-other example, it is questionable whetherall microorganisms in the Listeria mono-cytogenes species are capable of causingdisease in humans. Even though eachdistinct genetic lineage of the species ap-pears to harbor the known virulencegenes, mounting phylogenetic evidenceindicates that virulence characteristicsdiffer (Rasmussen et al., 1995; Wied-mann et al., 1997).

New scientific information providesthe opportunity to better target thepathogenic subpopulations within a spe-cies. Present day approaches to controlmust be modified to be consistent withthe information presently available. Asimproved detection and identificationmethods enable scientists to differentiatebetween virulent and avirulent organ-isms, we should be able to allocate ourrisk management resources more effi-ciently, focusing only on pathogenic or-ganisms in our food supply.

Virulence

If pathogenicity is a microorganism’sability to cause disease, then virulencecan be considered the degree of pathoge-nicity. Some pathogens are particularlyefficient at causing clinically significantdisease (highly virulent), while others

may cause only minor effects in a smallnumber of cases (less virulent). Viru-lence is related to the severity of diseaseand the number of ingested pathogens(infectious dose) required to cause ill-ness (see Fig. 4). Within the Escherichiacoli species, for example, there is greatcontrast in virulence. Enterohemorrhag-ic E. coli (e.g., O157:H7) can cause sig-nificant and severe illness even if verysmall numbers of cells (10 or less) areingested, whereas enterotoxigenic E. colirequire an estimated 100 million to 10billion cells to cause a relatively mild setof symptoms (FDA/CFSAN, 2002). Theunderlying reasons for observed differ-ences in virulence among various patho-gens, and even within a single species ofpathogen is difficult to state with abso-lute certainty, as at least two dynamicsare operative, the microorganism and thehost. Even less virulent pathogens cancause serious illness in debilitated hosts.Likewise, it is probable that fewer patho-gens need to be ingested by a debilitatedhost than a healthy host to cause infec-tion if all else is equal.

The infectious dose is different foreach pathogen. As stated above, E. coliO157:H7 is infectious in very low num-bers. There may be several possibleexplanations for this, including:(1) O157:H7 is more acid tolerant andtherefore fewer cells are killed by gastricacidity, (2) O157:H7’s virulence may beinfluenced by intestinal flora via quorumsensing (see sidebar, p. 16), (3) its com-bined virulence factors simply make itmore adaptable to the intestinal lumen,or (4) its virulence factors are more po-tent.

The situation among Salmonella ismore complex. Prior to the 1980s, con-ventional wisdom held that large num-bers of Salmonella were necessary for in-fection. Human feeding trials of volun-teers from penal institutions with several

different Salmonella sub-types certainly suggestedthat numbers >100,000cells were required for ill-ness (D’Aoust, 1985).However, outbreaks in-volving cheddar cheeseand chocolate were ap-parently caused by veryfew cells (<10 total cellsfor cheddar and 50-100total cells for chocolate).It has been theorized thatthe lipid content of ched-dar cheese and chocolate

facilitated the pathogen’s transit throughthe stomach acid. Very little is knownabout specific food matrices and the ef-fect on microbial survival and virulence.Some salmonellae appear to be acid sus-ceptible, and others appear acid tolerant,but additional research will be requiredto better answer this question. Antacidconsumption has been shown to affectthe apparent virulence of certain patho-gens such as Salmonella and L. monocy-togenes, and this suggests that stomachacid, and a microorganism’s tolerance tostomach acid, plays some role in definingthe infectious dose.

Nowhere in current food microbiol-ogy is the issue of infectious dose ascontentious as in the case of L. monocy-togenes. Currently, the detection of L.monocytogenes in ready-to-eat foods isgrounds for removing the foods from themarketplace. L. monocytogenes only veryrarely affects healthy individuals, but im-mune-compromised individuals are sus-ceptible to illness and sometimes death.Of all the known pathogens, L. monocy-togenes has one of the highest mortalityrates, approaching 20% (Mead et al.,1999).

The question is whether there is aningested dose of L. monocytogenes that istolerable by the vast majority of the pop-ulation (i.e., is there a realistic dose thatcan be tolerated in food?). Studies on vir-ulence gene and ribotype pattern haveseparated L. monocytogenes into threelineages, and each lineage appears to dif-fer in its pathogenic potential for hu-mans (Wiedmann et al., 1997). Studieson L. monocytogenes isolated fromsmoked fish using the same methods fur-ther suggest that some subtypes of L.monocytogenes in ready-to-eat foods mayhave limited pathogenic potential for hu-mans (Norton et al., 2001). If these vari-ations in pathogenic potential exist assuggested, it implies that more special-

Fig. 4. Virulence and Foodborne Illness

Number ofOrganisms

Few Many

Severe/Many

Severity of Illness/Number of Cases

Minor/FewLow

Virulence

High


ized testing might be prudent beforefoods are rejected for the presence of L.monocytogenes that may not be patho-genic.

The inherent ability to cause diseaseis the result of virulence factors encodedat the genetic level (Finlay and Cossart,1997; Finlay and Falkow, 1997). Manydiverse characteristics are considered vir-ulence factors. If these factors are miss-ing, the microorganism would be expect-ed to be less virulent or avirulent. Someexamples of virulence factors include:• toxins (molecules secreted by thebacteria that affect host cell processes),such as cholera toxin;• adhesins (molecules that enablepathogens to adhere to host surfaces),such as fimbriae; and,• invasins (molecules that enablepathogens to actively enter into a hostcell (invasion) where they can exist as anintracellular pathogen), such as thoseused by Shigella and Salmonella.

Most pathogens have a variety of vir-ulence factors that assist in host coloni-zation and disease. The repertoire of

Quorum SensingNot long ago, bacteria were

thought to lead a solitary existenceand either live or die as a single cell. Ithas been established that intercellularcommunication is fairly commonamong bacteria, and that this intercel-lular communication can lead to co-ordinated activities once thought tobe the exclusive domain of multi-cel-lular organisms. It was not surprisingthen that researchers asked whetherintercellular communication was in-volved in virulence, and it was alsonot surprising that genetically wellcharacterized foodborne pathogenssuch as Salmonella and Escherichiacoli would be investigated.

Intercellular communicationamong bacteria is carried out bysmall molecules called autoinducers.The theory is that at very low popu-lation densities, there is insufficientautoinducer in the environment tobe detected by those bacteria present,but that when some “threshold”number of bacteria is reached, auto-inducer is present in sufficient con-centration to trigger some activity inbacteria capable of detecting the au-

toinducer. The terminology to describethese events is quorum sensing.

Quorum sensing has been demon-strated in E. coli and Salmonella Typh-imurium (Surette and Bassler, 1998),and more recently, the role of quorumsensing in the virulence of enterohem-orrhagic (EHEC) and enteropathogen-ic (EPEC) E. coli has been elucidated(Sperandio et al., 1999). The hallmarkintestinal lesion caused by EHEC andEPEC is called attaching and effacing,and is coded by the Locus of Entero-cyte Effacement (LEE). This pathoge-nicity island codes for a type III secre-tion system, as well as other virulencefactors such as the intimin intestinalcolonization factor and the translocat-ed intimin receptor protein. It is possi-ble that the low infectious dose of E.coli O157:H7 is in part because thepathogen is induced to colonize the in-testine by quorum sensing of signalsfrom resident, nonpathogenic E. coli inthe intestine of the host.

Quorum sensing appears to regu-late the virulence factors of a wide vari-ety of plant and animal pathogens(Day and Maurelli, 2001). Unlike oth-er enteric pathogens, the signalling sys-

tem in Shigella flexneri does not reg-ulate virulence. There may besound ecologic reasons why someenteric pathogens regulate virulencewith quorum sensing systems andothers do not. Quorum sensing sys-tems are not limited to the gram-negative bacteria, and notable gram-positive bacteria, such as the patho-gen Staphylococcus aureus and itsnumerous virulence factors are un-der the control of intercellular sig-nals (de Kievit and Iglewski, 2000).

Obviously it is important thatwe gain a better understanding ofthe regulation of virulence via quo-rum sensing systems. One obviousquestion is: Does quorum sensingoccur in or on foodstuffs, and if so,is it a factor in the virulence of food-borne pathogens? Many pathogenicbacteria have evolved a chemicallanguage, and it would behoove usto learn and understand that lan-guage. It may be possible to exploitthese intercellular communicationpathways to reduce virulence, or usethem as targets of novel antimicro-bial substances (de Kievet and Ig-lewski, 2000).

functional virulence factors will dictatewhich host a pathogen will colonize,which tissues will become infected, andultimately which disease symptoms willoccur.

Many bacterial pathogens are genet-ically similar to common bacteria thatinhabit the host but do not cause dis-ease under ordinary circumstances(commensal organisms). Thus, apathogen is often a “genetically en-hanced” organism: a common organ-ism that contains a specialized collec-tion of virulence factors. This compli-cates efforts to control foodbornepathogens by diffusing resources acrossa large group of microorganisms whenonly a subset can cause foodborne ill-ness. The ability to share the geneticmaterial that encodes virulence factorswithin food-producing animals, the en-vironment, or the human gastrointesti-nal system significantly complicatescontrol (see Evolution, p. 19). Each ofthe several steps in the progression ofpathogenic foodborne disease requiresone or more virulence factors.

Toxins are perhaps the best under-stood family of bacterial virulence fac-tors, presumably because they are oftensecreted into the area surrounding thebacteria where they can be isolated andstudied. Toxins have specific mechanismsto recognize host cell surface receptorsenabling them to transport themselvesinto the host cell. Some toxins can mod-ify specific cellular targets to affect fluidsecretion, cytoskeletal structure, or evennerve functions. Others insert them-selves into the host cellular membrane,resulting in cell disintegration or disso-lution (lysis). Despite the vast numberof cellular targets, toxins can be classifiedinto families based on their structureand function.

To cause disease, most pathogens re-quire the ability to adhere to host surfac-es following ingestion. Mammalianhosts have several nonspecific defensesdesigned to prevent colonization by in-hibiting pathogen attachment, such asperistalsis, the mucocilliary system, andeven cell sloughing (see section on hu-man host, p. 28). Bacterial pathogens

17EXPERT REPORT

have a wide variety of adhesins, usuallyglycoproteins or glycolipids, on their sur-face that target specific host cell mole-cules. These adhesins are usually fila-mentous- or hair-like structures (fimbri-ae or pili) or proteins without the hair-like fimbriae (afimbrial adhesins) on thebacterial surface. Pathogens usually en-code several adhesins that help dictatetissue and host specificity. For example,Salmonella use a combination of at leastsix different adhesins that contribute tointestinal colonization. Many pathogensalso have the ability to adhere to the areaoutside and surrounding the cell (extra-cellular matrix), which contributes tohost colonization. Recently, it has be-come apparent that some pathogens,such as E. coli O157:H7, actually inserttheir own receptor into host cells, whichthen enables them to adhere to the out-side surface of the host cell (Kenny et al.,1997).

Many foodborne pathogens have theadditional capacity to invade host cells.For example, Salmonella, Shigella, Yersin-ia, and Listeria are invasive organisms,capable of penetrating the intestinal epi-thelial barrier as part of their pathoge-nicity. The ability to invade provides the

pathogen with additional environmentsto colonize that protect the pathogenfrom host defenses present within the lu-men of the intestinal tract. The virulencefactors that facilitate invasion may be assimple as a single surface protein, such asthose used by Yersinia (invasin) or Liste-ria (internalin). In other cases, the fac-tors may be complex, requiring the deliv-ery of several proteins into the host celland resulting in profound cell surfacechanges such as membrane ruffling andmacropinocytosis before bacterial uptakeoccurs.

Numerous bacterial pathogens, in-cluding Salmonella and Shigella, have acomplex secretion apparatus, known as aType III secretion system, that deliversnumerous bacterial proteins into the gel-like fluid inside the host cell (cytosol) tomediate invasion. Type III secretion sys-tems are major virulence mechanismsfor many gram-negative pathogens(Hueck, 1998). Salmonella actually havetwo systems (one for invasion, one forintracellular survival), while Yersinia useone to avoid isolation and destructionwithin the cell (phagocytosis).

Although invasion into a host cellmay protect the pathogen from some

host defenses such as antibodies, it alsoexposes the pathogen to another mecha-nism that kills most bacteria: fusionwith the lysosome. Lysosomes are smallsacks within the cell that contain en-zymes that “digest” substances—such asold DNA, proteins, or lipids—into smallpieces for reuse by the cell. Invasivepathogens, including bacteria and pro-tozoan parasites, have devised severalstrategies to avoid lysosomal fusion fol-lowing invasion. For example, a secondType III secretion system alters the tar-geting of the membrane-bound vacuolesurrounding Salmonella, causing thevacuole to diverge from the standardpathway that would deliver it to a lysos-ome. The salmonellae then live in theprotected intracellular environment ofthe vacuole (see sidebar below). In con-trast, Shigella and L. monocytogenes useenzymes to degrade the vacuolar mem-brane, thereby releasing the pathogeninto the host cytosol. Once present inthe cytosol, the pathogens use the struc-tural proteins in the cell around them topropel themselves within the cell andinto a neighboring cell. Finally, patho-gens such as Yersinia and enteropatho-genic E. coli use their Type III systems to

Virulence ofSalmonella

Recently, scientists have made sig-nificant advances toward understand-ing the molecular mechanisms of thepathogenicity of Salmonella. Similarto pathogenic E. coli, Salmonella havea collection of virulence factors, in-cluding factors needed to adhere tointestinal surfaces, invade host epithe-lial cells, and survive within phago-cytic cells. It has been estimated thatSalmonella contain more than 200virulence factors, encoded in at leastfive pathogenicity islands in additionto a virulence plasmid and manypathogenicity islets.

The complex mechanisms usedby Salmonella species to adhere tointestinal cells continue to be studiedand debated. Many adhesins havebeen reported, and because theyappear redundant, defining the roleof each in the disease process hasbeen difficult.

Significantly more is known

about the mechanisms Salmonella use toenter epithelial cells. These intestinalcells are not phagocytic, and thus do notnormally ingest microorganisms. How-ever, Salmonella use a Type III secretionsystem to inject several bacterial factorsinto the host cell. These molecules affectnormal cellular processes, includingthose that control the actin cytoskeletonand other signal transduction pathways.The end result is significant actin rear-rangement beneath the adherent bacteri-um, membrane ruffling, macropinocyto-sis, and engulfment of the bacterium intoa membrane-bound vacuole. UnlikeShigella, Salmonella remain within thevacuole to survive and proliferate. Sal-monella redirect vacuole movementwithin the host cell so that the vacuole itinhabits does not fuse with lysosomes.

Salmonella also form specializedstructures that enable it to survive andreplicate within phagocytic cells. Salmo-nella species have another Type III secre-tion system that encodes several factorsneeded for survival in this intracellularcompartment. Finally, Salmonella have a

virulence plasmid that provides thefactors needed for prolonged survivalwithin the host.

S. Typhi is different from most S.enterica serovars in that it has a capsuleand does not encode the virulenceplasmid. Because of its systemic infec-tious nature, S. Typhi presumably hasadditional virulence factors that con-tribute to the different nature of thedisease. Although Salmonella speciesare fairly close relatives of E. coli, theyhave many additional genes that pre-sumably account for their virulence.Scientists know little about the role formost of these additional genes and thefactors they encode.

Significant progress has been madewith vaccines for S. Typhi. Two currentvaccines—one live-attenuated strain,the other a component vaccine—havesignificantly decreased the levels of ty-phoid fever worldwide. However, littlesuccess has been made toward control-ling nontyphoidal salmonellae, whichcontinue to cause increasing numbersof illnesses worldwide.


avoid being pulled into the cell andtherefore blocking phagocytosis, allow-ing them to remain attached to the exte-rior of the cell (extracellular pathogens).

Because expression of virulence fac-tors at the correct time and place is criti-cal, bacterial virulence factors are tightly

regulated by a variety of control mecha-nisms and regulatory circuits. Successfulpathogens are very adept at sensing theirmicroenvironment, and virulence factorexpression often relies on environmentalparameters to ensure coordination andappropriate expression.

The continual genetic exchange be-tween bacteria ensures that pathogenswill continue to evolve as they acquiredifferent combinations of virulence fac-tors. This evolution is a natural process,although it can be enhanced and facili-tated by human actions.

Pathogens Are MoreThan Just Bacteria

Although bacteria are perhapsthe first type of microorganism thatcome to mind when discussing mi-crobiological food safety, they are byno means the only pathogenic food-borne microorganisms.

Human Enteric Viruses

Viruses of concern to humanhealth that are known to be transmis-sible through foods are shed in the fe-ces of infected humans and transmit-ted via the fecal-oral route. Of theseviruses, hepatitis A virus causes themost serious recognized foodborneviral infection, whereas the Norwalk-like gastrointestinal viruses (NLVs)are the most prevalent. According torecent epidemiological estimates, theNLVs account for over 60% of cases,33% of hospitalizations, and 7% ofdeaths among all of the illnesses thatare attributable to known foodbornepathogens (Mead et al., 1999). Hu-man enteric viruses have propertiesthat make them quite different fromthe common bacterial agents of food-borne disease. As obligate intracellu-lar parasites, they require live mam-malian cells to replicate. To protectthe viral genome from inactivationoutside of infected cells, virus parti-cles have properties that make themenvironmentally stable to the ex-tremes of pH and enzymes present inthe human gastrointestinal tract.This stability enables virus particlesto survive a variety of food produc-tion, processing, and storage condi-tions making virtually any type offood product a potential vehicle fortransmission of viral pathogens(Jaykus, 2000a). The inability of hu-man enteric viruses to replicate infoods and the fact that they are gener-

ally present in low numbers does not en-sure product safety, as the infectious doses(100-102 infectious units) are presumed tobe low (Iversen et al., 1987; Jaykus, 2000b;Moe et al., 1998).

Three major routes for viral contam-ination of foods have been recognizedand include: (1) shellfish contaminatedby fecally polluted marine waters;(2) human sewage pollution of drinkingand irrigation waters; and (3) ready-to-eat and prepared foods contaminated asa result of poor personal hygiene of in-fected food handlers (Jaykus, 2000b). Inaddition, the NLVs have been shown tobe spread by aerosolization of vomitusand through fomites (Marks et al., 2000;Patterson et al., 1997).

Protozoan Parasites

Like viruses, parasitic protozoa repli-cate in the intestines of infected hostsand are excreted in the feces. However,their host range is wider than viruses, be-ing able to replicate in human and non-human animal hosts. Since they aretransmitted primarily by the fecal-oralroute, the major source of contamina-tion for foods and water is through con-tact with human and animal fecal pollu-tion. This contamination may occur di-rectly, through contaminated meat car-casses or poor personal hygiene practicesof infected food handlers, or indirectly,via contact with fecally contaminatedwaters or other cross-contaminationroutes. Like the viruses, the parasiticprotozoa (in the cyst or oocyst form) areenvironmentally inert, they do not repli-cate in foods, are extremely environmen-tally stable, resistant to many of the tra-ditional methods used to control bacteri-al pathogens, and have notably low infec-tious doses (DuPont et al., 1995; Haas,1983). Although transmitted by the fecal-oral route, direct person-to-persontransmission of parasitic protozoa is un-likely because excreted oocysts requiredays or weeks under favorable condi-

tions to sporulate and become infectious.Since 1981, enteric protozoa have be-

come the leading cause of waterborne dis-ease outbreaks for which an etiologicalagent could be determined (Moe, 1996).Although considerably less information isavailable about their importance in food-borne disease, their potential for trans-mission by foodborne routes is increas-ingly recognized (Bean et al., 1990). Forinstance, from 1988-1992, seven food-as-sociated outbreaks of giardiasis, compris-ing 184 cases, were reported in the UnitedStates (Bean et al., 1996), and protozoanparasites such as Giardia and Cryptospo-ridium species may be present in shellfish-growing waters as a result of contamina-tion with animal farm runoff or as a re-sult of treated and untreated sewage in-put. The ability of bivalve molluscs toconcentrate Giardia and Cryptosporidiumhas been demonstrated by Toro et al.(1995). Cryptosporidium oocysts havebeen isolated in Eastern oysters harvestedfrom commercial sites in the ChesapeakeBay (Fayer et al., 1998).

The most common human entericparasitic infections in the United Statesare caused by Cryptosporidium parvumand Giardia lamblia. Cyclospora is alsoan emerging enteric protozoan that hasrecently been associated with the con-sumption of contaminated fruits (Ortegaet al., 1993). Large, community-widewaterborne outbreaks of parasitic proto-zoa are usually associated with surfacewater supplies that are either unfilteredor subjected to inadequate flocculationand filtration processes (Moe, 1996).Two large waterborne outbreaks have oc-curred in the United States within thelast 10 years (Hayes et al., 1989; MacKen-zie et al., 1994), including the largest re-corded waterborne disease outbreak inU.S. history (MacKenzie et al., 1994).

Marine Biotoxins

Marine biotoxins are produced byseveral dinoflagellate and diatom spe-

19EXPERT REPORT

cies (Epstein, 1998; Plumley, 1997),most of which are produced by the pro-liferation of algae in the form of harm-ful algal blooms (HABs) (Steidingerand Penta, 1999). Of the several thou-sand identified marine microalgae, atleast 80 species are known to be toxic orharmful (Baden et al., 1995). Reports ofseafood toxicity due to these biotoxinsdate back to the 1600s, have occurredworldwide, and recent evidence suggeststhat HABs are increasing. This indi-cates that marine biotoxins may indeedbe considered emerging pathogens(Anderson, 1994; Shumway, 1989).

The toxins produced by dinoflagel-late and diatom species are often classi-fied according to their mode of actionand resulting disease syndromes. Themost common marine HAB toxins arechemically characterized as alkaloids,polyethers, or substituted amines andare the end products of elaborate bio-chemical pathways (Plumley, 1997).Most have been classified as neurotox-ins although some hemolytic substanceshave also been identified (Baden et al.,1995). Although marine toxins arepharmacologically diverse, most exerttoxic effects through perturbations ofvoltage-gated sodium channels locatedin excitable membranes of neurons(Catterall, 1985; Manger et al., 1995).Binding of these toxins to receptor sitesleads to conformational changes of theion pore of the channel, thus alteringthe flow of ions controlling nerve sig-naling (Baden et al., 1995).

In general, the marine biotoxins aresmall, chemically complex, and highlypotent substances that tend to accumu-late in finfish and/or shellfish. Whenthe seafood is consumed by humans, itacts as a passive carrier of the toxins.The toxins are tasteless, odorless, andmost often heat and acid stable, whichmeans that routine food safety inspec-tion and food preparation techniqueswill not detect contamination, inactivatethe toxins, or prevent human disease

upon consumption (Baden et al., 1995).Symptoms of seafood-borne intoxica-tions may include gastrointestinal dis-tress, vomiting, headaches, neurologicdysfunction, paralysis and muscularpain. The degree of human toxicity isaffected by the health of the victim, theamount of toxin ingested, the rate oftoxin elimination, and the biotransfor-mation of toxins by enzyme systemswithin the body (Steidinger and Penta,1999).

Mycotoxins

A large group of mycotoxins, knowncollectively as trichothecenes, have beenassociated with acute human illness(Pestka and Casale, 1990). The most se-vere manifestation is alimentary toxicaleukia. Initial symptoms after a singlecontaminated meal include throat in-flammation, vomiting, diarrhea and ab-dominal pain; with continued ingestion,the illness can progress to oral hemor-rhaging, pneumonia, bone marrow de-pletion, and potentially death. Humanoutbreaks of gastrointestinal illness havebeen linked to millet contaminated by avariety of Fusarium that was capable ofproducing T-2 toxin and other trichoth-ecenes and also wheat and barley infect-ed with Fusarium that produced the tri-chothecenes deoxynivalenol (vomitoxin),nivalenol and fusarenon-x (Bhat et al.,1989; Luo, 1994).

Other mycotoxins have been associ-ated with acute disease. Aflatoxin inges-tion has caused acute liver inflamma-tion in Kenya (Ngindu et al., 1982) andIndia (Krishnamachari et al., 1975). Inthe early 20th century, cardiac beri-beri—an acute cardiac disease that pro-duces a rapid pulse, abnormal heartfunction, and low blood pressure lead-ing to respiratory failure and death—that occurred in Japan and other Asiancountries was linked to contaminatedrice containing the mycotoxin citreovir-idin (Ueno, 1974).

In general, except when largenumbers of people are affected, medi-cal professionals are unlikely to rec-ognize mycotoxicoses because of in-sufficient knowledge about these dis-eases and the lack of appropriate di-agnostic methods. Unlike microbialagents that can be detected by culturingand/or PCR amplification, mycotoxinsare metabolized and may no longer bepresent in the tissues of patientsshortly after the onset of acutedisease.

Mycotoxins produced by a widevariety of molds including Aspergil-lus, Penicillium and Fusarium maycause chronic toxicosis (Coulombe,1993). Chronic effects often resultfrom prolonged ingestion of low tomoderate levels of toxin that do notproduce symptoms of an acute illness,making the chronic effects difficult toattribute to contaminated food. Thelink between aflatoxin and human liv-er cancer has been well establishedthrough the use of clinical biomarkers(aflatoxin adducts), but other organs(kidney, spleen and pancreas) alsomay be affected. Another chronic dis-ease link that has been the subject ofconsiderable study is those diseasesinduced by the fumonisins, zearale-none, and trichothecene mycotoxins.Fumonisin levels in corn-based foodshave been statistically associated withan increased risk of human esoph-ageal cancer. Zearalenone has beenassociated with premature puberty.Trichothecenes modulate immunefunction, meaning that over timemycotoxicosis could reduce immuneresistance to infectious diseases, facil-itate tumor growth through reducedimmune function and cause autoim-mune disease (Bondy and Pestka,2000; Jackson et al., 1996). OchratoxinA is becoming closely linked to thekidney disease Balkan endemic neph-ropathy as well as tumors in the uri-nary tract and kidney.

Although foodborne pathogens havedeveloped a vast repertoire of virulencefactors, several common themes enablescientists to place many of these factorsin related families based on structure,function, and other characteristics. Un-derstanding why a pathogen is virulent

and how it overcomes host defenses andcauses illness creates opportunities forpreventing the infection or for mitigatingthe illness. The use of similar kinds ofType III secretion systems among manydiverse pathogens is one example of acommon mechanism that might be ex-

ploited for potential therapeutics andcontrol strategies.

Evolution

Pathogen evolution is a continuousbiological process that is influenced by


environmental factors. Historically, sci-entists thought that pathogens arosefrom sequentially cumulative muta-tions, gradually changing from avirulentto pathogenic. In the past few years,new evidence has shown that evolutionof pathogenicity progresses in quantumleaps that are driven by acquisition offoreign DNA (see Fig. 5). Analysis ofsome pathogen genomes has supportedthis new understanding and even ex-tended it by demonstrating that geneticexchange plays a larger role than previ-ously thought in genome composition.Comparing genomes of related patho-gens reveals many recently acquiredgenes scattered throughout the genome.

As discussed above, the ability tocause disease is attributed to specific vir-ulence factors. The genes encodingthese virulence factors are often foundon pathogenicity islands, that is, clus-tered together at specific loci on thechromosome or plasmids (Hacker et al.,1997). Evidence indicates that these ge-netic regions have evolved independentof the rest of the microbe’s genetic in-formation, i.e., they developed in a dif-ferent organism and were acquired as aset. These genetic regions usually have adifferent G+C content of DNA, oftenhave repetitive ends, and are often in-serted into or near tRNA genes. Particu-lar pathogenicity islands encode specificvirulence factors that in turn dictatewhich disease the pathogen may cause.

In gram-negative bacteria, Type IIIsecretion systems are often encodedwithin these pathogenicity islands. Forexample, enteropathogenic (EPEC) andenterohemorrhagic E. coli (EHEC),which both cause diarrhea, contain theLocus of Enterocyte Effacement (LEE)island, which encodes a Type III systemand other virulence factors essential fordisease (McDaniel et al., 1995). Howev-er, uropathogenic E. coli (which causeurinary tract infections) have a com-

pletely different pathogenicity islandinserted at exactly the same site(Hacker et al., 1997), which encodesan adhesin (P fimbriae) and a toxin(hemolysin), virulence factors need-ed for urinary tract colonization.Yersinia and Shigella species encodeType III systems on their virulenceplasmids rather than in their chro-mosomal DNA, but they have othervirulence attributes that are chro-mosomal and not in islands. In ad-dition to pathogenicity islands,smaller pieces of DNA (pathogenici-ty islets) also appear to move be-tween bacterial pathogens.

Bacteriophages, viruses that in-fect bacteria, also play a major rolein the movement of virulence factorsbetween pathogens. For example,Shiga toxin is a key virulence factorfor Shigella dysenteriae, which causesdysentery. The genes for toxin pro-duction are encoded on a phage thathas been incorporated into the chro-mosome. E. coli O157:H7 also con-tains genes for Shiga-like toxin(s), whichcause hemorrhagic colitis and contributeto disease progression to hemolytic ure-mic syndrome, sometimes characteristicof infection with this pathogen (Kaperand O’Brien, 1998). It is thought that thephage encoding this toxin infected anEPEC strain of E. coli and created a newpathogen, an EHEC. Under certain cir-cumstances, the phage DNA replicatesand breaks out, forming a new phageand releasing the toxin. Another exampleof the role that phage play in the evolu-tion of pathogens is found within Vibriocholerae (Waldor and Mekalanos, 1996).The cholera toxin is encoded within abacteriophage capable of moving be-tween strains, and, in this case, the recep-tor for this phage is one of the pathogen’smajor adhesins (a pilus). This arrange-ment ensures that the phage encodingthe toxin only infects bacteria that al-

ready possess an essential adhesin, thusensuring virulence.

Genomics has greatly facilitated ourunderstanding of pathogen evolution.For example, comparing the recentlysequenced E. coli O157:H7 genome to anonpathogenic E. coli reveals some sur-prising information (Perna et al., 2001).As expected, both strains share a com-mon genetic “backbone” of about 4.1million similarly arranged base pairs(see Fig. 6). However, the O157:H7 ge-nome contains an additional 1,387 newgenes in 1.37 million base pairs. Fur-thermore, the nonpathogenic E. coli has0.53 million base pairs (528 genes) thatare not in O157:H7. Perhaps the mostsignificant finding is that the additionalDNA in the pathogenic E. coli is distrib-uted among 177 different packets, eachof which was likely inherited throughan independent event. Researchers cal-culated that these two E. coli strainsshared a common ancestor approxi-mately 4.5 million years ago. Althoughscientists know the function of twopathogenicity islands (the LEE and theShiga toxin phage) within O157:H7 thatencode virulence factors, the functionof the additional genes within O157:H7remains to be determined, as does theircontribution, if any, to virulence. Evenamong O157:H7 strains, significantvariability in virulence can be detected,indicating that genome diversification

Gradual evolution

Avirulent Moderately virulent Highly virulent

Genetic exchangeof virulence factors

Type III secretionsystem acquired

Toxin-productiongene acquired

Fig. 5. Contrasting Views of Pathogen Evolution

Shared genetic material(4.1 million base pairs)

Genetic material only in E. coli O157:H7(1.37 million base pairs)

Genetic material not in O157:H7(0.53 million base pairs)

E. coli O157:H7 NonpathogenicE. coli

Fig. 6. Genetic Material in E. coli

21EXPERT REPORT

occurs rapidly.Genetic exchange between bacterial

species is obviously frequent and can sig-nificantly affect the evolution of patho-gens. The acquisition and spread of an-tibiotic resistance has been an easy wayto follow genetic exchanges, and the effi-ciency of this spread is demonstrated bythe pervasiveness of resistance in manybacteria not previously resistant to anti-biotics. As discussed above, the geneticmaterial that encodes virulence factorscan move between bacteria, with the po-tential to rapidly create a new pathogen,even from a commensal organism. Amajor unanswered question is oftenwhere these virulence factors originated.Despite the prevalence of Type III secre-tion systems within several bacterialpathogens, scientists have not found thesystem’s ancestor.

As in all organisms, the evolution ofpathogens is continuous. In some cases,classifying pathogens based on their vir-ulence factors is a more logical methodthan classification based on serotype orsome other trait unrelated to virulence.The evolution of pathogens will becomeclearer as scientists sequence and studyadditional genomes of related species.However, even seemingly related organ-isms can contain significant diversity

Evolution of SalmonellaS. enterica shows significant diver-

sity in its host specificity, but the rea-sons for host specificity remain unde-fined. Some serovars, such as S. Typhi,are very human specific while othersare animal specific (e.g., S. Pulloruminfects chickens). Others, such as S.Typhimurium and S. Enteritidis, havea wide host range. S. Typhimuriumtypically causes a diarrheal disease inhumans, but a typhoid-like disease inmice. Because many strains of Sal-monella cannot tolerate the low pH ofthe gastric environment, the infec-tious dose for ingested exposure toSalmonella is usually (but not always)quite high (greater than 100,000 bac-teria required to produce illness).However, if delivered intravenously, amere 10 organisms of strains that re-quire a large oral infectious dose cankill a mouse.

The incidence of antibiotic resis-tance among Salmonella and some

other foodborne pathogens continues toincrease. In some Asian countries, morethan 90% of Salmonella isolates are re-sistant to the most commonly used hu-man antibiotics. Globally, the three maincauses of antimicrobial resistance havebeen identified as use of antimicrobialagents in agriculture, overprescribing byphysicians, and misuse by patients. Thecombination of increased antibiotic re-sistance and an apparent increase in vir-ulence has resulted in strains, such as S.Typhimurium DT104, that continue tocause much concern. S. TyphimuriumDT104, which has a broad host reservoir,is usually resistant to five antibiotics (i.e.,ampicillin, chloramphenicol, streptomy-cin, sulphonamides, and tetracyclines)and can be resistant to others (e.g., fluo-roquinolones). Reporting on an out-break of quinolone-resistant S. Typh-imurium DT104, Molback et at. (1999)stated that because fluoroquinolones re-main a standard treatment for suspectedextraintestinal Salmonella infections and

serious gastroenteritis, the occur-rence of quinolone-resistant Salmo-nella in animals is a great concern.

Although the genomic sequenc-es of S. Typhimurium and S. Typhihave been recently completed, sci-entists still have more questionsthan answers about Salmonella. Itis not known how they cause diar-rhea, or why some are constrainedby host specificity. Because of thecomplexity of their virulence fac-tors, little progress has been madein converting the available knowl-edge into therapeutics. Good agri-cultural and manufacturing prac-tices, appropriate food handling,and adequate water treatmentremain our best preventive mea-sures for most Salmonella infec-tions, although the typhoid vac-cines are effective against S. Typhiin humans, and vaccines for severalother serovars have shown promisein food animals.

overlaid on a common genetic backbone.Also, many avirulent bacteria can carryinactive forms of genes common to viru-lent strains.

Selection

Evolution and selection are closelyrelated. Evolution produces microor-ganisms that are distinctly differentfrom previous generations; selectiongives these mutant strains an advantageand causes them to become prominent.Together, these forces have a profoundimpact on the emergence of foodbornepathogens.

The proposed step-wise model (Fenget al., 1998) for evolution of E. coliO157:H7 (as opposed to a gradual evo-lution) points out some rather interest-ing features of the process by which newpathogens emerge. First, mobility ofgene segments appears to be a limitingfactor in the rate at which microorgan-isms can test new gene combinations.New high-throughput genome sequenc-ing methods will produce data to developa much better estimate of the frequencyof such events and an improved under-standing of their underlying mecha-nisms. Second, the common virulencegenes shared by the O157:H7 and EHEC

O26:H11/O111:H8 lineages of enterohe-morrhagic E. coli demonstrate that ac-quisition of the same virulence gene ele-ments has occurred on multiple occa-sions, even through parallel paths (Reidet al., 2000). This discovery raises a sig-nificant question regarding the selectivepressures that cause the abrupt rise todominance of particular virulence genecombinations within a species.

In infectious diseases that are pri-marily carried by humans and trans-mitted person-to-person, the appear-ance of new alleles of virulence genes isassociated with the rise and spread ofmore virulent clones (Musser, 1996).This process likely also occurs amongfoodborne pathogens. However, food-borne pathogens must overcomeunique hurdles, including survival inthe pre-harvest environment, as well assurvival during food processing, stor-age, and preparation.

Many of the hurdles in food pro-duction and processing are a differentset of selective pressures than those ex-erted by the human host’s gastrointesti-nal tract. Successful foodborne patho-gens, such as Salmonella and L. monocy-togenes, have acquired not only viru-lence characteristics, but also physiolog-ical and ecological characteristics that


allow them to propagate in food pro-duction and processing environmentsand overcome hurdles. Using the newtool of genomics, researchers are exam-ining how changes in food production,processing, storage, and preservationmethods can impose new selective pres-sures on foodborne microorganismsand how such pressures might affectvirulence in known pathogens or emer-gence of new pathogens.

Shifts in Known Pathogens

A cause-effect relationship has beendocumented between changes in foodproduction methods and shifts in popu-lations of a pathogenic species favored inthe food production environment. Per-haps the best example is the displace-ment of Salmonella serovar Gallinarumby the Enteritidis serovar in poultry pro-duction environments. Scientists theorizethe shift was caused by programs toeradicate S. Gallinarum, a poultry patho-gen (Bäumler et al., 2000). Combina-tions of mathematical modeling, epide-miologic investigations, and populationgenetic studies suggest that the popula-tion shift and spread of S. Enteritidistook independent but parallel paths inEurope and North America. Geneticallydistinct subpopulations of S. Enteritidisrose to dominance on the two conti-nents, in theory, due to competitive ad-vantages over S. Gallinarum in occupy-ing the poultry environment (Rabsch etal., 2000).

Although this displacement likely didnot involve biological changes in thepoultry host, it illustrates the capacity forchanges in veterinary practices to havesignificant impact on the populations ofmicrobial species that inhabit the pro-duction environment. In addition, itdemonstrates that such population shiftshave the potential to change the relativerisk of foodborne illness to humans.Bioinformatics uses computationalmethods to analyze large sets of biologi-cal data, such as genome sequences, or tomake predictions, such as protein struc-tures. Given the tools of genomics andbioinformatics, it may be possible to un-derstand why such shifts occur when epi-demiologic studies fail to identify the ac-tual selection pressure.

Emergence of New Pathogens

Scientists know relatively littleabout how the microbial controls im-

posed during food production andprocessing affect the emergence of newpathogens. However, hurdles imposedin food processing—such as pH, os-molarity, and temperature—are allknown to affect physiological charac-teristics and, in some cases, virulencecharacteristics of pathogenic microor-ganisms. A good model from which tobegin drawing conclusions might be toexamine the distribution of Shiga tox-in-converting phages among strains ofE. coli. Clearly, the Shiga toxins play apivotal role in the pathogen’s ability tocause illness. However, Shiga toxin-producing E. coli (STEC) can be foundin many environments, and only cer-tain serotypes of STEC appear to com-monly cause disease in humans, indi-cating that the Shiga toxins alone areinsufficient to confer virulence. Theconvergence of Shiga toxin genes andgenes conferring other virulence char-acteristics—such as the ability to at-tach to host cells—is necessary for theemergence of a new and successfulpathogen. Considering the hurdles im-posed in the food production and pro-cessing environment, it seems likelythat convergence of such genes ontophysiologically robust genome back-bones favored the spread of EHEC lin-eages. Research is needed to identifyhow selective pressures in the foodproduction and processing environ-ment affect the potential for newpathogens to emerge or for subpopula-tions of known pathogens to increasein dominance.

By focusing on the results of selec-tion in recently emerged foodbornepathogens, scientists can begin to ad-dress these issues. The approach is anadaptation of the method used by ge-neticists to identify the genes involvedin a particular biochemical pathway.First, scientists induce mutations atrandom, and mutants of interest areselected based on phenotypic traits.Mutations that block the pathway un-der study are traced back to specificgenes. Once the genes are marked andidentified, scientists use biochemicaland molecular techniques to under-stand how the genes function in thepathway.

In the case of foodborne pathogens,the strategy is similar but reversed. Here,the goal is to identify the genes or allelesthat have been selected in food produc-tion environments, elucidate their func-tion, and pinpoint the selective advan-

tage conferred. The combination of ge-nomics and population genetics will pro-vide methods for identifying genetic al-terations that correlate with the descentof specific populations of foodbornepathogens. However, scientists must be-gin to devise strategies for identifyingwhich genetic alterations, among themany different gene sequences that de-fine a subpopulation, confer selective ad-vantage. Identifying and understandingthese genes may enable scientists to pin-point the selective forces at work.

Stress

Each microbe prefers a specific setof environmental conditions. When en-vironmental parameters are significant-ly different from the desired range, themicrobes undergo stress. To be morespecific, stress is defined as chemical orphysical parameters that impair thefunction of the macromolecular ma-chinery of the microorganism. Exam-ples of stress for certain microorgan-isms might include high and low tem-perature, acidic pH, low water availabil-ity, and presence or absence of oxygen.Specific genes are activated, producingproteins that protect the bacterium fromstress. This process, known as an adap-tive response, improves the microorgan-ism’s ability to survive under the stress-ful conditions. Although the responsesimprove the range of conditions the mi-croorganisms can tolerate, these re-sponses also require energy, so they areonly expressed when needed. Undernormal conditions, bacteria that do notturn off their stress responses would beoutgrown by those that reroute that en-ergy to other cellular processes. Stressresponses are of particular interest infoodborne pathogens because they canrender bacteria tolerant to traditionalfood processes or the intrinsic parame-ters of the food.

Bacteria have evolved elaborate net-works to protect against or repair dam-age caused by detrimental conditions.Bacterial responses to stress are variedand complex, including both structuraland physiological changes. For mostbacteria, these responses are modulatedby specific sigma (F) factors (Grossmanet al., 1984; Lange and Hengge-Aronis,1991) or regulators (Christman et al.,1989) that direct the activation of specificgenes that comprise regulons (largenumbers of coordinately controlledgenes) and encode for the proteins re-

23EXPERT REPORT

sponsible for cellular protection. Theproteins produced in response to stressenhance bacterial survival in the envi-ronment outside the host, including infoods (Cheville et al., 1996; Humphrey etal., 1993; Jenkins et al., 1988; Leyer andJohnson, 1993; Miller and Kaspar, 1993;O’Neal et al., 1994).

Studies of adaptive responses tostress can be classified into two distinctareas: the response itself and the abilityto generate the response. Of particularinterest in the response is how it miti-gates the physiological consequences ofstress conditions. On the other side isthe perception of stress, specifically howcells communicate the physical andchemical signals of ensuing stress condi-tions to the regulatory machinery thatgoverns the response. Understanding theresponse mechanics will provide the in-formation necessary to finely tune pro-cessing conditions to avoid triggering thestress protection mechanisms. In addi-tion, this knowledge can be used to de-velop rational targeting strategies toidentify novel antimicrobials for use inpre-harvest settings.

Responses to Stress

The cellular responses to stress cangenerally be divided into two catego-ries—general and specific stress re-sponses. In many instances, a certainstress response gene may be part ofboth specific and general stress re-sponse pathways. This is usually be-cause the gene has multiple regulatoryelements that are recognized by the dis-tinct machinery that coordinates thegeneral or specific response. It is there-fore difficult to separate the contribu-tion of general and specific stress re-sponse pathways to cellular viability un-der any given stress condition. Rather,the combination of the two, and specifi-cally the combined fine-tuning of eachpathway, dictates many of the survivalcharacteristics of the species.

General Stress Response

The general stress response (GSR)regulon is a large group of genes thatcollectively comprise several differentfunctions that facilitate growth and sur-vival under different conditions, such asosmotic shock, thermal stress, pH stress,oxidative stress, and nutrient depletion(Hengge-Aronis, 1996; Hengge-Aronis,2000; Lee et al., 1995). The GSR is a

complex system; more than 50 genes inE. coli are responsible for its GSR andare coordinately regulated by the prod-uct of the rpoS gene encoding Fs, an al-ternative sigma-subunit of RNA poly-merase (Loewen et al., 1998). Many ofthe genes in the GSR regulon appear tohave obvious functions for mitigatingspecific types of stress, such as compati-ble solute transporters that facilitatetransport of solutes in response to os-motic stress (Loewen et al., 1998), whileothers may confer general protectiveproperties under multiple stress condi-tions.

The GSR for one stress may inducechanges that improve the organism’s sur-vival under other stress conditions, a phe-nomenon known as cross protection. Forexample, it has been demonstrated in lab-oratory media that heat-shocking Salmo-nella Enteritidis (shifting the temperaturefrom 20 C to 45 C) results in an approxi-mate 3-fold increase in the D values (thetime required to inactivate 90% of the or-ganisms) at a pH of 2.6 and a greater than10-fold increase when the temperature israised to 56 C (Humphrey et al., 1993). Inthis example, the original stress (exposureto heat) increases protection to both heatand acid even when the bacteria had notbeen previously exposed to low pH.

Specific Stress Response

One of the best-characterized exam-ples of specific stress response systemsis the heat shock response. Like the GSR,the heat shock response involves an al-ternative sigma factor, F32, as a primaryregulator. When the temperature in-creases, F32, which is normally degradedrapidly, becomes more stable and istranslated at a higher rate, resulting in atransient accumulation of the F32 pro-tein and a corresponding increase in therate of transcription from heat shockpromoters that are recognized by F32

RNA polymerase (Morita et al., 2000).Approximately 30 proteins belong to theheat shock regulon. Basal levels of theheat shock proteins are produced at alltemperatures, but at higher temperaturesthe microorganism needs a greater con-centration of these proteins to remain vi-able (Gross, 1996). Induction of the heatshock response is somewhat more specif-ic than the GSR; however, there are othertriggers of the response, such as exposureto ethanol (Gross, 1996). Heat shock andthe production of associated proteinsprotects the cell from the detrimental ef-

fects of heat, and, as noted above, resultsin cross protection to other stresses. In E.coli, a second heat shock system, con-trolled by FE (F24), also has been identi-fied (Erickson and Gross, 1989, Wangand Kaguni, 1989). This system recentlyhas been shown to comprise a mecha-nism for sensing and coordinating re-sponses to the effects of thermal stress inthe periplasm (Mecsas et al., 1993). Thus,with F32 and FE, E. coli has separate andhighly specialized systems for adapting tothermal stress in the cytoplasmic andperiplasmic compartments.

Spore Formation

In addition to general and specificstress response pathways, some bacteriaand other microorganisms also haveevolved highly sophisticated pathways forstress adaptation, such as formingspores. Spores are metabolically inactiveor dormant and are much more resistantto adverse environmental conditions,e.g., extremes of temperature, low wateractivity, and radiation.

In organisms that form spores, theadaptive response pathways are somewhathierarchical and, depending on the envi-ronmental conditions, can be triggeredalone or in combination. Bacillus subtilis,the model organism for studying sporeformation, relies on the general stress re-sponse, several stationary phase and tran-sition state pathways, the development ofcompetence, and differentiation into thedormant endospore. The pathways forspore formation, competence, and normalgrowth are mutually exclusive, but eachone can be used in combination with thegeneral stress response. Before the organ-ism commits to a major step such as nor-mal growth, competence, or spore forma-tion, sophisticated signal transductionpathways measure environmental nutri-tional and chemical signals, as well as thestate of cellular processes such as DNAreplication.

Role of Stress Adaptation

Scientific interest has recently focusedon determining the role of stress adapta-tion pathways not only in the food matrixbut also in a host or host cell. To some ex-tent, virulence genes can be considered anadaptive response to the stresses encoun-tered during entry into the host. Studieshave shown that components of specificand general stress responses are some-times necessary to survive entry into a


host cell. Specific examples include therole of the general stress response regu-latory protein rpoS in survival of Salmo-nella inside phagosomal vacuoles (Fanget al., 1992), the role of protease/chaper-one proteins in the intracellular survivalof S. enterica and L. monocytogenes(Buchmeier and Heffron, 1990; Johnsonet al., 1991; Rouquette et al., 1996), andthe role of acid tolerance genes in intrac-ellular survival of L. monocytogenes(Marron et al., 1997). Because thesemolecules facilitate survival in both thefood matrix and entry into a host cell,inducing these responses in the food ma-trix could therefore “prime” the patho-genic microorganisms, increasing theircapacity to survive entry into a host celland establish infection.

Moreover, conditions in food pro-cessing environments that subjectpathogens to sublethal stress may fur-ther select pathogen subpopulationswith increased survival efficiency (seesublethal injury, p. 63). Over time, thesemechanisms could increase the relativepotential for a species to cause disease.Using genome-based methodologies infood processing research centers, futurescientists will be able to examine thepopulations of species that survive foodprocessing conditions.

Signal Perception and Induction

In addition to different stress re-sponses, microorganisms have evolvedmultiple and unique mechanisms (path-ways) for sensing and transducing physi-cal and chemical signals to the regulatorymachinery that coordinates stress re-sponses. It should be noted, however,that distinguishing between the impactof the regulatory machinery and slightvariations in the actual responses is diffi-cult. Examining the mechanics of signalperception and transduction can yieldnew insights into optimizing the safety offood production processes and can pro-vide specific targets for design of antimi-crobial agents.

Detailed analyses of distantly relatedbacteria reveal that similar stress re-sponses may be modulated by very dif-ferent regulatory machinery. For exam-ple, gram-negative enteric bacteria andlow G+C gram-positive organisms usedifferent mechanisms to trigger similarstress responses.

In gram-negative bacteria, the GSRpathway is modulated primarily by a pro-tein called F38 or RpoS, which is an RNA

polymerase subunit. F38 accumulates dur-ing several different stress conditions andactivates the target genes that produce thegeneral stress response. The rate of syn-thesis for the RpoS protein increases littleduring stress conditions; it accumulatesrapidly after stress because the degrada-tion rate is slowed. Unlike most responseregulators, which directly modulate genetranscription by increasing the amount ofsignal protein produced, this responseregulator functions by controlling the sta-bility of the protein and thereby changingits rate of degradation.

In the instance of the GSR of gram-positive bacteria, much is known aboutB. subtilis. As in the gram-negative bac-teria, the GSR in B. subtilis is modulatedprimarily by an alternative sigma sub-unit of RNA polymerase, in this caseknown as FB. However, FB is controlled

primarily by its accessibility, not by itsrate of synthesis or degradation. Whenunder stress, the organism produces an-other protein (the anti-sigma factor) thatbinds with the FB protein, making it nolonger accessible (Benson and Halden-wang, 1993). The anti-sigma factor pro-tein is controlled by a complex cascade ofsignal transduction proteins, which ap-pears to form branched pathways of sig-nal flow and provides several points ofentry for different types of signals. Be-cause many of these signal transductionproteins appear to bind to ribosomes, sci-entists have hypothesized that stress in thecytoplasm is measured by increases in ri-bosome dysfunction.

Despite differences in regulatory ma-chinery, there is striking similarity be-tween the stress protection system(s) ofdistantly related bacteria such as E. coli

F38 Regulated Proteins Initially identified and character-

ized in E. coli (Hengge-Aronis, 2000),F38 homologues with analogous func-tions have been identified in other en-teric and nonenteric gram-negativebacteria (Fujita et al., 1994; O’Neal etal., 1994). The rpoS gene encoding forF38 was initially identified as the mas-ter regulator of the phenotypic prop-erties associated with stationaryphase-reduced size and tolerance toa variety of physical and chemicalchallenges (Hengge-Aronis, 1993;Jenkins et al., 1988; 1990; Matin et al.,1989). The general stress tolerance in-duced by stationary phase/starvationis primarily due to the effects of F38-regulated proteins, although the con-comitant morphological and physio-logical changes likely contribute tothe stress-tolerance phenotype(Hengge-Aronis, 1993; Kolter et al.,

1993; Matin et al., 1989).These protective proteins are likely

involved in the ability of a pathogen tosurvive the gastric acidity and otherhost defenses (Fang et al., 1992; Priceet al., 2000). Moreover, F38 mediatesexpression of the SpvR virulence oper-on in Salmonella (Robbe-Saule et al.,1997) and the esp genes of pathogenicE. coli that encode for a Type III secre-tory system (Beltrametti et al., 1999)and consequently the virulence ofthese bacterial pathogens. Consideringthe important functions of F38-regulat-ed proteins, the finding of variations inthe rpoS allele in stationary-phase cul-tures of E. coli (Zambrano and Kolter,1996) is of particular importance tothe emergence of new strains of patho-gens with enhanced survival or viru-lence properties. These changes couldresult in enhanced production of theseprotective proteins and greater toler-ance to stress.

Function

Metabolic changes

Protection

Repair

Example

otsBA operon, trehalose metabolism

Oxidative stress protection by dps

Oxidative stress protection by katE,catalase HPII

aidB, repairs methylation damageof DNA

Reference

Hengge-Aronis, 2000

Altuvia et al., 1994

Mulvey et al., 1990

Landini et al., 1996

Table 4. Functions of F-Regulated Proteins

25EXPERT REPORT

and B. subtilis. Several genes in the RpoS-mediated stress protection system in E.coli have homologues in B. subtilis thatare part of the F32 stress protection sys-tem regulon. The shared genes of the re-spective stress protection systems likelyconstitute a core collection of importantfunctions that confer the general protec-tive properties needed for exploiting soiland intestinal environments.

Despite the similarity of the E. coliand B. subtilis stress protection systemregulons, there are also some clear differ-ences. For example, several heat shockgenes in the B. subtilis stress protectionsystem are governed independently ofthe stress protection system in E. coliwhere they are modulated primarily bythe heat shock regulons.

Comparing the FB-mediated stressprotection system in the closely-relatedspecies B. subtilis, L. monocytogenes,and S. aureus, reveals clear instances ofdivergent evolution despite the similar-ity of the regulatory machinery. In B.subtilis and L. monocytogenes, the sevenregulatory genes that modulate FB ac-tivity are collinear and comprise anoperon including the FB gene itself.Despite the collinearity in position,there is not a corresponding collineari-ty in structure; the distal gene of theoperon shares only distant similaritywith its homologue in the other speciesalthough the upstream genes are gener-ally much more similar to one another

(Becker et al., 1998). Given that thisgene, known as rsbX, is the first regula-tor in the signal transduction pathway,the divergence could be intimately as-sociated with the specialized physiolo-gy of the two species. A second exam-ple of the divergence is observed in S.aureus, which lacks three of the sevencollinear regulatory genes (Kulick andGiachino, 1997; Wu et al., 1996). Al-though the precise meaning of theseexamples of divergence in regulatorymolecules is not known, one mightgenerally conclude that the differencesreflect fine-tuning of the regulatorymachinery to the physiological needsof the species.

The function of the regulatory ma-chinery also is likely to be fine-tuned inboth signal perception and the targetgenes of the regulons. Studies with B. sub-tilis and L. monocytogenes have shown thatthe magnitude of the stress protection sys-tems to different types of signals is spe-cies-specific. For example, osmotic shocktriggers a relatively minor response in B.subtilis but in L. monocytogenes is one ofthe most potent inducers (Becker et al.,1998). Consequently, the stress protec-tion system contributes little to the growthof B. subtilis in conditions of high osmo-larity, but for growth of L. monocytogenesthe contribution is significant.

Collectively, stress protection sys-tems play pivotal roles in the survival,dissemination, and virulence of bacteri-

al foodborne pathogens. Understandingthem may enable scientists to predictfuture foodborne pathogens with great-er accuracy.

Driving Forces in Pathogenicity

Primary drivers of microbialpathogenicity are the growth in the hu-man population and the proportion ofthe population that is immunocom-promised either because of age, preg-nancy, underlying disease, or immuno-suppresive treatments. With higherdensities of humans, microbes can ex-ploit a multitude of routes to transferfrom an infected person to other hu-mans. The concentration of humans inurban settings can select for microbesthat are transmitted from person-to-person or through contaminated air,food, and water. The same case can bemade for high-density farms raisingmeat animals or food crops. World-wide human travel and the global dis-tribution of foods facilitate the intro-duction and flow of pathogens and ex-otic microbial genes into human andanimal populations.

In the absence of a properly func-tioning immune system, microbes thatare harmless to a majority of the popu-lation can cause life-threatening infec-tions in immunocompromised individ-uals. People older than 65 years of agegenerally have reduced immunity that

Emergence of Viruses,Parasitic Protozoa andMarine Biotoxins asFoodborne Pathogens

Although pathogenicity of someof the marine biotoxins has beencharacterized, there is a paucity ofinformation available regarding thedisease mechanisms or virulencefactors for enteric viruses and para-sitic protozoa. The reasons for thisare several-fold; for many of theseagents, in vitro cultivation and/oranimal models are nonexistent. Be-cause they have had relatively lessscientific emphasis over the last fewdecades, fewer research dollars havegone into understanding these

agents as compared to the emergingbacterial agents such as L. monocytoge-nes or E. coli O157:H7. Finally, theability to work with many of theseagents has been restricted by method-ological limitations that have been par-tially overcome by the introduction ofroutine molecular biological tech-niques.

New knowledge has emerged thathighlights the unique nature of theseagents. This information has been ob-tained largely through: (1) increasedepidemiological surveillance; (2) im-proved detection methods; and (3) in-creased research funding in food safety.Such initiatives have helped scientistsunderstand infectious doses, the roleof ever-increasing internationalizationof the food supply and the increased

impact of environmental pollution asa contamination source.

The prevalence of viral gastroen-teritis in the United States and world-wide has been drastically underesti-mated for many years. For instance,public health officials have consis-tently failed to report and investigateoutbreaks of mild gastrointestinaldisease, in part because of a lack ofresources. In the absence of reliablelaboratory methods, there has been ageneral reluctance on the part of pub-lic health officials to classify food-borne outbreaks as viral solely on thebasis of epidemiological criteria(Bean et al., 1990). Today, clinicallabs are using molecular biologytechniques such as the polymerase

Continued on next page


chain reaction to facilitate diagno-sis of infected patients, althoughthese methods are not yet adaptedto the routine detection of virusesin contaminated foods. Thesesame methods are being used toidentify genetic relatedness be-tween human caliciviruses andsimilar viruses detected in stoolsamples obtained from farm ani-mals, sparking the debate that ani-mals may be a reservoir for theNLVs and concern over the poten-tial for zoonotic transmission (vander Poel et al., 2000). Other inves-tigators have focused their effortson tracking epidemics in bothspace and time, concluding a win-ter-spring seasonality of NLV out-breaks; the presence of many ge-netically different variants, sug-gesting that most outbreaks are in-dependent events; and on occa-sion, the presence of a common,predominant strain, without obvi-ous epidemiological link, thatemerges, spreads and then disap-pears (Fankhauser et al., 1998).There are also ongoing researchinitiatives to ascertain the infec-tious dose of representative NLVssuch as the Norwalk agent and theSnow Mountain agent. Taken to-gether, these factors will continueto contribute to the emergence ofthis important group of foodbornepathogens.

We also do not yet understandthe importance of foods as vectorsfor parasitic protozoan disease, al-though it is likely that this will bebetter defined in the coming de-cades. Unlike the viruses whichare only transmitted by humans,animal fecal pollution, and associ-ated runoff from farms may con-tribute substantially to the con-tamination of water and subse-quently crops. International tradeissues have certainly impacted theemergence of C. cayetanensis, butthe importance of poor waterquality as opposed to direct con-tamination by local wildlife hasnot yet been determined. Thisdoes, however, bring up the criticalissue of water quality and food

handler hygiene, which is likely tobe less advanced in developing na-tions and hence may contribute di-rectly or indirectly to the safety offoods imported into the UnitedStates. As with the viruses, humanchallenge studies are currently un-derway in an effort to better under-stand the infectious doses of theparasitic protozoal pathogens. Al-though detection methods exist andrefinements are being reported, theroutine implementation of thesescreening methods requires highlytrained personnel, and scientists areunable to detect parasitic protozoain contaminated foods at thepresent time.

With respect to the marinebiotoxins, much has yet to belearned. While the mechanisms ofpathogenesis of some of the knownbiotoxins has been elucidated, theemerging agents such as Pfiesteriahave not been characterized. In fact,the purified toxins cannot always beisolated. In many instances, scien-tists do not fully understand thestimulation required for the pro-duction of HABs. While manyHABs may be associated with nor-mal fluctuations in nutrient inputand water temperature in the estua-rine environment, it is likely thatnutrient loading associated with or-ganic and inorganic pollution maycontribute to their increased preva-lence and perhaps to the emergenceof new toxic algal species. In thesoutheastern United States, somehave cited intensive animal agricul-ture practices and/or increased landdevelopment with associated popu-lation density increases as providingthe necessary environmental forces.Certainly, ongoing epidemiologicalstudies will help ascertain the truepublic health impacts of these newtoxic algae. In all instances (viruses,parasitic protozoa, and marinebiotoxins), continued emphasis onresearch and vigilant surveillancewill likely result in reports of in-creased prevalence, and hence“emergence,” of these agents as as-sociated with human foodbornedisease.

Continued from previous page

continues to decline with age. By 2050,the U.S. human population will reachan estimated 400 million, and, of thispopulation, 80 million will be 65 yearsof age or older. This growth in immu-nocompromised populations will cer-tainly affect the number of cases offoodborne illness associated with op-portunistic pathogens, which will be-come much more prevalent.

In terms of the environment, mi-crobes continue to evolve to gain acompetitive advantage, and, as advanc-es are made to eliminate or control onepathogen, another organism willquickly occupy the niche that has beenvacated, known as “niche filling”. Insome cases, the new organism that oc-cupies this niche may be more patho-genic than the original pathogen ormay employ a different mechanism ofvirulence that is more detrimental tohuman hosts (see shifts known patho-gens, p. 22).

In addition, microbiological ecolo-gy involves the interplay of climatechanges, pollution, and genetic ex-change that selects for and perhapsdrives the generation of new microbialstrains. Changes within an ecosystemwhereby the micro- and macro-popu-lations of organisms are out of bal-ance, selecting for new variants with acompetitive advantage, is one theorythat has been proposed to explain theemergence of new variants of existingmicrobes.

In terms of increased virulence inpathogens, two themes should be em-phasized. First, many of the stress re-sponse systems that contribute to sur-vival in the food matrix also contributeto survival during passage through thegastrointestinal tract and the invasionof host cells. If this phenomenonproves to be a significant feature of“virulence” for a species, then pre-har-vest environments and food processingconditions should be designed to avoidimposing sublethal stress and henceselection of resistant bacterial popula-tions. Secondly, the diversity of stressregulatory response systems and regu-latory molecules holds promise for ra-tional design and targeting of antimi-crobial agents that eliminate pathogenpopulations while minimizing the dis-ruption of the total bacterial popula-tion. Such agents could be used in pre-and post-harvest settings, such as feedsand carcass washes to facilitate elimi-nation of unwanted species.

27EXPERT REPORT

Pathogenicity ofE. coli O157:H7

E. coli O157:H7 is typical of whatmight be expected in terms of anemerging pathogen.

Nomenclature

E. coli O157:H7 belongs to agroup of E. coli (enterohemhorrhagicE. coli, EHEC) that cause hemorrhag-ic colitis (severe bloody diarrhea) and,in a small portion of the cases,hemolytic uremic syndrome (HUS).These strains are within a largergroup of Shiga-toxin-producing E.coli (STEC), but EHEC possess viru-lence factors that other STEC do not.EHEC have been the source of manyfood- and water-borne outbreaks. Ithas been estimated that these patho-gens cause about 75,000 cases of diar-rhea and several hundred deaths an-nually in the United States (Mead etal., 1999). Young children and olderadults are particularly susceptible toHUS, while people of all ages can getthe diarrhea.

The designation O157:H7 is basedon serological analysis: lipopolysac-charide (O) type 157, and flagellar (H)type 7. However, scientists have sincediscovered that several related E. colistrains with different O serotypescause similar disease, such asO26:H11. Although O157:H7 is themost predominant EHEC serotype inNorth America, Japan and the UnitedKingdom, different serotypes of EHECsuch as O26:H11 and O111:NM dom-inate in other areas of the world, nota-bly central Europe and Australia. Be-cause of the conservation of virulencefactors, but not serotypes, classifica-tion schemes of EHEC strains shouldprobably be based on virulence factorsrather than the variant serotype. How-ever, O157:H7 (EHEC 1) andO26:H11/O111:NM (EHEC 2) clearlycomprise two distinct genetic lineages.

Virulence

E. coli O157:H7 and relatedstrains have the capacity to persist incattle without causing disease becausecattle lack a receptor for the illness-producing Shiga toxin (Pruimboom-

Brees et al., 2000). As many as half of allcattle carry O157:H7 at some time intheir lives, and some observers suggestthat nearly all cattle have been exposedto EHEC. The organisms are introducedinto the environment through the feces,including manure used as fertilizer forfood crops. Rainwater runoff can thenspread them to water reservoirs andwells. Alternately, fecal contamination atslaughter may result in meat contamina-tion. In addition to cattle, other rumi-nants, such as goats and sheep, and wildruminants, such as deer, can carry thisorganism, and it has now been found inbirds, flies, and in the food-producingenvironment.

Tolerance to low pH facilitates pas-sage through the stomach, making it pos-sible for E. coli O157:H7 to cause diseaseat a low infectious dose (10-100 bacte-ria). The organism’s acid tolerance alsoallows it to survive within acidic food,which was a major factor in outbreaks ofillness traced to unpasteurized applejuice, a product with a pH of approxi-mately 3.5 that usually inhibits the lessvirulent strains of E. coli.

In the past few years, scientists havemade significant advances in under-standing the underlying virulence mech-anisms of EHEC strains. Two major vir-ulence pathways contributing to diseasehave been identified, although there areprobably many others yet to be discov-ered. To cause disease, EHEC must pos-sess a Shiga toxin gene and genes withinthe LEE pathogenicity island that enablethe bacteria to adhere to epithelial cellsand form a pedestal on the epithelial sur-face upon which the bacteria reside. Theattaching and effacing genes also arefound in enteropathogenic E. coli(EPEC), a common cause of watery diar-rhea in the developing world, but these E.coli lack the Shiga toxin gene and, possi-bly, other virulence factors present inEHEC.

The Shiga toxins are comprised oftwo components, A and B subunits, thatstructurally resemble other toxins suchas cholera toxin. The B subunit conferstissue specificity, enabling the toxin toadhere to a specific glycolipid receptor,globotriaosylceramide (Gb3), on cellsurfaces. The active (A) portion of thetoxin is then delivered into the host cellwhere it inhibits protein synthesis, ulti-mately killing the host cell. The toxins

target certain cells, including the en-dothelial cells of blood vessels; thedead cells accumulate and plug thekidney, causing HUS. Shiga toxin isencoded within a mobile genetic ele-ment, a bacteriophage, that enables itto move to different strains of bacteria.As discussed below, it is believed thatthe acquisition of this toxin by E. coli isa relatively recent genetic event.

In addition to producing Shigatoxin, EHEC adhere to the large bowel,the pathogen’s preferred site in thebody, using a variety of virulence fac-tors contained within the LEE patho-genicity island. This 38-kb region ofDNA, inserted near a tRNA gene, con-tains all the genes necessary for bind-ing to epithelial cells and causing apedestal to form on their surface. Ped-estals are created when actin in the cy-toplasm is accumulated and polymer-ized beneath the pathogen. In addi-tion, effacement of the microvilli oc-curs, giving rise to the term “attachingand effacing E. coli”. Pedestal forma-tion is a complex process that utilizes aType III secretion system, several TypeIII E. coli secreted proteins (Esp’s), anda key molecule, Tir (Translocated in-timin receptor), that is delivered tohost cell membranes. Once in the hostcell membrane, Tir binds to intimin, abacterial outer membrane protein, re-sulting in intimate bacterial adherence.Because Tir spans the host membrane,it is also able to recruit host cytoskele-tal proteins to cause actin accumula-tion and pedestal formation.

Emergence

E. coli O157:H7 is believed to havearisen from a series of fairly recent ge-netic events. O157:H7 was first report-ed as a foodborne pathogen followingan outbreak associated with contami-nated hamburgers in 1982. Subse-quent studies of diarrheal samplesfrom prior outbreaks and sporadiccases revealed only a single E. coliO157:H7 isolate in the Center for Dis-ease Control and Prevention’s collec-tion. This isolate had been obtainedfrom an individual in the mid-1970s.A closely related strain of E. coli calledenteropathogenic E. coli (EPEC) causesdiarrhea in children (but does not

Continued on next page


cause HUS). EPEC contains theLEE locus, but not the Shiga toxin.It is believed that EHEC arose whenan EPEC-like organism (containingthe LEE island) acquired a Shigatoxin via a bacteriophage. Thisevent has occurred on more thanone occasion, leading to the two dis-tinct EHEC lineages. Experimentalevidence comes from studies done inrabbits, where a LEE-encoding E.coli was altered to also encode theShiga toxin. The resulting pathogenproduced a diarrhea that resembledhemorrhagic colitis, which was anEHEC-like disease.

Treatment

Therapies against EHEC infec-tions are extremely limited. Treat-ment with antibiotics is thought toworsen the illness, presumably bybreaking up the bacteria, which re-leases more toxin and increases tox-

in expression. Kimmitt et al. (2000) re-ported that some antimicrobial agents,particularly quinolones, trimethoprim,and furazolidone, were shown to inducetoxin gene expression and should beavoided in treating patients with poten-tial or confirmed STEC infections. Theseinvestigators also reported, however, thatresults of available studies conflict withregard to the influence of antibiotics,noting that age group, timing of antibiot-ic therapy, and range of agents usedcomplicate the analyses. Further, Kim-mett et al. (2000) reported that their ob-servations suggest that the complex in-terplay of infection stage, number of or-ganisms present at the time antibioticsare administered, and the environmentalconditions of those microorganisms,coupled with time-concentration profile,and bactericidal effect of the drug, couldrender an antibiotic clinically beneficial,neutral, or disadvantageous in differentsituations.

One potential therapeutic is current-ly in phase III clinical trials. New thera-

pies use an inert substance that mim-ics the toxin’s glycolipid receptor. In-gestion of the mimicking substanceshould bind excess toxin and therebylimit disease progression. Experi-ments indicate it may be effective, butonly when taken very early after in-fection. Alternate therapies are beingexplored.

In addition, significant effort isfocused on developing treatmentssuch as vaccines and probiotics toreduce carriage of O157:H7 by cat-tle. Decreasing the level of O157:H7in cattle would significantly decreasethe potential for food and watercontamination. Similarly, childhoodvaccines are being developed, butgiven the low incidence of disease,questions remain about whetheruniversal vaccination should be em-ployed, were an effective vaccine tobe developed. Experimental ap-proaches to block virulence factorssuch as the Type III secretion systemalso are being studied.

Continued from previous page

Humans as Hosts of Foodborne Disease

A number of factors that relate to the

human host have a major impact on the

occurrence and severity of foodborne

disease. The host’s age, gender, place of

residence, ethnicity, educational

background, underlying health status,

and knowledge, attitudes, and practices

related to health and diet all have

important bearing on foodborne

illness. The health of the host affects

the individual’s susceptibility to

infection and illness, and the host’s

dietary and hygiene practices affect

exposure to pathogens. From medical

and behavioral perspectives, human

host factors can be altered by modifica-

tion of susceptibility or elimination of

exposure. For example, assuring

proper handling and cooking of

ground beef contaminated with

Escherichia coli O157:H7 could

eliminate that food safety risk even in

the presence of continuing contamina-

tion of raw ground beef. Obviously,

risk communication is an essential tool

for preventing foodborne illness.

However, successful control requires

effective interventions at all stages of

the food system.

Manifestations of Clinical Disease

As the name implies, foodborne dis-eases—including intoxication, infection,and toxicoinfection—are illnesses ac-

quired by consumption of food contain-ing pathogens or their toxins. The patho-gens or their toxins can damage or de-stroy host cells or processes, or they caninduce a host response to their presencethat is harmful to the human host. Food-borne illness is caused by: viral, bacterial,or parasitic infections (e.g., Norwalk-likeviral gastroenteritis, Campylobacter en-teritis, toxoplasmosis); toxins producedduring microbial growth in food (e.g.,Clostridium botulinum, Staphylococcusaureus, Bacillus cereus, and Aspergillusflavus); and toxins produced by algal andfungal species (e.g., ciguatera fish poi-soning) (see Table 5).

Foodborne infections occur whenpathogenic microorganisms are ingested,colonize the intestine, and sometimes in-vade the mucosa or other tissues. Food-borne toxicoinfections arise when a mi-croorganism from ingested food grows

29EXPERT REPORT

in the intestinal tract and elaborates atoxin(s) that damages tissues or inter-feres with normal tissue/organ function.Foodborne microbial intoxications oc-cur by ingestion of a food containingharmful toxins or chemicals produced bythe microorganisms, usually during theirgrowth in the food.

Despite our best efforts, foodbornediseases remain common. Based on theavailable data, the Centers for DiseaseControl and Prevention (CDC) has esti-mated that 76 million cases of foodborneillness occur annually resulting in325,000 hospitalizations and 5,000deaths (Mead et al., 1999).

Most infectious foodborne illness ischaracterized by acute symptoms that arelimited to the gastrointestinal tract, in-cluding vomiting and diarrhea. These ill-nesses are generally limited in both dura-tion and severity, and most patientswithout underlying illnesses or malnu-trition recover without medical treat-ment or require only modest supportivecare. These illnesses can be especially dif-ficult to quantify because medical treat-ment is not sought. For example, B.cereus—linked to a wide variety of foodssuch as meat, milk, vegetables, fish, andrice products—may cause diarrhea andabdominal cramps and pain that last ap-proximately one day. C. perfringens,which may be present in meat, meatproducts and gravies, can cause intenseabdominal cramps and diarrhea thatalso generally resolve within a one-dayperiod. Norwalk-like viruses, responsiblefor an estimated 66.6% of illnesses at-tributed to known foodborne pathogens

(Mead et al., 1999), result in nausea,vomiting, headache, diarrhea, and ab-dominal pain. The symptoms usuallysubside within a day or two, and the in-fections rarely come to the attention ofpublic health workers.

However, some pathogens cause gas-trointestinal symptoms that are more se-vere and take longer to subside, especial-ly in immunocompromised individualssuch as young children, the elderly orpeople with AIDS. For example,Cryptosporidium parvum is a parasiticprotozoan that causes severe watery diar-rhea and sometimes coughing, fever, andintestinal distress. Symptoms may lastfrom four days to three weeks. Campylo-bacter jejuni, usually associated with rawchicken and raw milk, can trigger waterydiarrhea, fever, abdominal pain, andnausea that last for days to weeks. Of theestimated illnesses attributed to knownfoodborne pathogens, Campylobacterspp. are responsible for more than 14%(Mead et al., 1999).

Not all foodborne disease is limitedto the gastrointestinal tract. Some food-borne pathogens invade deeper tissues orproduce toxins that are absorbed andcause systemic symptoms, including fe-ver, headache, kidney failure, anemia,and death. Salmonella has been linked tonumerous foods, but especially rawmeats, poultry, and eggs. The symptomsof salmonellosis include nausea, vomit-ing, abdominal cramps, fever, and head-aches with a duration that ranges fromdays to weeks. Nontyphoidal Salmonellaare responsible for an estimated 30.6%of deaths caused by known foodborne

pathogens. E. coli O157:H7 is most com-monly associated with undercookedground beef and causes severe crampingand bloody diarrhea. After the hemor-rhagic colitis caused by E. coli O157:H7and other enterohemorrhagic E. coli(EHECs), some children develop a char-acteristic set of kidney disfunction andanemia called hemolytic uremic syn-drome (HUS). In the United States, HUSis the leading cause of acute kidney fail-ure in children.

The major effects of some foodbornepathogens are outside the gastrointesti-nal tract. Listeria monocytogenes causesserious illness in pregnant women, theirfetuses, or newborns, often resulting inspontaneous abortion or stillborn ba-bies. An estimated 92% of foodbornecases of listeriosis result in hospitaliza-tion, and 20% result in death. C. botuli-num produces toxins that attack the cen-tral nervous system, resulting in weak-ness, vertigo, double vision and difficultyswallowing and speaking.

In addition to acute illness, somefoodborne pathogens cause chronic ill-ness, often within sensitive subgroups ofthe population. For example, hepatitis Avirus (HAV) causes fever, headache, an-orexia, malaise, nausea, abdominal dis-comfort and sometimes jaundice. Thesesymptoms usually take weeks to resolve.Within a genetically predisposed sub-group, prolonged HAV infection may bethe precipitating event in the onset of au-toimmune chronic active hepatitis(Bogdanos et al., 2000; Naniche and Old-stone, 2000; Rahyaman et al., 1994.

The foodborne parasite Toxoplasmagondii causes birth defects. In addition,chronic toxoplasmic encephalitis attrib-uted to T. gondii infection may occurwhen an individual’s immune system isimpaired. Toxoplasmal encephalitis,characterized by dementia and seizures,has become the most commonly recog-nized cause of opportunistic infection ofthe central nervous system in AIDS pa-tients. Activated macrophages, lympho-cytes, and cytokines play a major role incontrol of both the acute infection andmaintenance and/or prevention of thechronic stage (Cohen, 1999; Tenter et al.,2000).

Biotoxins also cause foodborne dis-ease, with both acute and chronic clinicalmanifestations. Because these com-pounds can be resistant to processingand cooking, biotoxins can be present ina food even in the absence of viable cellsof the causative agent. The target organs

Type of Causative Agent

Bacterial infection

Viral infection

Parasitic infection

Bacterial toxin

Algal toxin

Mycotoxin

Prions*

Inorganic contaminants*

Organic contaminants*

Example(s)

Campylobacter jejuni

Norwalk-like viruses

Toxoplasma gondii

Clostridium perfringens,Bacillus cereus

Ciguatera fish poisoning

Aflatoxin

BSE

Heavy metals

Pesticide residues

Frequency (in the U.S.)

Common

Very common

Relatively common

Relatively common

Less common

Less common

None

Less common

Less common

* Not addressed within this report.

Table 5. Causes of Foodborne Illness


for these toxins vary and can include theliver, kidney and gastrointestinal tract aswell as the immune, nervous, and repro-ductive systems. Biotoxins include toxinsproduced by bacteria (e.g., botulism tox-in), fungi (e.g., aflatoxin and fumonisin),marine organisms such as dinoflagellates(ciguatoxin), and plants (phytotoxins,which are not discussed in this report).

The chronic sequelae of foodborneinfections in particular often focus onextra-intestinal systems, although food-borne microorganisms also may play arole in chronic enteropathies such as in-flammatory bowel disease. Guillain-Bar-ré Syndrome (GBS) is a disorder of theperipheral nervous system that occursworldwide and is a common cause ofneuromuscular paralysis. Victims losethe ability to write or speak and experi-ence motor paralysis with mild sensorydisturbances. Cases of severe GBS havebeen linked to a previous infection withC. jejuni, although other enteric patho-gens also may trigger the disorder. Evi-

dence indicates GBS is an autoimmunedisease, but the immunologic mecha-nisms that produce GBS after infectionwith C. jejuni are complex. Studies sup-port the hypothesis of molecular mimic-ry, since peripheral nerves may shareepitopes with surface antigens of certainstrains of C. jejuni. Some data suggestthat patients share genetic traits (Smith,1995). Although it is clear that GBS is anautoimmune phenomenon, evidence in-dicates that infections with C. jejuni, acommon foodborne pathogen, frequent-ly start the pathologic process (Allos,1997; Shoenfeld et al., 1996).

Another example of chronic illnessrelated to a foodborne infection is reac-tive arthritis. Triggered by infection withYersinia enterocolitica, Yersinia pseudotu-berculosis, Shigella flexneri, Shigella dys-enteriae, Salmonella spp., C. jejuni, and E.coli, reactive arthritis is an acute sterileinflammation of the joints. In additionto joint pain, Reiter’s syndrome, a sub-type of reactive arthritis, affects eyes and

the urogenital system. A genetic predis-position to developing post-infectiousreactive arthritis has been documented inpersons who share certain genetic traits.Other genes acting in concert apparentlydetermine the clinical presentation.Chronic sequelae are thus related to ge-netically determined host risk factors incombination with an environmental trig-ger (Kobayashi and Ando, 2000; Parkerand Thomas, 2000). It is important tonote that as human and microbial ge-nome sequencing projects progress, sci-entists should gain increasing insightsinto the bacterial virulence factors andthe host factors that interact in the pro-duction of these chronic, autoimmunepathologies. Hopefully, these insightswill eventually result in rational thera-pies to prevent or treat these types of dis-eases.

Although the evidence is not com-plete, foodborne infections may play arole in the development of inflammatorybowel disease (IBD). IBD is the collective

Pfiesteria piscicidaand Pfiesteria-likeMicrobes As PotentialFoodborne Pathogens

An association between seafood-borne illness in humans and the oc-currence of Pfiesteria in the marineenvironment has not been estab-lished, and much remains unknown.Pfiesteria piscicida and Pfiesteria-likeorganisms, discovered in 1988, havecaught the attention of researchersand funding agencies interested incharacterizing the biological func-tions and effects of previously unrec-ognized toxic dinoflagellates (Flem-ing et al., 1999; Glasgow et al., 1995;Smith et al., 1988). Initially present-ing as the cause of massive fish killsin North Carolina in the late 1980s,Pfiesteria-like organisms also havestimulated public concern over thepotential threat these organisms mayhave on human health and seafoodsafety. Because of difficulties in iden-tification of the Pfiesteria species,these organisms are currently charac-terized on the basis of morphologyand hence referred to as morphologi-

cally related organisms (MROs). MROsare believed to occur in waters fromDelaware to the Gulf of Mexico (EPA,1998). Currently, the number of MROsis unknown, and while some are toxic,others are not (Steidinger and Penta,1999).

It has been proposed that the Pfieste-ria dinoflagellate, unlike most toxic di-noflagellates, excretes its toxin(s) into theestuary rather than retaining the toxinwithin its cell (Burkholder et al., 1995;Glasgow et al., 1995). Following expo-sure to toxic P. piscicida and other toxicMROs, fish appear to be narcotized andfrequently die (Burkholder et al., 1995;Glasgow et al., 1995; Noga, 2000). How-ever, the toxin (Pptx) is relatively unsta-ble in the marine system, and in spite ofcontinuing research efforts on the life cy-cle and physiology of Pfiesteria, very littleis known about the toxin(s) produced bythe dinoflagellate. Attempts to obtainpurified toxin(s) have been unsuccessful.Mechanisms of action and chemicalstructure are currently undetermined.Investigators believe that the toxin con-sists of both water-soluble fractions,which may be responsible for the allegedneurotoxic effects in humans, and highly

lipophilic components, which may beresponsible for fish morbidity andmortality (Fleming et al., 1999; Noga,1997). Reported human impacts in-clude respiratory irritation, skin rash-es and possible neurocognitive disor-ders (Glasgow et al., 1995). Althoughanimal models have been developedto investigate neurocognitive effects(Levin et al., 1997), and epidemiologi-cal studies to evaluate the potentialhuman health effects (Gratten et al.,1998; Savitz, 1998) are ongoing, re-sults are inconclusive to date.

It is important to note that all ofthese studies investigate transmissionroutes other than through the con-sumption of contaminated seafood.Nonetheless, consumer confidence inseafood is adversely affected byevents such as fish kills and healthalerts. Extensive media coverage ofPfiesteria, closings of recreationaland commercial waters, as well as agrowing list of scientific unknownsregarding the organism’s occurrence,toxin(s), and effects, have generatedimmediate food safety concerns inseafood consumers, despite the ap-parent safety of these foods.

31EXPERT REPORT

term for Crohn’s disease (CD) and ul-cerative colitis (UC), which are bothchronic inflammatory diseases with aprolonged clinical course. Abdominalabscesses are a complication of CDwhile in UC abdominal perforationsmay lead to peritonitis. The cause ofIBD, and the mechanism(s) for sponta-neous exacerbations and remissions re-main unresolved. Controversial reportsby some investigators about the poten-tial association of Mycobacteriumparatuberculosis, the etiologic agent ofJohne’s disease in ruminants, with CDhave appeared in the literature for years.A conference convened by the NationalInstitutes of Health Division of Micro-biology and Infectious Disease (NIH/DMID, 1998) and the European Com-mission’s Scientific Committee on Ani-mal Health and Animal Welfare (EC,2000) concluded that there is insuffi-cient evidence to prove or disprove thatM. paratuberculosis is the cause of CD.Studies to determine how M. paratuber-culosis could be transferred from ani-mals to humans have focused on milkand results have been conflicting(CAST, 2001). Both the NIH/DMIDconference and the EC committee rec-ommended further study of this issue.An association between IBD and thebacterial L-forms of Pseudomonas, My-cobacterium, Enterococcus faecalis andE. coli also has been suggested (Her-man-Taylor et al., 2000; Korzenik andDieckgraefe, 2000).

The preceding examples are repre-sentative, but they do not present a com-prehensive discussion of foodborne dis-ease. The differing symptoms, durationand severity make diagnosis and com-prehensive tracking of foodborne illnessdifficult. The large number of foodbornepathogens—each with its own virulencefactors—produce an astonishing arrayof illnesses, and the pathogens continueto evolve. Millions of illnesses and thou-sands of deaths occur each year as a re-sult of contaminated food in developedcountries, and the situation is muchworse in the developing world.

Resistance to MicrobialFoodborne Disease

Humans are protected from infec-tious foodborne disease by a variety ofnonspecific (innate) and specific im-mune system mechanisms. When all ofthese systems are functioning optimally,the chance of foodborne illness is re-

duced. Many factors cause these systemsto function below optimal levels, increas-ing the likelihood of illness. In addition,some foodborne pathogens have foundways to evade or trick the body’s defen-sive mechanisms.

Immune Response

Foodborne pathogens and theirtoxins enter human tissue via the gas-trointestinal (GI) tract. The GI tract isapproximately 30 feet long and in-cludes cells that produce acid, mucus,antibodies, and other substances thatprotect the host from foodbornepathogens (Burke, 1985). Food enter-ing the mouth is chewed, breaking itinto smaller pieces and mixing it withsaliva. When swallowed, the food trav-els via the pharynx and esophagus tothe stomach. Upon entering the stom-ach, the food is broken down by gastricjuices that contain pepsin, lipases andacids.

From the stomach, the partially di-gested food enters the small intestine.The intestine is approximately 20 feetlong and has an irregular, folded liningthat provides a very large surface areathat facilitates digestion and absorp-tion of nutrients. It also facilitates ex-posure to microorganisms that sur-vived passage through the stomach.The tissues immediately below the epi-thelial lining cells contain blood capil-laries to absorb monosaccharides andamino acids and deliver bloodbornedefenses, and lymph capillaries to ab-sorb fatty acids and glycerol. Enzymesand hormones secreted by the liver andpancreas assist digestion and absorp-tion in the small intestine. Because ofits length, surface area (equivalent to asingles tennis court), and digestive ca-pacity, more than 80% of absorptionoccurs in the small intestine. Undigest-ed food material enters the large intes-tine (colon), through which it travelsuntil it finally exits the anus. Anatomic,physiologic, and pathologic changes inthe GI tract influence the level of pro-tection from the effects of pathogens.These GI tract changes also affect thepopulation of nonpathogenic microor-ganisms living in the GI tract, whichinfluences the likelihood of foodborneinfections.

The large surface area, presence oflarge amounts of nutrients, and ab-sorptive capacity of the alimentary ca-nal make it particularly prone to pene-

tration by microbes and their toxins.Fortunately, the body defends itselfwith an extraordinary array of non-specific and specific immune mecha-nisms.

Nonspecific Immune Mechanisms

Nonspecific or “innate” immunity isthe front line of host defense against mi-croorganisms in the gut and other sites(Elwood and Garden, 1999; Pestka,1993; Takahashi and Kiyono, 1999). Theepithelial barriers prevent absorption ofmore than 99% of the proteins in the in-testinal tract (Newby, 1984). These epi-thelial lining cells are constantly re-newed to ensure that damaged villi donot provide a location vulnerable to in-fection. The continued movement of thegut contents also keeps the microbialpopulations (microflora) in the smallintestine at lower levels compared to thelarge intestine microflora, which existsin a more static environment.

Digestive secretions are another ma-jor form of nonspecific immunity in theGI tract. For example, the high acidityand pepsin content in the stomach workto destroy microbial pathogens and theirtoxins. Enzymes in bile acids and pan-creatic secretions also can protectagainst microbial pathogens. Gastricand intestinal epithelia are covered by amoving layer of continually replacedmucus, a protein that contains sugarresidues to protect against proteolytic at-tack and microbial attachment. In addi-tion to functioning as a lubricant andprotecting the stomach and intestinefrom acidic pH, the mucus is a vehiclefor antibacterial substances (e.g., secre-tory immunoglobulin A and enzymes)and prevents large molecular weight ma-terials from passing into enterocytes, theepithelial lining cells.

As a stable ecosystem, the normal in-testinal microflora diminish opportuni-ties for pathogenic microbial infection.By occupying the available binding siteson the enterocytes, decreasing the pH ofthe gut lumen, producing volatile fattyacids and selective antibiotics known asbacteriocins, and increasing motility ofthe gut contents, these nondisease-caus-ing (commensal) microbes provide animportant element of the nonspecificdefense system.

Microbial agents or antigens thatmanage to penetrate the epithelial barri-er may encounter mononuclear phago-cytes (blood monocytes or tissue mac-


rophages) and polymorphonuclear ph-agocytes (PMNs or granulocytes) thatdefend the rest of the body from thingsthat get through the superficial defenses.PMNs, a primary defense against infec-tious agents, can travel via blood vessels.Macrophages travel to an inflamed sitewhere they attempt to kill the intruder.Certain blood proteins also can serve asbackup nonspecific defense mechanisms(Pestka and Witt, 1985). Interferonformed by virus-infected cells can inhibitreplication of unrelated viruses. Kininsare a group of peptides which, when acti-vated, are involved in inflammation andblood clotting. Finally, the complementsystem, a series of proteins and enzymat-ic reactions, can destroy invading cells.

The nonspecific mechanisms de-scribed here act together to prevent infec-tion by enteric microorganisms or entryof large microbial toxins. Thus, undernormal conditions, relatively large num-bers of microorganisms would be re-quired for a few to survive the defensesand initiate infection. A variety of factorsmay depress nonspecific immunity, suchas decreased gastric acidity caused by in-gestion of antacids, diminished nativemicroflora following treatment with an-tibiotics, or damage to the epithelial bar-riers. When nonspecific immunity is de-pressed, the likelihood that small num-bers of a pathogen will cause an infec-tion is increased. However, even when in-nate protection fails, specific defensemechanisms can prevent infection anddisease.

Specific Immune Mechanisms

In addition to the nonspecific im-mune defenses described above, ingestedmicrobes face other compounds that cir-culate in the blood or are secreted intothe lumen of the GI tract that are specificto certain microbes or related groups ofmicrobes. This “acquired” immunity rec-ognizes characteristics or components ofthe microorganism, called antigens, andthen inactivates, removes or destroys themicroorganisms that possess these anti-gens. To do this, the immune systemmust be able to recognize small differ-ences in the chemical structure of an an-tigen and “remember” these chemicalstructures for long periods of time. Anti-gens are typically high molecular weight(>10,000 Daltons) proteins or polysac-charides. Parts of the pathogen—such asthe cell wall, flagella, capsule and tox-ins—serve as excellent antigens, in part

because they are multivalent, meaningthey have more than one chemical struc-ture that can be recognized by the im-mune system.

Specific responses can be functional-ly divided into phases: (1) recognition,(2) activation, and (3) effector (Abbas etal., 1997). In the recognition phase, for-eign antigens bind to specific receptorson existing lymphocytes. Lymphocyterecognition of specific antigens triggersthe activation phase. Activation eventsinclude development of antigen-specificlymphocytes and a shift from recogni-tion to defensive functions. In additionto antigens, activation requires “helper”or “accessory” signals from other cells.Finally, the effector phase implements anactive defense based on antigen recogni-tion and lymphocyte activation.

The immune system has numerouspossible effector responses to an antigen-ic stimulus. First, one or more compo-nents of the specific immune system canwork to remove the antigen. Second, spe-cific and nonspecific immune mecha-nisms can interact to enhance the host’sability to kill invading microorganisms.Third, an antigenic stimulus can induce“tolerance,” which is a “specific” type ofunresponsiveness. Thus, a host can rec-ognize and tolerate the host’s own pro-teins, known as self antigens. The abilityof the immune system to develop amemory allows the host to both preventfuture reinfection by an invading organ-ism and to avoid mounting a self-de-structive immune response.

Cells of the Immune System. Manyhighly specialized cells carry out the crit-ical functions of specific humoral (anti-body-mediated) and cell-mediated im-mune reactions that influence a host’s re-sistance to infection and serious disease.These cells are derived from stem cells inthe bone marrow and become the lym-phocytes, granulocytes, macrophages,dendritic cells and other specialized pro-tective cells during a process called he-matopoiesis. To be responsive to thepresent needs of the immune system,many aspects of leukocyte developmentare regulated by cell-to-cell interactionsand by cytokines, soluble protein factorsthat influence cell growth, differentiationand maturation. In most cases, the celltypes involved in generalized systemicimmunity also play key roles in gas-trointestinal immunity.

Lymphocytes carry out critical regu-latory and effector activities in specificimmunity. B lymphocytes are responsi-

ble for humoral immunity (antibodyproduction) and carry immunoglobu-lins (antibodies) on their surface. T lym-phocytes can have both effector and reg-ulatory functions. T cells control thematuration of both effector T and Bcells. T cells also are involved in cell-me-diated immune responses such as cyto-toxicity and delayed-type hypersensitivi-ty. Some B and T cells reside in specificareas in the “secondary” lymphoid or-gans such as the spleen and gut-associat-ed lymphoid tissue (GALT) to facilitatecontact between lymphocytes and circu-lating antigens.

In addition to B and T cells, accesso-ry cells (macrophages, monocytes, anddendritic cells) can ingest and destroy in-fectious particles and function in antigenpresentation that influences the strengthand type of antibody response. Mast cellscan respond to various antigens andgenerate a hypersensitivity response.Mononuclear cells known as “killer” cellscan bind to antibodies and facilitate lysisof tumor cells and cells infected with vi-ruses. Other cell types with the ability tospontaneously dissolve or disintegrateneoplastic cells have been called naturalkiller (NK) cells.

Gut-Associated Lymphoid Tissue(GALT). Differentiating generalized sys-temic immunity from mucosal immuni-ty is useful. The systemic immune systemincludes all the tissue involved in pro-tecting the body’s interior from invadingmicroorganisms. The mucosal immunesystem consists of the lymphoid tissuethat borders the external environment ofthe gut lumen or other sites such as thelungs and nose. While this classificationis helpful in analyzing diverse functions,many of the specific activities of lym-phoid tissue in the systemic and mucosalcompartments overlap and can affect thefunction of each other.

GALT is made up of aggregated andnon-aggregated tissue (Elwood and Gar-den, 1999). The aggregated componentincludes mesenteric lymph nodes, lym-phoid nodules, and groups of nodulescalled Peyer’s patches that occur right un-der the epithelial cells that line the lumenof the intestine. These sites contain a fullcomplement of the immune cells neces-sary to launch an immune response. Thenon-aggregated tissue includes lympho-cytes, macrophages, and mast cells in thelamina propria (connective tissue beneaththe epithelium) and the intraepitheliallymphocytes in the gut wall.

Antigen Uptake in the Gut. In gener-

33EXPERT REPORT

al, large molecules in the lumen of theintestines are digested into small compo-nent parts before they are absorbed.However, high molecular weight antigenscan move from the gut lumen into theblood. Prior to uptake, the antigens mustresist proteolytic activity in the lumenand penetrate the mucus layer so theycan interact with the various absorptivecell types. Factors that disrupt the mu-cosal barrier function and facilitate theuptake of antigens include immaturegastrointestinal function, malnutrition,inflammation, and immunoglobulin de-ficiencies (Walker, 1987). At least two dif-ferent mechanisms may result in uptakeof these macromolecules (Stokes, 1984).In the first, the intestinal epithelial cellcan incorporate macromolecular aggre-gates through endocytosis and deliverthese to the subepithelial space by exocy-tosis (Walker, 1987). In the second, anti-gens can be deliberately “sampled” by thespecialized epithelial cells (M cells) thatcover Peyer’s patches, which bring the in-tact antigen into the underlying lym-phoid tissue to trigger a comprehensivespecific immune response.

Specific Humoral Responses in theGut. Humoral immunity is mediated byhighly specific proteins known as anti-bodies, which are secreted in response tothe antigen that originally stimulates theantibody formation. Antibodies aresometimes called immunoglobulins(Igs). There are five major classes or “iso-types” of immunoglobulins, each ofwhich functions slightly differently. Ofthese, IgA is of predominant importancein local immunity in the gut becausemuch of it is secreted into the gut lumenwhere it can interact with microorgan-isms before they invade deeper into thebody. In fact, IgA accounts for 60% of to-tal daily antibody production in humans(McGhee et al., 1992). IgA is found bothin mucus secretions (secretory IgA orsIgA) of the gut and as a circulating Ig.Antigens in the gut, including those onmicrobes, are most likely to encountersIgA before any other Ig. Peyer’s patchesare usually considered to be the source ofmost IgA.

Primary roles that have been suggest-ed for sIgA are antigen exclusion, inhibi-tion of adherence of microorganisms, in-tracellular virus neutralization and ex-cretion of IgA immune complexes. Secre-tory IgA induces antigen removal by tak-ing advantage of the normal clearing ac-tivities of the gut (Newby, 1984). Thus,working with the nonspecific immune

system, sIgA is able to inhibit entry ofsoluble antigens and restrict epithelialcolonization by bacteria and viruses.

Specific Cell-Mediated Responses inthe Gut. Cytotoxic T cells also can defenda host against living antigens, such as vi-rally infected cells or intracellular patho-gens. In such a response, a target host cellbearing an antigen of the pathogen on itssurface interacts directly with a cytotoxicT cell resulting ultimately in the lysis ofthe target cell. The killing is unidirection-al and thus cytotoxic T cells can kill nu-merous target cells.

The Common Mucosal Immune Sys-tem. It appears that antigenic stimulationin the gut may result in IgA secretion atother mucosal sites such as salivaryglands and genitourinary sites, leading tothe concept of a “common mucosal im-mune system” (McDermott and Bienen-stock, 1979). While primarily demon-strated in experimental animals, evi-dence exists for a common mucosal im-mune system in humans, based on detec-tion of gut antigen-specific IgA at ana-tomically remote sites. Furthermore, an-tigen-specific IgA producing cells can befound in blood following oral immuni-zation and preceding their appearance insaliva and tears (Russell et al., 1991). Theadvantage of a common mucosal re-sponse relates to the mobilization of hu-moral and cellular immune elements tovarious sentinel sites (e.g., mouth, eye,genitourinary tract) that can prevent in-fection at all of these sites upon subse-quent reexposure to the pathogen.

Stimulation of Specific Gut Immunity.Stimulating the specific immune responsewithin the gut to protect against variousmicrobial illnesses helps prevent disease.However, achieving long-term memorywhen immunizing orally is difficult be-cause ingested antigens tend to be degrad-ed by acidic pH and proteolysis in the gut(Stokes, 1984). Oral immunization withlive organisms is generally more effectivethan nonreplicating organisms for induc-tion of IgA responses, implying that colo-nization and/or replication in the GI tractis required (McGhee et al., 1992). Further-more, particulate antigens function muchmore effectively than soluble ones. Thus,close contact with key components of thegut is required to induce a GI immune re-sponse.

Responses to Infectious Microbes.Many different bacteria, parasites and vi-ruses cause gastroenteritis or penetratethe gut as an entry point to cause sys-temic infection. The capacity to override

the GI defensive barriers depends on sev-eral factors, such as the number of mi-croorganisms and virulence factors (seevirulence, p. 15). Thorne (1986) outlinedfive pathogenic mechanisms for bacterialdiarrheal diseases: (1) bacteria producetoxin but do not generally adhere andmultiply (e.g., B. cereus, S. aureus, C. per-fringens, C. botulinum); (2) bacteria ad-here to the lining of the intestine andproduce toxin (e.g., enterotoxigenic E.coli, Vibrio cholerae); (3) bacteria adhereand damage the villi that make up thebrush border (e.g., enteropathogenic E.coli); (4) bacteria invade the mucosal lay-er and initiate intracellular multiplica-tion (e.g., Shigella spp.); and (5) bacteriapenetrate the mucosal layer and spreadto lamina propria and lymph nodes (e.g.,Yersinia). With each increasing level ofaction, the pathogen’s focus moves fromthe mucosal to the systemic compart-ment, and, hence, the specific immuneresponse must escalate.

The antigen-sampling process itselfmay become a major portal of entry forpathogens (Owen and Ermak, 1990).Wells et al. (1988) hypothesized that, insome instances, a motile phagocyte mayingest an intestinal bacterium, transportit to an extraintestinal site, fail to accom-plish intracellular killing, and then liber-ate the bacterium at the extraintestinalsite. This hypothesis was based on theobservation that the intestinal bacteriathat most readily translocate out of theintestinal tract are facultative intracellu-lar pathogens. Secondly, intestinal parti-cles without inherent motility (e.g., yeast,ferritin, starch) move out of the intesti-nal lumen within hours of their inges-tion. Thirdly, the rate of translocation ofintestinal bacteria can be altered withagents that modulate immune functionssuch as phagocytosis. Thus, systemic in-fection by translocating intestinal bacte-ria could be a result of the antigen-sam-pling process that evolved to regulate theimmune response to intestinal antigens.

Low Molecular Weight Toxins

As discussed above, high molecularweight toxins (proteins, polysaccharides)produced by microbes are cleared by theimmune system. However, the immunesystem does not respond to low molecu-lar weight, nonpolar compounds, such asmycotoxins, which can be rapidly ab-sorbed through the gastrointestinal tract.Higher vertebrates have developed the ca-pacity to metabolize mycotoxins and oth-


er foreign materials (xenobiotics) via aprocess known as biotransformation (deBethizy and Hayes, 1994). The liver is theprimary organ of xenobiotic biotransfor-mation because of its size and its centrallocation in systemic circulation. However,specific limited biotransformation capaci-ties can be found in other tissues and inthe microflora of the intestine.

Biotransformation can be dividedinto two distinct phases. Phase I reactionsadd specific functional groups to the toxinthat are used for subsequent metabolismby phase II enzymes. Phase II reactionsare considered biosynthetic. Biotransfor-mation changes hydrophobic toxins tomore polar, readily excreted compounds.Examples of phase I reactions include ox-idation, reduction, hydration, and hydrol-ysis. Examples of phase II reactions in-clude glucuronidation, sulfate conjuga-tion, glutathione addition, methylation,and acetylation.

Although biotransformation is an im-portant mechanism of host defense, insome cases, biotransformation can makexenobiotics more toxic. Notably, aflatoxinB

1 is converted to a reactive epoxide that

can react with nuclear DNA and causemutations that ultimately result in livercancer, although the original mycotoxinchemical structure is not carcinogenic.

Susceptibility to MicrobialFoodborne Disease

The extent to which the human host issusceptible to disease influences the likeli-hood of foodborne illness. Many factorsplay a role in the level of susceptibility.

Infectious Disease

Susceptibility to infectious disease isthe inability of the host’s body to preventor overcome invasion by pathogenic mi-

croorganisms. Susceptibility to infectiousdisease is increased by conditions that al-ter the host defenses and suppress thefunction of the immune system. Alteredhost defenses and immunosuppressioncan be caused by an infection, anotherdisease, aging, poor nutrition, or certainmedical treatments (see Table 6). Thesefactors have all been implicated in the in-creased risk of infection or increased se-verity of illness caused by many food-borne pathogens including Cryptospo-ridium, Toxoplasma, Campylobacter, Sal-monella, L. monocytogenes, and Giardia(see sidebar, p. 35).

As described above, humans possess anumber of general and specific host de-fenses against foodborne disease. Generaldefenses include normal indigenous mi-croflora, the acidic pH of the stomach,and the antibacterial effect of the variouspancreatic enzymes, bile and intestinal se-cretions. The constant movement of theintestine (peristalsis) helps maintain thebalance of normal flora and purge the in-testinal tract of harmful microorganisms.Factors that alter these general parameterscan increase susceptibility to infection.For example, Salmonella infection is morecommon in patients with decreased stom-ach acidity from medication or after gas-trectomy. Slowing peristalsis with bella-donna or opium alkaloids prolongssymptoms of shigellosis. Similarly, treat-ment of typhoid fever with antibioticsprolongs the carrier state for SalmonellaTyphi, and some evidence indicates thatantibiotic treatment of E. coli O157:H7increases the risk of HUS. Additionally,altering the bowel microflora with broadspectrum antibiotics can lead to over-growth of pathogenic organisms (e.g., Sal-monella).

In the United States, end-stage can-cer, renal disease, end-stage AIDS, liverdisease, and alcoholism are the mostcommon underlying illnesses that di-minish cellular immune response. Im-munosuppression often accompaniesdrug or radiation therapy. Corticoster-oids, chemotherapeutic agents used incancer and organ transplantation, andtotal lymphoid irradiation all suppressthe cell-mediated immune (CMI) func-tion. Organ transplant patients receivingcombined immunosuppressive therapy(corticosteroids, azothiprine, and cy-closporin) face an increased risk of in-fection (or reactivation of quiescent in-fections) with a variety of opportunisticpathogens including L. monocytogenes,Salmonella, T. gondii, Cryptosporidium,

Specific Factors

Age less than 5 years

Age greater than 50 or 60 years(depending on pathogen)

Pregnancy

Hospitalized people

Possession of certain human antigenicdeterminants duplicated or easilymimicked by microorganisms

Concomitant infections

Consumption of antibiotics

Excessive iron in blood

Reduced liver/kidney function(alcoholism)

Surgical removal of portions of stomachor intestines

Immunocompromised individualsincluding those on chemotherapy orradiation therapy; recipients of organtransplants taking immunocompro-mising drugs; people with leukemia,AIDS, or other illnesses

Reasons

Lack of developed immune systems,smaller infective dose-by-weightrequired

Immune systems failing, weakenedby chronic ailments, occurring asearly as 50 to 60 years of age

Altered immunity during pregnancy

Immune systems weakened by otherdiseases or injuries, or at risk ofexposure to antibiotic-resistantstrains

Predisposition to chronic illnesses(sequelae)

Overloaded or damaged immunesystems

Alteration of normal intestinalmicroflora

Iron in blood serving as nutrient forcertain organisms

Reduced digestion capabilities,altered blood-iron concentrations

Reduction in normal defensivesystems against infection

Immune system inadequate toprevent infection

Table 6. Factors That Increase Host Susceptibility (adapted from CAST, 1994)

GeneralFactors

Age

Sensitivepopulations

Underlyingmedicalconditions

35EXPERT REPORT

and Trichinella spiralis. In one study, in-dividuals with chronic heart disease hadan increased risk for listeriosis (Schuchatet al., 1992). Cellular immunity declinesduring pregnancy, which may accountfor the severity of certain infections.

Evidence indicates that immune defi-ciency not only increases the number ofcases but also the severity of infectionfrom a wide variety of foodborne patho-gens. For example, studies conducted inLos Angeles, San Francisco and NewYork City during the mid-1980s demon-strated that patients with acquired im-mune deficiency syndrome (AIDS) hadrates of Campylobacter and Salmonellainfection that were 19 – 94 times the gen-eral population rates in the same cities(Celum et al., 1987; Greunewald et al.,1994; Sorvillo et al., 1994). In addition,16% of Campylobacter infections and44% of Salmonella infections resulted inbacteremia in these compromised pa-tients, much higher rates of severe dis-ease than occurred in the general popu-lation. San Francisco residents infectedwith the human immunodeficiency virus(HIV) also have been shown to have inci-

dence rates of Shigella infection 30 timesgreater than the HIV-free population(Baer et al., 1999), and using an immunesuppressive medication was identified as arisk factor for sporadic E. coli O157:H7infections in a FoodNet case-controlstudy (Kassenborg et al., 1998).

It is known that the neonatal, pediat-ric, adult, and elderly immune systemsdiffer. The fetus and neonates are highlysusceptible to infection with a variety ofpathogens, presumably as a result of animmature immune system. The develop-ment of the immune system begins earlyin fetal development, but children are notimmunologically mature until puberty,putting them at increased risk for food-borne illness.

Improvements in health care and nu-trition in this century have increased thelife expectancy for most people. One re-sult is that the elderly are the fastest-grow-ing segment of our population. Elderlypeople experience significantly greatermorbidity and mortality from infectiousdiseases than the general population. Thisapparent susceptibility to infection in theelderly has been attributed to a decline of

immune function with age, termed “im-mune senescence.” The data regarding theeffects of aging are confusing and some-times conflicting. In general, cell-mediatedimmunity declines, including both func-tional and quantitative cell counts. Su-perimposed and interrelated with thisgeneralized impairment are age-relateddecreases in organ structure and function.Nutritional abnormalities in macro- andmicronutrients are common in the elderlyand may compound immune senescence.The presence of other illnesses and envi-ronmental factors also may contribute tothe decline.

Susceptibility to Biotoxins

The variability of human susceptibil-ity to mycotoxins and other biotoxins canbe attributed to physiologic and environ-mental factors, host genetics, and the pres-ence of infection and inflammation.

Physiologic Factors

A number of factors can influence aperson’s ability to detoxify ingested

CryptosporidiosisCryptosporidium was first de-

scribed in the early 1900s but was notconsidered to be medically or eco-nomically important. During the ear-ly 1970s, Cryptosporidium was linkedto diarrhea in calves, and case reportsdescribing its appearance in a varietyof animal species began to appear inthe literature. In 1976, Cryptosporidi-um was first associated with severewatery diarrhea in a patient who hadbeen receiving immunosuppressivechemotherapy for 5 weeks. The diar-rhea resolved 2 weeks after discontin-uation of the therapy. Following thisreport, additional case reports of se-vere, persistent diarrhea in immuno-suppressed or immunodeficient indi-viduals appeared in the literature (Pit-lik et al., 1983).

During the early 1980s, two seriesof observations began to shape theemerging epidemiology of humancryptosporidiosis. First, cases ofcryptosporidiosis were reportedamong persons who had normallyfunctioning immune systems and ex-

posure to infected calves. While thesepatients typically had self-limited illness-es of mild severity, they demonstratedthat calves with diarrhea were a potentialsource of human infection. The secondobservation was the occurrence ofchronic protracted diarrhea due toCryptosporidium in patients with ac-quired immune deficiency syndrome(AIDS). In many of these patients withsevere cell-mediated immune defects,cryptosporidiosis was unresponsive totherapy and culminated in death (Navinand Juranek, 1984).

During the mid-1980s, the first out-breaks of cryptosporidiosis in child daycare centers and the first waterborne out-breaks of cryptosporidiosis were reported.As clinicians and laboratories becamemore aware of Cryptosporidium, its role asa cause of community-acquired diarrheaemerged. This growing awareness of thepublic health importance of cryptospo-ridiosis culminated with the occurrence ofa massive waterborne outbreak in Mil-waukee, Wis., in 1993. More than 400,000illnesses were attributed to contamination

of water distributed by one of two wa-ter treatment plants in Milwaukee(MacKenzie et al., 1994).

Cryptosporidiosis is now recog-nized as an important cause of diar-rheal illness, and is estimated to cause300,000 illnesses each year in the Unit-ed States (Mead et al., 1999). Geno-typing methods have been developedto discriminate between strains of hu-man and bovine origin, although hu-mans are susceptible to infection withbovine strains. Application of thesemethods to outbreak investigationsand surveillance data will improve ourunderstanding of the epidemiology ofcryptosporidiosis. Interestingly, al-though manure runoff from dairyfarms and effluent from beef slaughterplants were suspected to be likelysources for the Milwaukee outbreak,Cryptosporidium oocysts recoveredfrom outbreak-associated cases wereof the human genotype (Sulaiman etal., 1998). Thus, effluent from a planttreating human waste was a more like-ly source.


biotoxins. For example, biotransfor-mation enzyme activity can vary dur-ing perinatal and postnatal develop-ment (deBethizy and Hayes, 1994).When the individual is older, environ-mental forces play a role in host re-sponse: the nutritional quality of thediet, the presence of chemicals in thediet, and the intake of prescription orelicit drugs can affect the profile ofbiotransformation enzymes. Finally,environmental toxins in air (e.g., ciga-rette smoke) and water can increase ordecrease the activity of biotransforma-tion enzymes toward specific mycotox-ins or other biotoxins.

Genetic Polymorphisms

Wide differences in the human ca-pacity to biotransform biotoxins appearto relate to genetic background that in-fluences the presence, amount and acti-vation of various enzyme systems thatmetabolize ingested biotoxins into chem-ical derivatives that are less or sometimesmore toxic than the original mycotoxin(Daly et al., 1994; Kalow, 1993). For ex-ample, the level of expression forCYP1A2, a cytochrome P-450 (CYP)-de-pendent monooxygenase that metaboliz-es aflatoxin B

1 varies considerably in the

human liver (Eaton et al., 1995). The ac-tivity of microsomal epoxide hydrolase,which acts coordinately with CYP1A2,can vary up to 40-fold in human tissue(Seidegard and Ekstrom, 1997). It hasbeen suggested that epoxide hydrolasepolymorphisms may alter the riskof aflatoxin-associated liver cancer(McGlynn et al., 1995). Specifically, en-zymes may vary in both the amountpresent and their effectiveness/activity,resulting in differing risks of a negativeoutcome from aflatoxin ingestion.

Infection and Inflammation

The simultaneous presence of an in-fectious microbe with attendant inflam-mation can increase the sensitivity of ahost to mycotoxic disease. Epidemiolog-ic studies have demonstrated that hepati-tis B infection predisposes humans whochronically ingest aflatoxins to developprimary liver cancer (Pitt, 2000). In ex-perimental animals, gram-negative bac-terial endotoxin can cause a predisposi-tion to acute liver injury from aflatoxinB

1 (Barton et al., 2000, 2001) and T-2

toxin (Tai and Pestka, 1988) and to de-pletion of lymphoid tissue by deoxyni-

valenol due to leukocyte cell death (Zhouet al., 1999; 2000).

Other Factors

Other significant host factors, apartfrom immune suppression per se, are as-sociated with increased risk of both acutefoodborne disease and chronic sequelae.Genetic predisposition and underlyingchronic disease have been cited as poten-tial risk factors. For instance, septicVibrio vulnificus infections are mostcommonly seen in men over 50 years ofage with liver and/or blood disorders(Desenclos et al., 1991; Tacket et al.,1984). These underlying conditions fre-quently result in elevated serum iron lev-els that play a role in V. vulnificus diseasepathogenesis, although the exact mecha-nism is not fully understood (Wright etal., 1981). Similar evidence is availablefor yersiniosis in iron-overloaded pa-tients treated with deferioxamine, al-though again the pathogenic mecha-nisms are not yet clear (Mandell andBennett, 1995). As previously discussed,genetic predisposition plays a role in thedevelopment of chronic reactive arthritis.Recent evidence from a quantitative hu-man challenge study for the Norwalk-like virus indicates a two-phase dose-re-sponse relationship that appears to beseparately associated with prior exposure(antibody titer) and individual suscepti-bility, neither of which are associatedwith any recognized specific host factors(Moe et al., 1999). When taken together,this body of evidence suggests that manyfactors apart from immune suppressioninfluence host susceptibility to food-borne disease agents.

Individual Choices that AffectDisease Risk

Which foods are consumed and howthose foods are prepared affect an indi-vidual’s risk of foodborne disease. De-spite education efforts, consumer behav-ior continues to play a significant role inexposure to foodborne pathogens (seeTable 7).

Behavior Changes

The 1990s saw a tremendous increasein public awareness of food safety issuesin the United States. This awarenessarose in part because of the continuinginterest in personal health and well be-ing, a phenomenon that occurred

throughout the world’s developed coun-tries. Schools, education programs, me-dia communications, and the Internethave made foodborne and waterbornediseases important concerns to manyconsumers. Outbreaks of foodborne ill-ness that would have gone unnoticed adecade ago are now the subject of rapid,in-depth news coverage. The increasedpublicity about infectious foodbornehazards appears to reinforce food safetymessages and to increase motivations tohandle foods safely.

Proper hygiene and sanitation relatedto food handling and preparation, ap-propriate methods of refrigeration andfreezing, and thorough cooking of foodscomprise a very effective approach topreventing foodborne illness. However,these behaviors are just one aspect of ahealthy life-style. Additional behavioralchanges—such as consuming probiotics,eating a balanced diet, and exercisingregularly to maintain a healthy weight—foster proper functioning of the immunesystem that may heighten resistance tooccasional pathogens in the food supply.

Consumer awareness of food safetyissues has placed additional pressure onthe food service and food processing in-dustries to improve their efforts to en-sure the safety of the products they pro-vide both domestically and internation-ally. Hazard Analysis and Critical Con-trol Points (HACCP) implementationhas become commonplace in the foodprocessing and delivery process. Com-panies have a strong economic incentiveto prevent outbreaks of foodborne ill-ness associated with their product or res-taurant.

While many host factors that influ-ence infection, occurrence and severityof illness are associated with humanphysiology, the factors that influence ex-posure to foodborne pathogens are oftentied to human behavior, specifically con-sumption, food handling, and prepara-tion behaviors.

Eating outside the home in restau-rants and other foodservice venues hasbeen identified as a risk factor for certainfoodborne diseases (Friedman et al.,2000), and the number of meals thatAmericans eat away from home contin-ues to increase. In the 1990s, food eatenoutside the home accounted for almost80% of reported foodborne illness out-breaks in the United States (Bean et al.,1996). Because of the larger number ofpeople involved, these outbreaks aremore likely to be recognized and, there-

37EXPERT REPORT

fore, reported to health officials, so theextent of the role of foodservice in food-borne disease may be overstated. None-theless, the role of foodservice has be-come more significant as the percentageof the food budget spent on eating outhas increased during recent decades(Manchester and Clauson, 1995).Quick-service restaurants and salad barswere rare 50 years ago but are primarysites for food consumption in today’sfast-paced society (Manchester andClauson, 1995).

Regardless of what has or has nothappened to commodities on their wayfrom farm to table, the final commonpathway for food involves storage, prepa-ration, and serving prior to the time ofconsumption. Unfortunately, foods thatare free of foodborne pathogens and tox-ins early in the food chain can becomeunsafe if not handled properly during thefinal stages of preparation and service.

Food handling behaviors such as in-adequate hand washing, unsafe storagetemperatures that permit the growth oflow levels of pathogens, incompletecooking of potentially hazardous foods,and cross-contamination of fresh andcooked foods are a problem whether theytake place inside or outside the home.Based on data from a 1993 nationwidesurvey, an estimated 37% of food han-dlers did not wash their hands after han-

dling raw meat (Altekruse et al., 1995).As many as 42% of survey respondentsdid not cook hamburgers at home untilwell done (Albrecht, 1995). In a 1992study, 23% of respondents reported eat-ing raw shellfish (Timbo et al., 1995).Fortunately, some high-risk food-han-dling and consumption behaviors, al-though still common in 1995 and 1996,did show improvement (CDC, 1998a).

The reasons why some consumersrespond to food safety information andadmonitions to choose safe food andhandle it properly while other consumersdo not are poorly understood. Demo-graphic factors such as gender, age andeducation level are associated with high-risk behaviors. For instance, all behaviorsassociated with increased risk of food-borne diseases were more prevalent inmen than in women. (Albrecht, 1995; Al-tekruse et al., 1995). The 1993 FDAHealth and Diet Survey indicated thatmen were less likely than women to washtheir hands after handling raw meat orpoultry (53% versus 75%) (Altekruse etal., 1995). Results of several studies indi-cate that younger people have a higherprevalence of a number of risky foodhandling, preparation, and consumptionpractices (Altekruse et al., 1995; Klontz etal., 1995; Timbo et al., 1995). Educationefforts are complicated by decreased op-portunities for food safety instruction

both in school and at home. Health edu-cators in secondary schools emphasizeprevention of other important healthconcerns (e.g., HIV infection, obesity)over consumer safety issues includingfood safety education (Collins et al.,1995). In addition, the trend toward two-income families and eating away fromhome leaves fewer opportunities to passfood safety information from parent tochild (Manchester and Clauson, 1995).

Travel, another factor in foodborneillness, has increased dramatically duringthe 20th century. Five million interna-tional tourist arrivals were reportedworldwide in 1950, and the number isexpected to reach 937 million by 2010(Paci, 1995). Travelers may become in-fected with foodborne pathogens un-common in their nation of residence,thus complicating diagnosis and treat-ment when their symptoms begin afterthey return home. In 1992, for example,an outbreak of cholera caused 75 illness-es in international airline passengers; 10persons were hospitalized, and one died(Eberhart-Phillips et al., 1996). Patho-gens also may be carried home and infectfamily members and other close personalcontacts (Finelli et al., 1992).

Changes in Food Consumption Behavior

Potential exposure to pathogens isnot only a function of how food is han-dled, but also what foods individualschoose to eat. Similar to the link betweenpoor sanitation practices and foodborneillness, consumption of certain foods in-creases the risk of illness.

Changes in food consumption havebrought to light previously unrecognizedor underestimated microbial hazards.Fresh fruit and vegetable consumption,for example, increased nearly 50% from1970 to 1994 (BC/USDC, 1996). Produceis susceptible to microbial contamina-tion during growth, harvest, and distri-bution (see section on microbial ecology,p. 40), which is of special concern forfoods eaten fresh and not cooked. Patho-gens on the surface of produce (e.g., mel-ons) can contaminate the interior duringcutting and multiply if the fruit is held atroom temperature (Reis et al., 1990). Inthe United States from 1990 to 1997, aseries of foodborne outbreaks were asso-ciated with produce such as sliced canta-loupe (Reis et al., 1990), green onions(Cook et al., 1995), unpasteurized cider(Besser et al., 1993), fresh-squeezed or-ange juice (Cook et al., 1996), lettuce

Specific Factors

Stress

Poor hygiene

Geographic location

Nutritional deficiencies either throughpoor absorption of food (mostly ill orelderly people) or unavailability ofadequate food supply (starving people)

Consumption of antacids

Consumption of large volume of liquidsincluding water

Ingestion of fatty foods (such aschocolate, cheese, hamburger)containing pathogens

Reasons

Body metabolism changes, allowingeasier establishment of pathogens, orlower dose of toxin required for illness

Increased likelihood of ingestion ofpathogens

Likelihood of exposure of endemic virulentstrains; limited and/or compromised foodand water supply; variable distribution oforganisms in water and soil

Inadequate strength to build up resistanceand/or consumption of poor-quality foodingredients, which may contain pathogens

Decreased stomach acidity (increased pH)

Dilution of acids in the stomach and rapidtransit through the stomach

Protection of pathogens against stomach

Table 7. Factors That Increase Risk of Foodborne Disease (adapted from CAST, 1994)

GeneralFactors

Life-style

Diet


(Ackers et al., 1996), raspberries (Her-waldt, 1997), alfalfa sprouts (Mahon etal., 1997), sliced tomatoes (Wood et al.,1991), and frozen strawberries (CDC,1997a).

In addition to relative changes inquantities, the past few decades haveseen dramatic changes in the diversityof foods available to the American pub-lic. The growing wealth of Americansand the profitability of fresh producehas led to the introduction and in-creased availability of a wide variety ofproduce items—from kiwi, mangoesand papayas, to alfalfa sprouts, specialtylettuces and fresh-cut, packaged pro-duce. Imported produce has played asignificant role in increasing diversity.Fresh produce has also become a main-stay of restaurant fare; dinner saladsand salad bars have become main-stream. In addition, ethnic cuisines thatfeature fresh produce ingredients, suchas Chinese, Mexican, Thai, and MiddleEastern, have become popular.

Individual Choices

Americans pride themselves ontheir individual freedoms. In a culturebased on these individual rights, we al-low people to engage in high-risk be-havior and offer products that some-times cater to these risks. Steps are tak-en to mitigate risk but ultimately cer-tain behaviors are inherently risky.Sometimes people knowingly engage inhigh-risk behavior. If people are awareof the risks and continue to engage inrisky behavior such as eating raw oys-ters or eggs, it is appropriate to consid-er to what lengths our society shouldgo to protect them.

Surveys conducted from 1998-1999as part of the FoodNet Active Surveil-lance Program for foodborne diseaseshave documented certain aspects ofconsumer behavior. With regard toconsumption of fresh produce that isknown to be at particular risk for mi-crobial contamination, 19% of respon-dents reported eating a mesclun lettucemix in the 7 days before the interview,and 8% reported eating alfalfa sprouts,although these eating habits are highlyregional (CDC, 1999a). Among otherpotentially risky food exposures, 25%of the people who had eaten eggs hadchosen to eat eggs that were runny;11% of persons who consumed ham-burgers ate burgers that were still pinkinside; 4.4% drank unpasteurized ap-

ple juice or apple cider; 3.4% drankunpasteurized milk; and 2.5% ate freshoysters.

Cultural Differences

Diet selection can create subpopula-tions at greater risk for certain food-borne illnesses. Some reports of food-borne illnesses involve transmission viafoods consumed primarily by immigrantgroups. Outbreaks of trichinosis havebecome relatively rare in the UnitedStates because cooking pork thoroughlyhas become a widespread cultural prac-tice. An exception occurred in 1990,when Laotian immigrants in Iowa pre-pared and ate undercooked pork, a tradi-tional food, as part of a wedding celebra-tion (Stehr-Green and Schantz, 1986).Other reports involve foods more com-monly consumed by ethnic populations.Y. enterocolitica outbreaks are also rare,but several outbreaks in African-Ameri-can communities were associated withpreparation and consumption of pig in-testines (Lee et al., 1990). The epidemiol-ogy of human brucellosis in Californiahas shifted from an occupational diseaserelated to animal husbandry to a food-borne disease most frequently affectingHispanics who often consume raw milkand cheeses made with raw milk whileabroad (Chomel et al., 1994). Consump-tion of rare hamburgers—a risk factorfor E. coli O157 infection—is more com-mon in U.S. Caucasians than in any oth-er racial/ethnic group (CDC, 1998a).

Dietary Recommendations

A host of organizations have issueddietary recommendations and providedinformation to assist in health promo-tion and chronic disease prevention (U.S.PSTF, 1997). These recommendationsmay advocate increased or decreasedconsumption of certain types of foods,outline specific circumstances for con-sumption, or caution individuals withcertain medical conditions. Each type ofrecommendation has consequences forfoodborne illness. In recognition of theimportance of food safety, the latest edi-tion of the federal government’s dietaryguidelines contains a section on foodsafety.

Counseling the general population tolimit dietary intake of fat and emphasizefoods containing fiber (i.e., fruits, vegeta-bles, grain products) has increased con-sumption of foods such as fresh produce

and leaner meats such as chicken. Recentincreases in the consumption of health-promoting fresh fruits and vegetableshave resulted in increased likelihood ofexposure to certain diseases like hepatitisA, shigellosis and salmonellosis fromcontaminated produce (Tauxe et al.,1997). The dietary shift toward increasedconsumption of chicken may have con-tributed to the high incidence of C. jeju-ni infection (Friedman et al., 1992),which now exceeds Salmonella as themost common bacterial cause of food-borne illness (Mead et al., 1999). A recentstudy demonstrated that only 17% ofAmericans ate five or more servingsof fresh fruits and vegetables per day(Thompson et al., 1999). Thus, publichealth marketing campaigns are likely toincrease fruit and vegetable consump-tion in years to come.

Medical and public health advisorybodies also have advised certain subpop-ulations to use special caution in diet se-lection and food preparation. For exam-ple, the American Academy of Pediatricshas recommended that children shouldnot drink unpasteurized milk or eat un-pasteurized cheese, undercooked eggs,raw or undercooked meat or meat prod-ucts (AAP, 2000). FoodNet populationsurveys (CDC, 1999a) demonstrate thatthese types of dietary recommendationsdo have an impact on consumer behav-ior, although final analyses are not com-plete. While 13.5% of adults aged 20-39who consumed hamburgers ate ham-burgers that were pink, only 4.4% ofchildren under 10 years of age did so.Similarly, adults were 2–3 times morelikely than children to drink unpasteur-ized milk, or eat alfalfa sprouts or runnyeggs (CDC, 1999a).

Ignoring recommendations aboutpreparation practices such as adequatecooking places consumers at greater riskfor foodborne illness. Consumer educa-tion is an important part of foodborneillness prevention. As the statistics abovedemonstrate, the media attention, safehandling labels, public health advisories,and public information and educationcampaigns to date have left a substantialpart of the population unprotected. For avariety of reasons, food safety and otherpublic health messages fail to reach theirintended audience, are misunderstood,or are disregarded. For those chargedwith preventing foodborne disease, thereare two inescapable lessons in this data:first, we need to know a great deal moreabout risk communication, education,

39EXPERT REPORT

and motivation before consumer andfood worker education programs can beconsidered credible parts of our overallfood safety strategy; and second, as longas pathogens are delivered to home andcommercial kitchens, some foodbornedisease will occur.

Modification of Susceptibility

Because of the significance of the hu-man host in foodborne illness, it is ap-propriate to look for host-related oppor-tunities for control or mitigation of ill-ness. Certainly, behavioral and demo-graphic issues influence susceptibilityand exposure. If the host could be ren-dered immune to infectious agents, ill-ness would cease to occur despite micro-bial contamination in the food supply.Although complete protection is current-ly out of reach for many pathogens, ap-proaches such as immunization and pro-biotics can decrease human susceptibili-ty. Coupling these approaches with ef-forts to mitigate exposure would furtherboost our ability to control foodborneillness of infectious origin.

Immunization

Vaccines use the host’s own immunesystem to combat disease. Knowing howthe immune system functions enablesscientists to investigate methods to en-hance its effectiveness or trigger its pro-tective effects without causing illness.The medical community has had greatsuccess with vaccination for some infec-tious diseases, and potential vaccines toprevent foodborne illness are the subjectof considerable research.

Although no vaccines are currentlyavailable for most enteric bacterialpathogens, experimental approaches areunder investigation. In general, vaccina-tion strategies to control enteric bacterialdiseases are complicated by many factors,not the least of which is the complexityof host immunity as well as a general ab-sence of appropriate animal models fororal challenge and subsequent diseasepresentation for many of these patho-gens. For example, an effective broad-spectrum vaccine against enterohemor-rhagic E. coli (EHEC) will likely need totarget systemic immunity against theShiga toxins as well as local intestinalimmunity against intestinal colonizationfactors (Nataro and Kaper, 1998). None-theless, genetic manipulation has en-abled researchers to produce bacterial

mutants that have an impaired ability tosurvive in vivo, and some of these havebeen used as live oral vaccines to immu-nize against subsequent infection withwild-type virulent strains (Fairweather etal., 1990). An alternative vaccination ap-proach has focused on the use of puri-fied antigens, although this usually re-sults in a comparatively poorer immuneresponse when administered orally (Fair-weather et al., 1990).

Scientists have had greater successdeveloping effective vaccines against thehuman enteric viruses that are common-ly spread through food and waterborneroutes. For instance, we are in the finalstages of worldwide eradication of thepoliovirus, an accomplishment madepossible by widespread immunization(Jacob, 2000). Vaccines for HAV are an-other success story. Licensed in the mid-1990s, these vaccines consist of forma-lin-inactivated organisms and are welltolerated; they produce durable immuni-ty persisting for more than 20 years(Cuthbert, 2001). Recent evidence alsoindicates that the HAV vaccine effectivelyprevents secondary HAV infection andmay be appropriate for administration toindividuals in frequent personal contactwith infected persons, replacing thewidely used immunoglobulin (Saglioccaet al., 1999). Although there has been in-terest in mandating routine HAV vacci-nation for food handlers, there is cur-rently no overwhelming support for thisproposal, in part due to the expense ofthis type of approach. Perhaps more effi-cacious would be recommendation ofroutine vaccination of individuals withunderlying chronic active hepatitis dueto infection with hepatitis B or C viruses,since this subpopulation is particularlysusceptible to very serious disease mani-festations if concurrently infected withHAV.

In 1998, FDA also licensed a live at-tenuated rotavirus vaccine (Rotashield,Wyeth-Lederle Vaccines and Pediatrics,Philadelphia, Penn.) for oral administra-tion to infants (AAP, 1998). Unfortu-nately, in the summer of 1999, CDC re-ported a clustering of cases of an intesti-nal complication in the weeks after vacci-nation, eventually leading to the volun-tary withdrawal of the product from theU.S. market and an uncertain future forthe vaccine (Weijer, 2000). Someprogress has been reported in vaccina-tion against the human gastrointestinalcaliciviruses. However, protection againstthis important group of foodborne in-

fectious agents remains challenging be-cause immunity is poorly understood,the group has tremendous antigenic andgenetic diversity, and the viruses cannotbe cultivated in model animal or labora-tory systems.

Probiotics

Probiotics represent another oppor-tunity to use our understanding of thegut microflora and the immune responseto facilitate human health and decreasesusceptibility to illness. Although origi-nally used to describe substances pro-duced by one protozoan that stimulatedanother (Fuller, 1989), the definition forthe term “probiotic” now generally im-plies a viable microbial supplement thatbeneficially affects the host (human oranimal) by improving or maintaining adesirable microbial balance in the gut.The reduction of harmful enteric micro-organisms is only one of numerous po-tential health benefits of maintaining ahealthy gut microflora. Probiotic cul-tures are consumed in foods or capsulesor are facilitated by ingesting prebiotics(compounds that enhance the prolifera-tion of beneficial indigenous bacteria).Currently, dairy foods such as yogurthave been the most popular vehicle ofchoice to deliver viable probiotic cul-tures. Intestinally-derived lactobacilliand bifidobacteria predominate in thisrole (Hughes and Hoover, 1991).

The effect of probiotic cultures variesbased on numerous conditions. The hu-man gut contains 100 trillion viable bac-teria and other microorganisms repre-senting anywhere from 100 to 400 differ-ent species; the population dynamics arequite complex. The microflora of thehuman intestinal tract is affected by ge-netic or host factors, the composition ofmicrobial populations, and the metabo-lites produced by these microbes; thesefactors are in turn influenced by climate,diet, stress, drugs, age, and disease (Mit-suoka, 1990). One can maintain thatwhenever there is a change in the intesti-nal microflora from its normal state, thechange is detrimental or undesirable. Afoodborne intestinal infection that pro-duces diarrhea can be viewed as a periodof microbial imbalance or instability inthe gastrointestinal tract. However, by es-tablishing themselves in the human GItract in proportionally high numbers,acidulating bacteria, such as lactobacilliand bifidobacteria, may protect the gutagainst invasive pathogenic agents.


Thus, regular consumption of foodscontaining probiotics has a strong po-tential to help maintain a beneficial andstable intestinal microflora that pro-motes intestinal health. This is especial-ly true for those subpopulations withcompromised or underdeveloped gutflora, such as the elderly, infants, andpatients treated with antibiotics or che-motherapy. For example, in the elderly,there is a steady decline in numbers ofbifidobacteria and an increase in thenumbers of C. perfringens with age.With this shift in gut flora, there is acorresponding increase in putrefactivesubstances in the intestinal tract that areinherently toxic and impose a constantstress upon the liver (Mitsuoka, 1990).In addition to C. perfringens, the putre-factive organisms that convert aminoacids into amines and other toxic sub-stances include Salmonella, Shigella, andE. coli. As a means to ameliorate thesedetrimental conditions, elevated levelsof bifidobacteria and lactobacilli fromdietary probiotics reduce fecal pH todiscourage growth and colonization by

acid-sensitive enteric pathogens, wheth-er the pathogens are indigenous to theintestinal population or opportunisticcontaminants of food and water (Lang-hendries et al., 1995).

A similar case for administration ofprobiotic cultures would be for the new-born (Mitsuoka, 1989). Within a day ofbirth, bacteria commence colonizationand proliferation in the previously sterileintestinal tract. Initially, coliforms, en-terococci, staphylococci, and clostridiaappear, but in three to four days afterbirth, lactobacilli and bifidobacteria pre-dominate. Bifidobacteria soon dominateall other bacteria, whether the infant isbreast-fed or bottle-fed. However, in bot-tle-fed infants, populations of coliformsand enterococci are ten times higher thanin breast-fed infants, a fact that may en-courage the use of infant formula thatincludes probiotic cultures.

The feeding of probiotic cultures toprevent or treat disease is well estab-lished in the scientific literature. ElieMetchnikoff of the Pasteur Institute firstpromoted the use of probiotics nearly

100 years ago (Metchnikoff, 1908).More recent examples include Saavedraet al. (1994), who showed that supple-menting infant formula with Bifidobac-terium bifidum and Streptococcus ther-mophilus can reduce the incidence ofacute diarrhea and rotavirus sheddingin infants; Bernet et al. (1994), whofound that consumption of a greaternumber of lactobacilli provided in-creased protection against cell associa-tion by enterotoxigenic and entero-pathogenic E. coli and S. Typhimurium,and against cell invasion by entero-pathogenic E. coli, S. Typhimurium andY. pseudotuberculosis; and Okamura etal. (1986), who used a tissue culture in-fection assay to demonstrate that ad-ministration of Bifidobacterium infantisprohibited invasion and intracellularmultiplication of S. flexneri. In all, pro-biotic cultures have demonstrated aninhibitive or antagonistic effect againstalmost all foodborne pathogens, includ-ing Salmonella, Shigella, E. coli, Campy-lobacter, Clostridium, Yersinia, Vibrioand Candida (Fuller, 1992).

Microbial Ecology and Foodborne Disease

The complexity of the pre-harvest,

harvest, and post-harvest environ-

ments makes it impossible to control

all potential sources of microbial

contamination. Efforts at prevention

and control are implemented through-

out the food production and processing

system. Researchers are continually

searching for a better understanding of

the pathogens and their interaction

with the environment, leading to

improved control technologies. But at

the same time, the pathogens continue

to evolve, and human actions some-

times drive that evolution. Even small

environmental changes can have

unforeseen or even unforeseeable

impact on microbial populations.

Improved understanding of these

complex factors provides insight into

pathogen evolution and opens the door

to new and improved prevention and

control methods.

PRE-HARVEST ENVIRONMENT

Efforts to minimize microbial con-tamination of food begin in the pre-har-vest environment. Raw ingredients areone way in which pathogens are intro-duced into the processing environment.Unfortunately, pathogen control in theproduction agriculture environment isoften difficult.

Overarching Issues

In the pre-harvest environment, manyof the significant issues in microbiological

food safety are broad concerns that applyto many different commodities. Becausethese issues affect so many commodities,improvements in these areas would havesignificant food safety impact.

Global Food Trade

Globalization of the world’s foodsupply has contributed to changing pat-terns of food consumption and food-borne illness. A growing percentage ofthe U.S. food supply is imported. Thesheer volume of these imports adds tothe complexity of foodborne illnesses.

Global sourcing provides economicbenefits and a wider selection for con-sumers that improves nutrition world-wide. However, in terms of diseasecontrol programs, globalization mini-mizes traditional geographic barriersto emerging as well as traditionalpathogens. Developing economies rep-resent major sources of certain im-

41EXPERT REPORT

ports (see Table 8). For many of thesecountries, infectious diseases still rep-resent a significant burden of illness.Diarrhea remains among the ten lead-ing causes of death and disease burdenin the developing world (Murray andLopez, 2001).

Imported foods can introduce patho-gens previously uncommon in the UnitedStates, such as Cyclospora and new strainsof Salmonella. Raspberries from Guate-mala, cantaloupes and scallions fromCentral American countries, coconut milkfrom Southeast Asia, and a Middle East-ern snack food have all been implicated inrecent foodborne disease outbreaks in theUnited States. As the percentage of im-ported foods consumed in the UnitedStates increases, the importance of ensur-ing that these foods are safe increases aswell. Food safety therefore cannot beachieved by focusing on domestic prod-ucts exclusively (GAO, 1998).

As shown in Table 9, the importshare of some commonly consumedfoods is increasing. For example, in1995, one-third of all fresh fruitsconsumed in the United States wasimported, and this trend is likely tocontinue.

Manure

The widespread occurrence and useof animal manure as fertilizer is a grow-ing environmental concern, because itcontaminates: water for drinking, irriga-tion, aquaculture and recreation; thehides, coats, and feathers of farm ani-mals; and farm equipment and build-ings. In the United States, cattle, hogs,chickens and turkeys produce an esti-mated 1.36 billion tons of manure annu-ally (EPA, 2000), with greater than 90%attributed to cattle. Each year livestockcreate an estimated five tons of animalmanure per person living in the UnitedStates, meaning the amount of animalmanure is 130 times greater than theamount of human waste produced (U.S.Senate Agriculture Committee Demo-cratic Staff, 1998).

Many of the most prominent food-borne pathogens in the United States, in-cluding Campylobacter jejuni, Salmonellaand Escherichia coli O157:H7, are carriedby livestock and are principally transmit-ted to foods by fecal contamination. C.jejuni accounts for an estimated 2 mil-lion cases of foodborne illness annually,with poultry and unpasteurized milk asits principal vehicles (Mead et al., 1999).Salmonella causes an estimated 1.3 mil-lion cases of foodborne illnesses annual-ly, with eggs, poultry, beef, pork and pro-duce as primary vehicles (Mead et al.,1999). Both C. jejuni and Salmonella arecarried in the intestinal tract of appar-ently healthy poultry and livestock. Fecalcontamination of hides, feathers andskin occurs during poultry and livestockproduction and slaughter. This contami-nation can subsequently carry throughto processing. A case-control study ofpatients with Campylobacter infectionidentified the following risk factors forcampylobacteriosis: foreign travel; eat-ing undercooked poultry; eating chicken,turkey or non-poultry meat cooked out-

side the home; eating raw seafood; drink-ing raw milk; living on or visiting a farm;and having contact with farm animals orpuppies (Friedman et al., 2000). Fecalcontamination is a common source of C.jejuni contamination for each risk factor.For example, poultry feces frequentlycontain C. jejuni at populations of 105 to107 colony forming units of bacteria pergram (CFU/g), resulting in levels ofgreater than 103 CFU C. jejuni per gramof carcass in 60 to 90% of retail poultry.

E. coli O157:H7 causes an estimated73,500 cases of infection in the UnitedStates annually. Its principal vehicles oftransmission are beef, produce, water(both drinking and recreational), andcontact with cattle (Doyle et al., 1997;Griffin, 1998). Because E. coli O157:H7 iscarried in the intestinal tract of cattle, thepathogen’s most frequent origin is director indirect contact with cow manure. Ma-nure can contaminate food when used asa soil fertilizer, when it pollutes irrigationwater, when cattle defecate near produceor foods of animal origin, and when in-testinal contents or manure-laden hidescontact carcasses during slaughter andprocessing. Case-control studies of pa-tients with E. coli O157:H7 infections re-vealed several major risk factors for ill-ness: eating undercooked ground beef, liv-ing on or visiting a farm, and having con-tact with farm animals, especially cattle(Kassenborg et al., 1998). Depending onenvironmental conditions, E. coliO157:H7 can survive in manure for manyweeks, and in some instances for morethan one year. Similarly, the pathogen cansurvive well in lake water, with as little as a10- to 100-fold reduction occurring dur-ing 13 weeks at 8 C.

Increased proximity and animal den-sity during production contribute toproblems of pathogens in runoff waterbecause the difficulty of manure man-agement is increased with greater vol-ume. Another issue is composting of ma-nure by farmers, including organic farm-ers. The conditions that effectively de-stroy pathogens are not well defined.Also, manure handling in other coun-tries may be worse than in the UnitedStates, a serious concern for importedfoods. Human feces from field workerswithout access to adequate sanitation fa-cilities remains an issue as well.

Water

In the pre-harvest environment, wa-ter can be obtained from a variety of

Country

Mexico

Canada

Chile

Costa Rica

Netherlands

Guatemala

Other

Percent(by $ value)

51

15

12

4

3

2

13

Table 8. Sources of Imported Fresh andFrozen Produce (1997 data) (GAO, 1999)

Table 9. Percentage of Total U.S. Consumption Provided by Imports (GAO, 1998)

Import Item

Fish & shellfish

Fresh fruits

Fresh vegetables

Tomatoes for processing

Broccoli for processing

1980

45.3

24.2

7.6

1.4

9.1

1985

53.8

28.0

8.9

7.0

22.2

1990

56.3

30.7

8.4

5.7

57.8

1995

55.3

33.3

11.7

3.5

84.9

% Change(1980-95)

22.1

37.6

53.9

150.0

833.0


sources, although ground water and wellwater are perhaps the most widely used.Most farms do not provide specific treat-ment of water for use in agriculturalproduction, and the water sources rou-tinely used in agriculture can becomecontaminated by a number of means.Perhaps the most common source is ani-mal manure contamination of runoffwater; less common is contamination ofwater with untreated human sewage,which is largely under control in theUnited States and other developed coun-tries, but of considerable concern forfoods produced in developing countrieswith inadequate water resources. Watersources also may become contaminatedby fecal excrement from wild animals orby general contamination of soil, but thesignificance of these to food safety islargely unknown.

Associations have been made be-tween the presence of pathogens in wa-tering troughs and their subsequentprevalence in animals. For instance, in-vestigators have reported isolation of E.coli O157:H7 in water troughs sampledfrom cattle farms (Faith et al., 1996; Mid-gley and Desmarchelier, 2001; Sargeantet al., 2000; Shere et al., 1998), and othershave reported that the practice of flush-ing alleyways with water to remove ma-nure results in as much as 8-fold in-creases in animal carriage rates (Garberet al., 1999). Furthermore, the organismis able to persist for days at ambient tem-perature in both soil and water (Maule,2000; Rice and Johnson, 2000). Likewise,it is recognized that contaminated wateris a significant source of Campylobacterfor infection of commercial poultryflocks (Shane, 2000). Such contamina-tion is usually followed by rapid, intra-flock dissemination, which has been ex-acerbated by intensification of animalagricultural practices (Gibbens et al.,2001; Shane, 2000).

Recent evidence of foodborne diseaseoutbreaks associated with the consump-tion of fresh produce has promptedsome to consider the role of contaminat-ed irrigation and surface runoff waters.Irrigation water containing raw or im-properly treated human sewage can bethe source of many pathogens, with Shi-gella and the enteric viruses (hepatitis Avirus, Norwalk-like viruses, rotaviruses)being perhaps the most significant(Beuchat, 1996; Beuchat and Ryu, 1997).Irrigation water contaminated with ani-mal fecal matter can also be a source ofpathogens on fresh produce. Although

animal fecal material may contain a widevariety of potential human pathogens, itappears that the heartier survivors, suchas parasitic protozoan oocysts(Cryptosporidium spp.) are likely to posethe greatest risk (Beuchat, 1996; Jaykus,1997). Also, the relative importance ofcontaminated irrigation water as op-posed to direct fecal contact is unknownfor pathogens such as Campylobacter,Listeria monocytogenes, E. coli O157:H7,and Cyclospora cayetanensis.

Typical Pre-Harvest Environmentfor Foods of Plant Origin

The microbiological status of a foodproduct at the time of consumption is afunction of its history. What kinds of mi-croorganisms and how many exist onand in the food are a direct result of thecircumstances of its production andhandling. During the pre-harvest pro-duction period and the harvest process,many opportunities exist for microor-ganisms to contaminate food materials.During the past 50 years, farming prac-tices have changed considerably. In gen-eral, intensive farming practices have im-proved process control but also havecontributed significantly to the rapidspread of human and animal pathogensby creating more concentrated environ-ments for pathogens to multiply andevolve and by generating larger quanti-ties of subsequently contaminated food(Rangarajan et al., 2000). At the sametime, distribution networks have become

more complex. Complicating the situa-tion further, microorganisms rapidlyadapt to new, adverse environmentalconditions, allowing them to survive andreplicate under extreme conditions in-volving high and low temperatures, pH,osmotic pressures, and oxygen levels thatare inhospitable to most higher forms oflife (Jay, 2000; Kushner, 1980).

As discussed in the previous section,American diets contain increasingamounts of fresh fruits and vegetables.Produce is commonly consumed raw(unprocessed), which makes it impossi-ble—with the currently available tech-nologies—to guarantee it is free of con-taminating pathogenic microorganismswhen consumed.

To control foodborne illness, the en-tire food supply chain must be consid-ered (Baird-Parker, 2000). In the case ofSalmonella transmission (Baird-Parker,1990), direct contributing factors for thecontamination of pre-harvest produceinclude contact with manure, water, hu-mans, livestock, wildlife, pets, environ-mental pollution and effluent/sewage.The primary source is considered to becontact with human or animal feces. Asnoted above, water is a major concernbecause it is used so extensively in farm-ing.

Not all bacterial and fungal foodpathogens exist in the pre-harvest envi-ronment as a result of human or animalfecal contamination. For example, manysporeforming bacteria of food safetyconsequence are native to soil and water,

Production Practicesand Mycotoxins

A recent example of how changesin ecology and production practiceshave affected mycotoxin incidence inthe United States is the massive in-crease in Fusarium head scab in Mid-western wheat and barley during thelast decade (McMullen et al., 1997).Head scab is often accompanied by el-evated contamination by deoxynivale-nol (vomitoxin) and other trichoth-ecenes. Two factors seem to have driv-en the head scab epidemic. One is anincreased spring rainfall during earlywheat head formation. The uncharac-teristic increase in rainfall may be aresult of the prolonged El Niño dur-

ing the 1990s, long-term climaticchanges, or global warming from hu-man activities. The second causativefactor is the increased use of no-tillagriculture methods, which have beenimplemented to reduce soil erosion.Residual stubble left in a field duringwinter can provide a way for fusaria tocontaminate the following year’s crop.

As the environment changes,sometimes as a result of human ac-tion, the microbial populations adapt.Some environmental changes can in-crease pathogen levels by providingfavorable conditions; other changescan select for traits that result in resis-tant microorganisms that survive un-favorable conditions.

43EXPERT REPORT

e.g., Clostridium botulinum and Bacilluscereus. Mycotoxigenic varieties of fungiinclude Fusarium, Claviceps purpureaand aflatoxin-producing strains of As-pergillus. In addition to known patho-gens, the environment represents a sub-stantial reservoir for potential emergingpathogens.

Because fresh produce undergoesvery little processing, emphasis at thefarm level has been directed towards theprevention of microbial contaminationrather than relying on corrective actionsonce contamination has occurred. Al-though preventing contamination ofcrops by pathogenic microorganisms isimportant, it is very difficult to accom-plish consistently and reliably, given thelarge number of possible sources ofpathogens prior to harvest. Science-based farming guidelines, known asgood agricultural practices, have beendeveloped to control microbial contami-nation in an effort to improve the safetyof produce (FDA/CFSAN, 1998).

Typical Pre-Harvest Environmentfor Foods of Animal Origin

Meat animal production has in-creased dramatically since 1975 (USDA/NASS, 2000). The largest increase hasbeen in poultry production, which rosefrom approximately 10 billion pounds in1975 to 40 billion pounds in 1999. Cattleproduction has remained relatively steadysince the early 1970s at approximately 41billion pounds. In addition to cattle andpoultry, 24 billion pounds of pork areproduced annually. These productionquantities, coupled with limited space forlivestock on the farm, promote the dis-semination of microorganisms such assalmonellae. A higher prevalence ofpathogens in food animals increases thechances that meat will become contami-nated, providing a route for the pathogensto reach humans. Furthermore, to in-crease livestock health, feed efficiency, andgrowth rates in these confined conditions,antibiotics are often added to animal feed,potentially contributing to the develop-ment of antibiotic-resistance in microor-ganisms that live in animals (zoonotic mi-croorganisms) (Angulo et al., 2000; Witte,1998). Approximately half of the antimi-crobials produced today are used in hu-man medicine; most of the remainder isadded to animal feed (WHO, 2002). Theemergence of antibiotic-resistant humanpathogens like Salmonella and Campylo-bacter spp. limit therapeutic options avail-

able for treating invasive human infec-tions.

Use of Antibiotics

The widespread use of antibiotics inanimal production and in the treatmentof human illness both facilitate theemergence of antibiotic resistance. Mi-croorganisms can develop resistance toantimicrobials through gene mutationsor by acquiring transferable genetic ele-ments, such as plasmids and conjugativetransposons, that harbor resistancegenes. These mobile genetic elementsare important in horizontal transmis-sion of genes from the resident to tran-sient microflora of the intestinal tract(Levy et al., 1976). In addition to themobility of the genetic elements, the an-tibiotic-resistant bacteria can be trans-mitted to different animal hosts. A tet-racycline-resistant E. coli strain fromcattle was traced to humans, mice, pigs,and fowl found at the same location(Levy et al., 1976). The selective pressurecaused by antibiotic administrationcauses the microbial populations thatharbor the appropriate resistancedeterminant(s) to flourish (Levy, 1992).These antibiotic-resistant microbes canmake their way to humans through con-taminated foods or animal-to-humantransmission (Angulo et al., 2000;Holmberg et al., 1984), although thepublic health impact of the use of veteri-nary drugs is difficult to measure (How-gate, 1997).

The contribution of sub-therapeuticlevels of antibiotics in animal feed to theemergence of antibiotic-resistant patho-gens has been debated for years (Fein-man, 1998). A growing body of evidencefrom epidemiological data and trace-back studies indicates that agriculturaluse of antibiotics plays an importantrole in the emergence of some antibiot-ic-resistant bacteria (Angulo et al.,2000). A review of Salmonella out-breaks between 1971 and 1983 revealedthat antibiotic-resistant strains weremore likely to originate from animalsthan were strains without resistance(Holmberg et al., 1984). Additionally,the emergence of Salmonella with de-creased susceptibility to fluoroquino-lone paralleled approval of the veteri-nary use of enrofloxacin, a fluoroqui-nolone antibiotic, even though fluoro-quinolones had been used in humansfor the preceding six years with little im-pact on the development of resistant

Salmonella (Threlfall et al., 1997). Al-though there is no evidence that the useof antibiotics in feed is responsible forthe evolution of the multi-drug resistantstrain of Salmonella TyphimuriumDT104, the rapid dissemination of thisstrain in animals and humans indicatesthere is an advantage for strains with theantibiotic-resistance phenotype (seesidebar, p. 44). S. Typhimurium DT104 isthe second leading cause of human sal-monellosis in England and Wales(Anonymous, 1996) and the most com-mon Salmonella species isolated fromcattle (Hollinger et al., 1998). S. Typh-imurium DT104 is resistant to ampicil-lin, chloramphenicol, streptomycin, sul-fonamides, and tetracycline (known asR-type ACSSuT). The resistance genes tothe antibiotics are located on the chro-mosome rather than a plasmid, indicat-ing they are nonmobile and stable(Threlfall et al., 1995). In the UnitedStates, the ACSSuT resistant pattern waspresent in 28% of 976 S. Typhimuriumisolates collected nationally in 1995, asubstantial increase from 7% in 1990isolates (Hosek et al., 1997). Human in-fection by S. Typhimurium DT104 hasgreater morbidity and mortality than theother nontyphoid Salmonella infections(Wall et al., 1994). Because the accumu-lating data from molecular subtypingmethods and epidemiological investiga-tions suggest that the use of sub-thera-peutic levels of antibiotics in animalfeeds plays a role in the emergence of an-tibiotic-resistant foodborne pathogens,the prudent and judicious use of antibi-otics in both the agricultural and medi-cal sectors is needed.

Feeding Practices

Another animal production practicewith potential ramifications for microbi-ological food safety is farm animal dietcomposition. A change of diet, for exam-ple, can change the microbial ecology ofthe ruminant digestive system. Manystudies have evaluated the effect of di-etary changes on fecal shedding of E. coliO157:H7 or acid-tolerant E. coli by cattleor sheep, with conflicting results. Someinvestigators have determined that sheepor cattle fed hay shed E. coli O157:H7 intheir feces considerably longer than ani-mals fed grain (Hovde et al., 1999; Kudvaet al., 1997). In contrast, studies of acid-tolerant E. coli, which is a characteristicof E. coli O157:H7, revealed that a mostlygrain diet promotes shedding of acid-tol-


erant E. coli in comparison to hay-fedcattle (Diez-Gonzalez et al., 1998). Sub-sequent studies revealed that changingthe diets of cattle from grain to hay orfrom hay to grain distinctly reduced fecalshedding of Shiga toxin-producing E. coli(STEC) within the first week after chang-ing the diet, whereas the E. coli cell num-bers increased considerably thereafter(Richter et al., 2000). Overall, a majorchange in the composition (i.e., grain orroughage) of the ruminant diet appearsto decrease for a few days the number ofSTEC shed in feces. Thereafter, cellnumbers of the pathogen increase.

Recent evidence indicates that thetype of grain fed to cattle can influence

fecal shedding of E. coli O157:H7 by cat-tle (Buchko et al. 2000). Feces of cattlefed 85% barley were more frequently E.coli O157:H7-positive than those fromcattle fed 85% corn, although no majordifferences were observed in cell num-bers of E. coli O157:H7 in feces through-out most of the study. Before specificdiet and feeding practices can be practi-cally applied to farm production practic-es for pathogen control, considerablymore research is needed to elucidate theinfluences that different dietary practicescontribute to gastrointestinal carriageand fecal shedding of pathogens.

As with humans, “good” bacteria—insome cases nonpathogenic normal gut

flora—in the gastrointestinal tract of an-imals may be able to prevent coloniza-tion by pathogens (Nurmi and Rantala,1973). In chickens for example, the gutof the hatchling chick is sterile until it in-gests microorganisms from the environ-ment. If the chicks are exposed to adultbird fecal material, colonization of thegut occurs rapidly. Today’s productionpractices remove this route of coloniza-tion, and the hatchling may not acquire anormal gut flora for days or weeks(Spencer and Garcia, 1995). The lack ofa fully developed “normal flora” increas-es the chances of the chick gut becominga carrier for microorganisms, such asSalmonella, that are pathogenic in hu-

Development andDissemination ofResistant Organisms

Microorganisms develop resis-tance to antibiotics encountered inclinical and environmental settings.This fact has led to calls for the judi-cious use of antibiotics in humanmedicine and for restrictions on theuse of antibiotics in veterinary medi-cine and animal production.

There are at least three fundamen-tally different ways that exposure to an-tibiotics can promote the developmentand/or dissemination of resistantmicroorganisms: (1) mutations and se-lection of mutants capable of survivingin vivo exposure to the antibiotic (e.g.,fluoroquinolone resistance in C. jejuni),(2) mobilization and horizontal trans-fer of genetic elements containing resis-tance genes among different species ofbacteria (e.g., vancomycin resistanceamong enterococci), and (3) wide-spread dissemination of strains withpreviously developed resistance (e.g., S.Typhimurium DT 104). The differencesbetween these mechanisms have impor-tant implications for the preventionand control of antibiotic resistanceamong foodborne bacteria.

Resistance of C. jejuni to fluoroqui-nolones is conferred by a point muta-tion in the gyr gene (Engberg et al.,2001). Resistant organisms with thesame molecular subtype characteristicsas sensitive microorganisms have beenisolated from human patients after ap-

parent failure of treatment with ciproflox-icin. During the 1990s, an increased oc-currence of resistant C. jejuni isolates inMinnesota was associated with treatmentwith a fluoroquinolone antibiotic andforeign travel. However, a growing pro-portion of resistant isolates was not at-tributable to these sources. Surveys ofchicken at retail markets in Minnesotademonstrated a 20% prevalence of con-tamination with resistant C. jejuni strains.These strains showed considerable diver-sity based on polymerase chain reaction(PCR)-restriction fragment length poly-morphism (RFLP) (PCR-RFLP) subtyp-ing of the flaA gene, suggesting that muta-tions and selection of mutants was inde-pendently occurring among many C. jeju-ni strains. The overlap of molecular sub-types obtained from human and chickensources suggest that chicken was a prima-ry source of resistant C. jejuni for humansthat were not treated and did not travel(Smith et al., 1999).

In Europe, a glycopeptide antibiotic,avoparcin, was used as an antimicrobialgrowth promoter. Following its introduc-tion, resistance of enterococci to vanco-mycin (a similar glycopeptide antibiotic)was observed in hospitalized patients, ex-posed animals, and the human popula-tion outside of hospitals. Vancomycin-resistance genes were transferred hori-zontally between species of enterococci(Simonsen et al., 1998; van den Braak etal., 1998). The potential for vancomycin-resistance genes to be acquired by strainsof Staphylococcus aureus represents animportant public health threat

(Cetinkaya et al., 2000).The strains of S. Typhimurium

DT104 that became a global publichealth concern during the 1990s ap-pear to be highly clonal (Baggeson etal., 2000). Although DT104 appears tohave accumulated multiple resistancegenes through horizontal gene trans-fer, these genes were likely accumulat-ed before the widespread dissemina-tion of the resistant strains. Wide-spread dissemination of DT104 mayhave been facilitated by the use of an-tibiotics on farms, either because indi-vidual animals were treated withdrugs that DT104 was resistant to, orbecause use of antibiotics altered theherds’ microflora and increased ani-mals’ susceptibility to colonizationand infection (Besser et al., 2000).

Although fluoroquinolone resis-tance in C. jejuni appears to be a directresponse to clinical or environmentalexposure to the antibiotics, DT104 rep-resents the epidemic spread of a micro-organism that has already developedresistance. The origins and factors con-tributing to the dissemination of theseorganisms require public health mea-sures that address the differences. Try-ing to accomplish comprehensive con-trol of these different situations prima-rily through restrictions on the veteri-nary and agriculture use of antibioticshas created an adversarial relationshipbetween public health and animal pro-duction communities that has imped-ed the application of science-basedcontrol strategies.

45EXPERT REPORT

mans. The chicken is most susceptible tocolonization by human pathogens dur-ing the first week of life (Nurmi et al.,1992). The prevention of colonizationwith pathogens by normal gut flora iscalled competitive exclusion.

The use of probiotics in animals hasexpanded to include virtually all foodand food-producing animals, as well ascompanion animals. The concept is thesame as for the human use of probiotics,that is, maintaining a healthy gut floraenhances health and may prevent coloni-zation with pathogens. Probiotics areadministered to maintain health underthe stresses animals experience due tocrowding, transportation, overwork, andother external forces, and also to increasefeed efficiency. They are touted as possi-ble alternatives to antibiotics in some sit-uations. Most probiotics used in ani-mals currently are single microorgan-isms or defined mixtures of microorgan-isms. Current policy prevents unquanti-fied or unidentified (undefined mix-tures) of microorganisms to be used as“direct-fed microbial products” that canbe regulated as food under Food andDrug Administration (FDA) CompliancePolicy Guide (CPG) 689.100 (FDA/CVM, 1997). Many competitive exclu-sion products, particularly those thatclaim to exclude Salmonella, are regardedas drugs by FDA.

Wild-Caught Shellfish and Fish

Feral shellfish present a unique op-portunity to transmit foodborne diseasesof bacterial, viral and protozoan origin.Of particular concern are the edible bi-valve molluscs of the class Pelecypoda thatinclude the species commonly referred toas oysters, mussels, clams, and cockles.Since most of these organisms are filterfeeders, they use siphoning organelles andmucous membranes to sieve suspendedparticles from the aquatic environment asa source of food. If their surrounding wa-ter is contaminated by bacteria, viruses, orparasitic protozoa, these mucous mem-branes may entrap the pathogens, whichare then transferred to the digestive tractof the animal. Since these molluscanshellfish may be consumed whole andraw, they can act as passive carriers of hu-man pathogens.

The most recent Centers for DiseaseControl and Prevention (CDC) statistics(1988-1992) of the overall foodbornedisease burden in the United States esti-mate that 0.7-2.1% of all outbreaks, ap-

proximately 1% of all cases, and up to13.3% of food-related deaths are due tothe consumption of contaminated shell-fish (Bean et al., 1997), although the to-tal number of cases is likely to be un-derestimated (Wallace et al., 1999).Two general groups of pathogenic mi-croorganisms may be transmitted by fe-ral shellfish. The first group is termedindigenous pathogens because these or-ganisms are native to the marine envi-ronment, consisting predominantly ofmembers of the family Vibrionaceae, in-cluding the genera Vibrio, Aeromonasand Plesiomonas. The presence of theseorganisms is unrelated to fecal pollu-tion. The second group, referred to asnon-indigenous pathogens, are not nat-ural marine inhabitants, and their pres-ence in shellfish arises from either directfecal contamination by human or ani-mal reservoirs, or due to poor generalsanitation during harvesting, process-ing, or preparation of the food animals.Within these two major groups, we canfurther characterize microorganismscontaminating shellfish as bacterial, vi-ral or parasitic protozoan in nature.

The presence of Salmonella and Shi-gella species in feral shellfish is well doc-umented, even in the recent literature.These organisms are an excellent exam-ple of non-indigenous bacterial patho-gens, the presence of which is usually dueto fecal contamination of harvestingsites. Fortunately, the National ShellfishSanitation Program sponsors long-termprograms targeting the prevention ofshellfish-associated disease caused pri-marily by enteric bacteria. In the UnitedStates, this program has established bac-teriological standards for shellfish andtheir harvesting waters based on the fecalcoliform index. Such standards havebeen quite effective in preventing entericbacterial contamination of feral shellfish,and with the exception of species im-ported into the United States from coun-tries with less stringent standards, out-breaks of shellfish-borne disease associ-ated with Salmonella and Shigella are rel-atively uncommon in the United States.

This is, however, not the case for thenon-indigenous viral and protozoanpathogens that may be transmitted bycontaminated shellfish. Enteric viruses,most notably hepatitis A virus and theNorwalk-like viruses (NLVs) are excretedin the fecal matter of infected personsand hence their source in the marine en-vironment is usually the disposal of un-treated or inadequately treated human

sewage. For a number of reasons, the fe-cal coliform index is inadequate formonitoring the presence of viral contam-ination in shellfish or their harvestingwaters, and both outbreaks and sporadiccases of enteric viral disease associatedwith the consumption of contaminatedshellfish continue to occur in the UnitedStates (Jaykus et al., 2001). Protozoanparasites such as Giardia and Cryptospo-ridium species have recently been identi-fied in feral shellfish (Fayer et al., 1998),although the significance of this foodwith respect to disease transmission hasyet to be determined. Shellfish harvest-ing beds become contaminated with par-asitic protozoa as a result of contamina-tion with animal farm runoff or humansewage, both treated or untreated.

The indigenous bacterial pathogens(most notably Vibrio cholerae, Vibrioparahaemolyticus, and Vibrio vulnificus)are part of the estuarine microflora andhave excellent survival capabilities in themarine environment. In addition, V.cholerae from human sewage may con-taminate the marine environment. Whilethe transmission of V. cholerae by sea-food is well documented throughout theworld, the United States has not had amajor outbreak since 1911, althoughsporadic cases have occurred, usuallydue to the consumption of importedcrustacea such as crabs and shrimp(Bean et al., 1997). Historically, V. para-haemolyticus outbreaks are rare in theUnited States, but two highly publicizedrecent outbreaks have challenged thistrend (CDC, 1998b; CDC, 1999b). Shell-fish implicated in these outbreaks wereharvested from both Atlantic and Pacificwaters during these outbreaks, but inboth cases, the mean surface water tem-peratures were significantly higher (1-5C) than those reported in previous years.This phenomenon suggests the potentialfor emergence of this pathogen is associ-ated with natural changes in the temper-ature of the Gulf Stream (El Niño) orglobal warming from human activities.

V. vulnificus has been of greater con-cern in the United States in recent years,since this organism results in a syn-drome characterized by gastrointestinaldisease followed by primary septicemia,with mortality rates approaching 50%.Individuals with underlying liver dys-function, circulatory problems particu-larly related to diabetes, or those who areimmunocompromised are especially atrisk (Hlady and Klontz, 1996). Evidenceexists that other Vibrio species also can


result in septic disease, including V. para-haemolyticus, V. cholerae non-O1, andVibrio hollisae (Hlady et al., 1993), a find-ing that may indicate the emergence ofnon-vulnificus Vibrio species as sourcesof life-threatening shellfish-borne disease.V. vulnificus is a leading cause of food-borne disease-related deaths, particularlyin the southern states (Bean et al., 1997;Hlady et al., 1993). There is indicationthat the microorganism is capable of es-tablishing the so-called viable-but-non-culturable (VBNC) state, which means

that routine examination by conventionalcultural methods may provide negativeresults although viable and potentiallyvirulent cells may be present in highnumbers (Oliver and Wanucha, 1989).

Because regulation of seafood occurspredominantly by pre-harvest monitor-ing of the fecal coliform index of grow-ing waters, the efficacy of control is high-ly dependent upon the relationship be-tween this test and the presence of patho-gens. A significant relationship does notexist in shellfish between the presence of

the fecal coliforms and important hu-man pathogens such as enteric viruses,the pathogenic vibrios, and perhaps theparasitic protozoa. In addition, shellfishharvested from other countries with lessstringent standards than those of theUnited States may be contaminated, andsome countries have much higher do-mestic rates of shellfish-associated dis-ease. Pre-harvest control methods suchas relaying (movement of shellfish fromone harvest site to an alternative, pristinesite to allow the animals to purge them-

The Role of Microbio-logical Indicators inAssuring Food Safety

The term “microbiological indica-tor” refers to a microorganism, agroup of microorganisms, or a meta-bolic product of a microorganism,whose presence in a food or the envi-ronment at a given level is indicative ofa potential quality or safety problem.The selection of an appropriate indica-tor is highly dependent upon the mi-crobiological criteria for the foodproduct in question. Important con-siderations in indicator selection in-clude: possible sources of pathogenicmicroorganisms; their incidence on orin the product; production, harvestingand processing practices; survival andgrowth of pathogens in the product;and specific analytical methods avail-able for detecting the indicator.

Although frequently used inter-changeably, scientists sometimes makea distinction between the terms micro-biological indicator and index organ-ism. In general, index organisms aremarkers whose presence in numbersexceeding pre-established limits indi-cates the possible occurrence of eco-logically similar pathogens (Mossel etal., 1995). Indicator tests are often em-ployed to assess a process control at-tribute, such as using the extent of me-sophilic growth as an indicator of in-adequate refrigeration (Ingram, 1977).A given marker can function both asan index and as an indicator organ-ism, even in the same food.

The presence of indicator organ-isms does not necessarily guarantee

the presence of pathogens (Banks andBoard, 1983). Ideally, the absence or alow concentration of a specific indicatormeans the food has not been exposed toconditions that would permit contami-nation by a specific target pathogen orpresent the opportunity for its growth.The indicator concept can be used toevaluate raw product directly from thefield or farm, or after some decontami-nation or inactivation process.

Selection of an indicator that is rele-vant for a given food and a given targetpathogen continues to be a challenge. Be-cause no microbiological indicator is ide-al, food safety professionals are frequentlyleft with a limited choice of indicators thatare relevant to some, but not all, food-borne pathogens. The most widely usedindicators are the Enterobacteriaceae,coliforms, “fecal” coliforms and E. coli.Coliforms are gram-negative asporoge-nous rod-shaped bacteria that are envi-ronmentally ubiquitous and may be asso-ciated with the fecal material of animalsincluding humans. Coliforms are fre-quently used as an indicator of inade-quate sanitation and process control infood that receives a pasteurization heattreatment. Among the proposed or ac-cepted uses of E. coli are as an indicator offecal contamination, acceptably condi-tioned manure, acceptable quality of wa-ter for irrigation, shellfish safety and gen-eral environmental sampling. Some orga-nizations have used the entire Enterobac-teriaceae family as an indicator or indexof potential pathogen contamination.Other alternative indicators have includedthe coliphage group and the “fecal strepto-cocci” or enterococci. Each of thesegroups of microorganisms has shortcom-

ings as indicators of enteric pathogens.The relationship between the pres-

ence of the fecal indicators and thepresence of foodborne pathogens offecal origin, such as Salmonella andCampylobacter, has been questionedfor years. Recently, Kornacki andJohnson (2001) stated that “numerousstudies have determined that E. coli,coliforms, fecal coliforms and Entero-bacteriacae are unreliable when usedas an index of pathogen contamina-tion of foods.” This unreliability ap-plies to fresh produce (Anonymous2000; DeRoever, 1998; Nguyen-theand Carlin, 2000), fresh meats(Goepfert, 1976; Linton et al., 1976;Roberts, 1976; Tompkin, 1983), andferal shellfish (Jaykus et al., 2001). In-dicators for other pathogens such as L.monocytogenes have perhaps fared bet-ter, and many processors continue touse the absence of generic listeria as anindicator for the absence of L. monocy-togenes.

Despite the clear need for more re-liable indicator systems, all the candi-date replacements, such as coliphage,Bifidobacter spp., and enterococci, havetheir own limitations. Beyond thepathogenic bacteria, indicators also areneeded that are specific for human en-teric viruses such as human calicivi-ruses (Norwalk-like viruses) and par-asitic protozoa such as Cryptosporidi-um, Cyclospora, and Giardia, all ofwhich tend to be more resistant andpersistent than bacterial foodbornepathogens. As in so many other areasof food microbiology, additional workin the area of microbiological indica-tors remains essential.

47EXPERT REPORT

selves of pathogens prior to harvest) havebeen effective for the elimination ofmany of the non-indigenous bacterialpathogens, but are reasonably ineffectivein controlling viral and Vibrio contami-nation. In short, there is a need to fur-ther evaluate pre-harvest issues that im-pact the safety of feral shellfish con-sumed by the U.S. population.

Specific Production Methods

Sometimes the method used to pro-duce the food commodity has a signifi-cant effect on the microbiological safety offoods. Organic foods are of special inter-est, in part because this production meth-od relies on manure as a major source offertilizer. In addition, aquaculture meth-ods have some significant food safety dif-ferences from traditional fishing.

Organically Grown Foods

Organic foods are the product of afarming system that avoids the use ofmanmade fertilizers, pesticides, growthregulators, and livestock feed additives. In-stead, the system relies on crop rotation,animal and plant manures, some handweeding, and biological pest control. Theorganic food industry has been growingat an annual rate of 20 percent during thepast decade, and continued growth is pro-jected. Produce grown by about 12,000organic farmers nationwide grossed ap-proximately $6 billion in 2000. The surg-ing market for organic produce is in partattributed to some consumers’ belief thatsuch foods are healthier, better tasting,and produced using environmentallyfriendly methods. Many consumers whohave concerns about pesticides and herbi-cides cite safety as the main reason for useof organic foods.

In December 2000, the U.S. Depart-ment of Agriculture adopted marketingstandards for the processing and labelingof organic foods (USDA/AMS, 2000).The new standards do not allow the use ofirradiation, genetic engineering, and hu-man sewage sludge fertilizer for any foodlabeled as organic. In addition, syntheticpesticides and fertilizers cannot be usedfor growing organic food, and antibioticsand hormones are prohibited in livestockproduction. Furthermore, livestock agri-cultural feed products must be 100 per-cent organic. Products meeting the re-quirements for labeling as “organic” cancarry a USDA Organic Seal. Products witha smaller percentage of organic ingredi-

ents may be labeled as “made with organicingredients” but cannot carry the USDAseal. Although the available data do notindicate that organic foods are safer ormore nutritious than conventional food,consumer surveys indicate that foods la-beled with the USDA organic seal are per-ceived as being more healthful.

A significant public health concern thathas been largely overlooked by the organicmovement is the potential for greater prev-alence of contamination by foodbornepathogens that are carried by livestock andpoultry and shed in their feces. Manure, asignificant vehicle for pathogens, is a majorsource of fertilizer used for growing organ-ic produce. The available scientific infor-mation is insufficient to ensure that food-borne pathogens are killed during com-posting and soil application.

Furthermore, manure contaminateshides, feathers and skin during livestockproduction. During slaughter and pro-cessing, carcasses occasionally come incontact with manure and become con-taminated with pathogens. Because or-ganic production standards prohibit theuse of irradiation and chemical treat-ments during processing, the availablemethods to reduce pathogen contamina-tion are restricted, resulting in a greaterlikelihood that organic meat and poultrywill have higher levels of pathogen con-tamination than conventionally pro-cessed meat and poultry.

Very few studies have been conduct-ed to address the microbiological safetyof organic foods. A relatively small studywas done to determine the prevalence ofE. coli and Salmonella on selected organ-ic and conventionally produced vegeta-bles available at Atlanta grocery stores(Doyle, 2000). A total of 216 samples—half organic and half conventional—wastested over a 6-week period. E. coli waspresent on 22 organic samples and 18conventional vegetables, and Salmonellawas found in 3 organic samples and 2conventional vegetables. These data in-dicate that the organic produce sampledwas not safer from a microbiologicalperspective than the conventional pro-duce. A comprehensive survey of food-borne pathogens in organic foods (in-cluding produce, meats and poultry) isneeded to more fully evaluate the relativemicrobiological safety of such foods.

Aquaculture

One area where food productionhas undergone dramatic change is

aquaculture or commercial fish farm-ing. To meet the demand for seafoodproducts, the seafood industry is turn-ing to aquaculture to supplement in-creasingly limited natural supplies. TheFood and Agriculture Organization(FAO) of the United Nations definesaquaculture as the “farming of aquaticorganisms, including fish, molluscs,crustaceans and aquatic plants” (FAO,1995).

The available data (primarily fromtemperate zones) indicate a low inci-dence of enteric pathogens in fish andcrustaceans raised in unfertilized sys-tems, and few reports of human illnessassociated with the consumption ofaquacultured finfish and crustaceans(Jensen and Greenless, 1997). Howgate(1997) reported that there was no reasonto expect the risk of illness from farmedmarine fish to be greater than corre-sponding wild species. In fact, the mostchallenging public health risks fromaquaculture are those associated withshellfish grown in open surface waters(Jensen and Greenless 1997), whereaquaculture faces many of the same is-sues as wild caught.

Antibiotics are used in aquacultureto control disease in cultured species.Drug use is regulated by FDA, but thenumber of approved drugs is limited andoff-label use has been reported (JSA,1994). These antibiotic compounds areallowed only for certain species and lifestages, with designated withdrawal peri-ods, during which cultured fish andshellfish that have been treated with vet-erinary drugs cannot be harvested orsold (JSA, 1994). Nevertheless, bacteriaresistant to antimicrobials have been iso-lated in farmed catfish and from theirenvironment (DePaola et al., 1988), insediments located beneath net pens (Ker-ry et al., 1994), and from the intestines ofcultured fish in net pens (Ervik et al.,1994). A World Health Organization re-port on food safety issues associated withaquaculture products stated that the riskto public health is probably limited to in-direct exposure to antimicrobials (WHO,1999). However, the use and misuse ofantibiotics to control diseases in aquac-ulture is worldwide and will likely in-crease as intensive aquaculture systemsbecome more common.

Current shellfish management pro-grams—which include classification andmonitoring of growing waters, propersiting of aquaculture areas, enforcementof best management practices and train-


ing of employees and harvesters in sani-tation procedures—can help to reduce,but do not eliminate, microbiologicalfood safety issues.

HARVEST ENVIRONMENT

The harvest environment is relativelycommodity specific, because the meth-ods used and harvest locations dependon the commodity in question. The har-vest environment is especially importantfor foods such as fruits and vegetablesthat undergo minimal additional pro-cessing prior to consumption.

Produce

Many different approaches are usedtoday to harvest produce depending on avariety of determinants, including thetype of produce, site of growing opera-tion, and labor availability. However,from a microbiological safety perspec-tive, there are many common hazards forharvest operations. Water quality, fieldworker hygiene, field sanitation, trucksanitation, and temperature control areall food safety issues related to the har-vest environment (IFT, 2002).

Water as ice or in the liquid form canreadily transmit microorganisms to pro-duce if contaminated. Most fruits andvegetables are washed with water at leastonce, and many types of produce aretreated with water several times duringprocessing. In addition to being used forwashing, water is used for cooling, con-veying produce (flume water), and forapplying disinfectants and fungicides.Disinfectants are added to about 50% to60% of water used in packing facilities.Care must be taken to control the sani-tary quality of water.

Field worker hygiene is a major con-sideration because of the widespread useof human hands in cutting or pickingvegetables and fruits in fields or or-chards. Approximately 90% of farmsthat grow fruit or vegetables harvest pro-duce exclusively by hand (USDA/NASS,2001). The amount of human hand con-tact that occurs during harvesting variesdepending on the type of produce. Mel-ons are handled at most steps of the op-eration, whereas apples receive consider-ably less frequent contact. Vegetablessuch as leaf lettuce may be harvested,trimmed, sorted and bagged by hand. Itis very difficult to uniformly enforceproper hand washing and glove use

among workers.An interesting practice in recent

years has been the production of fresh-cut produce in the field. Presumablydone to maintain product freshness, thispractice essentially brings food process-ing into the harvest environment. For ex-ample, many fresh-cut processors arenow removing the outer leaves and cor-ing lettuce in the field. Some are evencutting lettuce where it is grown. Unfor-tunately, it is much more difficult to con-trol contamination in the harvest envi-ronment, and this practice may increasethe chances for contamination of theseproducts. Certainly contamination ofwash and rinse water would be of con-cern under these circumstances, aswould be issues associated with poor hy-giene of food handlers and inadequatesanitation of equipment.

Fruits and vegetables frequentlycome into contact with harvestingequipment (such as knives, machetes,clippers and scissors) and containers(such as bins, boxes, buckets, pans,trailers, and truck beds). Such equip-ment and containers should be properlywashed and disinfected, although stud-ies indicate that washing and sanitizingis done only about 75% and 30% of thetime, respectively (USDA/NASS, 2001).Packing equipment such as tables, con-veyor belts, flumes, and washing andcooling bins are washed and sanitizedabout 75% and 50% of the time, respec-tively.

Equipment and containers used dur-ing harvest operations are frequentlymade of materials, such as wood, that aredifficult to clean. Soil from the field,which may contain pathogens, often en-crusts equipment and containers. If notremoved, soil adhering to equipmentused for washing and disinfecting pro-duce will reduce the sanitizing capacityof the disinfectant. Produce operationsmust prevent accumulation of soil onequipment and containers to enable ef-fective disinfection of food contact sur-faces.

Proper temperature control of fruitsand vegetables is critical for both safetyand quality purposes. Optimal temper-atures vary according to the commodi-ty; however, the temperature range forstoring fruits and vegetables is usuallyquite narrow. Most pathogens, but notall, are inhibited by the cool tempera-tures at which produce is stored. In ad-dition, cool temperatures tend to pro-long the survival of viruses and para-

sites. Temperature control contributesmost to the safety of fruits and vegeta-bles that are cut; however, its effective-ness in controlling microbiological haz-ards is in general less significant thanthe hazard reduction that occurs by re-frigerating raw foods of animal origin(IFT, 2002).

Food Animals

Production livestock are naturallycontaminated with a variety of potentialhuman pathogens, both externally andinternally. These microorganisms origi-nate from the environment in which theanimals are produced, as well as fromfeedstuffs and the co-mingling of ani-mals from various sources. The animalsalso can readily transmit these microor-ganisms among themselves, once theyare moved from the production environ-ment to the transportation system. Be-cause of centralized slaughter establish-ments, production livestock may betransported considerable distances be-fore slaughter.

During transportation, livestock areconfined to prevent excessive move-ment. As the animals are in close physi-cal contact, microorganisms may betransferred from one animal to anothereither by contact with each other ortheir excreta. Aerosol transmission ofsalmonella also has been demonstratedin chickens and mice (Clemmer et al.,1960; Darlow et al., 1961). The upperrespiratory tract may be important intransmission, and the tonsils and lungsmay be important sites for the invasionand dissemination of Salmonella in pigs(Fedorka-Cray et al., 1995; Fedorka-Cray et al., 2000; Gray et al., 1996).

Even under the best of conditions,transportation produces measurablestress upon live animals. Transportstress may have several physiological ef-fects on the live animal that may ulti-mately impact food safety. Transporta-tion of animals may increase fecal shed-ding of potential human pathogens, suchas salmonella. This increase in sheddingcan contaminate the trucks or trailersused for transportation, and potentiallyincrease the population of foodbornepathogens in and on many of the ani-mals within the truck or trailer.

Once animals arrive at a slaughterestablishment, they may be unloadedinto holding pens prior to slaughter, de-pending on the animal species in ques-tion. Animals from several trucks or

49EXPERT REPORT

trailers may be co-mingled in the samepen, although there is considerablevariation between slaughter establish-ments. Co-mingling increases the op-portunity for the spread of potentialhuman pathogens between animalsfrom different sources, with the possibleoutcome of more animals becomingcontaminated either externally or inter-nally with these microorganisms. In ad-dition, these holding pens are difficultto adequately clean and sanitize, and itis not uncommon to detect potentialhuman pathogens in these pens afterthey have been cleaned and sanitized(Davies and Wray, 1997; Gebreyes et al.,1999). As a result, populations of bacte-ria may remain in these pens and con-taminate animals that are brought infrom other sources. The same is true ofpoultry crates, which are equally diffi-cult to clean and sanitize.

Aquaculture and Wild-CaughtFish and Shellfish

The length of time between harvestand refrigeration is critical to the micro-biological safety of fish and shellfish.Time/temperature abuse is of particularconcern for the pathogenic Vibrio spe-cies. Rapid harvesting, cooling and pro-cessing influences quality, but it mayalso improve food safety for certain fish-ery products by slowing the growth ofpathogens and reducing the formationof histamine that causes scombroid poi-soning in certain species.

Commercial finfish are harvested bya variety of methods. Fishing vessels maytransport their harvests back to shore forprocessing on a daily basis, or they mayremain at sea for several days, weeks oreven longer. Fishing vessels use severalmethods to preserve their catch (e.g., ic-ing, freezing, or refrigerated seawater).Finfish can be minimally processed onboard (e.g., whole or eviscerated form).Finfish also can be processed into filletsand steaks and other value-added prod-ucts, such as surimi (a washed, mincedfish product) on commercial factoryships.

Shellfish also may be processed onboard, or they may be brought to shorelive. Mussels, clams, and oysters areusually quickly transported shore sideand shucked or sold as live whole shell-stock. Scallops are usually processed onboard the fishing vessel and stored icedor frozen for several days or weeks. Onthe west coast, some crab species are

processed on board the fishing vessel,while on the east coast and in the Gulfof Mexico, crabs are usually broughtalive to shore for further processing.Lobsters are brought alive to shore, andshrimp are iced or frozen on board thefishing vessel.

All fishery products, regardless ofwhether they are wild harvested or cul-tured, can be contaminated with a vari-ety of human pathogens. Although thereis no reason to expect cultured species tobe any more hazardous than wild caught(Howgate, 1997), there may be differenc-es depending on the harvest area or cul-ture method. For example, FDA recentlyfound a higher percentage of Salmonellaspp. on farm-raised shrimp (7%) com-pared with wild-harvested shrimp (<1%)(FDA, 2001). Most of the shrimp con-sumed in the United States are imported(NMFS, 2000), and a high percentage ofthese shrimp are aquacultured in tropi-cal countries. Reilly et al. (1992) reportedthat Salmonella spp. is a component ofthe natural flora of pond culturedshrimp in tropical countries. This find-ing may be due to the fact that in manytropical countries, chicken manure is of-ten used as a fertilizer in aquacultureponds. To address food safety issues,FDA is currently developing GAP guid-ance documents to help reduce patho-gens associated with cultured fisheryproducts.

POST-HARVEST ENVIRONMENT

In the post-harvest environment,food safety becomes less commodity ori-ented, as the food moves through pro-cessing into the distribution and retailsectors. Microbiological food safety is-sues in the processing environment havethe potential to affect many differentfoods. Also, at the processing stage, in-gredients from many different commodi-ty sectors may be combined into a singleproduct. As ingredients are combined,the physical characteristics of the foodchange, and its microbiological profile isaltered.

The microbiological controls ap-plied in the post-harvest environmentare often designed to intentionally stressthe microorganisms present in the food.These stresses may be designed to be le-thal on their own or in combination.Environmental conditions also may bemodified to limit microbial growth,through such techniques as drying or

refrigeration. Each of these stresses hasan impact on the microbial populationin the food.

Food Animal Slaughter and MeatProcessing

Although food animal slaughtercould be considered a harvest activity,from a microbial ecology perspective, itoccurs in the post-harvest sector of thefood chain. Slaughter and meat process-ing take place in a carefully controlledprocessing atmosphere that is differentfrom the traditional harvest environ-ment. Because food animals are natural-ly contaminated with a variety of poten-tial pathogens, meat processors applymany microbiological control methodsduring the slaughter and processing ofmeat.

Role of Inspection

An important aspect of meat in-spection in the United States and otherdeveloped countries is antemortem in-spection. All animals presented forslaughter are inspected prior to slaugh-ter. At present, this inspection focuseson observable clinical illness, and noton general hygiene of the animals.However, because of the recognition ofthe impact of animal hygiene on meatcontamination, there have been someattempts to regulate this on the live ani-mal. For example, part of the antemor-tem inspection in the United Kingdomis an evaluation of the overall hygieneof the animal, and the animals may berejected for slaughter if they are too“dirty.” Efforts to reduce the level of ex-ternal carcass contamination of live ani-mals have typically focused on eitherwashing the entire animal or trimmingthe hair from the most heavily contami-nated areas of the animal (typically therump and midline of a steer or cow).These procedures must be balancedwith animal welfare issues, which pre-vent the introduction of excessive, need-less or unnecessary stresses on live ani-mals.

Source of Contamination

Contamination of animal carcassesduring slaughtering procedures is unde-sirable but unavoidable in the conver-sion of live animals to meat for con-sumption. It is assumed that the muscletissue of healthy animals entering the


slaughter establishment is free of micro-organisms (Ayres, 1955). However, in-trinsic bacteria, occurring in the deepmuscle tissue of healthy animals, havebeen reported for many animal species(Ingram, 1964; Ingram and Dainty,1971). The most frequently character-ized intrinsic bacteria are Clostridiumspp. (Canada and Strong, 1964; Jensenand Hess, 1941; Narayan, 1966; Zagae-vskii, 1973). Other potential humanpathogens, such as salmonella, havenot been reported as intrinsic bacteriain the muscle tissue of healthy animals.The assumption is bacteria that con-taminate the muscle tissue are from ex-trinsic sources (gastrointestinal tract,lymph nodes, external carcass surfaces,and environmental sources). The ma-jority of the microflora transferred tothe tissue surfaces, while aestheticallyundesirable, is nonpathogenic.

Carcass processing may be dividedinto processes that impact the externalsurfaces of the carcass and those thatopen the body cavity (evisceration).Examples of surface processes includethe scalding of poultry and the remov-al of beef carcass hides. These process-es in fact remove significant popula-tions of bacteria from the carcass, butalso expose the edible tissue to con-tamination. While the process of hideremoval may be thought of as a prima-ry source of contamination of beef car-casses, it also removes significant con-tamination from the carcass. In a simi-lar fashion, scalding and de-featheringof chickens removes significant con-tamination from the chicken carcass,although some still remains on the car-cass. The microflora deposited bythese processes is primarily from envi-ronmental sources, that which is on thecarcass from the livestock productionenvironment and transportation. Thismicroflora may include enteric patho-gens, such as salmonella and E. coliO157:H7, and microorganisms such asListeria and Clostridium. From a mi-crobiological perspective, contamina-tion of edible tissue by external surfaceprocessing is a relatively common oc-currence.

Processes that open the body cavityexpose edible tissue to bacteria fromthe gastrointestinal tract. The primarybacteria of potential public health con-cern from this source are the entericpathogens. In the case of red meat pro-cessing, this source of contamination isinfrequent, and usually confined to

leaking bungs or an occasional rup-tured gastrointestinal tract.

Impact of Interventions

Over the last twenty years, a numberof interventions have evolved specificallyto address microbial contamination ofanimal carcasses. These may be dividedinto physical methods and chemicalmethods. In practice, these methods areused in combination, resulting in a seriesof process interventions or “hurdles” toimprove the microbiological quality ofmeats. Among the physical methods,prevention of contamination has re-ceived considerable attention. From theperspective of microbial adaptation, it isbetter to prevent contamination than toaddress it through processing methods.However, there are practical limitationsto this, given the nature of the process(i.e., converting a live animal to food forhuman consumption). After preventionof contamination to the extent possible,physical interventions involve either re-moval of contamination (trimming),heat, or chemical treatments.

Effective physical removal of con-tamination is dependent on first identi-fying the area of contamination and thenremoving the affected area withouttransferring the contamination to otherareas. There are practical limits to theidentification of affected areas, as micro-organisms cannot be seen. Therefore,identification of an affected area requiresthat there be sufficient contamination(mud, manure, etc.) to become visible tothe operator. Once the area has beenidentified, the operator must then re-move the area in an aseptic manner. Theprobable outcome of this operation, un-der controlled conditions, is the completeremoval of the contamination with a vir-tually sterile surface remaining aroundthe affected area. However, while this iseasily accomplished under controlledconditions, it becomes much more un-likely in a processing environment. Theoperator must clean and sanitize theequipment (knife and hook) prior to theoperation, remove the affected area froma carcass which is frequently moving ona processing line, and then re-sanitize theequipment prior to trimming the nextaffected area. In practice, trimming issubstantially less effective in processingenvironments than in laboratory envi-ronments (compare Gorman et al.(1995) with Reagan et al. (1996)). Trim-ming as a process is limited to a specific

area of a carcass.In contrast to trimming, heat may

be applied as either a localized treat-ment or as a whole carcass treatment.The most common form of localizedtreatment currently in use is the steamvacuum. Steam vacuuming, as the nameimplies, applies a steam treatment toboth loosen contamination and killbacteria, along with a vacuum processto physically remove contamination.Steam vacuuming may be highly effec-tive in reducing microbial populationsunder controlled conditions, but as withtrimming, becomes less effective underprocessing conditions (Castillo et al.,1999; Dorsa et al, 1996). As an alterna-tive to a localized treatment, heat maybe applied as a whole carcass treatment.Common examples of this are singeingof hog carcasses, hot water washing andsteam pasteurization. Heat in singeingprocesses is commonly applied as openflame from gas jets, and while there is areduction in microbial populations onthe carcass surface, the primary func-tion of this process is to remove residualhair from the carcass. In contrast, hotwater washing and steam pasteurizationwere specifically developed as antimi-crobial processes, applied as a final op-eration before chilling. Hot water wash-ing applies hot water (>80 C) as a wholecarcass rinse, while steam pasteuriza-tion places the carcasses in a chamberand applies steam to briefly raise thetemperature of the carcass surface.Both hot water and steam pasteuriza-tion have been demonstrated to be ef-fective in controlling microbial popula-tions on animal carcasses (Barkate etal., 1993; Gill et al., 1995; Phebus et al.,1997). However, surviving pathogenswill have undergone stress that may ei-ther increase or decrease the likelihoodof their survival during the remainingtime before consumption.

Chemical interventions involve theapplication of food grade chemicals tothe carcass surfaces to inhibit or kill mi-croorganisms (Dickson and Anderson,(1992; Siragusa, 1996). Typically, themode of action of these antimicrobials ispH, with organic acids, such as lactic oracetic (low pH) and trisodium phos-phate (high pH), the most commonlyused. The concerns with the use of anychemical intervention process are boththe potential to induce resistance in po-tential human pathogens and the poten-tial to select for resistant organisms outof the overall microbial population. If re-

51EXPERT REPORT

sistance becomes widespread in a micro-bial population, more organisms willsurvive, making the process less effective.

Perhaps the chemical intervention ofgreatest interest and concern is the or-ganic acids and their potential to induceacid tolerance. In controlled experi-ments, acid-tolerant bacteria were nomore tolerant to organic acid rinse pro-cesses and were in general more sensitiveto heat than their homologous non-acid-tolerant strains (Dickson and Kunduru,1995). This suggests that the develop-ment of acid tolerance does not present aunique hazard with organic acid rinseson animal carcasses.

Acid adaptation has been shown toenhance the ability of salmonella to sur-vive in acidic food systems (Leyer andJohnson, 1992). These authors reportedthat when salmonella were briefly ex-posed to mildly acidic conditions (pH5.8), the survival of these bacteria wasdramatically enhanced in cheese. Anoth-er study reported that S. Typhimuriumbriefly exposed to mildly acidic condi-tions (pH 5.8) was significantly more re-sistant to strong acid conditions (pH 3.3)than the non-adapted parent strain (Fos-ter and Hall, 1990). Acid shock at pH 4has also been reported to enhance thethermotolerance of L. monocytogenes(Farber and Pagotto, 1992). Foster andHall (1990) reported that the adaptiveacid tolerance response of S. Typhimuri-um did not appear to induce cross pro-tection with hydrogen peroxide or heatshock. Rowbury (1995) discussed theimpact of a variety of environmental fac-tors on acid tolerance, and noted that or-ganisms attached to surfaces were moretolerant to acid than non-attached cells.Buchanan et al. (1999) reported that pri-or growth at acidic conditions increasedthe resistance of E. coli O157:H7 to ion-izing radiation at acidic pHs.

Acid-adapted microorganisms mayhave a competitive or ecological advantagein the human stomach, which could po-tentially impact pathogenicity. The theoryis that an acid resistant microorganismwould be more capable of surviving theacid pH of the stomach, and thereforemore organisms would enter the small in-testine. This could then result in a lowerinfectious dose for the microorganism.

Chilling

Rapid chilling of hot (body tempera-ture) animal carcasses is essential to re-duce the outgrowth of contaminating

microbes. The common methods of car-cass chilling involve either forced air,forced air and water, or water chilling.While the primary intention of chilling isto limit microbial growth, some chillingmethods do contribute to a reduction inmicroflora. For example, forced air chill-ing dries the carcass surface and may in-jure or kill some microorganisms by de-hydration. The initial combination chill-ing process using forced air and waterwas the Chlor-Chil process (Swift andCompany, 1973), which used chilledchlorinated water to simultaneously re-duce microbial populations and chill thecarcasses more rapidly. Although the useof chlorine in the water is no longerwidely practiced, the process now com-monly known as spray chilling is almostuniversally used in the beef industry. Al-though it does not have the benefit ofdrying the carcass, spray chilling reducesthe surface temperatures of carcassesmore rapidly than air chilling, and, sincevirtually all of the bacteria are on thecarcass surfaces, it effectively reduces mi-crobial growth to a greater extent thanair chilling. Water chilling, widely used inthe poultry industry, has evolved from asignificant source of contamination be-tween poultry carcasses to a potentialmicrobial intervention process, with theuse of counter-flow chillers and the addi-tion of processing aids.

Fabrication

The disassembly of chilled carcasses isreferred to as fabrication. The outputs offabrication are fresh meat to go to theconsumer, and meat intended for furtherprocessing. During fabrication, microor-ganisms may be transferred from carcasssurfaces to other parts of the carcass, andmay be transferred from one carcass toanother by common contact surfaces. Be-cause of the latter phenomenon, one car-cass has the potential to contaminate sev-eral other carcasses. As the fabricationprocess proceeds, the carcass identity islost, and the potential for trace back is se-riously diminished. Other than the physi-cal intervention of preventing or reducingcontamination, there are essentially nocommercially viable interventions in fab-rication at this time. This certainly pro-vides opportunities for research and de-velopment, but it must be kept in mindthat the objective of the slaughter opera-tion is to place the least contaminated car-cass into the chilling and fabrication pro-cesses. Assuming that the microbiological

objectives have been achieved in theslaughter process, and that the fabricationprocess is maintained at a reasonable levelof hygiene, further interventions may beunnecessary.

Ground Meats

Ground meats are a special categoryof fresh meats. Ground meats are madefrom the less desirable cuts of fresh meatand from the trimmings of the more de-sirable cuts of meats. Unlike a steak or achicken breast, which originate from asingle carcass, ground meat may containmeat from many different carcasses. Thenature of the process is to grind and mixthe meat, which increases the likelihoodthat a single contaminated carcass maycontaminate a larger quantity of meat. Anadditional factor is that the trimmingsfrom the intact cuts of meat are oftenfrom the surface of the cut. Because themost consistent source of contaminationon animal carcasses are the processeswhich affect the external surfaces of thecarcass, the raw materials used for grind-ing have a higher probability of beingcontaminated with bacteria of potentialhuman health significance. A conse-quence of these factors is that the micro-bial populations in ground meat are con-sistently higher than those on intact meat.Other than irradiation, there are essential-ly no commercially viable interventionsfor raw ground meats at this time.

A process that has also been used torecover edible tissue from bones is me-chanical de-boning. This process wasfirst developed in the poultry industry,but is also used in the red meat industry.Meat produced from mechanical de-boning processes resembles groundmeat, but is used almost exclusively as aprocessing ingredient for cooked meats.This meat does not generally reach theconsumer as fresh meat. Because of thenature of the process, mechanically de-boned meat commonly has a higher mi-crobial population than other meats, in-cluding ground meats.

Processing

Fresh meats are frequently furtherprocessed for specific flavor characteris-tics or for increased shelf life. The pro-cesses involved with meats are the pro-cesses commonly used with other foods,including thermal processing (cooking,canning), dehydration, and fermenta-tion. These processes do not result in any


unusual microbiological hazards uniqueto meat products, but some of the com-mon microbiological hazards are worthrestating.

Fermented meats, such as pepperoni,select for microorganisms that can toler-ate higher osmotic pressure and acidicpHs. Microorganisms that can surviveand cause human health concerns in-clude S. aureus and some enteric bacte-ria, including E. coli O157:H7. Althoughit is less likely that enteric bacteria willsurvive, cases of foodborne disease out-breaks from these bacteria have beendocumented.

Cooked, ready-to-eat meats are aspecial category of processed meats.These products undergo a thermal pro-cess that renders them fully cooked witha very low population of microorgan-isms, and they are typically vacuumpackaged and refrigerated. The refriger-ated shelf life of these products can reach120 days, which provides an extendedopportunity for psychrotrophic bacteriato grow. The specific human pathogen ofconcern is L. monocytogenes, which ac-counts for approximately 30% of thefoodborne deaths in the United States(Mead et al., 1999). Because L. monocyto-genes in this case is a post-processingcontaminant, the intervention (heat) hasalready been applied, and there are fewintervention strategies that are viable af-ter packaging.

A discussion of microbiological con-trols in meat processing would not becomplete without considering the use ofirradiation (for detail, see irradiation, p.56). Irradiation has the advantage of be-ing able to penetrate packaging materi-als. A product that is packaged and thenirradiated is protected from recontami-nation while the packaging material re-mains intact. Irradiation may be the onlyeffective microbial intervention processfor fresh products that, by definition,cannot be processed by conventionalprocesses. In a similar manner, irradia-tion may ultimately serve as an interven-tion for packaged, processed meats.

Future Challenges

The net result of processing changeshas been improved microbiological con-trol and reduced levels of enteric patho-gens. Since the adoption of these morecomprehensive systems for pathogencontrol, the prevalence of salmonellahas decreased as determined by theUSDA/FSIS monitoring program. But

newly recognized food safety issues,such as E. coli O157:H7 in cattle and C.jejuni in broilers, challenge the indus-try’s efforts. Continued improvementsduring slaughter may occur, but the bet-ter long-term strategy would be to min-imize the presence of human pathogenson the incoming live animals. This ap-proach would require changes in farmmanagement practices that are based onscientific research.

Post-Harvest Processing of OtherCommodities

Further processing of other com-modities occurs as well. For example,fresh fruits and vegetables are frequentlysent to packinghouses where they may bewashed, trimmed, or otherwise changedprior to packaging. The environment ofthe packinghouse, and anything thatcomes into contact with the fresh pro-duce (e.g. water, conveyor lines, packag-ing material) can thus contribute to themicrobial ecology of the produce, andcan contaminate the produce withpathogens. Since many fresh produceitems are ready-to-eat, the cleanliness ofthe packinghouse environment is veryimportant.

Produce may be more extensivelyprocessed as “pre-cut” fresh fruits or veg-etables. Pre-cut or shredded lettuce hasbecome a large industry serving thechain restaurant industry, as have otherpre-cut vegetables for salad bars. Pre-packaged salads are also a popular itemin grocery stores, and consumers fre-quently believe no further washing is re-quired when the items are brought intothe home for serving. Because the cut-ting exposes more surface area and dis-rupts natural barriers of the fruit or veg-etable, these items are more permissive ofmicrobial growth.

The seafood processing industry iscomplex, and to maximize food safety,processors and importers of seafood aresubject to HACCP requirements as man-dated by FDA (FDA, 1995). The process-ing of wild and aquacultured fisheryproducts can be as simple as washingmolluscan and crustacean shellstock orfinfish with potable water. It also mayinclude shucking, filleting, beheadingand peeling followed by chilling and/orfreezing, for sale to wholesalers, distribu-tors and retailers. In some instances, fin-fish, shellfish and crustaceans are soldlive at the retail market, or processed byhand or machine at commercial facilities

into fillets and other value-added prod-ucts. Fishery products may be consumedraw, or they may be processed intoready-to-eat products. Salmon, troutand other fish species are often processedinto cold smoked and hot smoked fisheryproducts. Fishery products also are usedin a variety of salads and spreads, andthey are cured, fermented, pickled, dried,and canned. Rapid harvesting, coolingand processing can reduce or preventthe occurrence of certain biological andchemical food safety hazards associatedwith some wild caught and cultured fish-ery species (e.g., molluscan shellfish andscombrotoxin susceptible species) (NAS,1991). In other instances, the use of foodadditives or other post processing inter-ventions (e.g., high hydrostatic pressure,irradiation, or thermal pasteurization)may be required to control food safetyhazards.

Regardless of the commodity, pro-cessing hazards can include pathogensthat may be naturally present on thefood, that may be introduced duringprocessing and handling, or that may in-crease to hazardous levels during distri-bution and storage. Strict attention togood manufacturing practices, sanitationcontrol procedures and hygienic practic-es of plant employees are effective incontrolling many of these hazards.

Water

Water is used extensively in the post-harvest processing environment, makingwater quality a significant concern. Ad-vances in water treatment over the last100 years have resulted in dramatic im-provements to the microbiological safetyof the public water supplies of developedcountries (Dawson and Sartory, 2000).Perhaps of greatest present-day concernin these countries are large community-wide waterborne outbreaks of parasiticprotozoa that are associated with eitherunfiltered or inadequately flocculated orfiltered water, such as the Cryptosporidi-um outbreak that occurred in Milwaukeein the early 1990s (MacKenzie et al.,1994; Moe, 1996). It is important to notethat, although drinking water is quitesafe in the United States, it remains a sig-nificant cause of morbidity and mortali-ty in developing countries, where waterremains a common source of bacteria,viruses, and parasitic protozoa that im-pact human health. If this water is usedin food processing, waterborne diseasesof developing countries can be passed on

53EXPERT REPORT

to the consuming public in more devel-oped regions of the world through im-portation of contaminated foods.

From the perspective of post-harvestissues for produce, wash water, rinse wa-ter, and ice can serve as a potentialsource for contamination of fresh pro-duce (Beuchat and Ryu, 1997). Perhapspredictably, recent evidence suggests thatpuncture wounds on fruit usually har-bor greater numbers of pathogens togreater depths than seen at other loca-tions of the intact fruit (Burnett et al.,2000). More significant, Burnett et al.(2000) found that a negative temperaturedifferential (application of cold rinse wa-ter to warm fruit) may increase the rela-tive attachment and infiltration of bacte-rial pathogens in intact fruit. This obser-vation has significance for the use of hy-drocooling technology, which has beenapplied to produce items such as straw-berries, cherries, and field crops. Al-though hydrocooling water is frequentlydecontaminated by chlorination, inade-quate control of pathogens in water re-circulated through hydrocoolers couldprovide a source of pathogens that theninfiltrate produce items because of thetemperature differentials.

Historically, water for food process-ing in the United States has originatedfrom municipal systems. However, withincreased volume, as well as stricter reg-ulations, water usage in the processingenvironment has increased dramaticallyover the last decade. Water reclamationand reuse has received considerable at-tention, with guidelines provided by theEnvironmental Protection Agency (EPA)(EPA/AID, 1992), although actual stan-dards remain the responsibility of stateagencies. These EPA guidelines addresswater reclamation for nonpotable urban,industrial and agricultural reuse; directpotable reuse is not practiced in theUnited States. Food processors have ex-pressed some interest in water reuse aswell, particularly in animal slaughter fa-cilities and the USDA/FSIS has issuedguidance for this application. In general,the reuse of water from any one locationin the slaughter process is restricted tocertain other point locations in theslaughter process, as provided by specificrecommendations of the agency. In al-most every instance, decontaminationsteps such as addition of chlorine, as wellas microbiological monitoring, must bedone on the reconditioned water, withvery specific standards recommended.Reuse water from advanced wastewater

treatment facilities can be used on edibleproduct only if the facility meets EPA re-quirements. Unfortunately, these facili-ties are extremely expensive and mostU.S. slaughter operations practice little ifany routine water reuse at the currenttime.

Alternative ProcessingTechnologies

Preservation techniques currentlyact in one of three ways: (1) preventingpathogen access to foods, (2) inactivat-ing them should they gain access, or (3)preventing or slowing their growthshould the previous two methods fail(Gould, 2000a). Traditional food pro-cessing has relied on thermal treatmentsto kill/inactivate microbiological con-taminants. Unfortunately, thermal pro-cessing induces physical and chemicalchanges in the food. Canned greenbeans do not have the same taste andtexture as fresh, despite having similarnutritional profiles. Chemical preserva-tives and naturally occurring antimi-crobial compounds also have been usedextensively for food preservation(Davidson, 1997). Again, a “pickled” oracidified food such as cauliflower doesnot have the same role on the menu as afresh stalk of cauliflower, despite theiridentical origin. Beyond the use of sin-gular food preservation techniques,many strategies employ a combinationof preservation techniques, e.g., refriger-ated storage under modified atmo-sphere, reduced heat treatment withsome acidification, and mild heat withreduced water activity (Gould, 2000b).Leistner (2000) states that the impor-tant preservation approaches often usecombinations of several factors to as-sure microbiological safety. These fac-tors, often called “hurdles,” include heat,acidity, water activity, redox potential,preservatives, competitive flora, lowtemperatures, and more than 40 otherpossible factors. Increasingly, the Amer-ican public has sought “fresh” products,fueling efforts to develop many alterna-tive processing technologies that resultin products that have minimal process-induced changes in sensory and nutri-tional characteristics. It is expected thatthese technologies will play an increas-ing role in food processing in the future.

Any discussion of emerging foodsafety issues must consider the impact ofthese alternative food processing tech-nologies. This analysis must consider

both the immediate impact and the po-tential ramifications further down thefarm-to-table chain. Altering any param-eter along the entire food chain can haveconsequences beyond the immediatechange, be they intended or unintended.For example, modifying the feed of afood animal may alter the microbial con-tamination of the final product that mayin turn contaminate the kitchen of theunaware consumer. A new method ofheating food (e.g., microwave) may speedup the heating and change the types ofmicroorganisms that survive the heatingprocess, whether the heating takes placein a food processing plant or the con-sumer’s kitchen.

Overview

A recent report generated by IFT forFDA (IFT, 2000b) defined alternativetechnologies, identified the pathogens ofpublic health concern that are the mostresistant to various technologies, de-scribed the mechanisms of pathogen in-activation including the inactivation ki-netics, identified ways to validate the ef-fectiveness of microbial inactivation,identified critical process factors, anddescribed process deviations and waysto handle them. The report also de-scribed synergistic effects among tech-nologies, when data were available, andarticulated future research needs foreach technology.

Kinetic parameters and models arefrequently used to develop food preser-vation processes that ensure safety. Theyalso allow scientists to compare the abili-ty of different process technologies to re-duce microbial populations. Kinetic pa-rameters, with their recognized limita-tions, use empirical coefficients experi-mentally determined from microbial re-duction measurements to document therelationship between microbial popula-tion decreases and different process con-ditions. Kinetic parameters for microbialpopulations exposed to thermal treat-ments have been assembled over a signif-icant period of time. Published literaturehas included kinetic parameters neededto control most process, product and mi-crobial situations (Pflug and Gould,2000).

IFT (2000b) used the models and ki-netic parameters to present and comparemicrobial inactivation data from ther-mal, pressure and electromagnetic pro-cesses. Thermal parameters apply to mi-crowave energy, electrical resistance


(ohmic), and other temperature-basedprocesses. Researchers have spent sub-stantial time studying how various mi-crobial populations respond to thermaltreatments. The scientific literature con-tains kinetic parameters for most pro-cess, product and microbial situations.These thermal parameters provide asound basis for development of micro-

wave energy and ohmic technologies.The parameters currently used for pres-sure or pulsed electric field (PEF) treat-ments should be applicable to other pro-cesses where pressure or electricity is theprimary critical factor in reducing mi-crobial populations. Given the scarcityof data, parameters for pressure orpulsed electric field treatments must be

estimated, highlighting the urgent needfor additional research. For several othertechnologies, the quantity of data de-scribing the treatment’s reduction of mi-crobial populations is insufficient for acomparison.

Scientists face limitations in inter-preting these parameters. When the pa-rameters are used to develop a process,

Limitations applicable to all or most of the technologies:

Pulsed Electric Field:

Few published reports

Kinetics based on 2-point survivor curves

Few reports state threshold field strengthfor inactivation

Treatment vessel design is major variableamong studies

Mechanisms of action need confirmation

Most resistant pathogens and appropriatesurrogates need to be identified

Development of resistance after sublethaltreatment needs to be tested

Kinetic models and critical process factorsaffecting kinetics are needed

Effective monitoring systems, uniformtreatment chamber design, electrodeconstruction, and other hardwarecomponents are needed to assureconsistent delivery of specifiedtreatment

Electrothermal:

Difficult to monitor temperature distributions andheating patterns in solids and particles

No well-developed models for process deviationsand use of alternative or variable frequencies ofprocessing energy

Some food factors influence process effectiveness




High Pressure Processing:

Influence of pressure on reduction of microbialpopulations using the proper experimentaldesign (statistically valid, collection of data atdifferent pressures and control of temperatureand product), so that z(P) (increase in MPa toreduce D value by factor of 10) and/or activationvolumes (V) are quantified. Synergistic effects

among pressure, temperature and othervariables unknown.

High capital costs of equipment

Questionable reliability of equipment

Solid food must be batch processed andpumpable foods only semi-continuous

Minimal inactivation of bacterial spores in low-acid foods unless mild heat is applied

Survival curves often nonlinear complicatingkinetics and calculation of process parameters

Food enzymes respond differently from eachother

Excessive pressure denatures proteins andchanges food




Some of the technologies present greaterlimitations than others or are at adevelopment stage that requires extensivefurther scientific research before they canbe commercially used. For thesetechnologies, data are insufficient tocalculate kinetic parameters. Forexample, high voltage arc discharge(application of discharge voltages throughan electrode gap below an aqueousmedium) causes electrolysis and highlyreactive chemicals. Although microorgan-isms are inactivated, more recent designs

Linear first-order survivor curve modelmay be inadequate. Appropriate model(s)would be beneficial to all preservationtechnologies.

Standardized experimental protocol(s) forobtaining statistically reliable kineticparameters to describe survivor curves formicrobial populations exposed to variousalternative technologies.

Different inactivation action/mechanism(s)among alternative technologies unidentifiedand undefined to date.

Synergism or antagonism of one alternativeprocess used with another and their combinedeffect on microbial inactivation efficiencyundetermined to date.

Potential formation of unpalatable and toxicby-products due to processing.

No reliable methods for measuring andmonitoring temperatures or other treatmentactions within individual, large, solidparticulates.

Possible new or changing critical processfactors and their effect on microbialinactivation.

need to be developed before consideration foruse in food preservation. Likewise, oscillatingmagnetic fields have been explored for theirpotential to inactivate microorganisms.However, the results are inconsistent; differentstudies have shown the level of microorgan-isms may increase, decrease or not be affected.Data on inactivation of food microorganisms byultrasound (energy generated by sound wavesof 20,000 or more vibrations per second) arescarce and limitations include the inclusion ofparticulates and other interfering substances.Ultraviolet light (UV) is a promising technique

especially in treating water and fruit juices. A4-log bacterial reduction was obtained for avariety of microorganisms when 400 J/m2

was applied. Apple cider inoculated with E.coli O157:H7 treated in that manner achieveda 5-log reduction. To achieve bacterialinactivation, the UV radiant exposure must beat least 400 J/m2 in all parts of the product.Critical factors include the transmissivity, thegeometric configuration of the reactor, thepower, wavelength and physical arrangementof the UV source, the product flow profile, andradiation path length.

Table 10. Limitations to Alternative Processing Technologies Currently Under Development (IFT, 2000b)

Others:

55EXPERT REPORT

care should be taken to compare the re-sistance of different microbial popula-tions or to identify the appropriate targetmicroorganisms. When a new alternativeprocessing technology is under review, itis essential to determine the pathogens ofgreatest public health concern.

When exploring the new preserva-tion technologies, their preservation levelshould be compared to that of classicalthermal pasteurization or commercialsterilization technologies. Thermal pas-teurization focuses on inactivating vege-tative cells of pathogenic microorgan-isms, i.e., microbial cells that are not dor-mant, highly resistant spores. However,to have a commercially sterile product,the process must control or inactivateany microbial life (usually targetingspores of C. botulinum) capable of ger-minating and growing in the food undernormal storage conditions. Commercial-ly sterile products generally require more

extensive treatment than those that arepasteurized. This goal must be consid-ered when evaluating alternative process-ing technologies as well.

Development of a number of alterna-tive processing technologies is underway,although many of these technologies re-quire additional research before they areready for commercial application (see Ta-ble 10). Validation of effectiveness is animportant step in the development of anynew technology (see sidebar below).

High Pressure Processing

To preserve food using high hydro-static pressure, the food (normally pack-aged) is submerged in a liquid (usuallywater) contained in a vessel that gener-ates pressure by pumping more liquidinto the pressure vessel or reducing thevolume of the pressure chamber.

In the 19th century, scientists realized

that applying pressure in this manner in-activated microorganisms and could beused to preserve foods (Hite, 1899).Larsen and coworkers (1914) confirmedthat high pressure processing (HPP) cankill microbial cells. Vegetative bacteriawere inactivated after 14 hours at 607MPa; bacterial spores were extremely re-sistant to pressure but could be inacti-vated at 1,214 MPa. Today, HPP of foodsuses pressures within the range of 300 to700 MPa.

Over the last fifteen years, use ofHPP as a food preservation method hasbeen pursued in an effort to produce safefoods with a reasonable shelf life with theuse of minimal heat. Reducing or elimi-nating high temperatures in food pro-cessing can deliver a food product withflavor, texture, appearance, and nutrientcontent very similar or identical to freshor raw food. The first commercial prod-ucts treated by HPP—fruit products

Validation of Treat-ment EffectivenessUsing MicrobiologicalSurrogates

The function of surrogate organ-isms is different from that of microbio-logical indicators. Surrogates are usedto evaluate the effects and microbial re-sponses to processing treatments. Themain difference between surrogates andindicators is that the latter is naturallyoccurring and the former is introducedas an inoculum. In the case of freshand fresh-cut produce where no tradi-tional processing inactivation steps areused (e.g., heat pasteurization), surro-gates could be used to assess and vali-date decontamination procedures. Inthe case of alternative processing tech-nologies, surrogates could be used tovalidate specific processing efficacy andtreatment delivery. Surrogates may beselected cultures prepared in a labora-tory and inoculated onto or into theproduct, or they may be an inoculumof naturally occurring microorganismsthat conforms to the requirements of asurrogate and has been confirmed toexist at adequate concentrations in thespecific product. Generally, surrogatesare selected from the population ofwell-known microorganisms that have

well-defined characteristics and a longhistory of being nonpathogenic. It can beespecially difficult to identify surrogatesthat are not pathogenic for highly suscep-tible subpopulations and that are unlikelyto undergo transformation into a patho-genic phenotype in the production envi-ronment. In selecting surrogates, the fol-lowing microbial characteristics are desir-able:• Nonpathogenic• Inactivation characteristics and ki-netics that can be used to predict those ofthe target organism• Behavior similar to target microor-ganisms when exposed to processing pa-rameters (for example, pH stability, tem-perature sensitivity, and oxygen toler-ance)• Stable and consistent growth charac-teristics• Easily prepared to yield high-densitypopulations• Once prepared, population is con-stant until utilized• Easily enumerated using rapid, sen-sitive, inexpensive detection systems• Easily differentiated from other mi-croflora.

The validity of an established or newpreservation or decontamination processis frequently confirmed using an inocu-lated test pack consisting of the foodproduct inoculated and tested under ac-

tual plant conditions, which includesprocessing and control equipment,product handling and packaging. Be-cause pathogens should not be intro-duced into the production area, surro-gate microorganisms should be used ininoculated pack studies, and their sur-vival or growth can be measured tovalidate the process. For instance, sur-rogates have been used for many yearsin the low-acid canning industry to es-tablish and validate the destruction ofC. botulinum spores. The use of non-pathogenic spores of the putrefactiveanaerobe Clostridium sporogenes, orspores of the flat-sour thermophilic or-ganism Bacillus stearothermophilus assurrogates for C. botulinum, havehelped the industry develop thermalprocesses that ensure products are safeand commercially sterile. Listeria in-nocua M1 has thermal resistance pro-files similar to L. monocytogenes but isdesigned for easy detection as a surro-gate and is not a pathogen (Fairchildand Foegeding, 1993). In addition,nonpathogenic strains of E. coli haveserved as surrogates for E. coliO157:H7. In all cases, the surrogateorganism is added to the food productand used to obtain quantitative infor-mation to determine and validate theefficacy of food processing or decon-tamination methods.


such as jams and jellies—reached themarketplace in Japan in 1991. An acidicpH is an important element in the safetyof HPP preserved foods because thepressures used in commercial applica-tion of the technology have limited effec-tiveness against bacterial spores.

Certain principles apply to HPP in-activation of pathogenic bacteria: (1) in-creasing the pressure magnitude or timeof pressurization will usually increasethe number of bacteria destroyed (withthe exception of bacterial spores); (2) anacidic pH or temperatures above ambi-ent enhance pressure inactivation rates;(3) gram-positive bacteria tend to bemore resistant to HPP than gram-nega-tive bacteria; (4) cells in exponentialphase are generally more pressure-sensi-tive than in stationary phase; and, (5) in-complete inactivation of bacteria usingHPP can result in injured cells that arecapable of recovery under optimalgrowth conditions, a common phenome-non known as sublethal injury.

Although increasing the pressurekills more bacteria in less time, higherpressures can cause far greater levels ofprotein denaturation and other detri-mental changes in sensory quality thataffect the food’s appearance and textureas compared to the unprocessed product.

The major factors affecting the effec-tiveness of HPP are: the type of bacteriapresent in the food; its growth condi-tions; the composition, pH and water ac-tivity of the food; and the temperature,magnitude, and time of pressurization(Hoover, 1993).

As described by LeChatelier’s Princi-ple, pressure enhances reactions that re-sult in a volume decrease and inhibitsthose reactions leading to an increase involume (Johnson and Campbell, 1945).For this reason, pressure alters the equi-librium of the interactions that stabilizethe folded three-dimensional shape ofproteins (Masson, 1992). The extent ofdenaturation by pressure depends uponthe structure of the protein, the pressurerange, and other external parameterssuch as temperature, pH and solventcomposition. Pressure will primarily af-fect the hydrophobic interactions of pro-teins; covalent bonds are not affected.Consequently, the extent of hydropho-bicity of a protein can significantly deter-mine the degree of protein denaturationat any given pressure (Jaenicke, 1981).Pressurized membranes normally showaltered permeabilities and it is believedthat denaturation of membrane proteins

is one of the primary reasons for loss ofmembrane integrity leading to injuryand death of the bacterial cell (Paul andMorita, 1971). Pressure-induced mal-functions of the membrane inhibit ami-no acid uptake that is probably due tomembrane protein denaturation, but ithas been shown that bacteria with a rela-tively high content of diphosphatidyl-glycerol (shown to cause rigidity inmembranes in the presence of calcium)are more susceptible to inactivation byHPP (Smelt et al., 1994), and those com-pounds that enhance membrane fluidityusually impart pressure resistance to anorganism (Russell et al., 1995). It is wellestablished that a loss of intracellularcomponents occurs when microorgan-isms are exposed to high levels of hydro-static pressure.

Some foodborne pathogens are moreresistant to inactivation by HPP thanothers. In work by Patterson et al. (1995),Yersinia enterocolitica was reduced 5 log

10

cycles when exposed to 275 MPa for 15minutes in a phosphate-buffered salinesolution. For comparable 5-log

10 reduc-

tions using 15-minute treatments, S. Ty-phimurium required 350 MPa, L. mono-cytogenes required 375 MPa, S. Enteriti-dis required 450 MPa, and E. coliO157:H7 and S. aureus required 700MPa. The bacteria showed more pres-sure resistance in ultra high temperatureprocessed milk than meat or buffer.Thus, the variability of pressure responseis related to bacterial differences and dif-ferent food substrates. By using treat-ment temperatures of 50 C instead ofambient, similar reductions could be ac-complished for E. coli and S. aureus at500 MPa instead of 700 MPa (Pattersonand Kilpatrick, 1998).

For most foods, 10-minute expo-sures to pressures in the range of 250-300 MPa (37,500-45,000 psi) result inwhat can be called “cold pasteurization.”Here, levels of inactivation represent areduction of microorganisms (primarilyvegetative bacteria) of approximately 4 to6 log

10 CFU/mL or g. Cold pasteuriza-

tion is a currently popular term usedwidely for any “nonthermal” food pro-cess or processes (such as HPP or irradi-ation) that do not depend on extensiveheat as the major mechanism of microbi-al inactivation. Foods that are cold-pas-teurized are not cooked or heated in theconventional sense and thus do not sig-nificantly lose sensory quality (e.g., ap-pearance, texture, and flavor) and nutri-ent content; however, there is a signifi-

cant reduction in microbial population.Pressures greater than 500 MPa are

usually required for greater microbial re-duction or consistent product steriliza-tion, but use of low pH and mild heattreatment (45 – 70 C) is often necessaryto attain commercial sterility. For exam-ple, in the case of green infusion tea,Kinugasa et al. (1992) produced a com-mercially sterile product using 700 MPaat 70 C for 10 minutes. These treatmentparameters were successful even whenthe tea was inoculated with 106 spores ofB. cereus, Bacillus coagulans and Bacilluslicheniformis. The tea’s flavor was un-changed as were the catechins, vitamin Cand amino acids in the tea. In anotherexample of using HPP in a combinedpreservative approach, Shearer et al.(2000) measured the reduction of sporesof Bacillus, Clostridium and Alicycloba-cillus in test foods using HPP and 45 Cin combination with such preservativesas sucrose laurates, sucrose palmitate,sucrose stearates, and monolaurin. Oth-er preservative combinations that in-clude HPP also used carbon dioxide(Haas et al., 1989) and acidification plusaddition of nisin (Roberts and Hoover,1996).

HPP continues to be developed as anonthermal food processing method;however, scientists still need to develop areliable method to predict the HPP pro-cess endpoint, the point at which allpathogenic bacteria are inactivated. Theheat resistance of a pathogen does notdirectly correlate to its pressure resis-tance, and the potential emergence ofpathogens with unusual pressure resis-tance is an issue to address. As a relative-ly new commercial food process, concernstill exists for the safety of some foodsprocessed using HPP, especially given theever-evolving nature of microorganisms.

Irradiation

Food irradiation, first commerciallyintroduced in the early 1960s, uses ioniz-ing radiation to decontaminate and dis-infect food and inhibit sprouting andripening. The ionizing radiation is usual-ly in the form of gamma rays producedby radionuclides such as 60Co (cobalt) or137Cs (cesium). Newer irradiation tech-nologies include e-beam, where ordinaryelectricity is used to produce a stream ofelectrons, or x-ray irradiation, where theelectron beam is bounced off metal tocreate x-rays (Farkas, 1998). To get awayfrom using the term “irradiation,” food

57EXPERT REPORT

irradiation is more frequently referred toas cold pasteurization, because the muchshorter wavelengths of gamma and x-rays and e-beams penetrate the food veryrapidly and little or no heat is produced.The purpose of food irradiation dependson specific applications; in general, foodirradiation is used to reduce the levels offoodborne pathogens on the food, inacti-vate the food spoilage microorganisms,and prolong the shelf life of fresh foodsby decreasing the normal biologicalchanges associated with growth and mat-uration processes, such as ripening orsprouting. Both gamma irradiation andx-rays can be used for thick foods as theypenetrate several feet, whereas e-beam ir-radiation can only penetrate severalinches. As well, both e-beam and x-rayirradiation are considered environmen-tally friendly, due to the absence of a ra-dioactive power source and the ability toswitch the system off and on at will.

FDA has expanded the use of x-rayand e-beam irradiation for the treatmentof prepackaged foods, and companies inthe United States have recently begun us-ing e-beam technology to pasteurizeground beef. In addition, companies inthe United States will soon begin usingx-ray systems for irradiation of packagedfoods. For example, a new x-ray test cen-ter has opened, allowing food producersto fine-tune x-ray irradiation protocolsfor a variety of foods.

Like other physical processes such ascooking and freezing, irradiation cancause some alteration of the chemicaland sensory profiles of a food. Treatmentwith ionizing radiation results in chemi-cal modification of extremely smallamounts of the major constituents offood (carbohydrates, proteins and lip-ids), and can also affect minor compo-nents of food, specifically vitamins andDNA. However, these changes are con-sidered insignificant with respect to nu-tritional adequacy. In general, most foodnutrients are unaffected by irradiation,with the exception of some vitamins forwhich minor decreases may occur. It isunlikely, however, that any vitamin defi-ciency would result from the consump-tion of irradiated foods (Diehl, 1995; Jo-sephson et al., 1978; Kilcast, 1994).

U.S. commercial production of irra-diated foods for food safety purposes isrelatively recent, although the technologyhas a longer history as a treatment formedical devices and for control of insectinfestation and sprouting in fresh pro-duce. Applications of ionizing radiation

accepted by FDA for food safety purpos-es include the more recent addition ofmicrobial control in fresh and frozenmeat and poultry.

To establish the safety of a proposedfood irradiation application, FDA re-quires data on the radiological, toxico-logical, and microbiological safety, aswell as the nutritional wholesomeness ofthe irradiated product (Pauli and Taran-tino, 1995). From the perspective of ra-diological safety, the energy produced byapproved radiation sources is too low toinduce radioactivity in foods. The issueof toxicological safety is more complexand has been thoroughly studied in thepast. Recent petitions in the meat andpoultry area have indicated that FDA’sprincipal current interest lies in specifi-cation of the conditions for food irradia-tion (such as temperature and packagingatmosphere) and their impact on micro-biological safety and nutritional adequa-cy (Olson, 1998). The two most impor-tant concerns related to the microbiolog-ical safety of irradiated foods are (1) thepotential to create highly virulent mutantpathogens, and (2) the potential that re-ducing the harmless background microf-lora could eliminate competitive micro-bial forces and allow uncontrolledpathogen growth.

The relative radiation resistance ofmicroorganisms can be summarized asfollows (in order of most resistant toleast resistant): viruses>spores>gram-positive bacteria>gram-negativebacteria>yeasts and molds>parasites(Monk et al., 1995). However, there aresome nonsporeforming bacteria, such asDeinococcus radiodurans and Acineto-bacter radioresistens, that are extremelyradiation resistant by virtue of their ge-netic make-up. Although the exact cel-lular mechanism(s) responsible for suchresistance is unknown, scientists believethat these radiation-resistant bacteriapossess particularly effective nucleic acidrepair mechanisms. Although these or-ganisms are not known to be pathogens,they may provide a genetic pool fromwhich pathogens could theoretically pickup resistance genes through mechanismsof genetic exchange. Furthermore, someresearchers have proposed concerns thatirradiation may produce mutant strainsand/or radiation-resistant pathogens bymeans of natural selection. To date, vari-ous studies and independent reviews byresearchers and international organiza-tions have found no indication of specif-ic bacteriological hazards associated with

food irradiation. However, as irradiationbecomes more widely used, the potentialfor the production of resistant pathogen-ic mutants could be magnified, so con-tinued surveillance is warranted.

Treatment with low to medium doses(i.e., non-sterilizing doses) of ionizingradiation greatly reduces, but does notnecessarily eliminate, bacteria and otherorganisms that may be present in thefood (Monk et al., 1995). Complete ster-ilization of foods is not the purpose nor,in most cases, even desirable for irradia-tion. Because non-sterilizing doses of ra-diation do not kill all bacteria, certainpathogenic bacteria (e.g., C. botulinum,Salmonella) may survive and, in the ab-sence of competition from harmless bac-teria, may multiply to potentially hazard-ous levels. Before using irradiation as afood safety measure, experimental evi-dence is required to demonstrate that theproposed treatment achieves the intend-ed microbiological control without al-lowing C. botulinum growth and toxinproduction.

Most of the recent interest in food ir-radiation has focused on its efficacy incontrolling bacterial pathogens such asSalmonella, E. coli O157:H7 and L.monocytogenes in muscle foods. As withheat, higher irradiation doses kill greaternumbers of bacteria (Olson, 1998).While different bacterial species andstrains demonstrate differences in rela-tive radiation resistance, recommendedirradiation doses for fresh meat andpoultry result in destruction of greaterthan 99.9% of Salmonella and L. mono-cytogenes, and more than adequate con-trol of E. coli O157:H7 (Farkas, 1998; Ol-son, 1998). The marine Vibrio patho-gens, of concern in bivalve molluscanshellfish, are also extremely radiation-sensitive (Rashid et al., 1992). The para-site T. gondii is readily inactivated bygamma irradiation at doses of 0.25 kGy(Dubey et al., 1986), whereas T. spiralis ismore resistant to gamma irradiation andrequires doses of 7-9.3 kGy to kill theparasite in situ and 0.18 kGy to stop de-velopment of larvae to the adult stage(Monk et al., 1995). Unfortunately,foodborne viruses such as hepatitis A vi-rus and the Norwalk-like viruses are re-sistant to radiation inactivation, and thisis not a promising control technology forthese foodborne pathogens (Bidawid etal., 2000; Mallett et al., 1991).

Besides having an effect on pathogenload, an added benefit of irradiation is areduction in the numbers of spoilage mi-


croorganisms. For instance, the gram-negative psychrotrophs, which are thepredominant spoilage organisms forfresh meat and poultry, are very suscep-tible to irradiation. In general, irradia-tion of raw meats results in a significantextension of shelf life, as much as twicethat of non-irradiated, refrigeratedproducts (Olson, 1998). However, re-gardless of the impact of food irradia-tion on reducing spoilage and bacterialpathogens, proper storage and handling,including temperature control, afterprocessing is necessary to ensure thatthe food will be safe. It also should benoted that the effect of irradiation treat-ments on the sensory qualities of foodsdepends largely on the type of foodproduct undergoing treatment, as wellas the dose of radiation used. In someinstances, the dose of radiation neces-sary to destroy pathogens produces un-desirable organoleptic changes in thefood product. For example, oxidation oflipids in the food can cause discolora-tion and rancidity. In short, irradiationis not an effective treatment for patho-gen control in all food products.

Irradiation represents one toolamong several (e.g., fumigants, carcassrinses, steam pasteurization, chemicalsprout inhibitors, food preservatives) forenhancing food safety. It is likely that inthe future, food irradiation would beused to complement rather than replacemany of the other techniques already inuse. Alternatively, combination treat-ments in line with the hurdle concept(e.g., irradiation plus MAP) may be-come more commonplace and offer anadditional level of control. In some cas-es, irradiation may be a safer alternative.For example, when used to reduce themicrobial load on spices, irradiation canserve as an alternative to fumigants suchas ethylene oxide or methyl bromide,which are particularly toxic to occupa-tionally exposed individuals.

A key advantage of food irradiationfor controlling pathogens is that it re-duces the microbial load at the point atwhich the product has been packaged,which increases the likelihood that theproduct the consumer receives will besafe. However, like other processes, irra-diation only protects against pathogensthat contaminate the product at the timeof processing; it does not protect againstfuture contamination that may occurduring handling, storage and prepara-tion of the food.

A number of additional issues must

be addressed as food irradiation facilitiesare established. For example, the occupa-tional health and safety of workers in ir-radiation facilities and the transport ofradioactive materials for gamma irradia-tion facilities merit consideration. Ifgamma irradiation of food becomes asignificant industry in the future, it willlikely entail the construction of more fa-cilities and the increased transport of ra-dioactive materials. These concerns mayincrease advancement of the more envi-ronmentally friendly irradiation technol-ogies, namely x-ray and e-beam irradia-tion.

There is also a need to do more workon indicators that one can use to deter-mine if foods have been irradiated and ifso, the level of irradiation given. At therecommended doses for specific applica-tions, there are no major chemical, physi-cal, or sensory changes in irradiatedfoods. Therefore, detection methodsmust focus on minute changes such asminor chemical, physical, histological,morphological, and biological changes inthe food. Some promising methods formeasuring these changes in food includehydrocarbon and cyclobutone for lipid-containing foods, electron spin reso-nance for bone-containing food, ther-moluminescence for foods containingsilicate minerals (Olson, 1998) and theDNA comet assay for analysis of foodswith low fatty acid content (Cerda et al.,1997).

Active and Intelligent Packaging

The primary purpose of food pack-aging is to protect the food from physical,microbial and chemical contamination.Therefore, the type of packaging usedplays an important role in determiningthe shelf life of a food. Today’s consum-ers increasingly demand mildly pre-served convenience foods with fresh-likequalities. In addition, advances in retailand distribution practices (e.g., central-ization of activities, Internet shopping,global procurement) lead to greater dis-tribution distances and longer storagetimes for a variety of products with dif-ferent temperature requirements, therebycreating immense demands for innova-tion from the food packaging industry.Active and intelligent packaging technol-ogies are used to extend shelf life, im-prove safety and improve the sensoryproperties of packaged foods. This isachieved by providing the best microen-vironment within the package through

optimal gas composition and humiditylevel. The challenge for food manufac-turers is to maintain product safety whileproviding these storage and preservationconditions.

Ideally, active packaging would sensethe microenvironment within the pack-age and modify conditions accordinglyto extend shelf life, improve safety or en-hance sensory properties. This active ap-proach is different but complementary tointelligent packaging, which provides in-formation about critical parameters suchas temperature, time, gas content, or mi-crobial contamination. New packagingadvances are starting to combine thesetwo concepts in the next generation offood packaging.

Active packaging is not one technolo-gy, but a collection suited to specificproblems. Active packaging concepts canbe divided into three major categories:modified atmosphere packaging, activescavenging (oxygen, ethylene, carbon di-oxide) and releasing concepts [emitters(carbon dioxide, ethanol, flavors, fra-grances), and microbial control systems(chlorine dioxide, sulphur dioxide)].

Modified Atmosphere Packaging

Modified atmosphere packaging(MAP) involves the creation of a modi-fied atmosphere by altering the normalcomposition of air (78% nitrogen, 21%oxygen, 0.03% carbon dioxide and tracesof noble gases) to provide an optimumatmosphere for increasing the storagelength and quality of food (Moleyar andNarasimham, 1994; Phillips, 1996). Theatmospheric modification can beachieved by using controlled atmospherestorage (CAS) and/or active or passiveMAP. Active modification creates a slightvacuum inside the package that is thenreplaced by a desired mixture of gases.Passive modification occurs when theproduct is packaged using a selected filmtype, and a desired atmosphere developsnaturally as a consequence of the prod-uct’s respiration and the diffusion of gas-es through the film (Lee et al., 1996; Mo-leyar and Narasimham, 1994; Zagory,1995). The choice of the film is an inte-gral part of this system, because gas dif-fusion rates vary greatly among films,and therefore films differ in their abilityto maintain the desired modified atmo-sphere. Also taken into consideration isthe storage temperature, which will alsoaffect gas diffusion rates.

The increased product shelf life re-

59EXPERT REPORT

sults from the effect of the modified en-vironment on microbial growth and, forrespiring food products such as fruitsand vegetables, on the product (Molin,2000). Low oxygen levels negatively af-fect aerobic microorganisms. Carbon di-oxide inhibits the growth of some micro-organisms, however, it is not inhibitoryfor all microorganisms, and it is impor-tant to understand that under modifiedatmospheres, the growth rates of somemicroorganisms will be reduced, whileothers may increase or stay the same.

MAP is not without safety concerns.Beneficial microorganisms may be in-hibited, potentially allowing certainpathogens to proliferate and causefoodborne illness. C. botulinum, L.monocytogenes and potentially patho-genic psychrotrophs are of primaryconcern. Psychrotrophs grow well at orbelow 7 C and have their optimumgrowth temperature between 20 and 30C, meaning they can multiply under re-frigeration conditions. If the oxygen lev-el of the modified atmosphere is toolow, the anaerobic environment couldfacilitate C. botulinum growth. Al-though studies indicate that many foodsbecome inedible before any toxin is pro-duced (Molin, 2000), foods that remainorganoleptically acceptable after toxinproduction are a significant publichealth concern. For example, recentstudies have shown that C. botulinumcan grow and produce toxin in productssuch as MAP pizza and English-stylecrumpets, while the products remainorganoleptically acceptable (Daifas etal., 1999a,b). With respect to meat andfish products, Molin (2000) concludedthat most products are judged inediblewell before any toxin is produced. How-ever, there have been studies where toxinproduction in MAP meats preceded orcoincided with the development of un-acceptable sensory characteristics(Lawlor et al., 2000), as well as reportsof botulism due to consumption of vac-uum-packaged smoked fish products(Korkeala et al., 1998). L. monocytoge-nes is a problem with products, such asready-to-eat products, fruits and vege-tables, that are not heated adequatelybefore consumption. Low concentra-tions of CO

2 (less than 10%) used in

MAP of some products may inhibit thenatural microflora and increase thegrowth rate of L. monocytogenes. Com-bined with storage at refrigeration tem-peratures, which can select for L. mono-cytogenes, a low-CO

2 MAP environment

may pose a food safety concern. A lowpartial pressure of carbon dioxide, com-bined with refrigerated storage, can alsofavor the growth of Aeromonas and theEnterobacteriaceae. MAP containing el-evated levels of CO

2 (70-100%) inhibits

the growth of L. monocytogenes in a va-riety of products (meat products, cot-tage cheese, turkey roll slices), whereas100% N

2 allows multiplication of the

pathogen (Phillips, 1996). Fresh-cutand whole fruits and vegetables, in par-ticular, do not usually tolerate CO

2 con-

centrations above 15%, well below theinhibitory level for L. monocytogenes.The CO

2 rich atmosphere, however, can

select for lactic acid bacteria, which havebeen shown to be inhibitory towards L.monocytogenes (Francis and O’Beirne,1998).

Active Scavenging Concepts

Active scavenging concepts includeoxygen, ethylene and taint scavengers aswell as moisture absorbers or humiditycontrollers.

The presence of oxygen in food pack-ages accelerates the spoilage of manyfoods. Although vacuum packaging andMAP have been somewhat successful inextending the shelf life and quality offood, aerobic spoilage can still occur be-cause of residual oxygen in the head-space. Oxygen residual remains due tooxygen permeability of the packagingmaterial, small leaks due to impropersealing, air enclosed in the food or inade-quate evacuation and/or gas flushing.Oxygen can cause off-flavors, colorchange, nutrient loss, and growth of mi-croorganisms. Oxygen scavengers areuseful for removing residual oxygen andcan be applied in different ways: sachetsand labels containing oxygen-scavengingcomponents, closures (mainly used forplastic beer bottles), and oxygen-scav-enging flexibles. One of the largest appli-cations of oxygen-scavenging systems ismold control in packaged baked goodsand cheese food packages.

The majority of current commercial-ly available oxygen scavengers are basedon the principle of iron oxidation. In-serting a sachet into the food package iseffective, but can meet with resistanceamong food manufacturers, because offear of ingestion of the sachet, notwith-standing the labelling, and the potentialfor the sachet to leak the contents intofood, resulting in an adulterated product.

The main advantage of using oxygen

scavengers is that they can reduce oxygenlevels to less than 0.01%, which is muchlower than the typical 0.3-3.0% residuallevels achievable with modified atmo-sphere packaging. As well, oxygen scav-engers are sometimes used in combina-tion with carbon dioxide scavenger sys-tems.

In terms of public health, the mainissue surrounding the use of oxygenscavengers is their potential to create anenvironment that may promote thegrowth of potentially harmful anaerobicbacteria. Similar concerns have been dis-cussed with respect to MAP conditionsthat result in a low oxygen atmospherewithin the package. The main organismsof concern in this situation are Clostridi-um species, mainly C. botulinum, but alsoC. perfringens, due to their growth andtoxin production in anaerobic environ-ments.

Active Releasing Concepts

Considerable research has focusedon the release of antimicrobials frompackaging materials to limit microbialspoilage, however the choice of antimi-crobial compound is often limited by itscompatibility with the packaging materi-al and by the ability to withstand the heatduring extrusion (film formation). Ex-amples of such compounds are chlorinedioxide, sulphur dioxide, ethanol, carbondioxide emitters, antimicrobials (zeolite,triclosan and bacteriocins) and antioxi-dants (BHT/BHA or vitamin E).

Intelligent Packaging Systems

Intelligent packaging can sense theenvironment and/or convey informationto the user about its contents. Althoughresearch in this area has been extensive,only a few technologies have made it pastthe research and development phase. Twomain strategies are employed in the de-velopment of a new intelligent packagingsystem: the use of novel materials andprocesses, and the application of elec-tronics and MEMS (microelectro-me-chanical systems). The first strategy isusually the least costly, but the develop-ment and specification process can betime-consuming. Electronic or MEMStechnology is costly, but far more versa-tile and flexible. The technology can beused to micro-fabricate sensors that willdetect pressure, acceleration, humidity,temperature, physical damage and expo-sure to radiation.


Key applications for intelligent tech-nology are in the areas of tamper evi-dence, quality monitoring (i.e., tempera-ture abuse), counterfeiting (i.e., holo-grams), theft protection and supplychain management and traceability (i.e.,automatic data capture coupled to theInternet). Traceability in the food chainis a high profile issue, because of its im-portance in determining which foods arepart of a potentially contaminated lot.Tracing technology is already being usedin many parts of the food chain, however,a number of gaps exist.

An example of quality managementpackaging is a newly developed methodof preparing packaging material for foodand other products that contain diagnos-tic properties. The technology involvesimmobilizing and protecting antibodies(or other ligands) on the polymer filmsurface. These substances react with tar-gets—such as food pathogens and tox-ins, pesticide residues or proteins in thefood—and create a visual sign on thefilm surface, alerting the consumer or re-tailer that the food may be contaminated.However, this technology would likelyonly be applicable for surface contami-nation. Other examples of quality man-agement technologies are temperatureindicators, time indicators (aging strip),time/temperature indicators, and micro-wave cook indicators.

Future Directions and Concerns

From a human health and safetystandpoint, one must consider the effectsof active packaging on the microbialecology and safety of foods. As previous-ly discussed, removing oxygen fromwithin packs of high water activity,chilled, perishable food products maystimulate the growth of anaerobic patho-genic bacteria. In addition, the modifiedatmospheres created with MAP and scav-enging technologies may sufficientlychange the competitive microbial envi-ronment, allowing for pathogen growth.To control undesirable microorganismson food, antimicrobial substances can beincorporated into the surface of foodpackaging materials. The major potentialfood applications for antimicrobial filmsinclude meat, fish, poultry, bread, cheese,fruits and vegetables. However, antimi-crobial films that only inhibit spoilagemicroorganisms without affecting thegrowth of pathogenic bacteria, will raisefood safety concerns that must be ad-dressed. A better understanding of the

effects that new technologies have on theinterrelationship among food, microor-ganisms and package is required. Ques-tions remain about the by-products ofoxygen-scavenging systems, the effective-ness of antimicrobial films on productswith irregular surfaces, and the spectrumof activity of the antimicrobial additives,to name a few examples. In addition toimproving packaging technology, atten-tion must be given to fully understand-ing the effects on food quality and safety.

Active and intelligent packaging of-fer great benefits to both consumersand manufacturers and are undoubt-edly one of the areas of future innova-tion in the food industry. However,there are still concerns regarding theregulatory approval of these technolo-gies. Some of the packaging concepts,such as the active release of antioxi-dants or antimicrobials, have the po-tential for these additives to migrateinto the food product. Food-contactapproval must be established beforeany form of active packaging is used,and labeling may be needed in caseswhere active packaging gives rise toconsumer confusion. As well, it is im-portant to consider environmental reg-ulations covering disposal of activepackaging materials. Public perceptionof antimicrobial or antioxidant use infood packaging must also be consid-ered. Consumers are seeking morenatural foods that are free of contami-nants and additives, and their accep-tance of packaging that incorporatesthese substances may be questionable.The use of natural antimicrobials, suchas those from plants, and natural anti-oxidants, such as vitamin E, in foodcoatings or packaging may be more ac-cepted by consumers.

In the future, we will most likely seedifferent combinations of active and in-telligent packaging, e.g., combining an-timicrobial films with MAP. Research isneeded to see how these new packagingtechnologies will impact on the spoilagemicroflora and survival/growth offoodborne pathogens.

With additional research, the combi-nation of active and intelligent packag-ing technology may emerge as perhapsthe most important preservation tech-nology of the 21st century.

Transportation and Storage

Following the harvest of nearly allcommodity types, raw foodstuffs are

normally transported to holding, ship-ping, or processing facilities. Even pro-cessed foods that have received microbi-cidal treatment are frequently transport-ed to other locations for bottling, pack-aging, and/or shipping. Transport con-veyances are thus a part of the foodchain where contamination can occur.

In 1994, an estimated 224,000 per-sons developed salmonellosis from a na-tionally distributed brand of ice cream(Hennessy et al., 1996). S. Enteritidis wasthe cause, and the most likely scenariowas that pasteurized ice cream premixwas transported by a tanker trailer thathad carried non-pasteurized eggs justbefore being loaded with the premix.Eggs are a known source of S. Enteritidis.The authors concluded that to preventmore such occurrences, food productsnot destined to be re-pasteurized beforeuse should be transported in dedicatedcontainers.

As more and more food commoditiesare grown and harvested in the UnitedStates and abroad, transportation willcontinue to be a factor in foodborne ill-ness and, in fact, may grow in impor-tance. This is not an easy issue to dealwith, as the types of conveyances are asvaried as the types of commodities theytransport.

The storage of foodstuffs also can bean entry point for pathogenic microor-ganisms or permit the growth of patho-gens if present. It is generally acceptedthat most foods need to be maintained atcold temperatures from harvest to con-sumption. This “cold chain” is subject toabuse at several steps, and temperatureabuse can contribute to the growth ofpathogens that can increase the likeli-hood of foodborne illness.

Storage of foodstuffs is carried out inwarehouses and specialized storage facil-ities, as well as in virtually all institutionsthat serve food, including hospitals,nursing homes, schools, restaurants, re-tail stores and the home. FDA has deter-mined that improper cold holding offood is the most frequent temperatureviolation for nearly all facility types. Forexample, in a survey of fast food restau-rants, 31% were out of compliance inthat potentially hazardous foods werebeing stored at temperatures above 41 F(FDA, 2000).

The situation is no better, and isprobably worse, in the home setting. In arecent survey of homes, 16% were hold-ing refrigerated ingredients at too high atemperature, and 55% of the participants

61EXPERT REPORT

that were improperly holding cold ingre-dients did not know at what temperaturea refrigerator should hold product. Theremaining 45% of participants were un-aware that their refrigerators were notholding products at the proper tempera-tures (Daniels et al., 2000). A previoussurvey showed that 23% of the 106households had refrigerators that heldfood at too high a temperature (Daniels,1998).

While improper transportation andstorage are hardly new or emergingfood safety issues, they appear to beproblems that do not go away. As such,they will continue to contribute to theburden of foodborne illness. As the

food chain becomes even more com-plex, perhaps their effect will becomeeven greater.

Retail and Food Service

The retail and food service environ-ments are part of the post-harvest envi-ronment. As more meals are eaten awayfrom home, this environment becomesincreasingly significant. No matter howwell or how poorly food safety measuresare applied prior to consumption, avoid-ing illness often depends on how wellfood is handled immediately prior toconsumption.

While the role of human handling

in foodborne disease has been recog-nized for years, recent food safety initia-tives have increased our awareness ofparticular risks. For instance, strongepidemiological evidence supports thetransmission of Salmonella and Campy-lobacter to ready-to-eat (RTE) foodproducts via cross-contamination withuncooked poultry (D’Aoust, 1989;Deming et al., 1987; Harris et al., 1986;Hopkins et al., 1984). Equally strongevidence exists for the transmission ofviral foodborne disease by poor person-al hygiene of infected food handlers,with recent data suggesting that 50-95%of confirmed viral foodborne diseaseoutbreaks are attributable to human

Outbreaks of Shigellasonnei Infection Associ-ated with Fresh Parsley

In August 1998, the Minnesota De-partment of Health (MDH) receivedmultiple independent complaints of ill-ness and reports of confirmed Shigellainfections among persons who had eat-en at two restaurants (CDC, 1999c).The restaurants were in different cities,had separate water supplies, and had noemployees in common. Preliminary re-sults of interviews with patrons andfood handlers at both restaurants sug-gested that ill food handlers likelyplayed a role in contaminating ice andfresh produce items.

S. sonnei isolates from ill patronsand food handlers at the two restau-rants were submitted to MDH for mo-lecular subtyping by pulsed-field gelelectrophoresis (PFGE). Results ofPFGE demonstrated that both restau-rant outbreaks were caused by the samestrain of S. sonnei and that it was astrain that had not previously been iso-lated in Minnesota. Strains with simi-lar PFGE patterns had been isolatedfrom travelers returning from Mexico.

Because the outbreaks at the restau-rants appeared to have a commonsource, food histories of restaurant pa-trons were re-evaluated by food ingre-dients rather than by menu item. Inone restaurant, uncooked choppedparsley was associated with illness(odds ratio 4.3, 95% confidence inter-val 2.4, 8.0). In the other restaurant,parsley was not associated with illness,

but a high proportion of cases ate dishesthat were served with chopped parsley.These results suggested that choppedparsley was the likely common source forthe two restaurant outbreaks.

In collaboration with CDC, otherpublic health agencies and public healthlaboratories in the United States andCanada were notified of the outbreaksin Minnesota. Six similar outbreaksthat occurred during July-August wereidentified, two in California, and oneeach in the states of Massachusetts andFlorida and the provinces of Albertaand Ontario. S. sonnei isolates wereavailable from five of these six out-breaks. All had the same PFGE patternsseen in the Minnesota outbreaks. Ineach, chopped parsley was sprinkled onfoods that were either implicated by theresults of a formal investigation, or eat-en by a high proportion of cases. Thus,in simultaneous outbreaks linked byPFGE subtype, a common food itemwas implicated.

Tracebacks to determine the sourcesof parsley in the outbreaks linked byPFGE were conducted by state and localhealth officials, FDA, and the CanadianFood Inspection Agency. One farm, inBaja California, Mexico was identified asthe likely source of parsley served in sixof the seven outbreaks. Field investiga-tions at the farm found that municipalwater used for chilling the freshly pickedparsley was unchlorinated and vulnera-ble to contamination. This water alsowas used to make the ice with which theparsley was packed for shipping.

The widespread distribution of the

outbreaks linked to this parsley sourcesuggested that the parsley was contam-inated in the field or during packing.However, results of these investiga-tions also suggested that handling ofthe parsley at the restaurants contrib-uted to the occurrence of the out-breaks. Food handlers at six of theeight implicated restaurants reportedwashing the parsley before chopping it.The parsley was usually chopped inthe morning and left at room tempera-ture during the day before beingserved to customers. Studies at theUniversity of Georgia Center for FoodSafety demonstrated that S. sonnei de-creased by 1-log CFU/g per week onrefrigerated parsley, but increased by3-log CFU/g in 24 hours on choppedparsley at room temperature. In addi-tion, in at least two of the restaurants,food handlers became infected withthe outbreak strain of S. sonnei andappeared to have contributed to ongo-ing transmission in those outbreaks.

These outbreaks demonstrate thecomplexity of foodborne diseasetransmission and outbreak investiga-tions. A contaminated food ingredientwas widely distributed, contaminationwas amplified by handling practices insome restaurants, and infected foodhandlers further amplified the out-break by contaminating ice and otherready-to-eat foods. Only by linkingthis series of apparently unrelated out-breaks by PFGE subtyping of theagent, were public health officials ableto identify the common source and theother contributing factors.


handling (Bean et al., 1997). Anotherparticularly hazardous behavior is theconsumption of high-risk foods (pre-dominantly raw or undercooked foodsof animal origin) (Beletshachew et al.,2000; Klontz et al., 1995).

While all of these behaviors can oc-cur at home, institutions and retail es-tablishments are more significant ven-ues with respect to large, recognizedfoodborne disease outbreaks. That is,single illnesses due to unsafe food han-dling at home are unlikely to be attrib-uted to food and to be reported, eventhough home food handling is an im-portant cause of foodborne disease.Poor handling practices are influencedby a large number of demographic fac-tors including age, gender, race, educa-tion and income (Beletshachew et al.,2000; Klontz et al., 1995). While target-ed food safety education programs havereported some success, they are onlyone component of a larger initiative toinform and motivate food handlersabout food safety (Meer and Misner,2000; Yang et al., 2000).

Recent outbreaks of viral gastroen-teritis illustrate some new discoveriesregarding the significance of the foodhandler in the initiation and propaga-tion of outbreaks. Parashar et al. (1998)reported on the role of an asymptomat-ic food handler in a viral gastroenteritisoutbreak associated with the consump-tion of contaminated sandwiches. Re-searchers discovered that NLVs could beshed in the feces for up to 10 days afterdiarrhea ended in the food worker or byasymptomatic food handlers, and,through poor personal hygiene, subse-quently contaminate food. Green et al.(1999), in describing a prolonged viralgastroenteritis outbreak at a large hotel,noted that toilet rims (72%) and carpets(70%) had a high incidence of contami-nation. These environmental surfacesremained important reservoirs for thepropagation of the outbreak. In per-haps the most interesting study, Beckeret al. (2000) reported a primary food-borne NLV outbreak associated with theconsumption of boxed lunches thatwere served to a North Carolina collegefootball team. During a subsequentgame in Florida the next day, manymembers of the North Carolina teamdeveloped diarrhea and vomiting.Twenty-four hours later, similar symp-toms developed in some of the opposingteam members, illustrating the role ofdirect person-to-person transmission in

the propagation of a primary food-borne viral disease outbreak.

Microbial Stress Responses toProcessing

Human efforts to control microor-ganisms in the food production, pro-cessing and distribution systems havechanged the environment for foodbornepathogens. As the scientific understand-ing of foodborne pathogens has becomemore sophisticated, so too have the con-trol methods. These control efforts areone of many driving forces in pathogenevolution, and as such, their impact onthe virulence and survival of foodbornepathogens must be fully considered (Ar-cher, 1996). Scientists study the effect ofprocessing technologies and otherchanges to the microbial environment toevaluate the effectiveness of control tech-nologies and also the potential that con-trol efforts will drive pathogen evolution.

The many varied processes used toprepare and preserve the wide range offoods are frequently designed to inacti-vate, inhibit or prevent the growth ofpathogenic or spoilage microorganismsin the specific product. Any of these mi-croorganisms that survive the processmay be damaged. Consequently, processevaluations and the microbial inactiva-tion kinetics on which they are basedmust consider sublethal injury of cellsand spores, as well as resuscitation ofcells and alternative germination path-ways of spores. These considerationsmust be built into hazard analysis andrisk assessment to adequately assess andcontrol emerging food safety situations.

Resistance to Controls

The efficacy of a preservation tech-nology is influenced by a number of mi-croorganism-related factors that are gen-erally independent of the technology it-self. These factors include the type andform of the target microorganism; thegenus, species and strain of microorgan-ism; its growth stage; selection by envi-ronmental stresses; and sublethal injury.

Among the foodborne microbiologi-cal hazards, bacteria are generally theprimary targets for most preservationprocesses, and bacterial susceptibility tosublethal cellular injury is of specialconcern. Processes designed to inacti-vate pathogens also must address the re-sistance properties in foods of other mi-croorganisms—such as viruses, yeasts,

molds, and parasites—that may persistor grow in foods even if they do not ex-perience sublethal injury. The activity ofthe entire microbial ecosystem influencesthe survival and subsequent pathogenici-ty of the target cells whether or not thetarget microorganism has been suble-thally damaged.

A few genera of foodborne bacteria(for example, Clostridium spp. and Bacil-lus spp.) are capable of existing in twoforms: active vegetative cells and dor-mant spores. These two forms often dif-fer in their resistance properties to heat,chemicals, irradiation and other environ-mental stresses. Similarly, spores aretypically more resistant than vegetativecells to the alternative processing tech-nologies. Pasteurization inactivates vege-tative cells of disease-producing micro-organisms. To have a commercially ster-ile product, however, the process mustinactivate all microbial spores (usuallytargeting spores of C. botulinum) that arecapable of germinating and growing inthe food under normal storage condi-tions.

Differences in microbial resistance tocontrol methods may be found not onlybetween genera and species but also be-tween strains of the same species. Forinstance, at the genus level, some bacteri-al strains with unique resistance to ther-mal inactivation, irradiation, and highpressure processing have been identified,such as D. radiodurans and the thermo-plasmals, making it possible that, in thefuture, a pathogenic “super bug” couldemerge. Within species, some strains ofcommon enteric pathogens such as E.coli and Salmonella are more resistant tothe effects of low pH and high tempera-ture than other strains of the same or-ganism. If a bacterial strain with resis-tance to multiple control technologiesemerged, it could be a potential foodsafety hazard that would be uncontrolla-ble with technologies that have producedsafe products for generations. If the mi-croorganisms proved to be pathogenic,the control process would have to be re-designed to specifically inactivate it. Al-ternatively, if the “super bug” were not apathogen or spoilage microorganism, itmight be very useful as a possible surro-gate during process development andvalidation.

Another factor that can affect bacte-rial resistance to preservation processesis the stage of growth. Cells in exponen-tial or log phase of growth are generallyless resistant than cells in stationary

63EXPERT REPORT

phase. The development of stress resis-tance proteins in stationary phase is acontributing factor to this phenomenon.

Selection by Environmental Stresses

Extreme environments usually killmost bacterial cells and can result in theselection of mutant cells that are resis-tant to the severe conditions (see selec-tion, p. 21). Studies have suggested thatbacterial stress may induce hypermuta-bility, which would in turn lead to a mi-crobial population of greater resistance(Buchanan, 1997a). Therefore, the expo-sure of cells to some form of stress mayinduce and allow the survival of micro-organisms with unusually high durabili-ty to a given inactivation process.

The responses to stresses in the foodsystem may play a major role in the emer-gence of pathogens (Sheridan and Mc-Dowell, 1998). Bacteria are capable ofadapting to an immediate environmentalstress, but the response is temporary, andthe genes involved are switched off whenthe stress is removed (see stress, p. 22).The stress response may be triggered by asingle parameter or by several simulta-neously, causing variations in response.

Some stress responses have particu-lar relevance to the food processing envi-ronment. For example, stress responsesto temperature shifts can affect L. mono-cytogenes attachment to food contact sur-faces (Smoot and Pierson, 1998).

The cross protective effect in whichexposure to one stress triggers resistanceto other stresses is a special concern infood processing environments. Mazotta(1999) found that the heat resistance ofacid- or salt-adapted, heat-shocked, orstarved E. coli O157:H7 cells was higherthan that of cells grown to exponentialor stationary phase under optimum con-ditions. To add an extra safety factor,Mazotta suggested using stress-inducingculture conditions when studying thethermal resistance of vegetative patho-gens in specific products. The cross pro-tective effects of the bacterial stress re-sponse vary. Lou and Yousef (1997) de-termined that sublethal stresses to etha-nol, acid, hydrogen peroxide, heat, or salthad variable effects on subsequent expo-sure of L. monocytogenes to normally le-thal levels of the same stressors. For ex-ample, heat shocking increased the resis-tance of the microorganism to ethanol,hydrogen peroxide and salt, but not toacid. Similarly, resistance to food antimi-crobials can be acquired through previ-

ous exposure or through cross protec-tion triggered by environmental or pro-cessing factors including stresses such asheat or acid (Davidson, 1999). Some re-searchers have attempted to differentiatevarious multiple stresses, e.g., lethal pHand water activity. Shadbolt et al. (2001)hypothesized that pH-induced stresscauses an energy drain that sensitizesthe cell to other environmental con-straints. Similarly, the effect of habitualexposure to reduced water activity in-creased the heat tolerance of Salmonellaspp. (Mattick et al., 2000). Another ex-ample of cross protection is the in-creased radiation resistance caused bythe induction of acid resistance in en-terohemorrhagic E. coli (Buchanan et al.,1999).

Because environmental stress canincrease a pathogen’s resistance to con-trol technologies, the potential for in-creased resistance must be factored intothe process. The analysis should consid-er: (1) if the food environments are like-ly to induce stress conditions in the mi-croorganisms; (2) whether stress-in-duced resistance could possibly occur atany point in the food processing opera-tion; and (3) if it did, whether it wouldsignificantly impact the inactivationprocess and lead to possible underpro-cessing. Considering that most foodprocessing systems are designed to ex-pose microorganisms only once to anygiven stress-inducing factor (for exam-ple, heat, acid, or antimicrobials), a re-sistant population is unlikely to develop.However, it is possible that sublethal in-jury to E. coli O157:H7 as it passesthrough the low pH of a cow’s gas-trointestinal tract and then through anacid rinse at slaughter may change its re-sistance to heat inactivation duringpreparation. Likewise, L. monocytogenescells in the production environment thatare sublethally injured by repeated ex-posure to sanitizers may have alteredsurvival or virulence characteristics. Anadditional exception is previously pro-cessed material that is reintroduced intothe process stream. In this case, in-depth studies of the impact of process-ing-induced stress are needed.

Sublethal Injury

The microbial ecology of food is in-fluenced dramatically by food process-ing and preservation technologies.Whether or not the actions are directlyor indirectly aimed at microorganisms,

many actions within the food processingsystem—decontamination, sanitation,and product formulation and preserva-tion—can injure cells that survive theevent (see Table 11).

Some of these processes are classicand well established as effective againstpathogens that have existed in the past,but new product formulations, newequipment, or modified regimens cancreate opportunities for some microor-ganisms to survive, frequently in a suble-thally injured state. New processes alsohave the potential to create sublethallyinjured pathogens.

It would be a rare genus of bacteriaof concern in the food industry that hasnot in some situation demonstrated thecapacity to be sublethally stressed or in-jured by a food-related physical or chem-ical insult. Many yeasts and molds asso-ciated with food also have shown suscep-tibility to sublethal damage. Early in thestudy of cell injury, lactic acid bacteriaused as starter cultures for dairy fermen-tations were shown to be cold-damaged,requiring special nutrients for normalrapid growth.

Sublethal injury may be demonstrat-ed or exhibited as more exacting require-ments for growth, greater sensitivities toantagonistic agents (e.g., selective agentsin media or preservatives in a food), in-creased resistance to subsequent inacti-vation treatments by the same or differ-ent agents, increased lag time before ex-ponential growth ensues, or changed vir-ulence as a pathogenic cell. The injuredcell may return to its initial native stateby repairing the cellular damage undersuitable conditions. This resuscitation inthe absence of antagonistic agents or inthe presence of appropriate substrateswill result in a cell with its original capa-bilities, resistances, and virulence. Inother words, the damage of sublethal in-jury is reversible—it is not a permanentgenetic change. This phenomenon is dif-ferent from selection of a resistant mu-tant with a permanent genetic change.

The effectiveness of a microbial inacti-vation process is often measured by enu-merating any surviving organisms in a se-lective medium. Because viability in mi-croorganisms is generally based on theability to increase in numbers to somemeasurable level, death may be defined asan irreversible loss of the ability to gener-ate progeny (Mackey, 2000). Ideally, a mi-croorganism exposed to processing con-ditions would be either viable or dead; inactual practice, the control technology of-


ten produces a continuum of effects in-cluding some degree of cellular injury.

Injured cells can be easily underesti-mated, resulting in misleading conclu-sions about the efficiency of the inactiva-tion method. If the cell damage is notrecognized, the loss of specific identifi-able characteristics as a result of suble-thal injury leads to faulty data. Injuryoccurs in vegetative cells and bacterialendospores of pathogenic and non-pathogenic microorganisms. A cell thatappears to be “dead” because it is unableto multiply and demonstrate its viabilitymay be able to repair itself under somespecial circumstances. So a food, pro-cess, or other microbial environmentcould appear to be pathogen-free only tobecome dangerous when the injured cellsrecover. Mackey (2000) has described thenature of sublethal injury, emphasizingthe many conditions that influence inju-ry, resuscitation, and recovery, and high-lighting the role that sublethal injurymay play in the design of preservationprocesses.

Another irregular or unnatural stateor condition that microorganisms mayenter is the viable (or metabolically active)but nonculturable (VBNC) condition inresponse to stress. These microorganisms

lose their ability togrow in even nonselec-tive environments thatnormally support theirgrowth but are consid-ered still viable be-cause the cells remainphysically intact anddemonstrate metabol-ic activity. The VBNCstate of microbial cellshas been reported tooccur in a number offoodborne pathogensas an outcome of envi-ronmental stress(McKay, 1992; Xu etal., 1982). The possi-bility exists that theseVBNC cells, which arenon-detectable by tra-ditional cultural meth-ods, may cause diseaseif consumed (Colwellet al., 1985). If they aredeveloped in a foodproduction environ-ment, a food safetyrisk may be presented.The very existence ofthe VBNC state has

been questioned, based on studying theconcept from alternative perspectives(Bloomfield et al., 1998; McDougald et al.,1998). New approaches to studying theVBNC phenomenon need to be taken, toarrive at an agreement regarding its oc-currence, and then to determine if thesemicroorganisms have any importance inthe food production environment and tofood safety and the public health.

Future Implications

The responses to stresses in the foodsystem may play a major role in theemergence of pathogens (Sheridan andMcDowell, 1998). Stress responses mayincrease the pathogen’s resistance to in-activation methods, improve its ability tosurvive in the food processing environ-ment, and enhance its ability to cause ill-ness when consumed by humans.

Scientists are studying the responseof cells and spores to multiple stresstreatments in an attempt to develop newand better control systems for specificsituations. Combined treatment meth-ods take advantage of stress-induced mi-crobial weaknesses. For example, pres-sure-damaged E. coli are more sensitiveto acid than native cells (Pagan et al.,

2001). Addition of 5% ethanol enhancesinactivation by organic acids and osmot-ic stress (Barker and Park, 2001). On theother hand, cross protection from micro-bial stress responses must be consideredwhen evaluating treatment effectiveness.Growth under low a

w conditions increas-

es the heat resistance of Salmonella spp.(Mattick et al., 2000). Combined ap-proaches are not new, nor are they limit-ed to vegetative cells. Apparent inactiva-tion and control of heat-damaged C. bot-ulinum spores was considerably less infood that contained lysozyme (Peck andFernandez, 1995).

As scientists continue to improvetheir understanding of microbial stressresponses, it is increasingly possible totry to anticipate potential stress-relatedproblems in the food processing envi-ronment. The food industry is adoptingor considering a variety of methods forsanitizing beef, pork, lamb, and poultrycarcasses and reducing or eliminatingpathogens in meat products (Mermel-stein, 2001). Whether it is steam pas-teurization, rinsing with various anti-microbials, e-beam or x-ray treatment,high pressure processing, or some yet tobe identified procedure, it is essentialthat the stress-induced responses beconsidered.

The same considerations are essentialfor the decontamination procedures usedwith fresh and fresh-cut fruits and vegeta-bles. High pressure processing of orangejuice (Zook et al., 1999), pulsed electricfield treatment of orange-carrot juice(Rodrigo et al., 2001), UVB treatment ofsurface waters (Obiri-Danso et al., 2001),sanitizer treatment of Salmonella attachedto apples (Liao and Sapers, 2000), anddisinfectants killing Alicyclobacillus aci-doterrestris spores prior to pasteurizingfruit juices (Orr and Beuchat, 2000) areonly a few examples of possibilities for thefuture. If sublethal injury were to be ig-nored in these very promising food safetydevelopments, new food safety problemsmay occur. These concerns also apply tomultiple or combination processes thatare intended to benefit from additive orsynergistic effects. For example, pulsedhigh pressure processing with modest ele-vated temperatures appears to inactivatespores with high heat resistance (Meyer etal., 2000).

New Tools for Pathogen Research

New tools are needed to more easilyand specifically monitor populations of

Table 11. Conditions That Can Produce SublethallyInjured Cells (Ray, 1989)

Environmental Stress

Moderate heat

Low temperature

Low water activity

Radiation

Low pH

Preservatives

Sanitizers

Pressure

Electric fields

Nutrient deficiencies

Processing Parameter

PasteurizationConcentration

Refrigeration/chillingFreezing

DehydrationHigh solutes (salt, sugar)

X-raysGamma raysUltraviolet rays

Organic or inorganic acids

SorbateBenzoate

ChlorineQuaternary ammonium compoundsShort-chain fatty acidsPeroxyacetic acid

High hydrostatic pressure

Pulsed electric fields

Very clean surfaces

65EXPERT REPORT

pathogens in the post-harvest stages offood production. Although much of theemphasis in microbiological methodsdevelopment traditionally has been onpathogen detection in the food, it is crit-ical to develop better ways of monitor-ing the food processing environment, sothat the factors that influence food con-tamination may be understood. Thisunderstanding is a necessary step in thedevelopment of rational control strate-gies.

Traditional identification methodsbased on microbiological culture aretime-consuming, prohibiting scientistsfrom accumulating the large amountsof data needed to develop an under-standing of the ecology of pathogens inthe food production environment. Thecommercial market has provided noshortage of rapid method “kits,” whichemploy either immunochemical or nu-cleic acid-based detection technology,but they are expensive and must bevalidated for each particular applica-tion.

Research on biofilms, a persistentproblem in the food plant environment(Wong, 1998), is slow because theunique methods and instrumentationneeded for their study are not widelyavailable (Zottola, 1997), and becausefood microbiologists generally have notbeen trained in these techniques. Bio-films are a growing colony or mass ofbacterial cells, attached to each otherand a surface, that entrap debris, nutri-ents, and other microorganisms (Zotto-la, 1994). To research such issues, foodmicrobiologists must adopt more toolsthat are traditionally the province of themicrobial ecologist.

Although many traditional culturemethods are available and many rapidmethods have been commercialized forscreening/detection of foodbornepathogens (FDA, 1998), the options aremuch more limited if a quantitative de-termination is desired. Sensitive quanti-tative methods are necessary for assess-ing pathogen growth, survival and inac-tivation, as well as for accurate risk as-sessment. Specific quantitative determi-nations are traditionally obtained by themost probable number (MPN) tech-nique, a sample dilution technique forstatistically estimating the number ofmicroorganisms, or if background pop-ulations are not too intrusive, by directplating (FDA, 1998). The MPN tech-nique is the most sensitive, because itprovides enrichment conditions that al-

low for the recovery of many of the in-jured cells (see sublethal injury, p. 63).In direct plating, the sample is placeddirectly on or in selective media, whichmay inhibit the growth of the injuredcells. The MPN technique, however, isextremely labor intensive. Moleculartechniques, such as the coupling of im-munochemical or nucleic acid-basedassays to the MPN, or nucleic acidprobe hybridizations of colony blots (deBlackburn and McCarthy, 2000; FDA,1998; Miwa et al., 1998), are increasing-ly being applied to traditional methodsfor obtaining results more quickly.PCR, originally developed for screen-ing/detection applications, has beenmodified for use as a quantitative assay(Nogva et al., 2000). Recombinant bi-oluminescent strains of food pathogenshave been used to quantitatively assesssurvival in foods and the food process-ing environment by measurement oflight emission (Ramsaran, et al., 1998;Siragusa et al., 1999). These techniquesand others need to be further developed,refined and validated as our need forquantitative information on pathogensincreases.

The development of genetic finger-printing techniques has provided theability to finely discriminate strainswithin a species of pathogen. Severalvariations of fingerprinting techniquesare available (Farber et al., 2001), andnew ones continue to be developed. Ge-netic fingerprinting has been extremelyuseful in epidemiological work andidentification of foodborne illness out-breaks (e.g., PulseNet) (CDC, 1999d).The technology has begun to be appliedto the study of pathogens in the foodplant environment (Norton et al., 2001).This novel approach allows questionsabout pathogen evolution, persistenceand routes of contamination in the foodplant to be examined.

New tools also need to be developedand applied for assessing stress and in-jury of pathogens in the food produc-tion environment. Sublethally injuredcells may have different survival char-acteristics during food storage or afterconsumption, and may show greater re-sistance to control measures andheightened virulence than the originalpopulation (Abee and Wouters, 1999;Bower and Daeschel, 1999; Gahan andHill, 1999). If the injured cells are notdetectable by the method chosen, thesafety of a food process may be as-sumed, when, in fact, a food safety risk

may be present. Conventional methodsfor determining microbiological injurygenerally involve plating on two typesof media, i.e., selective and non-selec-tive; the rationale being that injuredcells cannot survive selective cultureand can grow only on non-selectivemedia, whereas healthy cells can growon both types of media (Ray, 1979).This strategy is somewhat imprecise;with each additional selective agent in-corporated into the medium, greaterpercentages of injury in the populationcan be revealed, indicating the presenceof different subpopulations of cellshaving varying degrees or differenttypes of injury. Molecular methods ofanalysis, in addition to those based onculture, can provide useful insights forassessment of injury and viability ofpathogen cells. Molecular probes ofcellular functions, e.g. membrane per-meability, respiratory electron trans-port, membrane esterase activity, etc.,may be useful for broadly categorizingtypes of cellular injury (Breeuwer andAbee, 2000; McFeters et al., 1995; Porteret al., 1995). Expression assays thatbuild on technologies such as the re-verse transcription (RT)-PCR (Sheri-dan et al., 1998), the nucleic acid se-quence-based amplification (Blais etal., 1997; Simpkins et al., 2000), anduse of reporter gene constructs (Cha etal., 1999; Stewart et al., 1997) may beuseful for studying the expression ofgenes known to be associated withstress. However, because we do notknow what all of these genes may be,the development of new tools, such asgenomic microarrays (Graves, 1999)for food pathogens, is needed. A ge-nomic microarray or “gene chip” couldbe, for example, an ordered set of all ofthe known genes of a particular micro-organism, deposited in precise loca-tions on a small solid surface. A prom-ising use for genomic microarrays is inexpression analysis, in which changesin the pattern of expression of thou-sands of genes can be studied simulta-neously, known as “functional genom-ics” (Oh and Liao, 2000; Tao et al.,1999). Studying the effect of a particu-lar stress on a pathogen, then, no long-er needs be limited to expression of oneor a few stress response genes. Func-tional genomics will becomean invaluable tool for understandingand monitoring stress responses inthe food production and processingenvironment.


Ability of PathogensTo Survive in theEnvironment

The Vaudois University Hospi-tal Medical Center in Lausanne,Switzerland, generally diagnoses amean of three listeriosis cases peryear. However, a cluster of 25 liste-riosis cases (14 adults and 11 ma-ternal/fetal) was observed at thesame medical facility between Janu-ary 1983 and March 1984, with 15additional cases diagnosed at sur-rounding hospitals (Bille, 1988; Ma-linverni et al., 1986).

Epidemiology

This epidemic appeared atypi-cal because of the high number ofhealthy, immunocompetent indi-viduals affected, the high rate ofbrain-stem encephalitis, and amortality rate of 45%. The organ-isms isolated from clinical speci-mens from thirty-eight of the 40patients involved in the outbreakwere serotype 4b. A high number(92%) of the L. monocytogenes se-rotype 4b cultures were of twounique phage types, comparedwith only 44% of the serotype 4bcultures obtained during the pre-vious 6 years. This evidence indi-cated the outbreak might be tracedto a single source. A thoroughcase study did not identify thesource or mode of Listeria trans-mission, however, public health of-ficials initiated a prospective case-control study, assuming that asimilar listeriosis outbreak waslikely the following winter. Over-all, 16 additional cases were identi-fied between November 1984 andApril 1985.

After a 1985 listeriosis outbreakin California was linked to con-sumption of Mexican-style cheese,Swiss officials initiated a baselinestudy to detect Listeria spp. in a va-riety of dairy products. While sur-veying soft, semi-hard, and hardcheeses, L. monocytogenes was iso-

lated from five of 25 surface samples ofVacherin Mont d’Or, a soft smear-ripenedcheese manufactured from October toMarch and consumed primarily in theoutbreak region. Furthermore, all fiveisolates belonged to serotype 4b, andtwo of the phage types isolated from thecheese were identical to most clinicalstrains isolated during the 1983-1986epidemic period.

In 1987, a third case-control studydemonstrated that 31 of 37 individualswho became ill had consumed Vach-erin Mont d’Or cheese, as comparedwith only 20 of 51 people in the con-trol group. Investigators isolated theepidemic strain of L. monocytogenesfrom a piece of Vacherin Mont d’Orcheese that had been partially con-sumed by one of the victims. There-fore, Swiss officials halted productionof the cheese and recalled the productthroughout Switzerland. Between 1983and 1987, a total of 122 cases of listeri-osis resulting in 34 deaths were record-ed in the western part of Switzerland.

Several years following the out-break, the isolates were further typedusing a variety of new methods, andthe clinical and cheese isolates wereidentical. The epidemic strain had thesame phage type, enzyme type, ri-botype and PFGE type as strains iso-lated during the 1985 listeriosis out-break in California.

Management

Immediately following the recall,Swiss officials began investigating howthe cheese could have become contami-nated. The cheese implicated in thisoutbreak was produced at 40 differentfactories located in western Switzerland,and all contaminated cheese was report-edly prepared from Listeria-free bovinemilk. After coagulating the milk, the re-sulting curd was dipped into woodenhoops and allowed to drain for 1-2days. When drained, the cheese wastransported to one of 12 cellars locatedthroughout the area and ripened for 3weeks on wooden shelves, during which

time the cheeses were turned dailyand brushed with salt water. Onceripened, they were packaged andreturned to the cheese factory. Tovalidate suspicions that the con-tamination occurred during ripen-ing, investigators took samplesfrom the wooden shelves, brushesand the surface rind of the cheese.Analysis of these samples revealedfairly high levels of L. monocytoge-nes (10,000 to 1,000,000 bacteria).Further investigation showed thatalmost half of the 12 ripening cel-lars were contaminated with one orboth epidemic strains of L. mono-cytogenes, suggesting cellar-to-cel-lar spread of the pathogen, whichwould explain why cheeses in all 40plants were contaminated. Al-though first detected in 1983, inves-tigators speculated that the outbreakbegan several years earlier, based onevidence that the epidemic strainhad been isolated from a listeriosisvictim in 1977. Thus, this particularstrain of L. monocytogenes had es-tablished itself and survived in thefactories and/or cellars of variousplants for up to 10 years.

All 40 factories and 12 cellarswere thoroughly cleaned and sani-tized. The wood from the ripeningcellars was removed and burned,and the cellars were refitted withmetal shelves. Examination of ex-perimental batches of the cheeseproduced over a 2-month periodindicated the cleanup effort wassuccessful.

It is evident that L. monocytoge-nes can adapt to different plantconditions and persist in the man-ufacturing environment formonths and even years. From theseenvironmental niches/biofilms, theorganism can find its way into fin-ished product, and cause sporadiccases or outbreaks of foodbornelisteriosis. As a part of an overallcontrol strategy, aggressive envi-ronmental and product monitoringis needed.

67EXPERT REPORT

Application of Science toFood Safety Management

Effective application of our current

scientific knowledge is a vital part of

continuing efforts to improve food

safety. Our food safety policies are

based on the best scientific informa-

tion available at the time they were

created, but our knowledge continues

to improve. Flexible, science-based

policies should incorporate new

information as it becomes available.

Everyone benefits from science-based

policies, although factors such as

economic impact and other trade-offs

are part of the policy-making environ-

ment as well. Ideally, food policy will

draw on a variety of science-based tools

and be structured to provide the

flexibility to apply these tools as

science dictates.The final decades of the twentieth

century saw tremendous change in ourknowledge of pathogenic microorgan-isms, their toxins and their metabolites.Scientists identified a wide array of mi-crobial hazards, and foods previouslythought to be safe were found to be im-portant vehicles of foodborne disease.This new knowledge resulted in newpolicies to protect consumers. This sce-nario will continue into the new centu-ry.

Unfortunately, current systems can-not deliver a risk-free food supply. Thescientific knowledge, technology andequipment are not available to eliminateall microbial hazards from all foods.This vulnerability will continue to in-fluence regulatory policy as regulatoryagencies and industry seek solutions tothese hazards. The primary weakness intoday’s control system is the presence ofenteric pathogens on raw agriculturalcommodities (e.g., meat, poultry, milk,eggs, fruits, vegetables, and seafood).

As a result, an essential part of sci-

ence-based policies for food safety andpublic health protection is risk analysis.Risk analysis is generally considered tohave three components: risk assessment,risk management, and risk communica-tion. New scientific tools can provideways to assess risk, enabling decisions tobe based on data and fact. These deci-sions are not easy, but risk assessmentgives us the basis for managing theserisks in an informed, intelligent way. Aneffective food safety system integratesscience and risk analysis at all levels ofthe system, including food safety re-search, information and technologytransfer, and consumer education.

Risk Assessment

Broadly defined, risk assessment isthe use of scientific data to identify, char-acterize, and measure hazards; assess ex-posure; and characterize the risks in-volved with a food. However, risk assess-ments do not specifically determinewhether a product is “safe” or “unsafe.”

In terms of foodborne illness, risk isa function of the probability of an ad-verse health effect and the severity of thateffect. In other words, risk is a measureof the likelihood that illness will occurwithin a population as a result of a haz-ard in food and the severity of that ill-ness (Buchanan, 1997b).

When used in food safety regulationor policy development, risk assessmentreflects the expected impact of a particu-lar food safety problem, the expected im-pact of protective mitigation measures,and the levels of urgency and controver-sy surrounding an issue. Risk assess-ment has long been applied to assessingrisks associated with chemical exposurebut only recently has it been formally ap-plied to foodborne pathogens. There-fore, a specific format for these risk as-sessments is still in the definition pro-cess; currently used formats vary de-pending upon the scope and objective ofthe assessment.

Risk assessments also play an im-portant role in international trade by

ensuring that countries establish foodsafety requirements that are scientifi-cally sound and by providing a meansfor determining equivalent levels ofpublic health protection betweencountries. Without systematic risk as-sessment, countries could set require-ments unrelated to food safety, creat-ing artificial barriers to trade. Recog-nizing the importance of this science-based approach to fair trade, theWorld Trade Organization requireseach country’s food safety measures tobe based on risk assessment. The Co-dex Alimentarius Commission (Co-dex), which establishes internationalfood safety standards, has developedprinciples and guidelines for conduct-ing risk assessments (CAC, 1999).

Regardless of the specific format, arisk assessment has four main compo-nents:• Hazard identification involves identi-fying the hazard (e.g., pathogen), the na-ture of the hazard, known or potentialhealth effects associated with the hazard,and the individuals at risk from the haz-ard.• Exposure assessment describes theexposure pathways and considers thelikely frequency and level of intake offood contaminated with the hazard.• Hazard characterization explores therelationship between the exposure leveland the nature of the adverse effects,considering both frequency and severity.• Risk characterization identifies thelikelihood that a population of individu-als would experience an adverse healthoutcome from exposure to the food thatmight contain the pathogen. The riskcharacterization also describes the vari-ability and uncertainty of the risk andidentifies data gaps in the assessment.

These same four components areconsidered in both quantitative andqualitative risk assessments. Qualitativerisk assessments may be chosen to iden-tify, describe, and rank hazards associat-ed with a food. Quantitative risk assess-ments may be chosen when substantivescientific data are available for analysis,


and these risk assessments almost alwaysyield a numerical expression of risk.

Qualitative Risk Assessment

Qualitative risk assessment is gener-ally considered a valuable method to de-termine which hazards are associatedwith a particular food. This process is of-ten used by an expert panel. Also, quali-tative assessments are useful when manygaps in the available data limit the preci-sion necessary for a quantitative risk as-sessment. For instance, scientists oftenhave little exact information about therelationship between the quantity ofpathogen ingested and resulting frequen-cy and severity of adverse health effects,especially for susceptible subpopula-tions. Information about exposure—theprobability of contamination, the extentof pathogen growth in the food, and theamount of the food consumed by vari-ous populations—is sometimes limited.Qualitative risk assessments can be use-ful in identifying these data gaps and intargeting or prioritizing research thatwould have the greatest public health im-pact.

Qualitative risk assessments have nu-merous applications in food safety anal-ysis. These assessments can help compa-nies develop more effective Hazard Anal-ysis Critical Control Points (HACCP)plans based on scientific data. For in-stance, a qualitative risk assessment canhelp identify likely hazards, although theresult might only be to rank a hazard ashigh, medium, or low in terms of preva-lence or potential contamination level.The assessment will have an increasedpower to inform decision-makers whenthe risk of an adverse health effect canalso be described, at least qualitatively ashigh, medium or low.

A significant disadvantage of qualita-tive risk assessments is the inability tocompare the extent to which particularmitigating factors or risk managementoptions can successfully reduce risk. Aqualitative risk assessment can comparetwo options that both address the expo-sure to a hazard—e.g., two different dis-infecting agents—or it can compare twooptions that address the effect—e.g., twomethods to prevent susceptible individu-als from consuming the food. However,qualitative risk assessments present diffi-culties when trying to compare one op-tion that addresses only exposure andanother option that addresses only effect.

Notwithstanding the disadvantages

of qualitative risk assessments, they arefrequently used and serve a purpose inscience-based food safety management.Qualitative risk assessment will continueto play a role in the future when timeand money constraints prohibit a fullquantitative risk assessment. However,scientists must develop a better under-standing of the exact role that qualitativerisk assessments can play and a betteroverall defined structure for these riskassessments.

Quantitative Risk Assessment

Quantitative risk assessment is for-mally defined as the technical assessmentof the nature and magnitude of a riskcaused by a hazard. The technique, firstdeveloped in the 1950s to evaluate nuclearproliferation risks, has since been used toevaluate the toxicological risks to plants,animals, and public health posed bychemical exposure and more recently inmicrobiological food safety evaluation.

As the name implies, quantitative riskassessment ultimately provides numericalestimates of risk that can be used in regu-latory decision-making and risk manage-ment. Quantitative microbial risk assess-ment as a discipline has been under scru-tiny recently, purportedly because of alack of rigor and precision. Because theprocess inherently depends upon the in-put of many scientists from frequently di-verse disciplines, as well as the incorpora-tion of assumptions when definitive dataare lacking, some individuals believe thatthe process is little more than an academicexercise. Of all the limitations in the cur-rent risk assessment methodology(Jaykus, 1996), the need for more and bet-ter data is the most pressing.

After the hazard identification phaseis complete, the exposure assessment es-timates or directly measures the quanti-ties or concentrations of risk agents re-ceived by individuals or populations.As such, exposure assessment is de-signed to characterize the circumstanc-es, source, magnitude, and duration ofexposure, the final goal of which is toproduce a mathematical expression forexposure, generally in the form of theprobability of ingestion of the infec-tious agent through the food vehicle ofinterest. Common data sources includesurvey information on the prevalenceand levels of contamination for a par-ticular pathogen in a particular foodcommodity. Scenario analysis usingvarious computer software packages is

frequently done to model the many cir-cumstances that may surround expo-sure. And since foodborne pathogengrowth, inactivation and survival arehighly dependent upon intrinsic (food-related) and extrinsic (environmental)factors such as relative humidity andstorage temperature, mathematicalmodels such as USDA’s Pathogen Mod-eling Program can be used to estimatethe effect of these factors on microbialpersistence and levels in foods.

The relationship between the inges-tion of pathogenic microorganisms andpossible health effects may be describedas the quantitative relationship betweenthe intensity of exposure (dose) and thefrequency of the occurrence of illnesswithin the exposed population of hosts(response). For pathogenic microor-ganisms, this is dependent upon thenumber of units of infectious agent in-gested in the food, the infectivity andpathogenicity of the infectious agent,and the vulnerability of the host. Thepurpose of the hazard characterizationstep is to quantify or statistically de-scribe the relationship between the riskagent and the magnitude of the adverseeffect. Included in this step is a full de-scription of the severity and duration ofadverse effects that may result from theingestion of a microorganism or toxinin a food. The source of data used forhazard characterization is usually hu-man challenge studies, whereby a de-fined population of consenting adults isgiven various doses of the infectiousagent, and their response is measured asinfection or disease. For pathogenscausing mild disease, this is feasible; forthose associated with more severe dis-eases, or affecting particularly suscepti-ble populations, risk assessors must relyon animal models of disease or otherdata sources. The raw input data fromthese types of studies is used in con-junction with statistical methods thatdescribe the dose-response relationshipin mathematical terms. Some frequent-ly used dose-response models includethe Beta-Poisson, Exponential, andGamma-Weibull models.

Risk characterization is the final stepof risk assessment and represents the inte-gration of the exposure assessment andhazard characterization to obtain a riskestimate of the likelihood and severity ofthe adverse effects that would occur in agiven population. The final risk estimateshould incorporate information about thevariability, uncertainty and assumptions

69EXPERT REPORT

identified in all previous steps of the riskassessment. Statistical methods are usedto characterize variability associated witha well-characterized phenomenon, whileMonte Carlo and Bayesian approachescan be applied in an effort to better char-acterize uncertainty associated with apoorly characterized phenomenon. Sen-sitivity analysis is frequently done to bet-ter understand the contribution of the in-dividual factors that influence the overallrisk estimate. Because the risk assessmentprocess usually involves extensive data in-put, the use of various mathematicalmodeling approaches, and some degree ofassumption on the part of the assessors,risk assessment teams now seek to pro-vide “transparent” documents that givethe reader a full and detailed descriptionof the process.

At present, the first reports of micro-bial risk assessment in food safety haveappeared in the scientific literature, andseveral regulatory agency-driven risk as-sessments have been completed in theUnited States during the last 5 years: theUSDA/FDA Salmonella Enteritidis riskassessment for shell eggs and egg prod-ucts (USDA/FSIS, 1998a); the USDA/FSIS risk assessment for Escherichia coliO157:H7 in beef and ground beef(USDA/FSIS, 1998b); the FDA draft as-sessment of the relative risk to publichealth from foodborne Listeria monocy-togenes among selected categories ofready-to-eat (RTE) foods (FDA/CFSAN,USDA/FSIS, and CDC, 2001); and theFDA draft assessment on the publichealth impact of Vibrio parahaemolyticusin raw molluscan shellfish (FDA/CF-SAN, 2001). Other risk assessments havebeen completed by the Canadian FoodInspection Agency. Despite these verysignificant efforts, the application ofquantitative risk assessment techniquesto microbiological foodborne hazards isstill in its relative infancy.

Quantitative risk assessment pro-vides a formal, conceptual framework forthe evaluation of foodborne diseaserisks, one that effectively uses all avail-able information and expertise. Prior tothe advent of this discipline, decisionsabout food safety regulation and man-agement were much less systematic. Fur-thermore, recent risk assessments pro-vide a clear picture of the role of uncer-tainty in overall risk modeling, usingsome of the more robust methods, suchas Monte Carlo simulation, to character-ize uncertainty. Finally, risk assessmentis, by nature, an iterative process; if the

models are properly constructed, theycan be readily updated as additional in-formation becomes available.

Role of Expert Panels

The process used by expert panels isvery similar to that for qualitative risk as-sessment. The major difference is that ex-pert panels, in general, collect and reviewavailable information and develop guid-ance or recommendations for use bygovernment and/or industry. Thus, spe-cific risk management options are an ex-pected outcome from the expert panelprocess. In addition, the time frame foran expert panel would normally be indays or perhaps several months, the ex-tended time being used to prepare the re-port of the panel’s deliberations and rec-ommendations.

Expert panels can effectively gatherinformation, evaluate available data,and develop recommendations in a rel-atively short period of time. Expert pan-els can be used to address a variety ofcircumstances: when a rapid decision isneeded to address a newly recognizedconcern, when resources and/or datafor a quantitative risk assessment arelimited, or when few management op-tions exist. When epidemiological evi-dence indicates a hazard is not undercontrol, panels may identify ways to in-crease consumer protection. Also, pan-els may address concerns raised bychanges in food processing technolo-gies, food packaging, or distributionsystems. Expert panels can addresssuch situations and evaluate the risk,based on scientific evidence.

Government and international bod-ies have made extensive use of expertpanels (e.g., Food and Agriculture Orga-nization of the United Nations (FAO)/World Health Organization (WHO)consultations) to address concernsabout the safety of a particular hazard-food combination. In addition, industryexpert panels have considered the fac-tors leading to foodborne disease anddeveloped recommendations for theircontrol (e.g., the Blue Ribbon TaskForce established by the National Live-stock and Meat Board to address E. coliO157:H7) (NLSMB, 1994).

Expert panels rely on epidemiolo-gists, public health specialists, food mi-crobiologists, food technologists andothers with knowledge about the food-borne disease or the conditions of foodproduction, processing, distribution and

use. Often, the panel uses a process thatis very similar to a HACCP hazard analy-sis. In some instances, a rough estima-tion of the risks associated with differentlikely scenarios is sufficient. One ap-proach is to assign relative probabilityand impact rankings—such as negligible,low, medium, or high—to the likelihoodof exposure and adverse outcome. Thepanel must clearly define and justify therankings to enable people to use the finalresult without misinterpretation.

An expert panel considers variousaspects of the issue and provides its bestjudgment based on information avail-able at that time. Because the process issimpler, it is also faster than a quantita-tive risk assessment. Although the com-plexity of the risk evaluation may vary,the panel follows the standard four-stepformat for a risk assessment and shouldprovide information on the conditionsthat lead to hazardous food. In the end,the panel may recommend one or moremeasures to control a hazard or, if neces-sary, the expert panel may recommendbanning the product or process. An ex-pert panel also may recommend estab-lishing a Food Safety Objective when itwould be an effective means to enhancethe safety of the food under consider-ation.

Risk Management Using FoodSafety Objectives

A recently proposed risk manage-ment approach revolves around the con-cept of Food Safety Objectives (FSOs).As defined by the International Commis-sion on Microbiological Specificationsfor Foods (ICMSF, 1997, 2002), a foodsafety objective (FSO) is a statement ofthe maximum frequency and/or concen-tration of a microbiological hazard in afood at the time of consumption thatprovides the appropriate level of protec-tion. The FSO approach can be used tointegrate risk assessment and currenthazard management practices into aframework that can be used to achievepublic health goals in a science-based,flexible manner. Fig. 7 demonstrates howthe FSO concept (boxes with roundedcorners) can integrate with HACCP(shaded boxes).

Although the FSO concept is relative-ly new and is still evolving, its acceptanceis growing because it offers a practicalmeans to convert public health goals intovalues or targets that can be used by reg-ulatory agencies and industry. For exam-


Public health concern identified

Hazard(s) identified

Exposure assessed

Hazard(s) characterized

Risk characterized

Agricultural interventions implemented

GMPs/GHPs implemented

Baseline level of microbial hazard measured

Processing safety objective calculated(FSO minus growth during storage and distribution)

Performance criteria established(baseline minus processing safety objective)

Critical control points identified

Process/product criteria established

Critical limits established

HACCP system verified

Recordkeeping implemented

Pathogen growth prevented or minimized

Monitoring procedures developed and implemented

Food Safety Objective met

Risk Assessment

Risk Management

Hazard Control and Monitoring

Food Consumption

Food Distribution and Storage

Microbiological criteria established (if necessary)

Processing safety objective met

Corrective actions implemented (as needed)

Consider options, including a Food Safety Objective

Fig. 7. Framework forFood Safety Management

HACCPApproach

FSO Approach

Method(s) of control selected Process-specific hazard analysis conducted

71EXPERT REPORT

ple, a public health goal may be to reducethe incidence of foodborne illness attrib-uted to pathogen A by 50% from 20 to 10cases per 100,000 people per year. A reg-ulatory agency or manufacturer cannotdesign control systems that would becertain to meet such a goal. However, ifthis goal were translated into a numeri-cal measure of the microbiological haz-ard’s frequency or concentration (e.g.,less than 100 CFU/g of L. monocytoge-nes or less than 15 mg/kg of aflatoxin),then the regulatory agency could estab-lish inspection procedures and industrycould design control processes based onthe FSO.

The FSO concept can be a usefultool for creating policies that are consis-tent with current science. The limits im-posed in an FSO should reflect not onlythe best available scientific informationbut also nonscientific input from a vari-ety of sources. FSOs should reflect soci-etal values with regard to levels of con-sumer protection. The FSO develop-ment process should be transparent andfacilitate input, both scientific and soci-etal, from all affected parties.

The FSO approach integrates scien-tific data from risk assessment to setquantifiable standards that address spe-cific public health outcomes. Becausethe FSO must be met at the time of con-sumption, it is necessary to consider thepotential for pathogen growth duringstorage and distribution. The processingsafety objective is the FSO minus anyprojected pathogen growth. For exam-ple, if the FSO is less than 100 CFU/g ofL. monocytogenes and 1 log cycle ofgrowth is projected, the processing safe-ty objective is calculated as no morethan 10 CFU/g of L. monocytogenes. Ifno pathogen growth is projected, theprocessing safety objective is the sameas the FSO.

The processing safety objective isthen used to develop the performanceand process/product criteria and to es-tablish verification and acceptanceprocedures. Good hygienic practices(GHPs) and good manufacturing prac-tices (GMPs) are important to mini-mize the hazard and prevent recontam-ination after processing. HACCP man-ages the application of control meth-ods, ensuring that the process is effec-tive. Table 12 provides three examplesof how the FSO approach might beused to address specific issues of mi-crobiological food safety. Regulatoryagencies can use FSOs and processing

safety objectives to communicate thelevel of control expected in food pro-cesses and then to evaluate the adequa-cy of a facility’s control system.

FSOs differ from the microbiologi-cal criteria that have been traditionallyused to determine the acceptance offood products. Microbiological criteriaspecify details such as a sampling planand the method of sample preparationand analysis, but the criteria cannot bereadily used to evaluate a process. Mi-crobiological testing of finished productfrom a plant provides a snapshot for thetime the food was produced. Review ofthe same plant’s food safety manage-ment system using an FSO approachwould provide a more meaningful as-sessment of long-term control.

FSOs can be used to communicatefood safety requirements for food pro-cesses; whereas microbiological criteriaare used to determine the acceptabilityof specific lots of food with respect toquality and/or safety. The principles forthe establishment of microbiologicalcriteria for food have been described byCodex (CAC, 1997a) and have been re-cently further elaborated upon by theICMSF (ICMSF, 2002). For example, thefollowing components are recommend-ed when microbiological criteria are tobe established:• a statement of the microorganismsof concern and/or their toxins/metabo-lites and the reason for that concern;• microbiological limits consideredappropriate to the food at the specifiedpoint(s) of the food chain,• the number of analytical units thatshould conform to these limits;• a sampling plan defining the numberof field samples to be taken, the methodof sampling and handling, and the size ofthe analytical unit; and• the analytical methods for their de-tection and/or quantification.

In addition, a microbiological criteri-on should also state:• the food to which the criterion ap-plies;• the point(s) in the food chain wherethe criterion applies; and• any actions to be taken when the cri-terion is not met.

When used for food safety, microbi-ological criteria should reflect the sever-ity of the disease and whether risk islikely to decrease, remain the same, orincrease between when a food is sam-pled and when it is consumed. As crite-ria are established to address newly

emerging food safety concerns, theremay be little information available. It isthe responsibility of regulatory agenciesto establish criteria that can be used toassess the safety of foods. Preferably, thecriteria should be considered interimstandards that will be adjusted to bemore or less stringent as more informa-tion becomes available.

The FSO approach assumes that afood is distributed, stored, and preparedas intended and expected when the foodsafety system was designed. Deviations inhandling and storage after the foodmeets the processing safety objective atthe factory could trigger a failure to meetthe FSO. Proper food handling andpreparation practices are essential underthe FSO approach.


Once an FSO has translated publichealth goals into quantifiable limits, haz-ard control and monitoring practicesmust be developed. Although hazardcontrols were developed long before themore recent formal risk assessment andrisk management approaches, the appli-cation of hazard controls can be directedand informed by the broader perspectivethat results from an integrated food safe-ty framework. Mandatory hazard con-trol processes have traditionally been thefocus of regulatory agencies, but recentinitiatives to develop risk-based ap-proaches offer the opportunity for flexi-ble, science-based hazard control.

Often, many different approachesare combined to achieve the desired re-sult. In the farm-to-table approach tofood safety, good agricultural practices(GAPs) can provide ingredients withimproved microbiological safety. GMPs,also known as GHPs in the internation-al arena, set basic standards for facilitysanitation and hazard control. Perfor-mance criteria quantify the hazard con-trol results necessary to meet the pro-cessing safety objective, and process/product criteria define the process vari-ables and product characteristics thatwill achieve the performance criteria.HACCP establishes the critical controlpoints at which the process/product cri-teria are applied, further defining theconditions that must be met into a spe-cific set of critical limits for the process.HACCP also monitors and documentssuccessful implementation of the con-trol process. Finally, microbiologicalcriteria and testing may be used, if nec-


Managing Microbiological Food Safety

Public health concern(often based on epidemiological data)

Risk Assessment

Hazard identification

Exposure assessment

Hazard characterization

Risk characterization

Risk Management

Food safety objective


GMPs

Performance criteria

Hazard analysis, method(s) of control, and CCPs

Process/product criteria

HACCP implementation and verification

Microbiological criteria (if necessary)

Salmonella in Dried Milk

Consumed as reconstituted milk, particularly by children

Salmonella are heat-sensitive bacteria that are a leading cause ofdiarrheal disease worldwide, especially among the very young andelderly

Post-processing contamination in the factory rarely occurs, and theconcentration of Salmonella in dried milk is low

Considering reconstituted dried milk is often consumed by children,assume worst case scenario: one Salmonella cell per 10 g serving ofmilk may cause illness

Based on limited data, probability of illness is <1 in 108 servings

To maintain the current estimated level of risk, FSO may be establishedas <1 Salmonella per 108 gram at the time the dried milk is reconsti-tuted for consumption

GMP to control recontamination after processing

Processing to achieve at least 8-log reduction for salmonella

Pasteurization

Time and temperature specifications forpasteurization

Monitor environment after pasteurizer by testing for indicator organismsand Salmonella to document effectiveness of GMPs

Testing product is not recommended

Table 12. FSOs in the Food Safety Management Framework (Derived from ICMSF, 2002)

essary, to further verify that the process-ing safety objective has been met.

Good Agricultural Practices

It is impossible to guarantee thatcrops will be free of all harmful microbio-logical contamination because disease-causing organisms have too many oppor-tunities to enter the food system throughthe production sector. Nonetheless, it ispossible to minimize the food safety risksand take preventive steps to protect pro-

duce and other agricultural commoditiesat the farm level. The Guide to MinimizeMicrobial Food Safety Hazards for FreshFruits and Vegetables, published by FDA’sCenter for Food Safety and Applied Nu-trition (FDA/CFSAN, 1998), lists basicprinciples to guide farmers and farmworkers on pre-harvest safety.

Management and control of manurehas become a critical issue in GAPs.Properly treated manure can be an effec-tive and safe fertilizer, but untreated, im-properly treated, or recontaminated ma-

nure can contain pathogens that canreach fresh produce in the field or nearbywater supplies. Evidence indicates thatsome pathogens can survive for extendedperiods of time in the soil and on pro-duce (Rangarajan et al., 2000). Manureshould be composted to effectively elimi-nate pathogens and applied appropriate-ly to minimize the possibility of patho-gen survival and subsequent crop con-tamination.

Irrigation water also is a key factor.The source of irrigation water should be

73EXPERT REPORT

L. monocytogenes in Ready-To-Eat Meats

Listeriosis has been associated with certainready-to-eat (RTE) meats, including hot dogs

Listeriosis is a rare but serious disease affecting immunocompromisedindividuals and the developing fetus

L. monocytogenes is a frequent contaminantin certain RTE foods (e.g., up to 5%)

The majority of cases appear to be associated with dose levels in excessof 104 CFU/serving of RTE meats (FAO/WHO, 2001)

Estimated median number of cases per serving: 3 x 10-6 for perinatalpopulations; 5 x 10-8 for elderly populations; and 5.9 x 10-9 forintermediate populations (FDA/CFSAN, USDA/FSIS, CDC, 2001)

Based on epidemiologic data. FSO set at no more than 100 CFU/g of L.monocytogenes in RTE meats when consumed

GMP to control recontamination after processing

Processing must result in 6-log reduction of L. monocytogenes

Thermal processing

Time and temperature for cooking

Emphasis on environmental testing to verify sanitation. Consider optionsto control pathogen growth if food should become contaminated.

Product testing is of little value in controlled environments but should beconsidered when control is uncertain and pathogen growth can occur inthe product

Escherichia coli O157:H7 in Ground Beef Patties

Outbreaks of illness associated with undercooked ground beef

E. coli O157:H7 infections can result in moderate to severe disease ordeath; children under 5 years and the elderly are the most sensitivepopulations

E. coli O157:H7 is frequently present on the hide and in the intestines ofcattle. The occurrence of E. coli O157:H7 in ground beef is estimated as<1% in the U.S.

Fewer than 100 cells can cause disease, especially among youngchildren

One estimate found 26 x 104 patties per year nationwide may containviable E. coli O157:H7 after cooking

To achieve a 25% reduction in the number of illnesses, the FSO may bea concentration of E. coli O157:H7 in ground beef of no more than 1/250g (equivalent to 1 cell per two 125g patties)

GMP to minimize contamination during slaughter and processing

Performance criteria cannot be specified at this time

Moist heat and/or acid sprays

Parameters for sanitation and control measures (prevention of contami-nation during slaughter and decontamination) may be defined

Sanitation and control parameters are monitored

Testing of raw materials may enable plants to select suppliers withdesired microbial quality. Lot testing may be conducted to identify highprevalence lots, but lots that test negative cannot be considered free ofthe pathogen or “safe”

known, and periodic testing may be ap-propriate.

For pathogen control, GAPs includerecommended practices prior to planting(Rangarajan et al., 2000). Where possi-ble, the field should be upstream of thefarm’s animal housings, and plansshould be put in place to prevent anyrunoff or drift from animal operationsfrom entering the field. Grazing livestockshould be located away from producefields, and traffic of wild and domesticanimals in produce fields should be min-

imized wherever possible.The hygiene of field workers should

be maintained, monitored and enforced.Employees should have clean restroomswith access to soap, clean water and sin-gle-use towels. All employees should beproperly trained to follow good hygienicpractices (FDA/CFSAN, 1998).

Good Manufacturing Practices

Current GMPs—the conditions nec-essary for each segment of the food in-

dustry to protect food while under itscontrol—are well defined and estab-lished in post-harvest food processing.These conditions and practices providethe basic environmental and operatingconditions that are necessary for the pro-duction of safe, wholesome food. Theseconditions and practices, many of whichare specified in federal, state, and localregulations and guidelines, are now con-sidered to be prerequisite to the develop-ment and implementation of effectiveHACCP plans (NACMCF, 1998). GMPs


cover sanitation issues, such as equip-ment design and cleaning, and pest con-trol.

Performance Criteria

FSOs are met using performancecriteria, which are the required out-come of a control step or a combina-tion of steps. At certain points in foodprocessing, control measures can beapplied to either prevent an unaccept-able increase in a microbiological haz-ard or reduce the hazard to an accept-able level. Chilling cooked meats andstews prevents the growth of Clostridi-um perfringens (i.e., increase of a haz-ard), and pasteurization of milk or fruitjuices eliminates enteric pathogens (i.e.,decrease of a hazard). The perfor-mance criteria are the reduction neces-sary (e.g., a 5-log reduction) to achievethe processing safety objective; they arecalculated from the baseline level of themicrobial hazard (see Fig. 8). Perfor-mance criteria also may address theprevention of pathogen growth (e.g.,less than 1-log growth).

Under the FSO approach, it is im-portant not to confuse performancecriteria and performance standards.

Some current food safety regulations(i.e., performance standards) mandatespecific pathogen reductions as a resultof processing, but this approach wouldnot ensure compliance with an FSO.For example, under the current system,a performance standard may require a5-log reduction in pathogen levels for araw agricultural commodity (e.g., freshjuice) (see Fig. 9). Although a food pro-cessor could design a system to achievethe required reduction, a higher base-line level of pathogens could result inhigher pathogen levels after processing.Under the FSO approach, the processorwould know the level of hazard that isallowed in the final product and wouldcalculate the performance criteria basedon the initial number of pathogens.

Process and Product Criteria

Performance criteria are imple-mented through application of processand/or product criteria, which are thevariables in the control process or thecharacteristics of the product thatachieve the necessary reduction or limitpathogen growth. A process criterioncould be the time and temperature of athermal process; a product criterion

could be a pH value. Process and prod-uct criteria may be used alone or incombination. Control of spores ofClostridium botulinum could be accom-plished with a process criterion of heat-ing low acid foods for a specified timeat a specified temperature, or with aproduct criterion of reducing the pHbelow a specified level for acidic foods.Often, more than one combination ofcriteria will meet the processing safetyobjective. The specific values for a par-ticular process are established as criti-cal control limits through the HACCPprocess (see below).

HACCP

HACCP is a management tool usedby the food industry to enhance foodsafety by implementing preventive mea-sures at certain steps of a process.When HACCP principles are properlyimplemented, microbiological hazardsthat have the potential to cause food-borne illness are controlled, i.e., pre-vented, eliminated or reduced to an ac-ceptable level.

The pathway to HACCP began in1959 because existing quality controltechniques could not provide the de-sired level of safety for food producedfor the space foods program (Bauman,1992). Traditional microbiological test-ing of finished product was impracticaland ineffective because of the smallquantities of food produced, and thehigh product cost limited the amountavailable for sampling. In addition, thefood industry had no uniform approachto managing food safety. HACCP wasdeveloped to meet this need.

Over the years, the fundamentalconcepts that comprise a HACCP pro-gram have been refined, and their appli-cation has become more practical. TheU.S. National Advisory Committee onMicrobiological Criteria for Foods(NACMCF)—a federal advisory com-mittee assembled to provide impartial,scientific advice to federal food safetyagencies for use in policy develop-ment—has revised and expanded theoriginal principles based on industryexperience (NACMCF, 1992; 1998). Thecommittee consists of experts in micro-biology, risk assessment, epidemiology,public health, food science, and otherrelevant disciplines. At the same time asthe NACMCF activities, Codex adopteda similar HACCP document (CAC,1997c). NACMCF adopted seven

Fig. 8. Establishing Performance Criteria

Foodprocessing

facility

Baseline

Processingsafety objective

106

101

5 log reduction = performance criteria

Foodprocessing

facility

Baseline


105

101

4 log reduction = performance criteria

106

Agriculturalintervention

1 log reduction affects facility baseline

Foodprocessing

facility

Baseline


101

101

no reduction required;performance criteria = limit/prevent growth

and/or contamination

75EXPERT REPORT

HACCP principles:Principle 1: Conduct a hazard

analysis.Principle 2: Determine the critical

control points.Principle 3: Establish critical limits.Principle 4: Establish monitoring

procedures.Principle 5: Establish corrective

actions.Principle 6: Establish verification

procedures.Principle 7: Establish record-

keeping and documen-tation procedures.

The fundamental HACCP principles,associated definitions, and principles ofapplication were intended to help the foodindustry implement food safety manage-ment systems. First, companies identifyhazards that have the potential to causeillness or injury to the consumer and de-termine whether or not these hazards aresignificant. For significant hazards, thecompany identifies critical control points(CCPs) in the food system, where possi-ble, and establishes critical limits for theircontrol. Critical limits are based on evi-dence that the control is effective in prac-tice. Next, CCPs must be monitored toassure critical limits are not violated. Thismonitoring must be a systematic proce-dure that verifies that the HACCP systemis functioning as intended and that it is ef-fective. Finally, the company must main-tain comprehensive records associatedwith the HACCP system.

HACCP is an essential part of imple-menting an FSO. Through the identifica-tion of CCPs, the process and/or productcriteria can be translated into process-specific critical limits. Monitoring ofthese critical limits ensures that the per-formance criteria are met.

Using the principles identified in theNACMCF and Codex documents as astandard, HACCP was widely adopted by

the food industry in the United Statesand throughout the world beginning inthe 1990s and continuing to the present.While many companies voluntarilyadopted HACCP as part of their respon-sibility to provide safe food, it also be-came mandatory for some U.S. compa-nies under new regulations. In 1995, theFood and Drug Administration (FDA)published a rule that required all seafoodprocessors and, in effect, those compa-nies exporting to the United States to de-velop and implement HACCP systems byDec. 18, 1997. In July 1996, USDA pub-lished a final rule that required all meatand poultry processors to implement aHACCP system. On Jan. 19, 2001, FDApublished a rule that requires large juiceprocessors to implement HACCP by2002, with later compliance dates forsmaller companies. These regulationsare playing a major role in HACCPadoption in the food industry.

Successful HACCP implementationhas not been limited to food processing.HACCP has been used worldwide to im-prove food safety in food service and re-tail, and food distribution.

The application of HACCP to pro-duction agriculture, however, is limited.Applying the HACCP principles revealsfew, if any, CCPs in the production ofraw agriculture commodities. A CCP is apoint in the process where control can beapplied and where the control is essentialto prevent or eliminate a food safety haz-ard or to reduce it to an acceptable level(NACMCF, 1998). The difficulty in pro-duction agriculture is two-fold: identify-ing the source of a hazard and finding aneffective control. Additional research isneeded to develop methods that effective-ly control hazards in the production ag-riculture environment. Currently, haz-ards in this part of the food system aremost effectively controlled by GAPs, suchas those recommended by FDA for fresh

fruits and vegetables (see GAPs, p. 72).For HACCP to be successful, it is

critically important that regulatory pol-icy be based on the best science current-ly available. In an effort to ensure dili-gent HACCP implementation, regulato-ry agencies may constrain the processand ultimately undermine HACCP’sscientific basis. Although it is an ex-tremely useful hazard management tool,HACCP is not appropriate for all situa-tions. Regulatory policies must allowthe flexibility to apply science in a prod-uct- and process-specific manner thatbest achieves the FSO. Some currentpolicies mandate the development of aHACCP plan, even when a scientificanalysis fails to identify a point in theprocess that meets the CCP criteria; it isimpossible to have a valid HACCP planwithout a critical control point. In addi-tion, certain policies generally prescribethe CCPs, without regard to the particu-lar circumstances at a given food manu-facturing facility. As a result, the foodindustry may pursue mere regulatorycompliance, following the form ofHACCP without its proper substance.HACCP is a science-based hazard man-agement tool, not purely an administra-tive system. When critics claim thatHACCP has failed, the failure is often inapplying science and the HACCP prin-ciples to managing food safety.

Microbiological Criteria and Testing

Routine microbiological testing canbe useful in certain applications. It canbe helpful for surveillance purposes andprocess verification, and it is sometimeshelpful for lot acceptance. However, mi-crobiological testing of finished productcan be misleading, and negative test re-sults do not ensure safety. Statistical lim-itations of microbiological testing are ofsignificant concern, especially when therate of contamination is very low. As thedefect rate in the product becomes low,emphasis should shift to improving theimplementation of food safety manage-ment strategies, such as HACCP, ratherthan relying on microbiological testing.

Statistical Limitations to Testing

To analyze food for microbiologicalagents, a sample of the product must betaken. Ideally, the sample or samples tak-en from a production lot of food will insome way reflect the whole of the lot, theunderlying concept in statistically-based

Foodprocessing

facility

Baseline

Pathogenlevel

108

103

5 log reductionFood

processingfacility

Baseline

Pathogenlevel

106

101

108

Agriculturalintervention

Fig. 9. Unequal Levels of Food Safety


sampling plans. So-called “lot accep-tance sampling plans” are in widespreaduse around the world to determinewhether or not a food product meets acertain set of specifications. These speci-fications can target maximum numbersof bacteria per unit size that are set by apurchaser of the food or by govern-ments. When applied to determiningfood quality, lot acceptance samplingplans in conjunction with 2- or 3-classsampling plans (ICMSF, 1986; 2002) areof some value.

The acceptance of a “quality-only de-fect” that will occur occasionally no mat-ter how rigid the sampling scheme isquite different from a situation in whichconsumer health is at risk. When sam-pling food for pathogens, sampling hassufficient inherent limitations to be ren-dered misleading. The following aresome examples of the possible conse-quences of rigorous sampling planswhen applied to making accept/reject de-cisions for safety reasons.

Most microbiological sampling plansinvolve anywhere from a single sample toas many as 60 samples per lot. Table 13indicates that:• If 10 samples are collected fromacross a lot of food that has a defect rateof 1%, there is a 90% probability that thedefect will not be detected and the lot willbe accepted.• If 300 samples are collected fromacross a lot of food that has a defect rate

of 1%, there is a 5% probability that thedefect will be not be detected and the lotwill be accepted.• If 10 samples are collected fromacross a lot of food that has a defect rateof 10%, there is a 35% probability thatthe defect will not be detected and the lotwill be accepted.

The implications of the table be-come apparent when the prevalence ofcontamination for various foods is con-sidered. For example, during the years1998-2000, the U.S. Department of Agri-culture (USDA) monitoring programfor salmonella in raw meat and poultrydetected a range of prevalence in vari-ous commodities (see Table 14). Con-tamination in some foods is much morelikely to be detected by sampling whenthe prevalence of the pathogen is high,as compared to foods with a lower de-fect rate. However, the prevalence ofcontamination for many foods is morelikely to be at the lower end of the scale,particularly in the case of ready-to-eat(RTE) foods. For example, the preva-lence rate for L. monocytogenes in RTEfoods during 1994-1998 was reported tobe from 1.08% to 4.91% (FDA/CFSAN,USDA/FSIS, CDC, 2001), and the prev-alence rate for L. monocytogenes in mostcategories of ready-to-eat meat andpoultry products was below 5% (Levineet al., 2001).

The above examples demonstratethat microbiological testing can have

utility for detecting pathogens in certaintypes of food; however, as the prevalencerates decrease, testing becomes less reli-able for detecting contaminated lots,even with large numbers of samples. Formany foods there is a favorable historyof safety, and testing for pathogens is notroutinely done. Thus, as more effectivecontrol measures are adopted by indus-try and the prevalence of contaminationdecreases, a point is reached where prod-uct testing is no longer practical or justi-fiable. At that stage, greater benefit can beachieved by shifting verification proce-dures to comprehensive analysis of con-trol systems that have been validated tocontrol the pathogens of concern.

The effectiveness of a sampling plan isinfluenced by a number of factors such aswhether random samples can be collectedfrom a lot of food, how samples are pre-pared to obtain analytical units, and thesensitivity and reliability of the analyticalmethod. Sensitive analytical methods donot exist for many of the pathogens re-sponsible for foodborne illness. This in-cludes the viruses that have been estimat-ed to be responsible for more than 50% ofall U.S. foodborne illness caused byknown pathogens (Mead et al., 1999).

Lot acceptance sampling plans as-sume the microbial population is ran-domly distributed throughout each lot offood that is to be sampled. In reality, thisis often not the case, particularly forfoods that are not liquids. Nonrandomdistribution of pathogens is a major con-tributing factor to the unreliability ofproduct testing to prevent contaminatedfood from entering the food supply. Thisis a particular problem for detecting E.coli O157:H7 in ground beef where theprevalence rate is less than 1%. In thiscase, the current USDA Food Safety andInspection Service (FSIS) sampling plan

Compositionof lot

Percentdefective

1

2

3

4

5

6

7

8

9

10

10

0.90

0.82

0.74

0.66

0.60

0.54

0.48

0.43

0.39

0.35

20

0.82

0.67

0.54

0.44

0.36

0.29

0.23

0.19

0.15

0.12

30

0.74

0.55

0.40

0.29

0.21

0.16

0.11

0.08

0.06

0.04

50

0.61

0.39

0.22

0.13

0.08

0.05

0.03

0.02

0.01

0.01

100

0.37

0.13

0.05

0.02

0.01

<

200

0.13

0.02

<

300

0.05

<

Number of Samples

Table 13. Probability of Acceptance (Pa) of Defective Product Using a 2-classSampling Plan with n=10 to n=300 and c=0 (ICMSF, 2002)

Product

Broilers

Market hogs

Cows/bulls

Steers/heifers

Ground beef

Ground chicken

Ground turkey

Samples

22,484

8,483

3,695

2,088

50,515

735

3,192

Positive

10.2%

7.0%

2.1%

0.3%

3.7%

14.5%

29.2%

Table 14. USDA Monitoring Program forSalmonella (1998-2000) (USDA/FSIS, 2000)

77EXPERT REPORT

Value of Test ResultsThe usefulness of sampling as ap-

plied in certain, more recent food-borne pathogen analysis situations hasbeen hotly debated. For highly infec-tious human pathogens, it is virtuallyimpossible to sample sufficient vol-umes of food to assure the total ab-sence of pathogens. Moreover, there isreal danger in assuming that if a foodis sampled for pathogen X, and thatpathogen is not found, the food is safe.Because sampling cannot assure safe-ty, other means to do so must befound or applied.

Relying on negative test results asan indicator of the food’s safety createsa disincentive to pursue additionalsafety measures. Consumers with afalse sense of security may relax theirvigilance with regard to food safety inpreparation and handling. Industryand regulatory agencies may hesitateto adopt new technologies that providea superior level of food safety, mistak-enly believing that the cost would notbe justified because negative test re-sults indicated present processingtechnologies were fully controlling thehazard.

Validity is defined as the ability ofthe test to do what it is intended todo—in this case, to detect the targetmicroorganism if it is present, and tonot detect it if it is absent. Two com-monly used measures of test validityare sensitivity and specificity. Sensi-tivity is the probability of a sampletesting positive if contamination istruly present, while specificity is theprobability of a sample testing nega-tive if the organism is truly absent.Both of these measures are inherentto the test itself, and fortunately, mosttests currently available for food-borne pathogens are highly sensitiveand highly specific, frequently in therange of 95% for both measures.

The predictive value of testing is anessential concept in the attempt to un-derstand the value of testing. Positivepredictive value is the probability thatthe product is indeed contaminatedgiven that the test is positive, while thenegative predictive value is the proba-bility that the product is free of con-tamination given a negative test result.While predictive value determinations

depend on the inherent sensitivity andspecificity of the test, they also dependupon the prevalence of contamination(see Fig. 10 ).

In the case of a contaminant thathas a high frequency of occurrence in agiven food, testing may have consider-able value because of the higher proba-bility that the contaminant will bepresent in a representative sample, and,given adequate test sensitivity and spec-ificity, the predictive value of positiveand negative test results will be high(i.e., in excess of 90%). Therefore, thetest is reasonably predictive of the truenature of contamination.

However, considering that the preva-lence of contamination by nearly allfoodborne pathogens is quite low inmost food commodities, the combinedeffect of low sampling plan efficacy andlow positive predictive value means thatthe value of testing is relatively minimal.The chance of obtaining a sample thathas the pathogen of interest is quitesmall, and nearly all presumptively posi-tive tests will be confirmed as negative af-ter additional testing. This means a tre-mendous expenditure for testing withvery little value with respect to detectingcontamination and potentially signifi-cant costs associated with product recalls

and holds. These costs are inevitablypassed on to the consumer, in ex-change for little public health benefit.

The relationship between contami-nation of food, pathogen test results,and public health impact is indeed acomplex one. In an ideal world, per-haps a better criterion of testing effica-cy would be based on the ability of thetest to accurately predict disease. Forthis to work, a number of critical fac-tors would need to be in place. Firstwould be the recognition that patho-gen contamination may occur as non-random, rare events that require verylarge samples to provide any confi-dence of finding positives if they exist.The second is the positive/negativepredictive value of the tests. Third istime-to-results and cost of tests.Fourth is converting true positive re-sults into a prediction of public healthimpact. And fifth is having to deal withthe public health impact of false-nega-tive results, which even for a test with95% sensitivity, will occur for 5% ofcontaminated samples anyway. In es-sence, when we establish a zero-toler-ance standard, it is more based on ourinability to predict the no effect levelthan it is our unwillingness to accepteven a single illness.

Fig. 10. Predictive Value (PV) of Test Results

PV(+) =(prevalence) (sensitivity) + (1-prevalence) (1-specificity)

(prevalence) (sensitivity)

Contaminant/food combination of relatively high prevalence

Campylobacter jejuni in raw poultry:60% prevalence;Test sensitivity and specificity of 95%

PV(+) = 0.966, or 96.6% of the time the test is positive, the sample is truly contaminated

PV(+) =(0.60) (0.95) + (0.40) (0.05)

(0.60) (0.95)

Contaminant/food combination of relatively low prevalence

E. coli O157:H7 in raw beef:1% prevalence;Test sensitivity and specificity of 95%

PV(+) = 0.161, or only 16.1% of the time the test is positive, the sample is truly contaminated

PV(+) =(0.01) (0.95) + (0.99) (0.05)

(0.01) (0.95)


involves 13 analytical units weighing 25geach and a very sensitive analytical meth-od. Aside from the low prevalence, thereis strong evidence indicating that thispathogen is not randomly distributedwithin production lots of ground beeffrom large commercial grinding opera-tions.

The criteria for most infectiousagents involve 2-class sampling plansand presence/absence testing. The strin-gency of the sampling plan is deter-mined by the number of samples ana-lyzed and the number of allowable posi-tive samples. Sampling plans that do notallow any positive sample units havebeen used for a variety of pathogens(e.g., salmonella and L. monocytogenesin RTE foods, and E. coli O157:H7 inraw ground beef). Some have referred tosampling plans that do not allow anypositives as “zero tolerance.” In reality,the zero tolerance can be made more orless stringent by increasing or decreas-ing the number of samples. Thus, asampling plan could specify 5, 10, 20 ormore samples. Although 25g analyticalunits are normally used, sampling planstringency also could be increased byincreasing the size of the sample unit,for example, to 50g.

Microbiological Criteria

Historically, attempts have beenmade to apply microbiological criteriafor the purpose of classifying foods aseither microbiologically acceptable, ormicrobiologically unacceptable. In1985, the Food and Nutrition Board ofthe National Research Council (FNB/NRC) addressed the subject of micro-biological criteria, and found that suchcriteria were of limited use, particular-ly if safety assurance is the goal, andthat HACCP should be applied wher-ever possible for safety assurance(FNB/NRC, 1985). The HACCP sys-tems envisioned by the FNB/NRC didnot rely on end-product testing forpathogens, but rather, if analyzed mi-crobiologically at all, analyses wereperformed at points along the food’sproduction chain, particularly to verifythat CCPs were under control.

Microbiological criteria consideredby FDA generally include standardplate count, coliform counts, yeast andmold counts, and E. coli (generic)counts. Coliforms and E. coli were be-lieved to be indicators of possible fecalcontamination, and therefore, it is pos-

sible that food containing either bacte-ria at a prescribed level was unsafe.

The “zero tolerance” for L. monocy-togenes in RTE foods was established asa safety-related criterion. The “zerotolerance” actually means that L.monocytogenes must be absent fromtwo 25g samples of foods under FDAinspection. The total absence require-ment was derived at a time when therewere no effective methods for findingL. monocytogenes in food (or any envi-ronment outside of the human or ani-mal). There was no understanding ofthe very widespread existence of L.monocytogenes throughout the envi-ronment, including food processingenvironments, nor an appreciation ofthe number of foods in which the bac-teria historically had been present.Thus, contemporary knowledge abouthuman exposure suggests that manyhumans are routinely exposed to thebacteria with no consequence to health,although L. monocytogenes does causeillness in sensitive subpopulations.The “zero tolerance” has acted as a dis-incentive for the application of quanti-tative (enumerative) methods, andthus, the body of human exposure datais incomplete.

Today, there is growing acceptanceof the need for management systemsbased on GMPs and HACCP to controlfood safety hazards. Microbiologicaltesting is sometimes valuable in verify-ing the effectiveness of GMP andHACCP systems and validating CCPswithin HACCP systems. There may bea role for testing certain ingredientswhen they can influence the safety of afinished product, but because the test-ing of ingredients faces the same weak-ness as end-product testing, auditingof suppliers’ control programs has in-creased to provide greater assurance.

In addition to product testing, en-vironmental testing may be necessary.Salmonella and L. monocytogenes havethe ability to become established asresidents in food processing establish-ments. Environmental sampling pro-grams assess the degree of control andindicate when corrective actions areneeded. They may or may not indicatea safety problem in the finished prod-uct.

The optimal regulatory approach toenvironmental testing would use ourunderstanding of basic human natureto encourage and reward diligence. De-pending on the type of food and the

processing conditions, it should be ex-pected that these pathogens will be pe-riodically introduced into the foodprocessing environment by variouspathways. To prevent the pathogensfrom becoming established and multi-plying, the sampling program shouldaggressively look for these pathogens.Finding a positive sample should betreated as a success, because correctiveactions can then be applied and con-sumer protection assured. Treating apositive sample as a failure and apply-ing a penalty decreases the desire to de-tect the pathogens, discouraging theaggressive nature of the environmentalsampling program.

Testing Methods

Disease surveillance and control ef-forts will benefit greatly from newpathogen detection methods that offergreater precision, rapid results, and de-creased cost. However, efforts to adaptthese new technologies to the challeng-es in the food environment are ongo-ing.

Many methods of detection are cur-rently available for foodborne patho-gens, but food microbiologists must of-ten choose between enumeration andidentification without the option ofboth. Enumerative methods are usuallybased on the ability of the normalhealthy bacterial cells to multiply in anutrient-rich medium. Although selec-tive agents are sometimes added to favorthe growth of a specific group of organ-isms, most of these methods are stillreasonably nonspecific. With respect topathogen identification, methods havehistorically relied on cultural enrich-ment to increase the numbers of the tar-get microorganism and allow resuscita-tion of injured cells. When followed byselective and differential plating, thesemethods provide discrimination of thetarget organism from the backgroundmicroflora, but are non-enumerative.For both enumerative and non-enu-merative methods, the combined effectof low levels of contamination and theneed for cultural growth results inlengthy assays, frequently extending be-yond four days for even preliminary re-sults.

Most rapid method developmentshave sought to shorten detection timeby replacing the selective and differen-tial plating steps with more rapid tech-nologies such as ELISA and DNA hy-

79EXPERT REPORT

bridization, but because these methodsremain hampered by less than optimalassay detection limits, lengthy culturalenrichment steps are still necessary.Furthermore, cultural confirmation forpresumptively positive results is gener-ally required for regulatory purposes.The limiting factor in making thesemethods truly rapid is predominantlythe lengthy incubation time required toincrease cell numbers.

Enzymatic nucleic acid amplifica-tion methods such as the polymerasechain reaction (PCR) offer several po-tential advantages for the rapid and reli-able detection of microbial pathogens infoods. The primary advantage of thistechnology is the theoretical replace-ment of cultural enrichment with spe-cific nucleic acid sequence enrichment,thereby decreasing total detection time.There are many reports of PCR-basedassays for the detection of foodbornepathogens and several companies mar-ket these systems, all of which are cur-rently in evaluation for AOAC approval.

In reality, rapid molecular detectionmethods for pathogens in food prod-ucts remain in the developmental stages.The significant methodological hurdlesyet to be addressed include the need to:(1) test larger, realistic sample volumes(at least 25 ml or g) instead of the smallvolumes (10-50 microliters) used inmolecular-based assays; (2) account forthe effect of residual food components

that inhibit PCR enzymatic reactions;(3) detect low levels of contaminatingpathogens; (4) assure detection of via-ble (infectious) pathogens; and (5) con-firm molecular amplification productswith more lengthy DNA hybridizationassays (Bej and Mahbubani, 1994).

With respect to the first three chal-lenges, the application of rapid methodscould perhaps be improved if pathogenswere separated, concentrated, and puri-fied from the food matrix before detec-tion (Swaminathan and Feng, 1994).None of the various bacterial concen-tration methods, including the mostwidely used immunomagnetic separa-tion, (Sharpe, 1997) is ideal, highlight-ing the need to increase research in thisarea. With respect to detection of viablepathogens, naked DNA and the DNA ofdead cells can persist for long periods oftime (Herman, 1997). Even the morestable 16S rRNA is not an ideal indica-tor of pathogen viability (McKillip et al.,1998). Although mRNA may be consid-ered a more promising target (Sheridanet al., 1998), key assay design issueswould need to be addressed. Finally,methods to simplify or even eliminatepost-amplification confirmation assays(McKillip and Drake, 2000; Norton andBatt, 1999; Sharma and Carlson, 2000)are expensive and have not been widelyapplied to food matrices (Koo andJaykus, 2000).

When taken together, the promise of

emerging molecular detection methods,including biosensors (deBoer andBeumer, 1999) and microarray technol-ogy (Epstein and Butow, 2000), must betempered with an appreciation for thecomplexity of the food matrix.

Surveillance for FoodborneHazards and Illness

One way to identify emerging patho-gens is surveillance of foodborne illness.Not only can scientists track the spreadand frequency of a pathogen by lookingfor its victims, they can quickly spotchanges in virulence or exposure. Sur-veillance data can be used for quick out-break response and also as the basis forqualitative and quantitative risk assess-ment. New scientific tools have signifi-cantly increased the speed and depth ofsurveillance information gathering, mak-ing it more effective.

Purposes and Mechanisms

Surveillance involves the systematiccollection of data with analysis and dis-semination of results. Surveillance sys-tems may be passive or active, nationalor regional in scope, or based on a senti-nel system of individual sites. Tradition-ally, human foodborne disease surveil-lance has been conducted for three rea-sons: (1) to identify, control, and preventoutbreaks of foodborne disease, (2) todetermine the causes of foodborne dis-ease, and (3) to monitor trends in occur-rence of foodborne disease.

By identifying outbreaks and theircauses quickly, surveillance can result inearly intervention to address hazards inthe food supply. Officials may be able toremove contaminated products from re-tail shelves (e.g., identification of Sal-monella Agona in contaminated cereal(CDC, 1998c)) or rectify inappropriatefood handling procedures (e.g., under-cooking of meats or cross-contamina-tion of vegetables from raw chicken).

The cumulative information ob-tained through surveillance and out-break investigation can reveal the magni-tude and trends of foodborne disease,helping policy makers identify optimalprevention strategies (Borgdorff andMotarjemi, 1997). Additionally, im-proved understanding of disease andhazard etiology can help researchers an-ticipate or recognize new problems, suchas toxins in one food that could pose aproblem in other foods or toxins that are

Testing for MycotoxinsIn the United States, raw agricul-

tural commodities are routinelyscreened for mycotoxins using specificFDA guidelines that are based on hu-man risk assessments. The testingmethods used must be validated bythe AOAC International (Gaithers-burg, Md.), which ensures reliabilityand reproducibility.

Some methods, such as commer-cial enzyme-linked immunosorbentassays (ELISAs), are particularly use-ful for rapid screening of commoditiesand take as little as 5 minutes to com-plete (Pestka et al., 1995). Becausecorn is susceptible to aspergilli thatproduce aflatoxins, it is screened at thegrain elevator and rejected if it exceedsthe FDA guideline. Thus, aflatoxin-contaminated corn is identified and

diverted early in the food processingchain. A similar approach is linkedwith an indemnification program forpeanuts, which are also highly suscep-tible to aflatoxin contamination.Analogous guidelines have been setfor other mycotoxins (trichothecenedeoxynivalenol and the fumonisins),and rapid tests are available.

In general, careful monitoring ofweather conditions and field testingfor mycotoxins will identify years inwhich there is increased potential forcontamination in specific commodi-ties. Screening efforts can be in-creased and targeted toward possibleproblematic materials and regions,such as aflatoxins in corn and peanutsduring an extended drought in thesoutheastern United States.


newly recognized as a human health haz-ard (see sidebar above).

HACCP systems rely on accurateknowledge of potential hazards (NACM-CF, 1998). Many of these hazards—spe-cific agents, food ingredients, or agent/food interactions—were originally iden-tified as a result of foodborne diseasesurveillance. Because food sources andfoodborne disease agents are constantlychanging, hazard analysis is an ongoingprocess that requires continuous supportfrom public health surveillance of food-borne disease.

Foodborne disease surveillance canalso supply important feedback on theeffectiveness of control strategies. Forexample, during the 1980s, the in-creased occurrence of sporadic S. Enter-itidis infections and outbreaks in NewEngland led to the identification of anew problem with S. Enteritidis con-tamination of grade A shell eggs (St.Louis et al., 1988). In the United States,USDA and FDA have worked with theegg industry to develop and implementa number of control strategies (Hogueet al., 1997). The incidence of S. Enter-itidis infections in FoodNet sites de-

clined 48% during 1996 -1999, suggest-ing that these control strategies are be-ginning to work (CDC, 2000a, b).

Current Surveillance Programs

Foodborne disease surveillanceconsists of four primary components:(1) identifying and reporting outbreaks,(2) monitoring for specific pathogens,(3) determining risk-factors for sporadiccases of infection with common food-borne pathogens, and (4) studying thepopulation to track gastrointestinal ill-ness, including trends in the requests forhealth care, food consumption and per-sonal prevention measures.

Improving pathogen-specific surveil-lance has been a major focus of the Na-tional Food Safety Initiative. Serotype-specific surveillance of Salmonella con-ducted by state health departments andCDC’s Public Health Laboratory Infor-mation System (PHLIS) has identifiedseveral large, multi-state outbreaks ofsalmonellosis. These outbreaks, causedby cantaloupes, tomatoes, and alfalfasprouts, were spotted because of unusu-al, time-related clusters of cases caused

by an uncommon Salmonella serotype.Epidemiological investigation of thesecases identified the source.

PHLIS has developed an automatedsurveillance outbreak detection algo-rithm (SODA), originally developed toaddress Salmonella, that uses the 5-yearmean number of cases from the samegeographic area and week of the year tolook for unusual case clusters (Hutwag-ner et al., 1997). S. Stanley and S. Agonainfections initially identified by individu-al state health departments were discov-ered to be multi-state outbreaks basedon disease clusters identified by SODA.

Because it compares current cases to5-year means, SODA appears to be moreeffective at detecting case clusters of un-common serotypes rather than commonserotypes, such as S. Typhimurium. Al-though the Minnesota Department ofHealth reported 11 confirmed outbreaksof S. Typhimurium infection from 1996-1998, SODA detected only three. Duringthe same time period, SODA identifiednine weeks where the number of S. Typh-imurium reports exceeded the 5-year av-erage. Six of these notifications were dueto epidemiologically unrelated S. Typh-

Outbreak Investiga-tions and NewFoodborne Pathogens

On April 29, 1991, local publichealth officials were notified of an out-break of foodborne illness among per-sons who celebrated “Secretary’s Day”at a local restaurant (Hedberg et al.,1997). Seventeen (89%) of 19 mem-bers of the index group developed di-arrhea and cramps 11 to 122 hours(median, 56 hours) after their meal.Fewer than half of cases reported nau-sea, myalgia, fever, or vomiting. Dura-tion of illness ranged from 4 to 7 days(median, 5 days). Similar illnesseswere also reported among other res-taurant patrons and among five (15%)of 34 food handlers at the restaurant.

The restaurant served a large hoteland conference center and featured anelaborate buffet with a variety of freshfruits, vegetables, salads, and gourmetfood items that combined cooked anduncooked foods. The apparent highattack rate of illness in the index group

and reports of illnesses among food han-dlers led to early concern that the restau-rant was experiencing a large outbreak ofviral gastroenteritis. However, as theclinical and epidemiologic features of theoutbreak emerged from interviews withpatrons, it appeared typical of previouslydescribed outbreaks caused by entero-toxigenic E. coli (ETEC).

The recognition that the outbreakmay have been caused by an uncommonfoodborne pathogen led to extensive ef-forts to obtain stool from ill patrons toconfirm the etiology. A lactose-negativenon-motile E. coli O39 was isolated from10 of 22 cases. No Salmonella, Shigella,Campylobacter, Yersinia, Vibrio, or Plesi-omonas species were isolated from ill pa-trons. Although the clinical and epide-miologic features of the outbreak sug-gested ETEC, the outbreak-associatedO39 strain did not possess ETEC heat-labile (LT) or heat-stable (ST) enterotox-ins. Extensive testing of the outbreak-as-sociated strain by a battery of geneprobes detected the presence of intimin(eae), an adherence factor associated

with enteropathogenic (EPEC) andenterohemorrhagic (EHEC) E. coli andEAST1 (astA), a heat-stable enterotox-in associated with enteroaggregativeadherent (EaggEC) E. coli. Thus, theoutbreak strain did not fit neatly intoany of the recognized categories of di-arrheagenic E. coli and would not havebeen identified as a pathogen had itnot been implicated in this outbreak.

The majority of diarrheagenic E.coli virulence factors are encoded onpathogenicity islands, transmissibleplasmids, bacteriophage, or trans-posons. Horizontal transmission ofvirulence factors led to the developmentof highly virulent E. coli O157:H7which has emerged as a major food-borne disease of public health impor-tance. This same genetic plasticitycould lead to emergence of other com-binations of virulence factors. Promptand thorough epidemiologic investiga-tion of outbreaks will be needed toidentify these novel emerging patho-gens and to further our understandingof their public health significance.

81EXPERT REPORT

imurium isolates with different pulsed-field gel electrophoresis (PFGE) patterns.One flagged the occurrence of two si-multaneous outbreaks and the other twowere due to a single outbreak that ex-tended over time (Bender et al., 2001).

Molecular subtyping schemes, suchas PFGE, can improve the investigationof outbreaks by distinguishing unrelat-ed sporadic cases from the main out-break-associated strain. The ability todistinguish specific subtypes among rel-atively common organisms, such as E.coli O157:H7 and S. Typhimurium, isthe basis of the National MolecularSubtyping Network (PulseNet), whichtakes advantage of recent advances inboth molecular biology and informa-tion technology. Highly reproduciblePFGE patterns are generated for apathogen implicated in an illness, andthe PFGE patterns can be electronicallyshared between participating laborato-ries. An outbreak of E. coli O157:H7 in-fections in Colorado was associatedwith consumption of a nationally dis-tributed ground beef product. Withindays, public health officials could com-pare the outbreak strain to PFGE pat-terns of E. coli O157:H7 isolatesthroughout the United States (CDC,1997). PulseNet has the potential to bethe “backbone” of a public health sur-veillance system that can provide trulynational surveillance for a variety offoodborne pathogens in a manner time-ly enough to be an early warning systemfor outbreaks of foodborne disease(Hedberg et al., 2001).

PulseNet’s usefulness is currentlylimited because not all public health lab-oratories are connected, not all clinicallaboratories routinely submit isolates topublic health laboratories, and manystates do not have sufficient epidemio-logic resources to investigate individualcases or clusters.

In contrast to PulseNet’s widespreadsurveillance area, FoodNet is a sentinel-site, active surveillance project designed totrack all diagnosed infections of impor-tant foodborne diseases and evaluate thelaboratory, physician and patient practicesthat cause an individual case to be diag-nosed. FoodNet’s initial surveillance area(13.2 million residents of Minnesota, Or-egon, and selected counties in California,Connecticut, and Georgia) was expandedin 2000 and 2001, adding sites in NewYork, Maryland, Tennessee, and Coloradothat brought the population under sur-veillance to 33.1 million persons. FoodNet

uses active surveillance, meaning publichealth authorities regularly contact clini-cians and laboratories to obtain case re-ports.

Most foodborne disease surveillanceuses passive reporting systems, in whichreports are voluntarily submitted byhealth clinics and laboratories. This sys-tem depends on the clinician’s ability todiagnose the illness and the willingnessof clinicians and laboratory personnel toreport the diagnoses to the appropriatepublic health authorities.

In general, active surveillance yieldsbetter data than passive systems but ismore expensive and limited in scope. Be-cause some cases of foodborne illnesswill remain unrecognized and go unre-ported, even active surveillance systemsare inherently incomplete (Potter andTauxe, 1997).

One of the most striking gaps in ourfoodborne disease surveillance is gener-ation of data about individuals whohave gastrointestinal illness but do notsee a physician. Furthermore, physiciansoften treat mild to moderate gas-trointestinal illness symptomaticallyand do not frequently culture speci-mens or conduct the wide range of di-agnostic tests necessary to identify allfoodborne agents.

Unknown Agents

Surveillance for foodborne diseasesis based on detection of specific patho-gens or the occurrence of illnesses, suchas diarrhea, in defined groups. An out-break may be recognized because a statepublic health laboratory detected the in-creased occurrence of a specific subtypeof E. coli O157:H7, or because half of thepeople who attended a specific event de-veloped vomiting and diarrhea shortlyafter the event. In the first case, surveil-lance is limited by what clinical laborato-ries routinely identify when processinghuman stool samples, whether by directexamination, culture, or use of non-cul-ture diagnostic tests. In the second case,surveillance has a better chance of identi-fying foodborne agents that are not partof routine clinical microbiology, but it isstill limited by the epidemiologic andlaboratory resources available to publichealth departments.

Published estimates of foodborne dis-ease occurrence highlight the limitationsof our current surveillance system. Of the76 million cases of foodborne illness esti-mated to occur each year in the United

States, 82% are attributed to “unknownagents” (Mead et al., 1999). Of the 28known foodborne pathogens included inthis estimate, routine passive surveillancesystems exist for only 17 (61%). Formany of the others, scientists must infertheir frequency from the occurrence ofoutbreaks or from a limited number ofpopulation-based studies that researchedthe causes of diarrhea. For example, clini-cal laboratories do not routinely identifyNorwalk-like viruses (NLVs), and no sur-veillance program tracks cases of NLV in-fection. Yet estimates attribute 11% of allepisodes of diarrheal illnesses to NLVs,based on a study from the Netherlands(Mead et al., 1999). The estimated propor-tion of foodborne illness that is caused byunidentified agents is bolstered by Centersfor Disease Control and Prevention(CDC) data (Mead et al., 1999). No agentwas identified in 1,873 (68%) of 2,751confirmed foodborne outbreaks reportedto CDC from 1993-1997 (Olsen et al.,2000).

A high percentage of the outbreaksattributed to “unknown etiology” areprobably outbreaks of viral gastroenteri-tis that were not confirmed either be-cause stool samples were not availablefor testing, or because public health lab-oratories did not perform the test neces-sary to detect viruses. For example, oneretrospective study confirmed the pres-ence of NLVs in 90% of a select group ofoutbreaks of non-bacterial gastroenteri-tis (Fankhauser et al., 1998). In Minne-sota, officials used the clinical and epide-miologic appearance of the outbreak tolink 120 (41%) of 295 confirmed food-borne outbreaks reported from 1981-1998 to NLVs, leaving only 26 outbreaks(9%) attributed to unknown agents (De-neen et al., 2000). The factors used in theclassification included a median incuba-tion period of between 24-48 hours, a12-60 hour duration of symptoms, and arelatively high proportion of cases expe-riencing vomiting (Hedberg and Oster-holm, 1993; Kaplan et al., 1982). In a ret-rospective review of foodborne out-breaks reported to CDC from 1982 to1989, almost half (48%) of the 712 out-breaks reported as having an undeter-mined etiology met the epidemiologiccriteria for outbreaks of NLV (Hall et al.,2001). Thus, although a high percentageof reported foodborne illnesses do nothave an identified cause, it appears thatmany are potentially identifiable causes.Additional surveillance would help re-searchers more frequently identify the


agent responsible for cases of foodborneillness and provide more reliable esti-mates of the true prevalence of variousfoodborne pathogens.

Integrated Surveillance

The modern concept of public healthsurveillance, first articulated by Alex-ander Langmuir, views surveillance as aprocess (Foege, 1996). As it relates tofood safety, the process is concerned notonly with outcomes in the human popu-lation but also with the occurrence offoodborne hazards in all types of foods,their sources, and the various stages intheir conversion to consumable food.Operationally, surveillance involves thesystematic monitoring of disease andhazard reports—in animal and plantpopulations, food production and pro-cessing environments, foods and ingredi-ents, and in human populations—through the systematic collection, analy-sis, and interpretation of outcome-spe-cific data, closely integrated with thetimely dissemination of these data tothose responsible for preventing andcontrolling disease or injury (Thackerand Berkelman, 1988).

Foodborne disease surveillance hastraditionally been viewed as a subset ofpublic health surveillance. The links be-tween surveillance for foodborne diseas-es in humans and surveillance for food-borne hazards in foods have only recent-ly received increased attention. Food-borne hazard surveillance monitors theconditions that can lead to foodborne ill-nesses (Guzewich et al., 1997). For exam-ple, hazard surveillance systems can de-tect microbial pathogens at various facil-ities that handle food (e.g., farms, meatand poultry processors, and restau-rants). Hazard surveillance typically in-volves the collection of data on food-borne hazards in food products andfood sources, follow-up data when haz-ards are present at unusual levels, andinformation that helps define the sourcesof hazards in foods.

Animal health surveillance as it re-lates to food safety is a component offoodborne hazard surveillance. Compre-hensive animal health surveillance sys-tems were nonexistent until the NationalAnimal Health Monitoring System(NAHMS) was implemented in 1983(King, 1990). Current resources limitthese surveillance programs; the NAHMSon-farm monitoring system does nar-

rowly focused studies of a single speciesin a particular segment of the produc-tion process.

Integrating the information from anon-farm monitoring program such asNAHMS with processing data, retail foodsurveillance, residue and antimicrobialresistance monitoring, and subsequentlywith FoodNet information will be criticalfor the implementation of a true farm-to-table approach to food safety surveil-lance. Not only will the data be more re-liable if a cohesive surveillance systemmonitors food from the farm to the table,but such a system will likely provide theimpetus for a more comprehensive sur-veillance system in domestic animals(Bush et al., 1990).

The awareness of the need to moni-tor pathogens in healthy food animals isfairly recent, and monitoring pathogens

in the food and water that food animalsconsume also may be appropriate(Tauxe, 1997). If farms show evidence ofincreasing pathogen prevalence, thenprompt intervention might prevent thepathogens from eventually being con-sumed by humans.

Effective surveillance for food safetyrequires the coherent assembly of infor-mation from different sources. Integrat-ing animal and environmental surveil-lance systems into established humansurveillance systems will greatly increaseour understanding of the epidemiologyand sources of foodborne disease. Inparticular, an independent molecularsubtyping system linked to PulseNet hasgreat potential value for evaluating thepotential public health significance ofpathogens isolated all along the foodprocessing continuum. For example, it

Animal Surveillance forE. coli O157:H7

Risk assessments for specificpathogens such as E. coli O157:H7 inground beef require measurements ofmany parameters at every step fromfarm to table. Surveillance programsgather these data and reveal useful in-formation about hazards. For exam-ple, E. coli O157:H7 is widely distrib-uted throughout beef and dairy cattleherds in the United States (Hancock etal., 1998). NAHMS addresses emerg-ing issues such as the association be-tween calf management practices andthe presence of E. coli O157:H7 in cat-tle herds (Garber et al., 1995; Losingeret al., 1995). The veterinary Diagnos-tic Laboratory Reporting System com-piles and analyzes reports from stateveterinary diagnostic laboratories toassess trends in infectious diseasesamong food animals (Salman et al.,1988).

To prepare for HACCP introduc-tion in the meat industry, USDA con-ducted a series of baseline surveys ofbeef slaughter plants and the groundbeef final product. At that time, only 4(0.2%) of 2,081 steer and heifer car-casses and none of 2,112 cow and bullcarcasses were contaminated with E.

coli O157:H7. Of 563 ground beefsamples, 78.6% were contaminatedby nonpathogenic E. coli but none byE. coli O157:H7 (USDA/FSIS, 1996).

More recently, with the aid ofmuch more sensitive detection meth-ods, researchers found EHEC O157(E. coli O157:H7 or O157:nonmo-tile) in the feces of 27.8% of cattle ata slaughter plant; 10.7% of hideswere contaminated, and 43.4% ofcarcasses were contaminated beforeevisceration (Elder et al., 2000). Only17.8% of carcasses were contaminat-ed post-evisceration, and 1.8% ofcarcass tissues contained EHECO157 after processing, which dem-onstrates the effectiveness of plantsanitation processes.

Despite this relative efficacy,USDA detected E. coli O157:H7 inground beef samples at a rate of ap-proximately 8.7 per 1,000 samples in2001, and E. coli O157:H7 contami-nation of ground beef resulted in 25recalls during 2001 (USDA/FSIS,2002). Regulatory agencies and foodprocessors need to work together toadvance the scientific understandingof the persistence and transmissionof these agents in food productionenvironments.

83EXPERT REPORT

would be extremely useful for a foodprocessor to be able to evaluate whethera particular environmental strain of List-eria isolated from a processing environ-ment had ever been associated with hu-man infections. The technology exists tocreate such a system. Data privacy andregulatory penalties will need to be mod-ified to encourage the food industry’sfull participation.

Future Methods: the Promise of Genomics

Since the advent of recombinantDNA technology, a better understandingof how and why pathogens do what theydo has emerged. Progress has beensteady. In many cases, one gene at a timehas given up its secrets, and, in the pro-cess of doing so, has presented new puz-zles to be solved. The science of genom-ics is simply the study of the genes of anorganism and their function. It is nowpossible to sequence entire genomes, andthis has been accomplished for some sig-nificant human pathogens. The processof whole genome sequencing has beenaccelerated by automation and the appli-cation of sophisticated computer tech-nologies (informatics).

Data gathered to date on pathogenicbacteria have already provided revealinginsights. For example, nearly one-half ofthe open reading frames (ORFs) se-quenced have no known function. It isclear that we have just begun to under-stand how these bacteria survive and reactto their environment. From comparisonsamong the complete sequences of bacte-ria, it is also clear that far more horizontaltransmission of genetic material has oc-curred than previously thought. Hori-zontal transmission of genes can rapidlytransform a commensal bacterium into apotential pathogen through the sharing oflarge numbers of virulence-related genes(pathogenicity islands) or genes encodingfor antibiotic resistance.

Comparisons also enable the genera-tion of hypotheses regarding a bacteri-um’s virulence potential that can then betested by other traditional laboratory ap-proaches, or by further genetic manipu-lation. Comparative genomics also maylead to new approaches to phylogeneticclassification. An adjunct to genomics isproteomics, the study of the completeprotein complement of an organism. Al-though it might appear that these tech-nologies have opened up a new level ofcomplexity, it is believed by many scien-tists that this very complexity may yield

new ways to control microorganismsthat have not yet been conceptualized.

Another exciting area with genomicsas the driving force relates to under-standing the bacterium’s global gene ex-pression while varying the bacterium’senvironment. This has been made possi-ble by the development of oligonucle-otide “chips” or cDNA microarrays thatenable “expression profiling,” that is, thestudy of the messenger RNAs, and whenand which ones are produced. Microar-rays may help unravel the function of thenumerous genes whose functions are notyet known. An excellent overview ofthese technologies was recently presentedby Schwartz (2000).

Using DNA microarrays, also knownas biochips, research scientists can analyzethe transcription profiles of the whole ge-nome for practically any microorganismof interest. For example, in organismssuch as Haemophilus influenzae andStreptococcus pneumoniae scientists haveused DNA biochips to map more than100 genes and their expression profiles.The current detection range has been be-tween one and five transcripts per cell, asconfirmed by conventional methods suchas Northern blot analysis. Scientists canlearn what gene(s) are turned on or offunder different conditions by putting(spotting) DNAs representing all ORFs ina bacterial genome and using differential-ly labelled cDNAs from both wild-typeand mutant bacteria.

Microarray technology affords un-precedented opportunities and ap-proaches to diagnostic and detectionmethods. For example, microarrays canbe used to develop rapid identificationsystems for both pathogenic and spoilagebacteria, to conduct mutation analyses,and to investigate protein-DNA interac-tion. RNA of a related species can bestudied using differential gene expres-sion under less stringent hybridizationconditions (referred to as virtual expres-sion arrays). In addition, microarrayswill be used in the future to detect organ-isms or foods modified using recombi-nant DNA biotechnology. For example,transcript mapping (imaging) of wild-type versus genetically modified organ-isms can monitor changes in risk-relatedfactors such as virulence genes.

Pathogen identification

New pathogen identification technol-ogies are faster, cheaper, more powerfuland increasingly automated. No single

method is appropriate for all circum-stances, so selection of the best method isimportant. This technology is changingrapidly, and the following informationprovides a brief overview of the progressin this area and the future potential.

Until recently, efforts to determinebacterial relatedness relied on techniquesthat assessed one or more phenotypicmarkers. These methods include sero-typing, phage typing, biotyping, antibiot-ic susceptibility testing and bacteriocintyping. Now, molecular typing methodscan identify different clones (geneticallyidentical organisms descended from asingle common ancestor) at the bacterialspecies level. These molecular tech-niques are used to physically characterizebacteria based on their DNA composi-tion (genotyping) or on production ofproteins, fatty acids, carbohydrates, orother biochemical content (phenotypingor chemotyping).

As technology has evolved, manypreviously complex processes have beenautomated, miniaturized and linked tocomputer control centers that guide theoperation, including data analysis. Asthe technologies have become wide-spread, researchers are now able to gen-erate more timely data at a lower unitcost. Of particular interest to those in-volved in the biochemical analysis of mi-croorganisms are procedures that havebeen adapted or are amenable to wholecell techniques, as they offer all of theconveniences of rapid and economicalanalysis. Huge libraries of customizablecomputer databases are now available toassist in pattern recognition for detectionand identification of microorganismsbased on analysis of whole cells, as wellas individual genetic elements or chemi-cal derivatives. The need for quick testresults have driven these advancements.Automated methods are now availablefor detection, identification, typing andanalysis of biological components orstructural changes that occur due to en-vironmental pressures or extraneous in-fluences. Bench top versions of sophisti-cated devices allow for more portabilityand efficient use of laboratory space.

Genotyping Methods

Genotyping has many advantagesover traditional typing procedures (Oliveand Bean, 1999; Spratt, 1999; Tompkins,1992; Versalovic et al., 1993). The majoradvantage lies in its ability to distinguishbetween two closely related strains. Oth-


er advantages of genotyping include:(1) DNA can always be extracted frombacteria so all strains are theoreticallytypeable; (2) analytical strategies for thegenotypic methods are similar and canbe applied to DNA from any source; (3)genotyping procedures do not generallyrequire species-specific reagents and (4)the methods are amenable to automa-tion and statistical data analysis (Arbeit,1995; Bingen et al,. 1994). Combina-tions of different genotypic methodscan be used to increase the discrimina-tory power of typing and fingerprintinganalyses. Furthermore, selection of theappropriate typing method can allowanalysis of groups of bacteria at the ap-propriate level and rate of change. To il-lustrate, ribotyping of Vibrio choleraeisolates responsible for the 1994 - 1995cholera epidemic in Ukraine indicatedthat only a single strain arising from in-troduction into that country was re-sponsible (Clark et al., 1998). Othermore discriminatory methods wereused to track the course of the epidemic.A combination of typing or fingerprint-ing methods may therefore be necessaryto fully characterize bacterial popula-tions important to public health.

The most common genotypic meth-ods currently used include:• chromosomal DNA restriction anal-ysis,• plasmid typing,• DNA probe-based hybridizations(such as ribotyping)• amplified fragment length polymor-phism (AFLP)• PFGE• PCR-based methods (such as ran-domly-amplified polymorphic DNA(RAPD), repetitive sequence-based PCR(rep-PCR), PCR-ribotyping and PCR-re-striction fragment length polymorphism(PCR-RFLP)), and• sequence-based methods, includingmultilocus sequence typing, flagellar lo-cus and flagellar short variable region se-quencing (e.g., for Campylobacter), andanalysis of DNA sequences of a numberof other genes.

PFGE has now been applied to awide range of microorganisms and hasbecome the genotypic method of choicefor many scientists because it is very dis-criminating, reproducible and broadlyapplicable. PFGE has recently been usedto help in the investigations of wide-spread foodborne outbreaks involving

Salmonella (Bender et al., 2001; VanBeneden et al., 1999), L. monocytogenes(Graves and Swaminathan, 2001; Ojeniviet al., 2000; Proctor et al., 1995), E. coliO157:H7 (Barrett et al., 1994) and virus-es (CDC, 2001), and is the method thatis currently being used by CDC as thebasis for its PulseNet system.

There is a trend to use new diagnos-tic assays to disclose the presence ofpathogenic bacteria in foods or humanpatients without isolation of the organ-ism. This presents a challenge to micro-biologists, in that many of the typing/fingerprinting methods currently in userely on large quantities of DNA isolatedafter amplification of the strain of inter-est. Methodologies based on DNA se-quencing after PCR amplification direct-ly from the source material may presentat least a partial solution to the problemscreated by the absence of an isolate. TheCDC PulseNet group has, for instance,recently provided funding to interestedstate laboratories for research into theappropriate genes to be sequenced to al-low differentiation of bacterial patho-gens of interest.

Biochemical and Chemical Methods

Numerous biochemical techniquesoffer an alternative to direct nucleic acidfingerprinting. Chemotaxonomy in-volves the application of chemical andphysical manipulations to the analysis ofthe chemical composition of whole bac-terial cells or their cellular componentsto arrive at some identification or taxo-nomic positioning. Even with accelerat-ed advances in technology that have al-lowed for miniaturization and automa-tion of analytical equipment, the bacteri-al growth period and chemical deriva-tions prior to analysis still remain the ul-timate limiting factor with respect torapid analysis. Thus, methods that areamenable or adaptable to whole celltechniques and require only minutequantities of sample and/or in situchemical derivations are of particular in-terest, including:• mass spectrometry (MS), especiallymatrix-assisted laser desorption/ionization-time of flight (MALDI-TOF)mass spectrometry• pyrolysis (Py),• gas or liquid chromatography (GCor LC), and• infrared spectroscopy (IR).

Future Issues

Although genome-based typingmethods are increasingly powerful toolsin molecular epidemiology, several issuesneed to be addressed if these methods areto be incorporated more routinely. First-ly, one should not forget about the ad-vantages of classifying organisms to thegenus and species level, as well as doingserotyping and phage-typing for somebacteria, before interpreting bandingpatterns resulting from molecular typ-ing. A good example of this is a 2001 sal-monellosis outbreak associated with rawalmonds that was detected only becauseboth serotyping and phage-typing of sal-monella isolates were done. As well, inthe absence of a “gold standard” bywhich to judge a typing method (vanEmbden et al., 1993), careful standard-ization of and adherence to laboratoryprotocols is essential, if individual meth-ods are to be accepted for classificationof strains. The lack of reproducibility ofcertain techniques is another contentiousissue. Consistent reproducibility is es-sential, if these methods are to be of val-ue in the long-term analysis and catego-rizing of bacterial strains. Another ex-tremely important issue is in the inter-pretation of minor (ca. 1 to 3) bandingdifferences between strains. Some scien-tists argue that a single difference in theproduction of an enzyme or the shift of asingle band on a gel is not enough to saythat two isolates are different, and thatclonality should be considered as a rela-tive concept (Arbeit, 1995). In addition,strain relatedness should only be judgedin the presence of other data, especiallyepidemiologic data.

The ultimate goal is an “ideal” molec-ular typing method; one that is easy toperform, cost-effective, relatively rapid,amenable to statistical analysis and auto-mation, able to type all possible strains,reproducible, and properly balanced be-tween increased discriminatory powerand applicability. Rapid advances in typ-ing methods based on whole organismDNA sequencing are helping us to ap-proach such an ideal method. In the fu-ture, the use of molecular typing in food-borne disease investigations will assist usin identifying the source of many moreoutbreaks, will lead to the earlier detectionof outbreaks, and will be beneficial inidentifying and eliminating areas of per-sistent contamination in food plants.

85EXPERT REPORT

Next Steps in Food Safety Management

Foodborne illness in the United States

is a major and complex problem that is

likely to become a greater problem as

we become a more global society. To

adequately address this complex

problem, we need to develop and

implement a well conceived strategic

approach that quickly and accurately

identifies hazards, ranks the hazards by

level of importance, and identifies

approaches that have the greatest

impact on reducing hazards, including

strategies to address emerging hazards

that were previously unrecognized.Because certain elements of patho-

gen evolution are inherently unpredict-able, it is impossible to predict, with ab-solute accuracy, the emerging microbio-logical food safety issues of the future.However, our knowledge of the currentissues and the complex factors that drivechanges in microbiological food safetydo provide us with a good sense of likelytrends. With this knowledge and under-standing, we can target our research andsurveillance efforts to spot emerging is-sues as they arise and be prepared to re-spond quickly and appropriately. Such aresponse requires a flexible regulatoryframework that is based on science.

Recent scientific advances have pro-vided tremendous insight into each ofthe three factors in foodborne illness,both separately and in combination.Knowledge of the pathogens themselvesand their interaction with the microbialenvironment creates opportunities forboth prevention and control. Foodborneillness surveillance systems can benefitfrom enhanced understanding of patho-gen evolution to spot new problemsquickly, and, in turn, the outbreak inves-tigation data can provide further insightinto the forces that drive pathogen evolu-tion. Policies based on risk assessmentand Food Safety Objectives (FSOs) willenable us to better consider the impact ofchanging population demographics andconsumer behaviors. This science-based

approach to food safety will require thatthe education of food safety profession-als become more multidisciplinary.These broadly educated professionalswill work in partnership with a widerange of experts across government, in-dustry and academia.

To achieve the maximum benefits,our food safety efforts and policies mustbe carefully prioritized, both in terms ofresearch and in application of controls.As scientific advances provide a betterpicture of pathogenicity, we must decidewhether to focus our efforts on thosepathogens that cause many cases ofmoderate illness or instead focus onthose pathogens with the greatest severi-ty, despite the relatively few number ofcases. In the move toward making deci-sions based on risk, our food safety poli-cies need to weigh these issues, and com-municate information about risk to allstakeholders, including the public.

Strategic Prioritization to ReduceFoodborne Disease

In an ideal world, gaps in data wouldbe quickly filled by data from high quali-ty research. In reality, our research needsare far greater than our willingness tofund research. To maximize our resourc-es, prioritization of research is essential.Similarly, our efforts at control and pre-vention should focus first on areas withthe greatest impact on public health.

Resource Priorities

The Centers for Disease Control andPrevention (CDC) estimates that 76 mil-lion cases of foodborne illness occur an-nually in the United States. A large num-ber of known pathogens are responsiblefor 14 million cases; for the other 62 mil-lion cases, the pathogen causing the ill-ness is not known. Of the known patho-gens, Norwalk-like viruses (9,280,000cases), Campylobacter species (1,963,000cases), and nontyphoid Salmonella(1,332,000 cases) are responsible formost illness, while Trichinella spiralis,Vibrio cholerae, and Vibrio vulnificus arethe known pathogens responsible for the

fewest cases of illness (approximately 50cases each) (Mead et al., 1999).

With the large number of pathogensresponsible for foodborne illnesses andthe apparent lack of a single, all encom-passing solution to foodborne disease,how should a public health organizationdetermine its priorities and distribute itsresources to have the greatest impact onfood safety? The reality is that publichealth priorities have always been influ-enced by a crisis like a recent outbreak orby the concerns of special interestgroups. However, as a society, we need tobalance these influences with a more sys-tematic approach that allocates scarce re-sources to have the greatest impact onfood safety.

Instead, a more strategic approach isneeded. Ranking hazards based on quan-titative hazard analysis—to identify, inorder of importance, those pathogens ofprincipal concern to public health—pro-vides a scientifically based approach forresource allocation. Criteria for suchhazard ranking must be established. Ex-amples of suitable criteria include: inci-dence and severity of illnesses, numberand predisposing conditions of high-riskpopulations, principal risk factors asso-ciated with illness, and prevalence andvirulence of the pathogen.

Efforts to prioritize public health- problems based on more objective crite-ria have been conducted in the past andmay serve as a model for food safety(Murray and Lopez, 1996).

Difficulties exist, however, in weigh-ing various components of public healthimpact when conducting quantitativehazard analysis for ranking pathogens.For example, Salmonella species are re-sponsible for an estimated 1.34 millioncases of illness and 553 deaths, with amortality rate of 0.04%, via a variety offoods of animal and plant origin. Mostcases involve mild diarrhea of only a fewdays duration. The young and elderlypopulations are at greatest risk for severesymptoms. On the other hand, V. vulnifi-cus causes an estimated 47 cases of food-borne illness and 18 deaths annually,principally via raw oysters. Forty percentof the cases involve fulminating septice-


mia that results in death. The populationat greatest risk is people with high levels ofserum iron. These two very different food-borne diseases illustrate the many factorsthat must be considered in prioritizing thehazard ranking. Severity of illness, whileimportant, may or may not be the mostimportant factor in the ranking.

Each agent responsible for foodborneillness has unique characteristics that in-fluence its transmission or ability to causeillness (Doyle et al., 1997). Transmissionof Norwalk-like viruses is controlled bypreventing contamination of food by hu-man feces, whereas transmission ofCampylobacter jejuni is often controlledby preventing contamination of carcassesby poultry feces. Because there are somany different factors influencing patho-gen contamination of foods, no single so-lution can be broadly applied to eliminatefoodborne illness; each agent must be ad-dressed on an individual basis with differ-ent procedures for control applied de-pending on the pathogen.

Creating such a policy frameworkwill not be an easy task. Tailored regula-tory responses that react to newly recog-nized hazards with the best science avail-able at the time may be criticized as pre-mature or arbitrary regulatory enforce-ment that creates uneven economic bur-dens within the food industry. But the al-ternative is waiting until there is signifi-cant scientific information and applyingit in a uniform manner to all foods,whether they pose public health hazardsor not. Although this approach may bepolitically easier, it fails to maximizepublic health protection.

Strategic Control Measures

Because each pathogen must be ad-dressed individually, a strategic approachto applying control measures is neces-sary. Within a strategic approach, inter-vention strategies identify points atwhich control measures will have thegreatest influence on providing safefoods. To identify and rank these points,microbial risk assessments are conduct-ed. The risk assessments involve system-atically collecting and analyzing expo-sure and dose-response data. Case-con-trol studies and other epidemiologic re-search approaches are helpful in identify-ing risk factors in sporadic infectionsand outbreaks.

The U.S. Department of Agriculture(USDA) and the Food and Drug Admin-istration (FDA) have used this general ap-

proach to develop critical food safety in-formation regarding the control of patho-gens in specific foods. For example, riskassessments of Salmonella Enteritidis ineggs, Escherichia coli O157:H7 in groundbeef, V. parahaemolyticus in raw mollus-can shellfish, and Listeria monocytogenesin ready-to-eat (RTE) foods have beendrafted. When sufficient data are avail-able, quantitative risk assessments can:identify what foods are of greatest riskand contribute most to specific foodborneillnesses, estimate the levels of pathogensin foods that are unsafe, and identify whatpoints within the food continuum havethe greatest influence on exacerbating orpreventing foodborne illnesses.

Examples of the use of case-controlstudies to identify risk factors for spo-radic illnesses include E. coli O157:H7,Campylobacter, and Cryptosporidium.Major risk factors associated with spo-radic cases of E. coli O157:H7 infectionin the United States are eating under-cooked ground beef and visiting a farm.Risk factors for E. coli O157:H7 infectionin Scotland are handling/preparing rawfood (40%), being involved in garden-ing/garden play (36%), living on or visit-ing a farm (20%), having direct/indirectcontact with animal manure (17%), hav-ing private water supplies (12%), and re-cent failures with high coliform countsof water supplies (12%) (Coia et al.,1998). Risk factors for sporadic Campy-lobacter infections in the United States,identified in a case-control study of sixFoodNet sites from January 1998through March 1999 involving 1,463 pa-tients with Campylobacter infection and1,317 controls included: foreign travel,eating undercooked poultry, eatingchicken or turkey cooked outside thehome, eating non-poultry meat cookedoutside the home, eating raw seafood,drinking raw milk, living on or visiting afarm, contact with farm animals, andcontact with puppies (Friedman et al.,2000). Risk factors associated withcryptosporidiosis cases in Minnesotafrom July 1-December 31, 1998, wereswimming in public pools (e.g., hotel orschool pools), drinking well water, visit-ing a farm, living on a farm for those lessthan age 6, and exposure to cattle and tomanure for those not living on a farm(Soderlund et al., 2000). The underlyingvehicle largely responsible for transmit-ting these pathogens to humans is con-taminated manure.

The vast quantities of manure pro-duced each year as a by-product of ani-

mal agriculture present a challenge. Cat-tle, hogs, chickens and turkey producedan estimated 1.37 billion tons of ma-nure in 1997 (U.S. Senate AgricultureCommittee, 1998). Because many of themost prominent foodborne pathogensin the United States, including C. jejuni,Salmonella, and E. coli O157:H7, arecarried by livestock and are principallytransmitted to foods by fecal contami-nation, the amount of manure createdin the United States is a growing envi-ronmental threat.

Manure-related food safety issueson the near term horizon include issuesrelated to fresh produce and organicproduce in particular. For example, re-cent outbreaks of E. coli O157:H7 infec-tion and salmonellosis have been asso-ciated with organically produced alfalfaand clover sprouts and mesclun lettuce.Use of contaminated cow manure is amajor concern. The lack of an estab-lished, proven composting protocol toassure elimination of pathogens andprevent recontamination contributes tothis concern. Another issue on the hori-zon is the importation of fruits and veg-etables from countries with poor agri-cultural practices, i.e., use of contami-nated irrigation water, improper prepa-ration and application of manure as fer-tilizer, and harvesting and washing pro-duce under unsanitary conditions.

Food irradiation has received con-siderable attention as a means to ad-dress food safety issues. The suggestion,however, that irradiation is a single so-lution to eliminating most pathogensassociated with fresh or RTE foods lacksfoundation. For some foods, irradiationresults in foods with unacceptable sen-sory characteristics (Olson, 1998). Irra-diation is a tool with broad applicability,but it is not a comprehensive solutionfor all infectious foodborne hazards inall foods.

Emerging Pathogens

Unfortunately, pathogens can be ad-dressed only after they evolve. Consider,for example, the relatively recent identi-fication of E. coli O157:H7. E. coliO157:H7 received relatively little atten-tion from food safety scientists and themedical community for more than a de-cade after its discovery. It was not until1993, following a large outbreak involv-ing more than 700 patients infected byeating undercooked fast-food hamburg-ers, that this pathogen rose to promi-

87EXPERT REPORT

nence as a major food safety issue (Doyleet al., 1997). USDA established a “zerotolerance” policy for E. coli O157:H7 inground beef, the first rule to “outlaw” thepresence of a pathogen in a raw food(Griffin, 1998). Although some im-provement has been made, the policyclearly has not resolved the problem, asillnesses associated with ground beefcontinue to occur. An estimated 73,500cases of E. coli O157:H7 infection (bothfood- and nonfood-related) occur annu-ally in the United States. Many outbreaksare associated with swimming in recre-ational lakes, drinking contaminated wa-ter, handling animals, and consumingcontaminated foods, including alfalfasprouts, lettuce, unpasteurized applejuice, coleslaw, and undercooked groundbeef (Doyle et al., 1997; Griffin, 1998).

Although the emergence of foodbornepathogens similar to E. coli O157:H7 can-not be anticipated, a well-conceived planshould be in place to address these issuesas they arise. Essential information need-ed to assess the significance and likely im-pact of the pathogen as an agent of food-borne disease is often unavailable for ahazard analysis. Scientists need to knowthe pathogen’s reservoir, prevalence, viru-lence, ability to survive in different envi-ronments, and association with humanillness. Also, sensitive methods to detectthe pathogen are often lacking. A frame-work is needed to identify and prioritizethe information required for a hazardanalysis and a subsequent quantitativemicrobial risk assessment. Public healthagencies should be prepared to quicklyobtain the essential information to com-plete a hazard analysis and, depending onthe degree of hazard and available data, arisk assessment.

Outbreak Investigation

A comprehensive system of food-borne disease surveillance must includea system for detecting and rapidly re-sponding to potential outbreaks. Out-breaks caused by specific foodbornepathogens, such as Salmonella and E. coliO157:H7, may be identified by detectingunusual case clusters or increased occur-rence of cases by routine surveillance ofcases reported by medical clinics andclinical laboratories. Recently, molecularsubtyping of isolates by pulsed-field gelelectrophoresis (PFGE) has increasedboth the sensitivity and specificity ofpathogen-specific surveillance for detect-ing outbreaks caused by relatively com-

mon pathogens, such as Salmonella Ty-phimurium (Bender et al., 2001). Rou-tine subtyping and transmission of sub-type patterns through electronic com-munication networks, such as PulseNet,creates the potential to detect widely dis-persed outbreaks that might not be rec-ognized in any individual state (Swami-nathan et al., 2001). Investigation ofthese outbreaks is required to determinethe source and mode of transmission ofthe outbreak-associated strain.

In addition to detecting outbreaksthrough laboratory-based surveillance,outbreaks may also be recognized be-cause of the occurrence of similar ill-nesses among persons who attended anevent or establishment together. Many ofthese outbreaks are recognized before acausative agent has been diagnosed.Thus, investigation of these outbreaksmust be conducted to identify the agentas well as the source and mode of trans-mission.

Outbreak investigations require theclose collaboration of epidemiologists,environmental health specialists and pub-lic health laboratories. Collection of stoolsamples and interviews of ill persons andhealthy comparison groups must be con-ducted rapidly. Epidemiologists must co-ordinate their activities with the publichealth laboratories to maximize the po-tential to isolate the agent. Informationcollected by epidemiologists can helpguide environmental health evaluation ofan establishment and interviews of foodservice workers. Results of environmentalhealth evaluations can further guide epi-demiologic investigations. Epidemiologicdata need to be analyzed and interpretedin light of the results of laboratory testsand environmental investigations. Strate-gies for outbreak control and preventionneed to be identified and implemented assoon as can be justified by the results ofthe investigation. Depending on the scopeof the outbreak and nature of the re-sponse, coordination with other state andlocal agencies, FDA, USDA, and CDCmay be needed.

Most outbreak investigations are ini-tiated by local or state health depart-ments. Because outbreak investigationsare complex activities that need to berapid, thorough, and well-coordinated,CDC issued a report (CDC/NCID/DBMC, 2000) intended to assist stateand local health departments assess theiroutbreak response capacities and to helpguide them in developing and strength-ening their foodborne disease surveil-

lance programs. The core componentsrequired for outbreak investigations in-clude: epidemiology, food protectionprograms, and public health laborato-ries. The essential element to improvingfoodborne outbreak investigations is thecapacity to respond quickly and compre-hensively to the occurrence of suspectedfoodborne illness.

National Initiatives

Foodborne illness has no easy solu-tions. However, major strides can bemade by developing and implementing awell-conceived strategic approach thatprioritizes the hazards and defines thestrategies that will most effectively reducehazards. This approach must include astrategy to address emerging hazards.This strategic approach should be a na-tional initiative that includes state, local,and international involvement, and per-haps reorganizes existing federal foodsafety agencies and programs.

The importance of an expanded sur-veillance system that covers animalhealth and the environment cannot beoverstated. The additional informationfrom an expanded and coordinated sur-veillance system would enable a broadervision of the flow of pathogens and po-tential pathogens throughout the foodchain, and it would fill some importantdata gaps in risk assessment. Coupledwith the new genetic tools, potentialfoodborne pathogens may be detectedbefore they cause confirmed human ill-ness. While it may be politically or eco-nomically impractical to respond vigor-ously to a likely pathogen before casesare identified and linked to the food/pathogen combination, prior knowledgeof the potential pathogen will decreaseresponse time and enable a more appro-priate first response. Expanded surveil-lance will require state and local partici-pation, coupled with leadership and co-ordination at the national level. Thus, ex-panded surveillance should be part ofany national initiative. These activitiesalso are consistent with international ef-forts being planned by the World HealthOrganization (WHO) towards establish-ing a coordinated, expanded, worldwidesurveillance system (Archer, 2001).

Strategies for the Future

The complex interrelationship of thepathogen, host, and microbial ecologyensures a role for everyone in food safety


management: government, industry,and consumers. A flexible, science-based approach that relies on all partiesto fulfill their role is our best weaponagainst emerging microbiological foodsafety issues.

Role of Government and Industry

Developing a strategic, science-basedapproach that prioritizes our resourceswill not be an easy task. Quantitative riskassessment must be based on data, butour current system does not effectivelyencourage data generation and sharing.The regulatory framework must be struc-tured to allow the food industry to gener-ate and share data and information withthe regulatory agencies. In addition, a sci-ence-based program will necessarily in-volve acceptance of some level of risk, be-cause zero risk is not achievable. Using theFSO approach will enable us to translateour public health goals into achievablestandards that are based on science.

Addressing consumer attitudes willpresent a substantial challenge. Natural-ly, we all desire the minimum possiblerisk that can be reasonably achieved.Achieving consensus on an appropriatelevel of risk will be difficult. Risk com-munication and modification of percep-tion and behavior will need to be consid-ered an important part of any move to arisk-based food safety policy.

Food manufacturers must accepttheir role in microbiological food safetyand achieving public health goals. Rapidresponse to a new food safety issue mayrequire investing money for controls be-fore the scientific data are complete. Inexchange for flexibility, food manufac-turers must be willing to work as part-ners with regulatory officials, sharing sci-entific information and data to developappropriate food safety policies.

Developing and implementing thesenew, risk-based policies will require foodsafety professionals with a broad under-standing of many scientific disciplinesand subjects. Changes in how we educatefood safety professionals will ensure theyhave the knowledge and the skills tomaximize the effectiveness of new toolsand methods.

Data Sharing and Cooperation

For risk assessments that are bebased on the best data available andtranslate into the soundest science-baseddecisions possible, ways need to be

found to access data from food manu-facturers. This is far from simple undercurrent conditions. From their ownquality assurance (in-line, and environ-mental) and/or finished product moni-toring programs, manufacturers gatherhuge amounts of data. If they were avail-able, these data could provide valuableexposure information to risk assessorsand information on the prevalence ofpathogens in various food processing en-vironments. Food manufacturers cur-rently do not often share such data oreven collect potentially useful informa-tion, because of potential regulatoryramifications or for product liability rea-sons. Ways must be found to collect andshare this information in a penalty-freemanner.

For example, food producers con-cerned with L. monocytogenes are hesi-tant to test below the genus level, suchthat data show only Listeria spp. Whilesomewhat useful, further speciation andsubtyping could yield even more usefuldata, but the finding of L. monocytogenesin a finished, RTE food would result inregulatory action under current policy ofboth FDA and USDA. Knowing how thepresence of other Listeria species in foodor the processing environment relates tothe possible presence of L. monocytogenesis thus not achievable.

Recent studies question whether allsubspecies of L. monocytogenes are viru-lent, or of equal virulence (Wiedmann etal., 1997). Perhaps this finding, whenfurther developed, will help define moreappropriate policies that foster collectionof good and meaningful data. Addition-ally, knowing that the presence of L.monocytogenes alone may not necessarilymean the food is potentially harmfulmay be an incentive to manufacturers tospeciate further, and to apply methodsthat determine or indicate virulence orlack of virulence.

Interdisciplinary Research. A growingarea in federal research funding has beenthe formation of interdisciplinary teamsto examine complex problems. This shiftfrom more traditional projects with a sin-gle researcher has reached all the majorfederal funding agencies. Some of thesenew interdisciplinary programs have beenhighly successful, most notably in the ar-eas of vaccine development, epidemiolog-ic surveillance, and genome sequencing.These different program structures havecreated new paradigms for the generationand sharing of data that provide addedbenefit for detecting the emergence of

pathogens, spotting new transmissionpatterns, or predicting the effects of newproduction technologies.

Pre-harvest Safety. In the last 10years, several teams have been formed tofocus on the microbiological aspects ofpre-harvest food safety. These teamstypically address animal production butrecently have also targeted produce. Themain emphasis of these teams has beento define the existing problems, typicallywith the use of epidemiologic surveys,and to then develop and test interventionstrategies. A good example is the effortsto develop pre-harvest interventions forE. coli O157:H7 in the beef productionindustry. Such teams usually compriseveterinary microbiologists, veterinarians,food microbiologists, animal scientists,and epidemiologists. The studies usuallyuse epidemiologic surveillance methodsto identify potential intervention pointsin current animal production methods.

These systematically designed studiesgenerate large microbial strain sets. Cur-rently, these strains are often logged andstored, without generating much addi-tional data, save for occasional heroic ef-forts to perform modest genotypingstudies. However, such strain sets andthe associated samples hold much infor-mation about the impact of different fac-tors on populations of pathogens andcommensal organisms. This informa-tion could be mined in collaborationwith population geneticists and genomeresearchers to examine the relationshipsbetween microbial population land-scapes, genome evolution, and ecology.

Sanitation Assurance. Surveillancestudies of pathogens and indicator or-ganisms in food production facilities area part of industry sanitation programs.These programs are designed to identifyin-house events, catching potential haz-ards before they develop or become es-tablished in the production line. A care-fully designed sampling regimen, cou-pled with the appropriate statisticalmethods for data mining, could providea tremendous amount of informationregarding the nature of hazardous eventsand the identification of previously un-known hazards. Moreover, inclusion ofhigh-throughput genome studies onpopulations of pathogens or indicatororganisms would again provide a wealthof added information regarding evolu-tion and ecology of microorganisms infood production settings. In this context,combining information from several dif-ferent producers could provide public

89EXPERT REPORT

health officials with further informationabout geographic and other factors asso-ciated with the emergence and spread ofpopulations of problematic organisms.

Food Safety Education

New integrative educational ap-proaches that directly link the basic andapplied sciences will be necessary to ef-fectively train the future food safety pro-fessional. For instance, tomorrow’s foodsafety professional must be knowledge-able in basic sciences such as microbiol-ogy and toxicology, yet also understandthe entire food safety continuum and beable to address issues wherever they oc-cur along that continuum. And becausemicrobiology and toxicology are inti-mately tied to newer disciplines such asmolecular biology, genomic sciences, andmathematical modeling, extensive sub-discipline training remains crucial.

This new approach will focus onpreparation of professionals who, in ad-dition to expertise in their primary disci-pline, also are grounded in supportingfood safety areas such as veterinary, ag-ronomic, environmental and publichealth practices. Although it is difficultfor a single campus to provide suchbroad training, yet through creative col-laborations, new food safety curricula al-ready are developing. These programsmake aggressive use of distance learningtechnologies; emphasize critical thinkingskills, professional development and eth-ics training; provide practical field expe-rience through summer internships; anddevote special attention to diversity is-sues to create a highly trained and well-represented work force (Jaykus andWard, 1999). Food science departmentsaround the country are uniquely posi-tioned to offer the strongest and mostcomprehensive leadership in the develop-

ment of the integrated graduate foodsafety education programs that will pro-vide the larger, multidisciplinary work-force needed to address emerging foodsafety issues.

Role of Consumer Understanding

For the last 30 years, the dominantfood safety message has been that theUnited States has the world’s safest foodsupply. As a result, most consumers be-lieve that there is in place an extensivesystem of controls applied throughoutthe food production and distributionsystem, guaranteed ultimately by govern-ment oversight, and that this system pro-tects them against well-recognized andemerging foodborne disease. One of theconsequences of this confidence is thatfood safety problems may be seen only asdefects of the system to be fixed bystrengthening the system of controls and

A CooperativeApproach to theSafety of Sprouts

Outbreaks of foodborne illnessassociated with the consumption ofraw vegetable sprouts are a recentlyemerged food safety issue addressedusing a combination of industry andregulatory action.

Sprouts can harbor large popula-tions of microorganisms, because theconditions used for sprouting seedsalso promote rapid microbial growth.If present on the seeds, pathogensgrow to high levels. In the 1980s and1990s, consumption of fresh, un-cooked vegetable sprouts becamepopular, and commercial sprout sup-pliers developed broad distributionsystems. The number of reported ill-nesses increased significantly. In1997-1998 alone, at least 7 docu-mented outbreaks of Salmonella andE. coli O157:H7 infections werecaused by consumption of varioustypes of raw vegetable sprouts. Oneof these outbreaks, which occurred inJapan, was the largest outbreak of E.coli O157:H7 ever recorded (Taormi-na et al., 1999).

The responsiveness and coordinatedefforts of our institutions are criticalfactors in understanding and gainingcontrol of an emerging food safety is-sue. In the case of sprouts, federal andstate government agencies worked co-operatively with industry and academicsectors to respond. CDC and FDA metwith sprout industry representatives in1995 to discuss food safety concerns. In1997, the National Advisory Committeeon Microbiological Criteria for Foods(NACMCF) was asked to review avail-able data and to formulate science-based recommendations to enhance thesafety of sprouts.

A 1998 public meeting gatheredsprout suppliers and trade organiza-tions, consumer groups, academic re-search scientists, federal and state gov-ernment research scientists, and publichealth officials to discuss the currentscientific data and possible food safetystrategies. At that meeting, participantsshared information about epidemiolog-ic and outbreak data, agricultural andsprouting practices, pathogen detectionmethodology, and disinfection and con-trol measures. In that same year, theCalifornia Environmental ProtectionAgency, acting on a request from the

California Department of HealthServices, issued a special local needregistration that allowed seeds to betreated with 20,000 ppm calcium hy-pochlorite to kill pathogens. Federaland state agencies issued consumeradvisories to warn at-risk popula-tions (children, elderly, and thosewith compromised immune systems)to avoid consumption of rawsprouts. Also in 1998, the NationalCenter for Food Safety and Toxicolo-gy, a food safety research consortiumcomprised of industry, FDA and aca-demic participants, formed a SproutsTask Force to identify data gaps andprioritize short-term research goals.

All of these activities, along withthe NACMCF report (1999), formedthe scientific underpinnings for regu-latory guidance (FDA, 1999) with rec-ommendations for the sprout indus-try that were based on the best avail-able scientific information at the time.To reduce safety risks associated withsprouts, the guidance documents pro-moted the combined approaches ofgood agricultural and manufacturingpractices, seed disinfection, and rapidpathogen testing of spent irrigationwater.


government action, rather than also asproblems with a strong component ofconsumer-based risk reduction or riskavoidance.

Studies indicate that 80% of con-sumers think food safety problems aremainly due to failures in food process-ing, food distribution and food prepara-tion in restaurants; in other words, con-sumers believe the failures are occurringin the most regulated parts of the foodsafety system where they have little directresponsibility. Relatively few consumersperceive food safety problems due to ac-tions in the home or in supermarkets—the final stage of the food safety sys-tem—or on farms—the beginning of thesystem (Levy, 1997; Penner et al., 1985;Williamson, 1991). To achieve a trulyfarm-to-table approach to maximizingfood safety, it is important to considerpotential contributions from all seg-ments of the food chain.

Since 1993, the Food Marketing Insti-tute Trends Survey has asked consumersan open-ended question about the great-est threats to food safety to measure top ofthe mind awareness of different possiblesources of food safety problems. Thenumber of respondents who mentionedimproper quality control/shipping/han-dling and storage rose from 9 percent in1993 to 34 percent in 1997. During thesame time period, the number of peoplementioning food preparation declinedfrom 12 percent to 2 percent.

Consumer Behavior

Research indicates that people con-sider themselves fairly knowledgeableabout food safety guidelines, and for themost part they are. However, as in otherareas of health and safety, knowledge andawareness does not always translate intobehavioral changes. Between 1988 and1993, indicators of concern about foodincreased significantly, suggesting anemerging public awareness and interestin food safety problems. At the sametime, data suggest that unsafe food con-sumption and preparation behavior ac-tually increased (Levy, 1997).

Behavioral surveillance systems canprovide data identifying people orgroups in which behaviors associatedwith foodborne diseases are more com-mon and who are at higher risk for food-borne illness, thus assisting in the devel-opment of food safety education pro-grams (Yang et al., 1998). Further, sur-veillance data can be used to evaluate the

progress of education programs (Altek-ruse et al., 1999; Yang et al, 1998). Datacollected through the Behavioral RiskFactor Surveillance Systems during1995-1996, which included 19,356 sur-vey participants, showed that several highrisk food handling, preparation, andconsumption behaviors were common,and some varied by gender, age, race/eth-nicity, education and income. For exam-ple, 50.2% of respondents reported eat-ing undercooked eggs, and 19.7% report-ed eating undercooked hamburgers. Allhigh risk food handling, preparation,and consumption behaviors were moreprevalent in men than in women. Theprevalence of reported consumption ofundercooked hamburgers decreased withage, increased with education, and in-creased with income.

Decisions about behavior frequentlyare guided by risk perception rather thanrisk awareness (Frewer et al., 1994). De-fining risk as “hazard + outrage,” Sand-man (1997) stated that when people mis-perceive hazards it is often because theyare outraged. Sandman noted that gen-erally, even when the hazard is serious,the public is apathetic and the least dan-gerous hazard often generates the great-est outrage. He said, “Too often, expertsfocus on the hazard and ignore the out-rage while the public focuses on the out-rage and ignores the hazard.” Underthese circumstances, the hazard cannotbe mitigated without addressing theiroutrage. Sandman suggested that expertsdetermine why the outrage is high andwhat can be done to lower it so that peo-ple want to hear or acknowledge the ex-tent of the hazard. As an example, Sand-man stated that “consumers know howto cook and generally will get angry ifyou tell them how to do something theyalready know about.”

If people do not recognize and accepttheir role in food safety problems, behav-ior change is unlikely. One way to breakthrough public misconceptions is to de-scribe the magnitude of food safetyproblems and challenge people’s under-standing of themselves as experts. Newdata from the FoodNet surveillance sys-tem may be the best way to challengepeople’s understanding of themselves asexperts (Levy, 1997).

Consumer Education

Recent federal initiatives have soughtto improve the safety of the U.S. foodsupply using a farm-to-table approach,

recognizing that food safety is not onlythe responsibility of the federal govern-ment, but is the shared responsibility ofall components of the food system fromprimary producers to consumers. Con-sumer education about risk reductionwill be a valuable component of an FSOprogram. Consumers will need to under-stand their role in preventing foodborneillness.

Numerous sources provide a wealthof information about food safety andother food-related issues in many for-mats to meet the needs of various audi-ences. Different information sourcesserve different needs, and the effective-ness is not equal.

Consumer Information Sources. Sur-veys indicate that people get most of theirinformation about food safety from elec-tronic and print news media; additionalinformation sources include labels andfood packages, regulatory agencies, andcookbooks (Hingley, 1997; Levy, 1997).

A national educational campaign ofthe Partnership for Food Safety Educa-tion (a public-private partnership of thefederal government, food industry, andconsumer organizations), FightBAC!TM,was created in 1996 to conduct broad-based food safety education designed toreach people of all ages. The FightBAC!TM

campaign has produced multiple educa-tional tools used through many informa-tion channels, i.e., public service an-nouncements, the Internet, point of pur-chase materials, and school and commu-nity outreach.

The National Food Safety Informa-tion Network formed in 1998 by FDA’sCenter for Food Safety and Applied Nu-trition (CFSAN) and USDA’s Food Safe-ty Inspection Service (FSIS) and Na-tional Agricultural Library reaches con-sumers with information on food-relat-ed issues and safe food handling viaUSDA’s Meat and Poultry Hotline, CF-SAN’s Outreach Information Center,USDA/FDA’s Foodborne Illness Educa-tional Information Center, thefoodsafety.gov web site, EdNet (foodsafety educators’ network), and theFoodsafe listserve. Considered the“gateway to government food safety in-formation,” the www.foodsafety.gov website provides the public access to advicepertaining to specific population sub-groups (e.g., children, people with im-mune diseases) as well as product spe-cific advice (e.g., refrigerated RTEfoods). CFSAN and FSIS also have con-ducted public awareness campaigns.

91EXPERT REPORT

For example, CFSAN developed materi-als (press kits, consumer brochure, vid-eo news release, and public service an-nouncement) explaining the risk thatunpasteurized or untreated juices maypose to vulnerable populations. Thematerials were targeted to a variety ofaudiences (senior citizen groups, daycare centers, elementary schools, andPTA offices). Similarly, FSIS launched a“Thermy the Thermometer” campaignin 2000 to encourage use of thermome-ters to ensure sufficient cooking of meatand poultry.

Consumer Trust in Information.Consumers place different levels oftrust in information from differentsources. A survey of more than 1000Americans conducted by the Universityof Kentucky found that more consum-ers “completely trusted” the accuracy ofthe information from government pub-lications and food labels than from anyother source (Buzby and Ready, 1996).Of those respondents who trusted foodsafety information from governmentpublications, 10.8% trusted the accura-cy completely. Of those who trustedfood safety information on food pack-aging and labels, 10.2 percent did socompletely. Consumers’ complete trustof store brochures and advertisementswas lower than other sources, 3% and1.4% respectively. The authors statedthat it is not surprising that these werethe least trusted of the seven sources offood-safety information, because peoplemay feel that advertisers have incentivesto make positive claims about theirproducts.

Addressing the data on how compet-ing motivations, risk perception, andtaste preferences affect hamburger prep-aration, Ralston et al. (2000) reportedthat several information channels ap-pear to be effective for communicatingthe risks of unsafe food preparation.Respondents who said they get their in-formation from magazines, television,cookbooks, or government hotlines had15-17% higher risk motivation, i.e.,were more risk averse, than those whodid not cite these sources of informa-tion. Respondents who said that theyget information from labels did nothave a higher risk motivation index af-ter accounting for other factors that alsoincrease awareness. Consumers whocited brochures as their informationsource had lower risk motivation thanrespondents who did not. Ralston et al.(2000) noted that it is difficult to sepa-

rate the effects of different forms of in-formation because consumers are ex-posed to several information sources si-multaneously and information sourcesmay interact in their effect on percep-tions. The authors concluded that theresults show that consumers who aremore aware of risk from undercookedhamburger are more likely to adopt safebehavior and thus contribute to a re-duction of foodborne disease.

Risk Communication

Risk communication is a necessaryand critical tool to appropriately defineissues and to produce the best risk man-agement decisions (FAO/WHO, 1998).Risk communication has been defined asthe interactive exchange of informationand opinions concerning risk and riskmanagement among risk assessors, riskmanagers, consumers and other interest-ed parties (CAC, 1997b). Others haveadded risk-related factors to the defini-tion, reflecting a wider risk communica-tion concept (FAO/WHO, 1998).

Several factors play a role in under-standing and communicating risk:whether a risk is voluntary or involun-tary, whether the distribution of riskand benefit is equitable, the degree ofpersonal control, the individual dreadof the adverse event, the catastrophicpotential of the event, and the extent oftrust in the risk managers (Covello etal., 1988; NRC, 1989; Sandman, 1987).A Consultation of the Food and Agri-cultural Organization (FAO) of theUnited Nations and WHO (FAO/WHO,1998) described these principles of riskcommunication: (1) know the audience,(2) involve the scientific experts,(3) establish expertise in communica-tion, (4) be a credible source of infor-mation, (5) share responsibility, (6) dif-ferentiate between science and valuejudgment, (7) assure transparency, and(8) put the risk in perspective. Barriersto effective risk communication can oc-cur within the risk analysis process (e.g.,inadequate access to vital informationand inadequate participation of inter-ested parties) or in a broader context(e.g., differing perceptions among par-ticipants, limited understanding of thescientific process, lack of credibility ofthe information source, and societalcharacteristics). Elements of effectiverisk communication include:• the nature of the risk (e.g., its magni-tude and severity);

• the nature of the benefits (e.g., whobenefits and in what ways);• uncertainties in risk assessment (e.g.,the methods used to assess the risk andweaknesses in the available data); and• risk management options (e.g., ac-tion taken to control the risk and actionindividuals may take to reduce personalrisk) (FAO/WHO, 1998).

A National Research Council Com-mittee on Risk Perception and Commu-nication (NRC, 1989) addressed ways toimprove risk communication. The com-mittee noted that it is a mistake to expectthe public to always want simple answersabout risk; often, at least part of the pub-lic desires considerable detailed informa-tion about risks. The committee con-cluded that successful risk communica-tion improves or increases the base of ac-curate information that decision makersuse, whether they are government offi-cials, industry managers, or individualcitizens, and, at the same time, satisfiesthose involved that they are adequatelyinformed within the limits of availableknowledge. Further, the committee ex-plained that because risk communica-tion is tightly linked to the managementof risks, solutions to the problems of riskcommunication often entail changes inrisk management and risk analysis. Inmoving toward risk-based food safetypolicies, risk communication with all in-terested parties, including risk assessors,risk managers, and the public, will be animportant part of the process (NRC,1989).

The FAO/WHO Consultation alsoidentified several considerations forframing risk communication strategies.From an international perspective, ad-dressing the critical role of effective com-munication in determining equivalenceof food control measures in differentcountries is a consideration. From anindustry perspective, labeling is a consid-eration. The consultation recommendedthat if consumer food handling, storageor other practices can assist in control-ling a foodborne illness or disease out-break, clear instructions using unambig-uous language should be presented. Theeffectiveness of labeling as a risk man-agement/communication strategy, how-ever, needs further study. The consulta-tion recommended that labeling—which has been used extensively to con-vey information such as product com-position, nutrition, weights and mea-sures, and health-related warnings—should not be used as a substitute for


consumer education. The consultationalso stated that because national gov-ernments are responsible for food quali-ty and safety and are the primary sourcesfor risk communication with the publicon food safety issues, the capability to ef-fectively communicate risks should beone of the highest priorities for theseagencies.

It is important that risk communica-tion involve effective dialogue, a two-wayexchange, among interested parties(NRC, 1989). Effective dialogue goes be-yond passively providing access to therisk message formation process, e.g., viapro forma public hearings, to includingearly in the process all interested and af-fected groups and comprehending therange of potentially contending view-points (NRC, 1989).

Addressing the importance of broadparticipation in communication of in-terested and affected parties, a NationalResearch Council (NRC) Committee onRisk Characterization reported (NRC,1996) that deliberation is intimatelyconnected with and as important asanalysis in understanding risks. Thecommittee stated that analysis and de-liberation can be thought of as twocomplementary approaches to gainingknowledge of the world, forming under-standings on the basis of knowledgeand reaching agreement among people.Defined by the committee as any formalor informal process for communicationand for raising and collectively consid-ering issues, deliberation is importantin risk decision-making for its role inconsidering conflicts of values and in-terests. A variety of techniques are usedfor deliberation and public participa-tion. These include citizen advisorycommittees and task forces, alternativedispute resolution, citizens juries andpanels, surveys, focus groups, interac-tive technology-based approaches, andcombinations of methods (NRC, 1996).

Risk communication takes place atlocal, national, and international levels.On an international scale, risk commu-nication on food safety occurs withinCodex Alimentarius Commission, itssubsidiary bodies, and its United Nationsparent organizations, FAO and WHO,and their expert advisory groups. TheSanitary and Phytosanitary MeasuresAgreement of the World Trade Organiza-tion encourages harmonization andplaces a strong emphasis on risk com-munication principles of transparencyand consistency in the development and

application of food safety measures(FAO/WHO, 1998) and refers to Codexstandards as international benchmarksfor nations.

Anticipating the Future: FoodSafety Issues on the Horizon

Looking ahead, and considering thecontent of this report, several food safetyissues are likely to come to the forefrontin the next decade.

Globalization of the Food Supply

FDA electronically screened all 2.7million entries of imported foods underits jurisdiction in fiscal year 1997, andphysically inspected 1.7%, or 46,000 en-tries. FSIS visually inspected all 118,000entries of imported meat and poultryunder its jurisdiction in calendar year1997, and conducted further physicalexamination of about 20% of entries.These numbers of entries will only in-crease in the future, and this brings intoquestion how the regulatory agencieswill handle the increases most effective-ly and efficiently. The shortcomings ofsampling and analysis for the ever in-creasing list of pathogens, natural tox-ins, or pesticide and industrial chemi-cals suggest that different approachesmust be sought. HACCP is gaining rec-ognition worldwide as a desirable sys-tem of safety assurance, but for interna-tional trade, mutual recognition ofHACCP systems must be sought.

The demand for year-round freshfruits and vegetables is firmly estab-lished in the United States. Again, thevolume of fresh produce being offeredfor entry into the United States will onlycontinue to grow, as will the variety ofproduce offered. Without mutual un-derstanding and application of good ag-ricultural practices, resulting in mutualrecognition of systems to assure safety,regulatory agencies will be furtherstressed.

Alternative Processing Technologiesand Novel Foods

Novel foods and alternative process-ing technologies will continue to appear.With each new introduction, we mustconsider the possible consequences, in-tended and unintended, within the foodsystem. Some technologies will reduceor eliminate microbiological hazards in-herently present in current food safety

systems. For example, treatment of araw vegetable with a new chemical dis-infectant might eliminate concern overcertain pathogens of manure origin.However, this treatment also might in-advertently select for unidentified mi-croorganisms that were previously in-consequential but that become hazard-ous without normal microbial competi-tion to keep their numbers in check.The complex relationship between vari-ous factors cannot be overlooked.

Similarly, the introduction of anynovel food requires a full analysis to as-sess the potential microbial hazards.The hazard analysis must be broadenough to consider all the intrinsic andextrinsic conditions that influencepathogens known to be associated withfoods, ingredients or processes relatedto the novel food being introduced. Theanalysis also must consider microor-ganisms unique to the new situationthat pose a threat to safety of the novelfood.

Similar considerations are essentialwith the introduction of any alternativetechnology or combination of variousalternative technologies and/or preser-vatives. It is essential to identify thepathogens that are most resistant to thealternative technology, determine mech-anisms of inactivation or control in-cluding required conditions and kinet-ics, identify validation procedures, anddescribe critical process factors.

Scientists continue to be challengedto adequately address all the parametersassociated with the introduction of anovel food or alternative processingtechnology. Once developed, new tech-nologies must be appropriately regulat-ed to ensure their proper applicationand acceptance by the public.

Increases in Organic Foods

Organic foods are becoming main-stream items in most grocery stores, andit is likely that this segment of the freshproduce industry will continue to grow.With or without facts to back up the as-sumption, consumers assume organicproduce is more nutritionally completeand safer than conventionally grownproduce. Recent outbreaks of salmo-nellosis and E. coli O157:H7 infectionassociated with organically producedsprouts and mesclun lettuce grown anddistributed in the United States are evi-dence of an emerging problem (Griffin,1998; Hilborn et al., 1999). Cow ma-

93EXPERT REPORT

nure is a well-documented vehicle forSalmonella and E. coli O157:H7, and itsuse in produce production must becontrolled to prevent contamination.With an estimated 1.2 billion tons ofmanure produced by cattle annually inthe United States (U.S. Senate Agricul-ture Committee, 1998), this voluminoussource of foodborne pathogens is likelyto be an influential factor in the trans-mission of foodborne illness for theforeseeable future. Methods are neededto reduce the shedding of pathogens inlivestock and poultry and to identify ef-fective procedures for eliminatingpathogens in manure before they con-taminate the environment and food.

Changes in Food Consumption

The global trade in food stuffs is onlyone force in people’s changing dietarypatterns. Certainly the variety of avail-able food has expanded drastically andwill continue to do so. Ethnic foods areincreasingly popular, and the percentageof foods prepared and/or consumed out-side the home will continue to rise. Fruitsand vegetables are likely to constitute agreater portion of the average diet, andthe consumer demand for “fresh” prod-ucts will lead to more minimally pro-cessed foods. Our control and preven-tion methods will need to be adapted tothese changing dynamics.

At-Risk Subpopulations

It is likely that the number of personsat higher risk for foodborne diseaseagents will continue to increase with time.The population of the United States is ag-ing, and clearly aging is a risk factor formore serious outcomes from agents such

as Salmonella. As people live longer, theymay develop more chronic underlying ill-nesses that predispose them to foodborneillness. Increasingly complex combina-tions of drugs to treat various conditionsin the elderly can have unpredictable ef-fects on susceptibility. Formerly fatal con-ditions, such as loss of major organ func-tion, are now survivable thanks to organtransplants. The numbers of transplantrecipients will likely increase with time,but it is important to remember that theseindividuals may be among the most sus-ceptible populations to certain foodbornepathogens.

Pathogen Evolution

Microbial evolution has always hap-pened and will always happen. Bacteria,for example, have an enormous capacityfor mutation, integration of new geneticmaterial, and recombination of geneticmaterial in order to assure survival. Bac-teria can sense and react to their envi-ronment and genetically change them-selves in response. Unfortunately, newlyevolved pathogens are first recognizedwhen they cause an outbreak of illness.Using new technologies and genomics,perhaps surveillance of food animalsand the environment for newly emergingmicroorganisms with pathogenic poten-tial will become a reality in the future,and there will be no need to wait for hu-man illness. Another hope for the futureis a better understanding of how humanactivities affect foodborne pathogens.For example, does the cross protectionafforded a pathogen by exposure to anenvironmental insult have a negative im-pact on further processing? Genomicsmay also provide a better snapshot ofhow a microorganism manifests viru-

lence, and even help determine why andwhat to do about it.

Consumer Understanding

Although consumers are only a smallpart of the food safety chain, as consum-ers we all need to take responsibility forour contribution to food safety. Thosethat have not already done so must ac-cept that zero risk is not a reality. Thesetwo concepts may be difficult for someconsumers to accept. Education and riskcommunication will be necessary to pro-vide consumers with a more accurateperception of food safety risks and to en-courage behavior modification, whereneeded.

Integrated Food Safety System

A farm-to-table food safety systemmust involve many interested parties work-ing together toward a common goal. Whenproperly applied, the FSO approach wouldincorporate input from all stakeholders indeveloping the appropriate levels of protec-tion. Although regulatory oversight is nec-essary to monitor and enforce the perfor-mance of the food safety system, foodmanufacturers must play an importantrole as well, because they have first-handinformation about food safety hazards andthe production environment. A partner-ship environment will enhance data shar-ing and provide a solid scientific basis forpolicies. An ideal system identifies hazards,institutes appropriate controls in a flexiblemanner through FSOs, and monitors theoperation of the system. The challenge is tobuild a system that applies science in a pre-dictable, consistent, and transparent man-ner to enable harmonization within andbetween countries.

ConclusionsHistory has demonstrated that science,

when appropriately applied through

food safety management policies, can

dramatically improve food safety. The

past century produced numerous

examples: refrigeration of perishable

foods, pasteurization of milk, and

commercial canning of low acid foods.

Our current level of food safety is the

result of effective implementation on

the part of industry, government, and

consumers. More recent approaches,

such as the development of Hazard

Analysis and Critical Control Point

(HACCP) systems, continue to be

refined as we strive to further improve

food safety.

This report articulates the science be-hind microbiological food safety, espe-cially as it relates to emerging microbio-logical hazards. Interpretation and anal-ysis of this scientific information pro-


vides insight into food safety policiesand management practices. At the sim-plest level, foodborne illness can be re-duced to three factors: the pathogen, thehost, and the environment in which theyinteract. Any efforts to improve foodsafety must address these factors.

Managing microbiological foodsafety is a complex task. Microbiologicalhazards are ever-changing, and theamount and complexity of data and theresidual unknowns are growing at arapid rate. Each new scientific advancegives us the opportunity to add to ourknowledge of foodborne illness usingnew techniques and researching newquestions. At the same time, humansusceptibilities are increasing, and ourability to link food to adverse healthoutcomes is improving. The humanhealth and economic consequences ofemerging microbiological food safety is-sues are immense. To address these cir-cumstances, food safety policies shouldbe developed as part of national initia-tives, with input from all stakeholders.

International coordination of foodsafety efforts should be encouraged.Globalization of the food supply hascontributed to changing patterns offood consumption and foodborne ill-ness. Global food trade has the poten-tial to introduce pathogens to new geo-graphic areas. In addition, we have gen-erally less knowledge about the growingconditions and processing and distribu-tion practices for imported foods thanfor foods produced domestically.

Scientific research has resulted insignificant success in improving foodsafety, but the current science underpin-ning the safety of our food supply is notsufficient to protect us from all theemerging issues associated with thecomplexity of the food supply. The bodyof scientific knowledge must be furtherdeveloped, with our research effortscarefully prioritized to yield the greatestbenefit. Food safety and regulatory poli-cies must be based on science and mustbe applied in a flexible manner to incor-porate new information as it becomesavailable and to implement new tech-nologies quickly. The food industry,regulatory agencies and allied profes-sionals should develop partnerships toimprove food safety management.

Human foodborne disease surveil-lance will continue to be very importantto: (1) identify outbreaks of foodbornedisease so they can be controlled andprevented; (2) determine the causes of

foodborne disease; (3) improve controlstrategies; and (4) monitor trends in oc-currence of foodborne disease. Com-prehensive, coordinated surveillance ac-tivities must be expanded to include an-imal health and the production andprocessing environment. Integratinganimal and environmental surveillancesystems into established human surveil-lance systems will greatly increase ourunderstanding of the epidemiology andsources of foodborne disease.

Enhanced surveillance will providedata that can be used in risk assessment,which is appropriately becoming afoundation for selecting food safetymanagement options. One of the pri-mary limiting factors for quantitativerisk assessment is the quality and suffi-ciency of available data. As an example,there is little information availableabout the relationship between thequantity of pathogen ingested and re-sulting frequency and severity of ad-verse health effects, especially for sus-ceptible subpopulations. Risk assess-ment is an iterative process, and assess-ments must be updated as additionalinformation becomes available. As riskassessments are refined with bettermethods and improved data, their newconclusions must be shared broadlywith all stakeholders to enlighten thepublic debate over appropriate levels ofprotection.

Our scientific understanding of themicrobiology of foodborne pathogenscontinues to improve. Scientists areonly just beginning to understand thefactors that cause a particular microbialstrain to be pathogenic while otherstrains of the same microorganism arenot, the ways by which some microor-ganisms adapt and evolve to becomepathogenic, and the mechanisms patho-gens employ to adapt to differing envi-ronments. Further research is essentialto understand microbial ecology andvirulence sufficiently well to anticipatefuture microbial hazards and constructbarriers to disease.

Some pathogenic microorganismsare significantly more virulent than oth-ers. Virulence may vary within species,subspecies, and even different strains.Understanding the many different viru-lence factors that microorganisms useto cause illness offers opportunities todevelop better controls and therapeu-tics. Further research will enable scien-tists to classify pathogens based on spe-cific virulence factors rather than based

on name, serotype or other traits unre-lated to virulence. This research willimprove our evaluation of safety, whichcurrently is focused on microbes thatmay or may not be pathogenic.

Recent advances in genomics havecontributed to the further understand-ing of virulence at the genus level (e.g.,Salmonella) and at the level of specificstrains within a species (e.g., Escheri-chia coli O157:H7). Genomics also hasgreatly facilitated our understanding ofthe continuous and sometimes rapidevolution of pathogens.

Adverse changes in the microbialenvironment and ecology may causebacteria to experience stress. Althoughmany bacteria die, some may survive,because bacteria have elaborate systemsto adapt to environmental stress. In ad-dition to tolerance of the originalstress, the surviving microorganismsalso may be tolerant to other unrelatedstresses, and these tolerant microor-ganisms may demonstrate increasedvirulence. Understanding these re-sponse mechanisms will provide theinformation necessary to refine foodprocessing conditions or to developother appropriate intervention strate-gies that enhance food safety.

Improved analytical systems areneeded to gather better data aboutpathogens in the food production envi-ronment to improve our understand-ing of the microbial ecology in thesesituations. Sensitive quantitative meth-ods are necessary for assessing patho-gen growth, survival, and inactivation,as well as for accurate risk assessments.

New processing and packagingtechnologies offer the potential forcontinued improvement in the organo-leptic quality of foods, extended shelflife, and enhanced microbiologicalsafety. However, these new processesand packaging technologies maychange the microbial ecology, resultingin potential positive and negative ef-fects that must be assessed along theentire food chain. Even an apparentlyinsignificant change in the microbialenvironment can trigger a food safetyconcern because of the complexity ofthe microbial environment and the in-terrelationship of various factors.

Combinations of food manufactur-ers’ efforts, regulatory programs, andconsumer actions have driven downrates of certain foodborne diseases, butcontinued efforts are necessary. Al-though not easy to accomplish, it is

95EXPERT REPORT

critically important that regulatory pol-icies be based on the best science cur-rently available. Regulatory policiesbased on sampling and testing may in-correctly imply an absence of patho-gens, causing some individuals to as-sume that it is unnecessary to engage inproper food selection and handlingpractices. Given the characteristics ofsome foods, available technologies, andour desire for year-round availability ofa diverse array of foods, it is unlikelythat the marketplace can be made freefrom the presence of pathogenic micro-organisms at all times.

The large-scale production of someready-to-eat (RTE) foods consistentlyfree of Listeria monocytogenes appearspractically impossible. A great deal ofprogress has been made during the pastquarter century to reduce the levels andfrequency of contamination of ready-to-eat foods during their manufacture,but consistently assuring the absence ofListeria has remained out of reach. Thisbacterium is commonly present in theenvironment and is constantly reintro-duced into the processing environmenton raw ingredients and via other means.Listeria survives well in the manufactur-ing and retail environment, and it growsat refrigerator temperatures. At present,a “limit-of-detection standard” existsfor L. monocytogenes in an RTE foodsuch that its mere detection is groundsfor legal action against the companydistributing the food and is the legal ba-sis for removing the food from the mar-ketplace by regulatory agencies. A truefarm-to-table food safety system wouldconsider downstream handling andconsumption patterns and epidemio-logic characteristics of cases; such a sys-tem would not destroy foods that areunlikely to cause illness in the generalpopulation. In addition, some subtypesof L. monocytogenes found in or onfoods have not been associated with ill-ness, and additional research may dem-onstrate that some subtypes are notpathogenic.

Current technologies are also un-likely to consistently satisfy the demandfor large volumes of fresh fruits andvegetables that are free of harmful mi-croorganisms. Because these raw agri-cultural commodities are often con-sumed without cooking, effective inter-ventions are needed that will diminishor prevent the presence of pathogens.Policies that result in minimizing con-tamination and preventing illness from

residual levels of pathogens will bemore likely to achieve public healthgoals than policies that haphazardly in-terdict some small percentage of con-taminated produce.

Although a great deal of progresshas been made in minimizing contami-nation of animal carcasses duringslaughter, the occasional presence ofpathogens on meat and poultry carcass-es is largely unavoidable. Prescribedmicrobial control processes and regula-tory standards—in combination with anumber of other important risk-reduc-ing measures, including educationalprograms—apparently have minimizedthe risk of E. coli O157:H7 infections.Some segments of the marketplace aresuccessfully using strict purchase speci-fications as part of sophisticated qualitycontrol programs. Unfortunately, thiscombination of factors is not in placefor all parts of the marketplace. Con-tinuing to focus on “limit-of-detectionpathogen standards” for some rawmeats in the absence of effective mea-sures in other parts of the farm-to-tablecontinuum is more likely to shift therisk to other parts of the marketplace(such as those with less strict purchasespecifications) than it is to achieve pub-lic health goals. Greater attention topreventing cross-contamination andundercooking may have more impacton the public’s health than further re-ductions in the already small numbersof E. coli O157:H7 occasionally presentin raw ground beef.

Improving the scientific basis offood safety programs will depend onfurther understanding the pathogenicityof microorganisms, including the infec-tious dose of foodborne pathogens un-der a variety of conditions, and a fur-ther reinforcement and implementationof proper hygienic and food handlingpractices of those responsible for pre-venting foodborne disease, includingfood producers and processors, publichealth professionals, retail food prepar-ers, and consumers.

Regulatory agencies should workwith other public health officials and in-terested parties, including industry andconsumers, to establish Food Safety Ob-jectives (FSOs). FSOs offer a means toconvert public health goals into valuesor targets that can be used by regulatoryagencies and food manufacturers. FSOs,which can be applied throughout thefood chain, specify the level of hazardthat would be appropriate at the time a

food is consumed. FSOs would enablefood manufacturers to design processesthat provide the appropriate level ofcontrol and that could be monitored toverify effectiveness. Establishing FSOs isa societal issue that will require inclu-sive participation of all sectors of soci-ety.

The FSO approach can be used tointegrate risk assessment and currenthazard management practices into aframework that achieves public healthgoals in a science-based, flexible man-ner. FSOs help translate the outcome ofrisk assessment into something that canbe used with HACCP programs. TheFSO approach will be successful whendirectly intertwined with a food proces-sor’s good manufacturing practices(GMPs) and HACCP systems.

Although HACCP is a science-basedmanagement tool, HACCP may not bean appropriate approach for all circum-stances. It is not possible to have a validHACCP plan when a scientific analysisdoes not identify any point that meetsthe critical control point criteria. HAC-CP implementation must remain flexi-ble to incorporate the scientific knowl-edge and data available in a product-and process-specific manner that bestmeets FSOs.

The application of HACCP to pri-mary production is particularly limited,because all the HACCP principles gen-erally cannot be achieved. Well-defined,science-based good agricultural practic-es should be further developed for spe-cific commodities where appropriate.Additional research will be necessary tobetter understand the microbial ecologyin these agricultural environments andto formulate science-based recommen-dations for pathogen control.

Routine microbiological testing isuseful for some purposes but not forothers. It can focus on pathogens of in-terest or on nonpathogenic microor-ganisms whose presence indicates con-ditions favorable to the presence ofpathogens. Testing is useful for surveil-lance and HACCP verification purpos-es. It also is used for validating and re-validating processes.

However, microbiological testing offinished product can be misleading, be-cause negative results do not ensuresafety. Testing has statistical limitationsbased on the amount of product sam-pled, the percentage of product that iscontaminated, and the uniformity of thedistribution of contamination through-


out the food. As the amount of contam-ination in the food decreases, the foodsafety emphasis should focus on fur-ther controlling processing conditionsthrough the application of science-based HACCP plans.

Because fresh produce undergoesvery little processing, preventing con-tamination is the primary emphasis forensuring the microbiological safety offresh fruits and vegetables. Thus, care-ful consideration must be given togrowing conditions—including soil,water, manure, livestock, wildlife, pets,environmental pollution and effluent/sewage, and humans—and their effecton food safety throughout the foodchain.

Many of the most prominent food-borne pathogens in the United Statesare carried by livestock and are princi-pally transmitted to food by fecal con-tamination. Manure, a significant vehi-cle for pathogens, is a growing sourceof fertilizer. Use of manure fertilizer isan increasing environmental concernbecause it may contaminate water fordrinking, irrigation, and recreation.Manure also is applied with or withoutcomposting to the soil used to growfood crops. Manure used in the pro-duction of food crops is of special con-cern because the available scientific in-formation is insufficient to ensure thatfoodborne pathogens are killed by cur-rent agricultural practices. Intensivefarming practices can contribute to therapid spread of human and animalpathogens by creating more concen-trated environments for pathogens tomultiply and evolve and by generatinglarger quantities of subsequently con-taminated food.

An examination of the science re-veals that foodborne illness is causedby a complex combination of factorsthat must be managed on a continualbasis. To achieve our public healthgoals, everyone along the farm-to-tablecontinuum must take responsibility fortheir role in food safety management.

Foodborne disease is widely recog-nized for acute effects on the gas-trointestinal tract but also includesother effects throughout the body. Inaddition, foodborne pathogens maycause chronic disease, which may occurindependently or accompany an acuteillness. Many of these chronic diseaseshave only recently been linked to food-borne pathogens. In addition, a grow-ing proportion of foodborne illnesses

are due to viruses, and improvementsare needed in testing methods for viralpathogens in patients and in foods.

The range of pathogens associatedwith foodborne illness continues to in-crease as new information identifiespathogen/food associations. When newfood vectors are identified, risk man-agement decisions must consider thebest approach to control and preven-tion. Application of controls duringfood production and processing may benecessary, although some hazards maybe better addressed at the consumerlevel through modification of exposureor susceptibility.

The person who serves as the hostfor the foodborne pathogen is a majorfactor in the occurrence and characterof foodborne disease. The individual’shealth, food consumption habits, andsanitation practices all directly affectthe risk of foodborne illness. Hygiene,food preparation, and food handlingand storage practices contribute topathogen exposure. Food selection alsocontributes to the likelihood of expo-sure. In addition, an individual’s under-lying health can have a significant im-pact on susceptibility to disease whenexposed.

An important contributor to micro-bial pathogenicity and human illness isthe changing human population and itsbehavior. The elderly portion of theU.S. population continues to grow, andlarge numbers of individuals have con-ditions necessitating the use of immun-osuppressive drugs or drug combina-tions with unknown effects, potentiallyincreasing their susceptibility to food-borne illnesses. Many factors causechanges in the immune system func-tion, such as age, health conditions (e.g.,AIDS, cancer), pregnancy, nutritionalstatus, and antacid/medication use.Factors that suppress the immune sys-tem increase the risk of foodborne ill-ness.

The increased understanding of in-testinal microflora and the immunesystem is providing opportunities forintervention strategies, such as probiot-ics, to facilitate human health and de-crease susceptibility to foodborne ill-nesses.

The combination of proper hygieneand sanitation related to food handlingand preparation, appropriate methodsof refrigeration and freezing, and thor-ough cooking of foods comprise a veryeffective approach to preventing food-

borne illness. After GMPs and HACCPprovide adequately safe foods, the indi-viduals preparing the food must useproper knowledge, attitudes, skills andpractices to achieve food safety.

Consumers are sometimes inatten-tive to their personal ability to reducethe risk of foodborne illness. The publichealth community has the responsibilityto discuss risk reduction strategies withconsumers. Current risk communica-tion is inadequate, and some consumerperceptions and behaviors are not con-sistent with reasonable expectations re-garding the safety of some foods. Com-munication with consumers to improvefood choices and handling practices willbe an essential component of strategiesfor the further reduction of foodborneillness. This approach has been success-ful in the education of sensitive popula-tions, an activity that will necessarilycontinue in the future.

Scientific data are a very substantiallimiting factor in enhancing food safety.Further research will continue to helpresolve complex problems and to pro-vide information to improve the deliv-ery of safe foods. Appropriate and ag-gressive data collection throughout thefood production and processing systemis essential for valid risk assessmentsand the resulting food safety improve-ments. Procedures must be implement-ed to obtain data from food manufac-turers in “penalty-free” environments sothe data can be properly evaluated bypublic officials and the results madeavailable to all interested parties.

It is difficult to conceive of a foodsafety system that responds effectivelyand efficiently to emerging microbio-logical food safety concerns that doesnot permit rapid changes in approachbased on advances in science. Flexibilityto respond to new information and newhazards will require unfettered datasharing. In addition, such a system can-not rely on the use of prescribed micro-bial control processes but instead mustemphasize validation and verification ofthe methods used to assure food safety.

The complex interrelationship of thepathogen, host, and microbial ecologyensures a role for everyone in food safe-ty management—industry, regulatoryagencies, public health officials, andconsumers. A flexible, science-basedapproach that relies on all parties tofulfill their role is our best weaponagainst emerging microbiological foodsafety issues.

97EXPERT REPORT

ReferencesA

AAP. 1998. Prevention of rotavirus disease: Guide-lines for use of rotavirus vaccine. American Academyof Pediatrics. Pediatr. 102: 1483-1491.

AAP. 2000. Potentially contaminated food products. In“Red Book: Report of the Committee on InfectiousDiseases.” 25th Ed. ed. L.K. Pickering. pp. 770-771. American Academy of Pediatrics, Elk Grove Vil-lage, Ill.

Abbas, A.K., Lichtman, A.H., and Pober, J.S. 1997.Cellular and Molecular Immunology. Saunders, Phila-delphia, Pa.

Abee, T. and Wouters, J.A. 1999. Microbial stress re-sponse in minimal processing. Intl. J. Food Microbi-ol. 50: 65-91.

Ackers, M., Mahon, B., Leahy, E., Damrow, T., Hut-wagner, L., Barrett, T., et al. 1996. An outbreak ofEscherichia coli infections associated with leaf lettuceconsumption, Western Montana. Proc. of the 36thInterscience Conference on Antimicrobial Agents andChemotherapy, (Abstr. K43) Sep. 15-18, New Or-leans, La. ASM Press. Washington D.C.

Albrecht, J.A. 1995. Food safety knowledge andpractices of consumers in the U.S.A. J. ConsumerStudies Home Econ. 19: 119-34.

Allos, B.M. 1997. Association between Campylobacterinfection and Guillain-Barré syndrome. J. Infect. Dis.175: 125S-128S.

Altekruse, S.F., Street, D.A., Fein, S.B., and Levy, A.S.1995. Consumer knowledge of foodborne microbialhazards and food-handling practices. J. Food Pro-tect. 59: 287-294.

Altekruse, S.F., Yang, S., Timbo, B.B., and Angulo, F.J.1999. A multi-state survey of consumer food-han-dling and food-consumption practices. Am. J. Prev.Med. 16: 216-221.

Altuvia, S., Almirón, M., Huisman, G., Kolter, R., andStorz, G. 1994. The dps promoter is activated byOxyR during growth and by IHF and ss in stationaryphase. Molec. Microbiol. 13: 265-272.

Anderson, D.M. 1994. Red tides. Sci. Am. 271: 62-68.

Angulo, F., Johnson, K., Tauxe, R., and Cohen, M.2000. Significance and sources of antimicrobial-resistant nontyphoidal Salmonella infections in theUnited States. Microbial Drug Resistance 6(1):77-83.

Anonymous. 1996. Salmonella typhimurium infectionsin humans. Commun. Dis. Rep. 6(41): 356.

Anonymous. 2000. Significance of coliforms and ge-neric Escherichia coli on fresh and minimally pro-cessed produce. NFPA White Paper. National FoodProcessors Association, Washington, D.C.

Arbeit, R.D. 1995. Laboratory procedures for the epi-demiologic analysis of microorganisms. In: “Manualof Clinical Microbiology,” 6th Ed. ed. P.R. Murray,E.J. Baron, M.A. Pfaller, F.C. Tenover, and R.H.Yolken. pp. 190-208. ASM Press. Washington, D.C.

Archer, D.L. 1996. Preservation microbiology andsafety: Evidence that stress enhances virulence andtriggers adaptive mutations. Trends Food Sci. Tech-nol. 71: 91-95.

Archer, D.L. 2001. WHO holds food safety strategicplanning meeting. Food Technol. 55(5): 22.

Ayres, J.C. 1955. Microbiological implications in thehandling, slaughtering, and dressing of meat animals.Adv. Food Res. 6: 109-161.

BBaden, D.A., Fleming, L.E., and Bean, J.A. 1995. Ma-

rine toxins. Handbook of Clinical Neurology 21: 141-175.

Baer, J.T., Vugia, D.J., Reingold, A.L., Aragon, T., An-gulo, F.J., and Bradford, W.Z. 1999. HIV infection asa risk factor for shigellosis. Emerg. Infect. Dis. 5(6):820-823.

Baggesen, D.L., Sandvang, D., and Aarestrup, F.M.2000. Characterization of Salmonella enterica serovartyphimurium DT104 isolated from Denmark andcomparison with isolates from Europe and the UnitedStates. J. Clin. Microbiol. 38: 1581-1586.

Baird-Parker, A.C. 1990. Foodborne salmonellosis.The Lancet 336: 1231-1234.

Baird-Parker, A.C. 2000. The production of microbio-logically safe and stable foods. Chpt. 1 In “The Mi-crobiological Safety and Quality of Food,” Vol. 1, pp.3-18. Aspen Publishers, Inc., Gaithersburg, Md.

Banks, J.G. and Board, R.G. 1983. The incidence andlevel of contamination of British fresh sausages andingredients with salmonellas. J. Hygiene 90: 213-223.

Barkate, M.L., Acuff, G.R., Lucia, L.M., and Hale, D.S.1993. Hot water decontamination of beef carcassesfor reduction of initial bacterial numbers. Meat Sci.35: 397-401.

Barker, C. and Park, S.F. 2001. Sensitization of Listeriamonocytogenes to low pH, organic acids, and osmot-ic stress by ethanol. Appl. Environ. Microbiol. 67:1594-1600.

Barrett, T.J., Lior, H., Green, J.H., Khakhria, R., Wells,J.G., Bell, B.P., Greene, K.D., Lewis, J., and Griffin,P.M. 1994. Laboratory investigation of a multistatefoodborne outbreak of Escherichia coli O157: H7 byusing pulsed field gel electrophoresis and phage typ-ing. J. Clin. Microbiol. 32(12): 3013-3017.

Barton, C.C., Ganey, P.E., and Roth, R.A. 2000. Li-popolysaccharide augments aflatoxin B1-induced liverinjury through neutrophil-dependent and -independentmechanisms. Toxicol. Sci. 58: 208-215.

Barton, C.C., Barton, E.X., Ganey, P.E., Kunkel, S.L.,and Roth, R.A. 2001. Bacterial lipopolysaccharideenhances aflatoxin B1 hepatotoxicity in rats by amechanism that depends on tumor necrosis factor al-pha. Hepatol. 33: 66-73.

Bauman, H.E. 1992. Introduction to HACCP. In “HAC-CP Principles and Applications.” Ed. M.D. Piersonand D.A. Corlett, Jr., Van Nostrand Reinhold. NewYork, N.Y.

Bäumler, A.J., Hargis, B.M., and Tsolis, R.M. 2000.Tracing the origins of Salmonella outbreaks. Science287: 50-2.

BC/USDC. 1996. Per capita utilization of selected com-mercially produced fresh fruits and vegetables: 1970to 1994. Statistical Abstract of the United States,116th Ed. p. 148. Bureau of the Census/U.S. Dept.of Commerce. U.S. Government Printing Office. Wash-ington, D.C.

Bean, N.H., Griffin, P.M., Goulding, J.S., and Ivey, C.B.1990. Foodborne disease outbreaks, 5-year summa-ry, 1983-1987. CDC Surveill. Summ. 39: 15-57.

Bean, N.H., Goulding, J.S., Lao, C., and Angulo, F.J.1996. Surveillance for foodborne disease outbreaks-United States, CDC Surveill. Summ. 45: 1-66.

Bean, N.H., Goulding, J.S., Daniels, M.T., and Angulo,F.J. 1997. Surveillance for foodborne disease out-

breaks-United States, 1988-1992. J. Food Protect.60: 1265-1286.

Becker, L.A., Cetin, M.S., Hutkins, R.W., and Benson,A.K. 1998. Identification of the gene encoding the al-ternative sigma factor sB in Listeria monocytogenesand its role in osmotolerance. J. Bacteriol. 180:4547-4554.

Becker, K.M., Moe, C.L., Southwick, K.L., and Mac-Cormack, J.N. 2000. Transmission of Norwalk virusduring a football game. N. Eng. J. Med. 343: 1223-1227.

Bej, A.K. and Mahbubani, M.H. 1994. Detection ofFood borne Microbial Pathogens. Chpt. X In “PCRTechnology: Current Innovations”. p. 341-365. CRCPress, Boca Raton, Fla.

Beletshachew, S., Yang, S., Cieslak, P., Vugia, D., Mar-cus, R., Koehler, J., Deneen, V., Angulo, F., and theFoodNet Working Group. 2000. Prevalence of high-risk food consumtion and food-handling practicesamong adults: A multistate survey, 1996 to 1997. J.Food Protect. 63: 1538-1543.

Beltrametti, F., Kresse, A.U., and Guzmán, C.A. 1999.Transcriptional regulation of the esp genes of entero-hemorrhagic Escherichia coli. J. Bacteriol. 181:3409-3418.

Bender, J.B., Hedberg, C.W., Boxrud, D.J., Besser,J.M., Wicklund, J.H., Smith, K.E., and Osterholm,M.T. 2001. Molecular subtype surveillance of Salmo-nella typhimurium, Minnesota 1994-1998. N. Engl. J.Med. 344: 189-195.

Benson, A.K., and Haldenwang, W.G. 1993. Bacillussubtilis sB is regulated by a binding protein (RsbW)that blocks its association with core RNA polymerase.Proc. Natl. Acad. Sci. 90: 2330-2334.

Bergthorsson, U. and Ochman, H. 1998. Distributionof chromosome length variation in natural isolates ofEscherichia coli. Molec. Biol. Evol. Jan. 15(1): 6-16.

Bernet, M.F., Brassart, D., Neeser, J.R., and Servin,A.L. 1994. Lactobacillus acidophilus LA1 binds tocultured human intestinal cell lines and inhibits cell at-tachment and cell invasion by enterovirulent bacteria.Gut 35: 483-489.

Besser, R.E., Lett, S.M., Weber, J.T., Doyle, M.P., Bar-rett, T.J., Wells, J.G., and Griffin, P.M. 1993. An out-break of diarrhea and hemolytic uremic syndromefrom Escherichia coli O157:H7 in fresh pressed applecider. J. Am. Med. Assn. 269: 2217-2220.

Besser, T.E., Goldoft, M., Pritchett, L.C., Khakhria, R.,Hancock, D.D., Rice, D.H., Gay, J.M., Johnson, W.,and Gay, C.C. 2000. Multiresistant Salmonella Typh-imurium DT104 infections of humans and domesticanimals in the Pacific Northwest of the United States.Epidemiol. Infect. 124: 193-200.

Beuchat, L.R. 1996. Pathogenic microorganisms asso-ciated with fresh produce. J. Food Protect. 59: 204-216.

Beuchat, L.R. and Ryu, J.-H. 1997. Produce handlingand processing practices. Emerg. Infect. Dis. 3: 459-465.

Bhat, R.V., Beedu, S.R., Ramakrishna, Y., and Munshi,K.L. 1989. Outbreak of trichothecene mycotoxicosisassociated with consumption of mould-damagedwheat production in Kashmir Valley, India. Lancet 1:35-37.

Bidawid, S., Farber, J.M., and Sattar, S.A. 2000. Inacti-vation of hepatitis A virus (HAV) in fruits and vegeta-bles by gamma irradiation. Intl. J. Food Microbiol.57: 91-97.


Bille, J. 1988. Anatomy of a foodborne listeriosis out-break. Foodborne Listeriosis. Proc. of a Symposium.Sept. 7, pp. 30-35. Wiesbaden, Germany.

Bingen, E.H., Denamur, E., and Elion, J. 1994. Use ofribotyping in epidemiological surveillance of nosoco-mial outbreaks. Clin. Microbiol. Rev. 7(3): 311-327.

Blais, B.W., Turner, G., Sooknanan, R., and Malek, L.T.1997. A nucleic acid sequence-based amplificationsystem for detection of Listeria monocytogenes hlyAsequences. Appl. Environ. Microbiol. 63: 310-313.

Bloomfield, S.F., Stewart, G.S.A.B., Dodd, C.E.R.,Booth, I.R., and Power, E.G.M. 1998. The viable butnon-culturable phenomenon explained? Microbiol.144: 1-3.

Bogdanos, D.P., Mieli-Vergani, G., and Vergani, D.2000. Virus, liver and autoimmunity. Dig. Liver Dis.32(5): 440-446

Bondy, G.S. and Pestka, J.J. 2000. Immunomodulationby fungal toxins. J. Toxicol. Environ. Health B. Crit.Rev. 3(2): 109-143.

Borgdorff, M.W. and Motarjemi, Y. 1997. Surveillanceof Foodborne Diseases: What Are the Options? WorldHealth Statistics Quarterly 50: 12-23.

Bower, C.K. and Daeschel, M. A. 1999. Resistance re-sponses of microorganisms in food environments.Int. J. Food Microbiol. 50: 33-44.

Breeuwer, P. and Abee, T. 2000. Assessment of viabilityof microorganisms employing fluorescence tech-niques. Intl. J. Food Microbiol. 55: 193-200.

Brown, J.R. and Doolittle, W.F. 1997. Archaea and theprokaryote-to-eukaryote transition. Microbiol. Mol.Biol. Rev. 61(4): 456-502.

Buchanan, R.L. 1997a. Identifying and controllingemerging foodborne pathogens: Research needs.Emerg. Infect. Dis. 3(4): 517-521.

Buchanan, R.L. 1997b. National Advisory Committeeon Microbiological Criteria for Foods: Principles of riskassessment for illnesses caused by foodborne biologi-cal agents. J. Food Protect. 60: 1417-1419..

Buchanan, R.L., Edelson, S.G., and Boyd, G. 1999. Ef-fects of pH and acid resistance on the radiation resis-tance of enterohemorrhagic Escherichia coli. J. FoodProtect. 62: 219-228.

Buchko, S.J., Holley, R.A., Olson, W.O., Gannon, V.P.,and Veira, D.M. 2000. The effect of different grain di-ets on fecal shedding of Escherichia coli O157: H7by steers. J. Food Protect. 63: 1467-1474.

Buchmeier, N.A. and Heffron, F. 1990. Induction of Sal-monella stress proteins upon infection of macrophag-es. Science. 248: 730-732.

Burke, S. 1985. “Human Anatomy and Physiology forthe Health Sciences.” 2nd Ed. John Wiley and Sons,New York, N.Y.

Burkholder, J.M., Glasgow, H.B., and Hobbs, C.W.1995. Fish kills linked to a toxic ambush-predator di-noflagellate: Distribution and environmental condi-tions. Marine Ecology Progress Series 124: 43-61.

Burnett, S.L., Chen, J., and Beuchat, L.R. 2000. At-tachment of Escherichia coli O157:H7 to the surfacesand internal structures of apples as detected by con-focal scanning laser microscopy. Appl. Environ. Mi-crobiol. 66: 4679-4687.

Bush, E.J., Lautner, E., McKean, J., and Miller, L.M.1990. SwineFutures Project, Final Report. 91-51-048. U.S. Dept. Agriculture/Animal and Plant HealthInspection Service, Washington, D.C.

Buzby, J.C. and Ready, R.C. 1996. Do consumerstrust food-safety information? Food Rev. Jan.-Apr.,pp. 46-49.

CCAC. 1997a. Principles for the establishment and appli-

cation of microbiological criteria for foods. In “GeneralRequirements (Food Hygiene): Supplement to Volume1B – 1997,” 2nd Ed., p. 27. CAC/GL 21-1997.

Joint FAO/WHO Food Standards Prgoramme. CodexAlimentarius Commission, Rome.

CAC. 1997b. Report of the 22nd Session of the CodexAlimentarius Commission. Joint FAO/WHO Food Stan-dards Programme. ALINORM 97/37. Codex Alimen-tarius Commission, Rome.

CAC. 1997c. Hazard Analysis and Critical Control Point(HACCP) System and Guidelines for Its Application. In“General Requirements (Food Hygiene): Supplementto Volume 1B-1997,” 2nd Ed., pp. 33-46. Annex toCAC/RCP1-1969, Rev. 3(1997). Joint FAO/WHOFood Standards Programme. Codex AlimentariusCommission, Rome.

CAC. 1999. Principles and guidelines for the conduct ofmicrobiological risk assessment. CAC/GL-30. CodexAlimentarius Commission, Rome.

Canada, J.C. and Strong, D.H. 1964. Clostridium per-fringens in bovine livers. J. Food Sci. 29: 862-864.

CAST. 1994. Foodborne pathogens: Risks and Conse-quences. Council for Agricultural Science and Technol-ogy, Ames, Ia.

CAST. 2001. Johne’s disease in cattle. Issue Paper No.17. Council for Agricultural Science and Technology,Ames, Ia.

Castillo, A., Lucia, L.M., Goodson, K.J., Savell, J.W.,and Acuff, G.R. 1999. Decontamination of beef car-cass surface tissue by steam vacuuming alone andcombined with hot water and lactic acid sprays. J.Food Protect. 62: 146-151.

Catterall, W.A. 1985. The voltage sensitive sodiumchannel: A receptor for multiple neurotoxins. In “ToxicDinoflaellates.” Ed. D.M. Anderson, A.W. White, andD.G. Baden. pp. 329-342. Elsevier, New York, N.Y.

CDC. 1997. Escherichia coli O157:H7 infections asso-ciated with eating a nationally distributed commercialbrand of frozen ground beef patties and burgers-Col-orado, 1997. Centers for Disease Control and Pre-vention, Atlanta, Ga. MMWR 46: 777-778.

CDC. 1998a. Multistate surveillance for food-handling,preparation, and consumption behaviors associatedwith foodborne diseases: 1995 and 1996 BRFSSfood-safety questions. Centers for Disease Controland Prevention, Atlanta, Ga. MMWR. 47(SS-4): 33-54.

CDC. 1998b. Outbreak of Vibrio parahaemolyticus in-fections associated with eating raw oysters-PacificNorthwest, 1997. Centers for Disease Control andPrevention, Atlanta, Ga. MMWR. 47: 457-462.

CDC. 1998c. Multistate outbreak of salmonella serotypeagona infections linked to toasted oats cereal-UnitedStates, April-May, 1998. Centers for Disease Controland Prevention, Atlanta, Ga. MMWR. 47(22): 462-464.

CDC. 1999a. Foodborne diseases active surveillancenetwork (FoodNet): Population survey atlas of expo-sures: 1998-1999. Centers for Disease Control andPrevention, Atlanta, Ga.

CDC. 1999b. Outbreak of Vibrio parahaemolyticus in-fection associated with eating raw oysters and clamsharvest from Long Island Sound-Connecticut, NewJersey, and New York, 1998. Centers for DiseaseControl and Prevention, Atlanta, Ga. MMWR. 48: 48-51.

CDC. 1999c. Outbreaks of Shigella sonnei infection as-sociated with eating fresh parsley: United States andCanada. Centers for Disease Control and Prevention,Atlanta, Ga. MMWR. 48: 285-9.

CDC. 1999d. PulseNet: The national molecular subtyp-ing network for foodborne disease surveillance. Cen-ters for Disease Control and Prevention, Atlanta, Ga.FoodNet News 1(3): 3.

CDC. 2000a. Preliminary FoodNet data on the inci-dence of foodborne illnesses- selected sites, 1999.Centers for Disease and Prevention, Atlanta, Ga.MMWR. 49: 201-205.

CDC. 2000b. Erratum. Centers for Disease and Preven-

tion, Atlanta, GA. MMWR. 49: 286.CDC/NCID/DBMC. 2000. Essential epidemiology and

laboratory components of a state foodborne diseaseprevention and control program. Letter to state andterritorial epidemiologists and state and territorial pub-lic health laboratory directors from Director, Divisionof Bacterial and Mycotic Diseases. National Center forInfectious Disease/Centers for Disease Control andPrevention. Feb. 7. Atlanta. Ga.

CDC. 2001. Norwalk-like viruses: Public health conse-quences and outbreak management. Centers for Dis-ease Control and Prevention, Atlanta, Ga. MMWR.50(RR09): 1-18

CDC. 2002. Toxoplasmosis. A Fact Sheet of the Cen-ters for Disease Control and Prevention, Atlanta, Ga.(http://www.cdc.gov/ncidod/dpd/parasites/toxoplas-mosis/factsht_toxoplasmosis.htm).

Celum, C.L., Chaisson, R.E., Rutherford, G.W., Barn-hart, J.L., and Echenberg, D.F. 1987. Incidence ofsalmonellosis in patients with AIDS. J. Infect. Dis. Dec.156(6): 998-1002.

Cerda, H., Delincee, H., Haine, H., and Rupp, H. 1997.The DNA comet assay as a rapid screening techniqueto control irradiated food. Mut. Res. 375: 167-181.

Cetinkaya, Y., Falk, P., and Mayhall, C.G. 2000. Vanco-mycin-resistant enterococci. Clin. Microbiol. Rev. 13:686-707.

Cha, J.J., Srivastava, R., Vakharia, V.N., Rao, G., andBentley, W.E. 1999. Green fluorescent protein as anoninvasive stress probe in resting Escherichia colicells. Appl. Environ. Microbiol. 65: 409-414.

Cheville, A.M., Arnold, K.W., Buchrieser, C., Cheng, C.-M., and Kaspar, C.W. 1996. rpoS regulation of acid,heat, and salt tolerance in Escherichia coli O157:H7.Appl. Environ. Microbiol. 62: 1822-1824.

Chomel, B.B., DeBess, E.E., Mangiamele, D.M., Reilly,K.F., Farver, T.B., Sun, R.K., and Barrett, L.R. 1994.Changing trends in the epidemiology of brucellosis inCalifornia from 1973-1992: A shift to foodbornetransmission. J. Infect. Dis. 170: 1216-1223.

Christman, M.F., Storz, G., and Ames, B.N. 1989.OxyR, a positive regulator of hydrogen peroxide-in-ducible genes in Escherichia coli and Salmonella typh-imurium, is homologous to a family of bacterial regu-latory proteins. Proc. Natl. Acad. Sci. 86: 3484-3488.

Clark, C.G., Kraveta, A.N., Dendy, C., Wang, G., andTyler, K.D. 1998. Investigation of the 1994-5 Ukraini-an Vibrio cholerae epidemic using molecular methods.Epidemiol. Infect. 121: 15-29.

Clemmer, D.I., Hickey, J.L.S., Bridges, J.F., Schliess-mann, D.J., and Shaffer, M.F. 1960. Bacteriologicstudies of experimental air-borne salmonellosis inchicks. J. Infect. Dis. 6: 197-210.

Cohen, B.A. 1999. Neurological manifestations of toxo-plasmosis in AIDS. Semin. Neurol. 19(2): 201-11.

Coia, J.E., Sharp, J.C., Campbell, D.M., Curnow, J.,and Ramsay, C.N. 1998. Environmental risk factorsfor sporadic Escherichia coli O157 infection in Scot-land: Results of a descriptive epidemiology study. J.Infect. 36(3): 317-21.

Collins, J.L., Small, M.L., Kann, L., Pateman, B.C.,Gold, R.S., and Kolbe, L.J. 1995. School health ed-ucation, J. Sch. Health 65: 302-11.

Colwell, R.R., Brayton, P.R., Grimes, D.J., Roszak,D.B., Huq, S.A., and Palmer, L.M. 1985. Viable butnon-culturable Vibrio cholerae and related pathogensin the environment: Implications for release of geneti-cally engineered microorganisms. Bio/Technol. 3:817-820.

Cook, K.A., Boyce, T., Langkop, C., Kuo, K., Swartz,M., Ewert, D., et al. 1995. Scallions and shigellosis: Amultistate outbreak of food traced to imported greenonions [Abst.]. Proc. of the 44th Annual Conferenceof the Epidemic Intelligence Service. Mar. 27-31, At-lanta, Ga, p. 36. Centers for Disease Control and

99EXPERT REPORT

Prevention, Atlanta, Ga.Cook, K.A., Swerdlow, D., Dobbs, T., Wells, J., Puhr,

Hlady G., Wells, J.G., Barrett, T.J., Phur, N.D., Lan-rette, G.A., Bodager, D.W., Toth, B.L., Genese, C.A.,Highsmith, A.K., Pilot, K.E., Finelli, L., and Swerdlow,D.L. 1996. Outbreak of Salmonella serotype hartfordinfections associated with unpasteurized orange juice.J. Am. Med. Assoc. 280(17): 1504-1509.

Covello, V.T., Sandman, P.M., and Slovic, P. 1988.“Risk Communication, Risk Statistics, and Risk Com-parisons: A Manual for Plant Managers.” ChemicalManufacturers Association, Washington, D.C.

Coulombe, R.A. 1993. Biological action of mycotoxins.J. Dairy Sci. 76: 880-91.

Cuthbert, J.A. 2001. Hepatitis A: Old and new. Clin. Mi-crobiol. Rev. 14: 38-58.

DDaifas, D.P., Smith, J.P., Blanchfield, B., and Austin,

J.W. 1999a. Growth and toxin production by Clostrid-ium botulinum in English-style crumpets packaged un-der modified atmospheres. J. Food Protect. 62:349-55.

Daifas, D.P., Smith, J.P., Blanchfield, B., and Austin,J.W. 1999b. Challenge studies with Clostridium botuli-num in high moisture bakery products packaged un-der modified atmospheres. Food Austral. 51: 507-511.

Daly, A.K., Cholerton, S., Armstrong, M., and Idle, J.R.1994. Genotyping for polymorphisms in xenobioticmetabolism as predictor of disease susceptibility. Envi-ron. Health Perspect. 102: S44-S61.

Daniels, R.W. 1998. Home Food Safety. Food Technol.52: 54-56.

Daniels, R., Daniels, B., Gilmet, P., and Noonan, D.2000. Audits International 2000 Home Food SafetyStudy Report. Audits Intl., Northbrook, Ill.

D’Aoust, J.W. 1985. Infective dose of Salmonella Typh-imurium in cheddar cheese. Am. J. Epidemiol. 122(4):717-720.

D’Aoust, J-Y. 1989. Salmonella. In “Foodborne Bacteri-al Pathogens.” pp. 327-445. ed. M.P. Doyle, MarcelDekker, New York, N.Y.

Darlow, H.M., Bale, W.R., and Carter, G.B. 1961. Infec-tion of mice by respiratory route with Salmonella typh-imurium. J. Hyg. 59: 303-308.

Davidson, P.M. 1997. Chemical preservatives and natu-ral antimicrobial compounds. Chpt. 29, In “Food Mi-crobiology: Fundamentals and Frontiers,” ed. M.P.Doyle, L.R. Beuchat, T.J. Montville. pp. 520-556.ASM Press, Washington, DC.

Davidson, P.M. 1999. Potential for emergence of resis-tance to antimicrobials used in the food industry.Proc. of Symposium on Food Microbiology. pp. 19-21. International Life Sciences Institute/ILSI NorthAmerica Technical Committe on Food Microbiology.(http://www.ilsi.org/file/iamfes99.pdf)

Davies, R.H. and Wray, C. 1997. Distribution of Salmo-nella on 23 pig farms in the U.K. In Proc. of the 2nd In-ternational Symposium on the Epidemiology and Con-trol of Salmonella in Pork. pp. 137-141. Copenhagen.

Dawson, D.J. and Sartory, D.P. 2000. Microbiologicalsafety of water. Brit. Med. Bull. 56: 74-83.

Day, W.A. and Maurelli, A.T. 2001. Shigella flexneriLuxS quorum-sensing system modulates virB expres-sion but is not essential for virulence. Infect. Immun.69(1): 15-23.

De Bethizy, J. and Hayes, J.R. 1994. Metaolism: A de-terminant of toxicity. In “Principles and Methods ofToxicology” 3rd Ed., ed. A.W. Hayes. Raven Press,New York, N.Y.

De Blackburn, W.C. and McCarthy, J.D. 2000. Modifi-cations to methods for the enumeration and detectionof injured Escherichia coli O157:H7 in foods. Intl. J.Food Microbiol. 55: 285-290.

De Boer, E. and Beumer, R.R. 1999. Methodology fordetection and typing of foodborne microorganisms.Intl. J. Food Microbiol. 50: 119-130.

De Kievit, T.R. and Iglewski, B.H. 2000. Bacterial quo-rum sensing in pathogenic relationships. Infect. Immu-nol. 68: 4839-4849.

Del Torre, M., Stecchini, M.L., and Peck, M.W. 1998.Investigation of the ability of proteolytic Clostridiumbotulinum to multiply and produce toxin in fresh Italianpasta. J. Food Protect. 61: 988-993.

Deming, M.S., Tauxe, R.V., Blake, P.A., Dixon, S.E.,Fowler, B.S., Jones, T.S., Lockamy, E.A., Patton,C.M., and Sikes, R.O. 1987. Campylobacter enteritisat a university: Transmission from eating chicken andfrom cats. Am. J. Epidemiol. 126: 526-534.

Deneen, V.C., Hunt, J.M., Paule, C.F., James, R.I.,Johnson, R.G., Raymond, M.J., and Hedberg, C.W.2000. The impact of foodborne calicivirus disease:The Minnesota experience. J. Infect. Dis. 181(Suppl.2): S281-283.

DePaola, A., Flynn, P.A., McPhearson, R.M., and Levy,S.B. 1988. Phenotypic and genotypic characteriza-tions of tetracycline and oxytetracycline resistant Aero-monas hydrophila for altered channel catfish (letaluruspunctatus) and their environments. Appl. Environ. Mi-crobiol. 54: 1861-1863.

Desenclos, J.A., Klontz, K.C., Wolfe, L.E., and Hoe-cherl, S. 1991. The risk of Vibrio illness in the Floridaraw oyster eating population, 1981-1988. Am. J. Ep-idemiol. 134: 290-297.

De Roever, C. 1998. Microbiological safety evaluationsand recommendations on fresh produce. Food Con-trol 9(6): 321-347.

Dickson, J.S. and Anderson, M.E. 1992. Microbiologi-cal decontamination of food animal carcasses bywashing and sanitizing systems. J. Food Protect. 55:133-140.

Dickson, J.S. and Kunduru, M.R. 1995. Resistance ofacid adapted salmonellae to organic acid rinses onbeef. J. Food Protect. 58: 973-976.

Diehl, J.F. 1995. “Safety of irradiated foods,” 2nd Ed.Marcel Dekker, New York, N.Y.

Diez-Gonzalez, F., Callaway, T.R., Kizoulis, M.G., andRussell, J.B. 1998. Grain feeding and the dissemina-tion of acid-resistant Escherichia coli from cattle. Sci-ence 281: 1666-1668.

Dorsa, W.J., Cutter, C.N., Siragusa, G.R., and Koohma-raie, M. 1996. Microbial decontamination of beef andsheep carcasses by steam, hot water spray washesand a steam-vacuum sanitizer. J. Food Protect. 59:127-135.

Downes, T. 1989. Food packaging in the IFT era: Fivedecades of unprecedented growth and change. FoodTechnol. 43(9): 228-240.

Doyle, M. 1998. Extending the shelf life of refrigeratedfoods: For better or worse? Food Technol. 52(2): 20.

Doyle, M.P. 2000. Reducing foodborne disease: Whatare the priorities? Nutrition 16(7/8): 647-694.

Doyle, M.P., Beauchat, L.R., and Montville, T.J. 1997.“Food Microbiology: Fundamentals and Frontiers.” pp.327-336. ASM Press, Washington, D.C.

Dubey, J.P., Brake, R.J., Murrell, K.D., and Fayer, R.1986. Effect of irradiation on the viability of Toxo-plasma gondii cysts in tissues of mice and pig. Am.J. Vet. Res. 47: 518-522.

Dubois, A. 1995. Spiral bacterial in the human stock:The gastric helicobacters. Emerg. Infect. Dis. 1(3):79-85.

DuPont, H.L., Chappell, C.L., Sterling, C.R., Okhuysen,P.C., Rose, J.B., and Jakubowski, W. 1995. The in-fectivity of Cryptosporidium parvum in healthy volun-teers. N. Eng. J. Med. 332: 855-859.

EEaton, D.A., Gallagher, E.P., Bammler, T.K., and Kunze,

K.L. 1995. Role of cytochrome P4501A2 in chemi-cal carcinogenesis: Implications for human variabilityin expression and enzyme activity. Pharmacogen. 5:259-274.

Eberhart-Phillips J., Besser, R.E., Tormey, M.P., Koo,D., Feikin, D., Araneta, M.R., Wells, J., Kilman, L.,Rutherford, G.W., Griffin, P.M. Baron, R., and Masco-la, L. 1996. An outbreak of cholera from foodserved on an international aircraft. Epidemiol. Infect.116: 9-13.

EC. 2000. Possible links between Crohn’s disease andparatuberculosis. Report of the Scientific Committeeon Animal Health and Animal Welfare, Adopted Mar.21. SANCO/B3/R16/2000. Directorate-GeneralHealth and Consumer Protection. European Commis-sion.

Elder, R.O., Keen, J.E., Siragusa, G.R., Barkocy-Gal-lagher, G.A., Koohmarie, M., and Laegreid, W.W.2000. Correlation of enterohemorrhagic Escherichiacoli O157 prevalence in feces, hides, and carcassesof beef cattle during processing. Proc. Natl. Acad.Sci. 97: 2999-3003.

Elwood, C.M. and Garden, O.A. 1999. Gastrointestinalimmunity in health and disease. Vet. Clin. North Am.Small Anim. Pract. 29(2): 471-500, vi-vii.

Engberg, J., Aarestrup, F.M., Taylor, D.E., Gerner-Smidt,P., Nachamkin, I. 2001. Quinolone and macrolideresistance in Campylobacter jejuni and C. coli: Resis-tance mechanisms and trends in human isolates.Emerg. Infect. Dis. 7: 24-34.

EPA. 1998. What you should know about Pfiesteria pis-cicida: A Fact Sheet. pp. 1-6. U.S. Government Print-ing Office, Washington, D.C.

EPA. 2000. Proposed regulations to address water pol-lution from concentrated animal feeding operations.Report No. 833-F-00-016. Office of Water, Environ-mental Protection Agency, Washington, D.C.

EPA/AID. 1992. Manual: Guidelines for Water Reuse.U.S. Environmental Protection Agency / Agency for In-ternational Development. (http://www.fsis.usda.gov/oppde/rdad/frpubs/san_update1.pdf).

Epstein, P. 1998. Marine Ecosystems: Emerging Dis-eases as Indicators of Change. pp. 4-75. Health,Ecological, and Economic Dimensions (HEED) of Glo-bal Change Program, Harvard University, Boston,Mass.

Epstein, C.B. and Butow, R.A. 2000. Microarray tech-nology: Enhanced versatility, persistent challenge.Current Opinions in Biotechnol. 11(1): 36-41.

Erickson, J.W. and Gross, C.A. 1989. Identification ofthe sE subunit of Escherichia coli RNA polymerase: Asecond alternate s factor involved in high-temperaturegene expression. Genes Dev. 3: 1462-1471.

Ervik, A., Tehorsen, B., Eriksen, V., Lunestad, B.T., andSamuelsen, O.B. 1994. Impact of administering anti-bacterial agents on wild fish and blue mussels (Mytilusedulis) in the vicinity of fish farms. Disease of AquaticOrganisms18: 45-51.

F

Fairchild, T.M. and Foegeding, P.M. 1993. A proposednonpathogenic biological indicator for thermal inacti-vation of Listeria monocytogenes. Appl. Environ. Mi-crobiol. 59: 1247-1250.

Fairweather, N.F., Chatfiled, S.N., Charles, I.G., Roberts,M., Lipscombe, M., Jiang Li, L., Strugnell, D., Com-erford, S., Tite, J., and Dougan, G. 1990. Use of liveattenuated bacteria to stimulate immunity. Res. Mi-crobiol. 141: 769-773.

Faith, N.G., Shere, J.A., Brosch, R., Arnold, K.W., An-say, S.E., Lee, M.-S., Luchansky, J.B., and Kaspar,C.W. 1996. Prevalence and clonal nature of Escheri-chia coli O157:H7 on dairy farms in Wisconsin. Appl.Environ. Microbiol. 62: 1519-1525.


Fang, F. C., Libby, S.J., Buchmeier, N.A., Loewen, P.C.,Switala, J., Harwood, J., and Guiney, D.G. 1992. Thealternative s factor KatF (RpoS) regulates Salmonellavirulence. Proc. Natl. Acad. Sci. 89: 11978-11982.

Fankhauser, R.L., Noel, J.S., Monroe, S.S., Ando, T.,and Glass, R.I. 1998. Molecular epidemiology of“Norwalk-like viruses” in outbreaks of gastroenteritis inthe United States. J. Infect. Dis. 178: 1571-1578.

FAO. 1995. Code of conduct for responsible fisheries.Food and Agriculture Organization of the United Na-tions. Fisheries Department. p. 41. Rome.

FAO/WHO. 1998. Application of risk communication tofood standards and safety matters. An expert consul-tation of the Food and Agricultural Organization of theUnited Nations and the World Health Organization,Feb. 2-6, Rome.

FAO/WHO. 2001. Risk characterization of Salmonellaspp. in eggs and broiler chickens and Listeria mono-cytogenes in ready-to-eat foods. A Joint FAO/WHOConsultation on Risk Assessment of MicrobiologicalHazards in Food. Apr. 30 - May 4. Food and Agricul-ture Organization of the United Nations/World HealthOrganization, Rome.

Farber, J.M. and Pagotto, F. 1992. The effect of acidshock on the heat resistance of Listeria monocytoge-nes. Lett. Appl. Microbiol. 15: 197-201.

Farber, J.M., Gendel, S.M., Tyler, K.D., Boerlin, P., Lan-dry, W.L., Fritschel, S.J., and Barrett, T.J. 2001.Typing and Differentiation. Chpt. 11 In “Compendiumof Methods for the Microbiological Examination ofFoods,” 4th Ed., ed. F. P. Downes and K. Ito, Ameri-can Public Health Association, Washington, D. C.

Farkas, J. 1998. Irradiation as a method for decontami-nating food: A review. Int. J. Food Microbiol. 44:189-204.

Fayer, R., Graczyk, T.K., Lewis, E.J., Trout, J.M., andFarley, C.A. 1998. Survival in infectious Cryptosporidi-um parvuum oocysts in seawater and Eastern oysters(Crassostrea virginica) in the Chesapeake Bay. Appl.Environ. Microbiol. 64: 1070-1074.

FDA. 1995. Procedures for the safe and sanitary pro-cessing and importing of fish and fishery products; Fi-nal Rule. Fed. Reg. 60: 65095-65202.

FDA. 1998. “Bacteriological Analytical Manual,” 8th ed.AOAC International, Gaithersburg, Md.

FDA. 1999. Guidance for industry: Reducing microbialfood safety hazards for sprouted seeds and guidancefor industry: Sampling and microbial testing of spentirrigation water during sprout production. Fed. Reg.64: 57893-57902.

FDA. 2000. Report of the FDA Retail Food ProgramDatabase of Foodborne Illness Risk Factors. Septem-ber 10. Food and Drug Administration, Washington,D.C.

FDA. 2001. Good agricultural practices meeting. Foodand Drug Administration, Orlando, Fla., June 28.

FDA/CFSAN. 1998. Guide to minimize microbial foodsafety hazards for fresh fruits and vegetables. U.S.Department of Health and Human Services, Food andDrug Administration/Center for Food Safety and Ap-plied Nutrition. Washington, D.C. (http://www.foodsafety.gov/~dms/prodguid.html).

FDA/CFSAN. 2001. Draft risk assessment on the publichealth impact of Vibrio parahaemolyticus in raw mol-luscan shellfish. Food and Drug Administration/Centerfor Food Safety and Applied Nutrition. (http://www.cfsan.fda.gov/~dms/vprisk.html).

FDA/CFSAN. 2002. Foodborne Pathogenic Microor-ganisms and Natural Toxins Handbook: The “Bad BugBook.” Food and Drug Administration/Center for FoodSafety and Applied Nutrition, Washington, D.C. Ac-cessed Jan. 29, 2002. (http://www.cfsan.fda.gov/~mow/intro.html)

FDA/CFSAN, USDA/FSIS, CDC. 2001. Draft assess-ment of the relative risk to public health from food-borne Listeria monocytogenes among selected cate-

gories of ready-to-eat foods. Food and Drug Adminis-tration/Center for Food Safety and Applied Nutrition,U.S. Dept. of Agriculture/Food Safety and InspectionService, and Centers for Disease Control and Preven-tion. (http://www.FoodSafety.gov/~dms/lmrisk.html).

FDA/CVM. 1997. CVM policy on competitive exclusionproducts. Feb. 21. Food and Drug Administration/Center for Veterinary Medicine. Washington, D.C.

Fedorka-Cray, P.J., Kelley, L.C., Stabel, T.J., Gray, J.T.,and Laufer, J.A. 1995. Alternate routes of invasionmay affect pathogenesis of Salmonella typhimurium inswine. Infect. and Immun. 63: 2658-2664.

Fedorka-Cray, P., Gray, J.T., and Wray, C. 2000. Sal-monella infections in pigs. In “Salmonella in DomesticAnimals.” pp. 191-207. ed. C. Wray and A. Wray.CABI Publishing, Wallingford.

Feinman, S.E. 1998. Antibiotics in animal feed- Drugresistance revisited. ASM News 64: 24-30.

Feng, P., Lampel, K.A., Karch, H., and Whittam, T.S.1998. Genotypic and phenotypic changes in theemergence of Escherichia coli O157:H7. J. Infect.Dis. 177: 1750-1753.

Finelli, L., Swerdlow, D., Mertz, K., Ragazzoni, H., andSpitalny, K. 1992. Outbreak of cholera associatedwith crab brought from an area with epidemic dis-ease. J. Infect. Dis. 166: 1433-6.

Finlay, B.B. and Cossart, P. 1997. Exploitation of mam-malian host cell functions by bacterial pathogens. Sci-ence 276: 718-725.

Finlay, B.B. and Falkow, S. 1997. Common themes inmicrobial pathogenicity revisited. Microbiol. Mol. Biol.Rev. 61: 136-69.

Fleming, L.E., Easom, J., Baden, D., Rowan, A., andLevin, B. 1999. Emerging harmful algal blooms andhuman health: Pfiesteria and related organisms. Toxi-col. Pathol. 27(5): 573-581.

FNB/NRC. 1985. An evaluation of the role of microbio-logical criteria for foods and food ingredients. Sub-committee on Microbiological Criteria, Committee onFood Protection, Food and Nutrition Board, NationalResearch Council. National Academy Press, Washing-ton, D.C.

Foege, W.H. 1996. Alexander D. Langmuir: His Impacton Public Health. A Tribute to Alexander D. Langmuir.Am. J. Epidemiol. 144(8 Suppl): S11-50.

Foster, J.W. and Hall, H.K. 1990. Adaptive acidificationtolerance response of Salmonella typhimurium. J.Bacteriol. 172: 771-778.

Francis, G.A. and O’Beirne, D. 1998. Effects of storageatmosphere on Listeria monocytogenes and compet-ing microflora using a surface model system. Intl. J.Food Sci. Technol. 33: 465-476.

Frewer, L.J., Shepherd, R., and Sparks, P. 1994. Theinterrelationship between perceived knowledge, con-trol and risk associated with a range of food-relatedhazards targeted at the individual, other people andsociety. J. Food Safety 14: 19-40.

Friedman, C.R., Neimann, J., Wegener, H.C ., andTauxe, R.V. 1992. Epidemiology of Campylobacter je-juni Infections in the United States and Other IndustrialCountries. In “Campylobacter jejuni: Current Statusand Future Trends. ed. I. Nachamkin, M.J. Blaser, L.S.Tompkins, pp. 121-138. ASM Press, Washington,D.C.

Friedman, C., Reddy, S., Samual, M., Marcus, R.,Bender, J., Desai, S., Shiferaw, B., Helfrick, D., Cart-er, M., Anderson, B., Hoekstra, M., and the EIPWorking Group. 2000. Risk factors for sporadicCampylobacter infections in the United States: Acase-control study on FoodNet sites. 2nd InternationalConference on Emerging Infectious Diseases, 2000.July 16-19. Abstr. No. 63, pp. 149-150. Atlanta, Ga.(http://www.cdc.gov/foodnet/pub/iceid/2000/friedman_c.htm).

Fujita, M., Tanaka, K., Takahashi, H., and Amemura, A.1994. Transcription of the principal sigma-factor

genes, rpoD and rpoS, in Pseudomonas aeruginosa iscontrolled according to the growth phase. Molec. Mi-crobiol. 13: 1071-1077.

Fuller, R. 1989. A review: Probiotics in man and ani-mals. J. Appl. Bacteriol. 66: 365-378.

Fuller, R. 1992. “Problems and prospects.” In Probiot-ics, The Scientific Basis, ed. R. Fuller. Chapter 14, pp.377-386. Chapman & Hall, London.

GGahan, C.G.M. and Hill, C. 1999. The relationship be-

tween acid stress responses and virulence in Salmo-nella typhimurium and Listeria monocytogenes. Intl. J.Food Microbiol. 50: 93-100.

GAO. 1998. “Federal Efforts to Ensure the Safety ofImported Foods are Inconsistent and Unreliable.”GAO/RCED-98-103, USGAO2. General AccountingOffice. Washington, D.C.

GAO. 1999. “Fresh Produce: Potential Implications ofCountry-of-Origin Labeling.” GAO/T-RCED-99-200.General Accounting Office. Washington, D.C.

Garber, L.P., Wells, S.J., Hancock, D.D., Doyle, M.P.,Tuttle, J., Shere, J.A., and Zhao, T. 1995. Risk fac-tors for fecal shedding of Escherichia coli O157: H7in dairy calves. J. Am. Vet. Med. Assn. 207: 46-49.

Garber, L., Wells, S., Schroeder-Tucker, L., and Ferris,K. 1999. Factors associated with fecal shedding ofverotoxin-producing Escherichia coli O157 on dairyfarms. J. Food Protect. 62: 307-312.

Gebreyes, W.A., Turner, M.B., Funk, J.A., Altier, C., andDavies, P.R. 1999. Salmonella prevalence, serotypes,and patterns of antimicrobial resistance in cohorts ofnursery and finishing pigs. In “Proc. of the 3rd Interna-tional Symposium on the Epidemiology and Control ofSalmonella in Pork”, pp. 250-251. Washington, D.C.

Gibbens, J.C., Pascoe, S.J., Evans, S.J., Davies, R.H.,and Sayers, A.R. 2001. A trial of biosecurity as ameans to control Campylobacter infection of broilerchickens. Prev. Vet. Med. 48: 85-99.

Gill, C.O., McGinnis, D.S., Bryant, J., and Chabot, B.1995. Decontamination of commercial, polished pigcarcasses with hot water. Food Microbiol. 12: 143-149.

Glasgow, H.B., Burkholder, J.M., Schmechel, D.E.,Tester, P.A., and Rublee, P.A. 1995. Insidious effectsof a toxic estuarine dinoflagellate on fish survival andhuman health. J. Toxicol. Environ. Health 46: 501-522.

Goepfert, J.M. 1976. The aerobic plate count, coliformand Escherichia coli content of raw ground beef atthe retail level. J. Milk Food Technol. 39: 175-178.

Goldblith, S.A. 1989. 50 Years of progress in food sci-ence and technology: From art based on experienceto technology based on science. Food Technol. 88-105, 286.

Gorman, B.M., Sofos, J.N., Morgan, J.B., Schmidt,G.R., and Smith, G.C. 1995. Evaluation of hand-trim-ming, various sanitizing agents, and hot water spraywashing as decontamination interventions on beefbrisket adipose tissue. J. Food Protect. 58: 899-907.

Gould, G.W. 2000a. Strategies for food preservation.Chpt. 2, In “The Microbiological Safety and Quality ofFood.” Vol. I. ed. B.M. Lund, T.C. Baird-Parker, andG.W. Gould, pp. 19-35. Aspen Publishers Inc., Gaith-ersburg, Md.

Gould, G.W. 2000b. New and emerging physical meth-ods of preservation. Chpt. 13, In “The MicrobiologicalSafety and Quality of Food.” Vol. I. ed. B.M. Lund,T.C. Baird-Parker, and G.W. Gould, pp. 277-293. As-pen Publishers Inc., Gaithersburg, Md.

Grattan, L.M., Oldach, D., Perl, T.M., Lowitt, M.H., Ma-tuszak, D.L., Dickson, C., Parrott, C., Shoemaker,R.C., Kauffman, C.L., Wasserman, M.P., Hebel, J.R.,Charache, P., and Morris, J.G. 1998. Learning and

101EXPERT REPORT

memory difficulties after environmental exposure towaterways containing toxin-producing Pfiesteria or Pfi-esteria-like dinoflagellates. Lancet 352: 532-539.

Graves, D.J. 1999. Powerful tools for genetic analysiscome of age. Trends Biotechnol. 17: 127-134.

Graves, L.M. and Swaminathan, B. 2001. PulseNetstandardized protocol for subtyping Listeria monocy-togenes by macrorestriction and pulsed-field gel elec-trophoresis. Intl. J. Food Microbiol. 65(1-2): 55-62.

Gray, J.T., Fedorka-Cray, P.J., Stabel, T.J., and Kramer,T.T. 1996. Natural transmission of Salmonella choler-aesuis in swine. Appl. Environ. Microbiol. 62: 141-146.

Green, J., Brown, D., Cheesbrough, J., and Wright, P.1999. The detection of environmental small roundstructured virus by nested reverse transcriptase poly-merase chain reaction. (Abst. S3-8). InternationalWorkshop on Human Caliciviruses, Mar. 29-31, Atlan-ta, Ga.

Griffin, P.M. 1998. Epidemiology of Shiga toxin-produc-ing Escherichia coli infections in humans in the UnitedStates. In “Escherichia coli and Other Shiga Toxin-Producing E. coli Strains.” ed. J.B. Kaper and A.D.O’Brien. pp. 15-22. ASM Press, Washington, D.C.

Gross, C.A. 1996. Function and regulation of the heatshock proteins In “Escherichia coli and Salmonella.”ed. F.C. Neidhardt. pp. 1382-1399. ASM Press,Washington, D.C.

Grossman, A.D., Erickson, J.W., and Gross, C.A. 1984.The htpR gene of E. coli is a sigma factor for heatshock promoters. Cell 38: 383-390.

Gruenewald, R., Blum, S., and Chan, J. 1994. Rela-tionship between human immunodeficiency virus infec-tion and salmonellosis in 20- to 59-year-old residentsof New York City. Clin. Infect. Dis. 18(3): 358-363.

Guzewich, J.J., Bryan, R.L., and Todd, E.C.D. 1997.Surveillance of foodborne disease I. Purposes andtypes of surveillance systems and networks. J. FoodProtect. 60: 555-556.

HHaas, C.N. 1983. Estimation of risk due to low doses of

microorganisms: A comparison of alternative method-ologies. Am. J. Epidemiol. 118: 573-582.

Haas, G.J., Prescott, H.E., Dudley, E., Dik, R., Hintlian,C., and Keane, L. 1989. Inactivation of microorgan-isms by carbon dioxide under pressure. J. Food Safe-ty 9: 253-265.

Hacker, J. and Kaper, J.B. 2000. Pathogenicity islandsand the evolution of microbes. Ann. Rev. Microbiol.54: 641-79.

Hacker, J., Blum-Oehler, G., Muhldorfer, I., andTschape, H. 1997. Pathogenicity islands of virulentbacteria: Structure, function and impact on microbialevolution. Molec. Microbiol. 23: 1089-97.

Hall, J.A., Goulding, J.S., Bean, N.H., Tauxe, R.V., andHedberg, C.W. 2001. Epidemiologic profiling: Evalu-ating foodborne outbreaks for which no pathogenwas isolated by routine laboratory testing, UnitedStates: 1982-1989. Epidemiol. Infect. In press.

Hancock, D.D., Besser, T.E., Rice, D.H., Ebel, E.D.,Herriott, D.E., and Carpenter, L.V. 1998. Multiplesources of Escherichia coli O157 in feedlots and dairyfarms in Northwestern USA. Prevent. Vet. Med. 35:11-19.

Harris, N.V., Weiss, N.S., and Nolan, C.M. 1986. Therole of poultry and meats in the etiology of Campylo-bacter jejuni/coli enteritis. Am. J. Publ. Health 76:407-411.

Hayes, E.B., Matte, T.D., O’Brien, T.R., McKinley, T.W.,Logsdon, G.S., Rose, J.B., Ungar, B.L.P., Word,D.M., Pinsky, P.F., Cummings, M.L., Wilson, M.A.,Long, E.G., Hurwitz, E.S., and Juranek, D.D. 1989.Large community outbreak of cryptosporidiosis dueto contamination of a filtered public water supply. N.

Eng. J. Med. 320: 1372-1376.Hedberg, C.W. and Osterholm, M.T. 1993. Outbreaks

of foodborne and waterborne viral gastroenteritis.Clin. Microbiol. Rev. 6: 199-210.

Hedberg, C.W., Savarino, S.J., Besser, J.M., Paulus,C.J., Thelen,V.M., Myers, L.J., Cameron, D.N., Bar-rett, T.J., Kaper, J.B., and Osterholm, M.T. 1997. Anoutbreak of foodborne illness caused by Escherichiacoli O39:NM, an agent not fitting into the existingscheme for classifying diarrheageneic E. coli. J. Infect.Dis. 176: 1625-1628.

Hedberg, C.W., Smith, K.E., Besser, J.M., Boxrud, D.J.,Hennessy, T.W., Bender, J.B., Anderson, F.A., andOsterholm, M.T. 2001. Limitations of pulsed-field gelelectrophoresis for the routine surveillance of Campy-lobacter infections. J. Infect. Dis. 184: 242-243.

Hengge-Aronis, R. 1993. Survival of hunger and stress:The role of rpoS in early stationary phase gene regu-lation in E. coli. Cell 72: 165-168.

Hengge-Aronis, R. 1996. Back to log phase: ss as aglobal regulator in the osmotic control of gene ex-pression in Escherichia coli. Molec. Microbiol. 21:887-893.

Hengge-Aronis, R. 2000. The general stress responsein Escherichia coli. In “Bacterial stress responses.”ed. G. Storz and R. Hengge-Aronis. ASM Press,Washington, D.C.

Hennessy, T.W., Hedberg, C.W., Slutsker, L., White,K.E., Besser-Wiek, J.M., Moen, M.E., Feldman, J.,Coleman, W.W., Edmonson, L.M., MacDonald, K.L.,and Osterholm, M.T. 1996. A national outbreak ofSalmonella enteritidis infections from ice cream. N.Eng. J. Med. 334: 1281-1286.

Herman, L. 1997. Detection of viable and dead Listeriamonocytogenes by PCR. Food Microbiol. 14: 103-110.

Herman-Taylor, J., Bull, T.J., Sheridan, J.M., Cheng, J.,Stellakis, M.L., and Sumar, N. 2000. Causation ofCrohn’s disease by Mycobacterium avium subspeciesparatuberculosis, Can. J. Gastroenterol. 14(6): 521-39.

Herwaldt ,B.L., Ackers, M.L., and the CyclosporaWorking Group. 1997. An outbreak in 1996 of cy-closporiasis associated with imported raspberries. N.Engl J. Med. 336(22): 1548-1556.

Hilborn, E.D., Mermin, J.H., Mshar, P.A., Hadler, J.L.,Voetsch, A., Wojtkunski, C., Swartz, M., Mshar, R.,Lambert-Fair, M.A., Farrar, J.A., Glynn, M.K., andSlutsker, L. 1999. A multistate outbreak of Escheri-chia coli O157:H7 infections associated with con-sumption of mesclun lettuce. Archiv. Intern. Med.159: 1758-1764.

Hingley, A. 1997. Rallying the troops to fight foodborneillness. FDA Consumer Magazine. Nov./Dec.(www.fda.gov/fdac/features/1997/797-food.html)

Hite, B. 1899. The effect of pressure in the preservationof milk. WV Univ. Agric. Exp. Sta. Bull. No. 146: 1-67. Morgantown, W.V.

Hlady, W.G. and Klontz, K.C. 1996. The epidemiologyof Vibrio infections in Florida, 1981-1993. J. Infect.Dis. 173: 1176-1183.

Hlady, W.G., Mulle, R.C., and Hopkins, R.S. 1993.Vibrio vulnificus from raw oysters: Leading cause ofreported deaths from illness in Florida. J. Fla. Med.Assn. 80: 536-538.

Hogue, A., White, P., Guard-Petter, J., Schlosser, W.,Gast, R., Ebel, E., Farrar, J., Gomez, T., Madden, J.,Madison, M., McNamara, A.M., Morales, R., Parham,D., Sparling, P., Sutherlin, W., and Swerdlow, D.1997. Epidemiology and control of egg-associatedSalmonella enteritidis in the United States of America.Rev. Sci. Tech. 16: 542-53.

Hollinger, K., Wray, C., Evans, S., Pascoe, S., Chappell,S., and Jones, Y. 1998. Salmonella typhimuriumDT104 in cattle in Great Britian. J. Am. Vet. Med.Assn. 213: 1732-1733.

Holmberg, S.D., Wells, I.G., and Cohen, M.L. 1984.Animal-to-human transmission of antimicrobial-resis-tant Salmonella: Investigations of U.S. outbreaks1971-1983. Science 225: 833-835.

Hoover, D.G. 1993. Pressure effects on biological sys-tems. Food Technol. 47(6): 150-155.

Hopkins, R.S., Olmsted, R., and Istre, G.R. 1984.Campylobacter jejuni infection in Colorado: Identifiedrisk factors. Am. J. Publ. Health 74: 249-250.

Hosek, G., Leschinsky, D., Irons, S., and Safranek, T.J.1997. Multidrug-resistant Salmonella serotypre typh-imurium- United States, 1996. MMWR. 46: 308-310.

Howgate, P. 1997. Review of the public health safety ofproducts from aquaculture. A report prepared for theFish Utilization and Marketing Service, Fisheries Utiliza-tion Division, Food and Agricultural Organization of theUnited Nations. Presented at the Joint FAO/WHOStudy Group on Food Safety Issues Associated withProducts from Aquaculture. June 1997. Bangkok,Thailand.

Hovde, C.J., Austin, P.R., Cloud, K.A., Williams, C.J.,and Hunt, C.W. 1999. Effect of cattle diet on Escheri-chia coli O157:H7 acid resistance. Appl. Environ. Mi-crobiol. 65: 3233-3235.

Hueck, C.J. 1998. Type III protein secretion systems inbacterial pathogens of animals and plants. Microbiol.Molec. Biol. Rev. 62: 379-433.

Hughes, D.B. and Hoover, D.G. 1991. Bifidobacteria:Their potential for use in American dairy products.Food Technol. 45(1): 74-83.

Humphrey, T.J., Richardson, N.P., Statton, K.M., andRowbury, R.J. 1993. Effects of temperature shift onacid and heat tolerance in Salmonella enteritidis ph-age type 4. Appl. Environ. Microbiol. 59: 3120-3122.

Hutwagner, L.C., Maloney, E.K., Bean, N.H., Slutsker,L., and Martin, S.M. 1997. Using laboratory-basedsurveillance data for prevention: An algorithm for de-tecting Salmonella outbreaks. Emerg. Infect. Dis. 3:395-400.

IICMSF. 1986. “Microorganisms in Foods 2. Sampling

for Microbiological Analysis: Principles and SpecificApplications.” 2nd Ed. International Commission on Mi-crobiological Specifications for Food. University ofToronto Press, Toronto.

ICMSF. 1996. “Microorganisms in Foods 5. Microbio-logical Specifications of Food Pathogens.” Internation-al Commission on Microbiological Specifications forFood. Blackie Academic and Professional, an imprintof Chapman and Hall, New York, N.Y.

ICMSF. 1997. Establishment of microbiological safetycriteria for foods in international trade. InternationalCommission on Microbiological Specifications forFood. World Health Statist. Quart. 50: 119-123.

ICMSF. 2002. “Microorganisms in Foods 7. Microbio-logical Testing in Food Safety Management.” Interna-tional Commission on Microbiological Specificationsfor Food. Kluwer Academic, N.Y.

IFT. 1989. Top 10 food science innovations 1939-1989. Food Technol. 43(9): 308.

IFT. 2000a. A Century of Food Science. IFT ResearchReport. Institute of Food Technologists, Chicago, Ill.

IFT. 2000b. Kinetics of microbial inactivation for alterna-tive food processing technologies. A report of the In-stitute of Food Technologists for the Food and DrugAdministration, Institute of Food Technologists, Chica-go, Ill. J. Food Sci. 65(8): 1S-108S.

IFT. 2002. Analysis and evaluation of preventive controlmeasures for the control and reduction/elimination ofmicrobial hazards on fresh and fresh-cut produce. Areport of the Institute of Food Technologists for theFood and Drug Administration, Institute of Food Tech-


nologists, Chicago, Ill. J. Food Sci. In press.Ingram, M. 1964. Feeding meat animals before slaugh-

ter. Vet. Record 76: 1305-1315.Ingram, M. 1977. The significance of the index and in-

dicator organisms in foods. Presented at the 10th Intl.Symposium of the IUMS Committee on Food Micro-biology and Hygiene, Sept. 5-10, Szczecin, Poland.

Ingram M. and Dainty, R.H. 1971. Changes caused bymicrobes in spoilage of meat. J. Appl. Bacteriol. 34:21-39.

Iversen, A.M., Gill, M., Bartlett, C.L.R. 1987. Two out-breaks of foodborne gastroenteritis caused by a smallround structured virus: Evidence of prolonged infectiv-ity in a food handler. Lancet 2(8558): 556-558.

JJackson, L.S., De Vries, J.W., and Bullerman, L.B.

1996. “Fumonisins in food”. Plenum Press, New York,N.Y.

Jacob, J.Q. 2000. The final stages of global eradicationof polio. N. Eng. J. Med. 343: 806-807.

Jaenicke, R. 1981. Enzymes under extreme conditions.Ann. Rev. Biophys. Bioeng. 10: 1-67.

Jay, J.J. 2000. “Modern Food Microbiology”. 6th ed.pp. 425-437. Aspen Publishers, Inc. Gaithersburg,Md.

Jaykus, L.A. 1996. The application of quantitative riskassessment to microbial food safety risks. CRC Crit.Rev. Microbiol. 22: 279-93.

Jaykus, L. 1997. Epidemiology and detection as op-tions for control of viral and parasitic foodborne dis-ease. Emerg. Infect. Dis. 3: 529-539.

Jaykus, L.A. 2000a. Detection of human enteric virusesin foods. In “Foodborne Disease Handbook: Viruses,Parasites, and HACCP.” Vol. 2, 2nd Ed. ed. Y.H. Hui,S.A. Sattar, K.D. Murrell, W.K. Nip, P.S. Stanfield. pp.137-163. Marcel Dekker, New York, N.Y.

Jaykus, L.A. 2000b. Enteric viruses as “emerging”agents of foodborne disease. Irish J. Agr. Food Res.39: 245-255.

Jaykus, L. and Ward, D.R. 1999. An integrated ap-proach: The future of graduate food safety educa-tion. Dairy Food Environ. Sanita. 19: 14-17.

Jaykus, L., DeLeon, R., and Toranzos, G.A. 2001. De-tection of bacteria, viruses and parasitic protozoa inshellfish. In “Manual of Environmental Microbiology,”2nd Ed. ed. C.J. Hurst. In press.

Jenkins, D.E., Schultz, J.E., and Matin, A. 1988. Star-vation-induced cross protection against heat or H

2O

2challenge in Escherichia coli. J. Bacteriol. 170: 3910-3914.

Jenkins, D.E., Chaisson, S.A., and Matin, A. 1990.Starvation-induced cross protection against osmoticchallenge in Escherichia coli. J. Bacteriol. 172: 2779-2781.

Jensen, L.B. and Hess, W.R. 1941. A study of hamsouring. Food Res. 6: 272-326.

Jensen, G.L. and Greenless, K.J. 1997. Public health is-sues in aquaculture. OIE Scientific and Technical Re-view. 16(2): 641-651.

Johnson, F.H. and Campbell, D.H. 1945. The retarda-tion of protein denaturation by hydrostatic pressure. J.Cell Comp. Physiol. 26: 4-7.

Johnson, K., Charles, I., Dougan, G., Pickard, D.,O’Gaora, P., Costa, G., Ali, T., Miller, I., and Hor-maeche, C. 1991. The role of a stress-response pro-tein in Salmonella typhimurium virulence. Mol. Microbi-ol. 5: 401-407.

Josephson, E.S., Thomas, M.H., and Calhoun, W.K.1978. Nutritional aspects of food irradiation: An over-view. J. Food Proc. Preserva. 2: 299-313.

JSA. 1994. Guide to drug vaccine and pesticide use inaquaculture. Publication No. B-5085. Texas Agricultur-al Extension Service. Texas A&M University, CollegeStation.

KKalow, W. 1993. Pharmacogenetics: Its biologic roots

and the medical challenge. Clin. Pharmacol. Ther. 54:235-241.

Kaplan, J.E., Gary, G.W., Baron, R.C., Singh, N.,Schonberger, L.B., Feldman, R., and Greenberg, H.B.1982. The epidemiology of Norwalk gastroenteritisand the role of Norwalk virus in outbreaks of acutenonbacterial gastroenteritis. Ann. Intern. Med. 96:756-761.

Kaper, J.B. and O’Brien, A.D. 1998. “Escherichia coliO157:H7 and other Shiga toxin-producing E. colistrains.” ASM Press, Washington, D.C.

Kassenborg, H., Hedberg, C., Evans, M., Chin, G.,Fiorentino, T., Vugia, D., Bardsley, M., Slutsker, S.,and Griffin, P. 1998. Case-control study of sporadicEscherichia coli O157: H7 infections in 5 FoodNetsites (CA, CT, GA, MN, OR). (Abstr.) In “Program andAbstracts of the International Conference on Emerg-ing Infectious Diseases,” Atlanta, Mar. 8-11. AmericanSociety for Microbiology, Washington, D.C.

Kenny, B., DeVinney, R., Stein, M., Reinscheid, D.J.,Frey, E.A., and Finlay, B.B. 1997. EnteropathogenicE. coli (EPEC) transfers its receptor for intimate ad-herence into mammalian cells. Cell 91: 511-20.

Kerry, J., Hiney, M., Coyne, R., Cazabon, D., NicGab-hainn, S., and Smith, P. 1994. Frequency and distri-bution of resistance to oxytetracycline in microorgan-isms isolated from marine fish farm sediments follow-ing therapeutic use of oxytetracycline. Aquaculture(123): 43-54.

Kilcast, D. 1994. Effect of irradiation on vitamins. FoodChem. 49: 157-164.

Kim, J., Nietfeldt, J., and Benson, A.K. 1999. Octamerbased genome scanning distinguishes a unique sub-population of Eschericia coli O157: H7 strains in cat-tle. Proc. Natl. Acad. Sci. 96: 13288-13293.

Kimmitt, P.T., Harwood, C.R., and Barer, M.R. 2000.Toxin gene expression by shiga toxin-producing Es-cherichia coli: The role of antibiotics and the bacterialSOS response. Emerg. Infect. Dis. 6(5): 458-465.

King, L.J. 1990. The National Health Monitoring Sys-tem: Fulfilling a Commitment. Preven.Vet. Med. 8:89-95.

King, W., Raposa, S., Warshaw, J., Johnson, A., Hal-bert, D., and Klinger, J.D. 1989. A new colorimetricnucleic acid hybridization assay for Listeria in foods.Intl. J. Food Microbiol. 8: 225.

Kinugasa, H., Takeo, T., Fukumoto, K., and Ishihara, M.1992. Changes in tea components during processingand preservation of tea extract by hydrostatis pres-sure sterilization. Nippon Nogeikagaku Kaishi 66(4):707-712.

Klontz, K.C., Timbo, B., Fein, S., and Levy, A. 1995.Prevalence of selected food consumption and prepa-ration behaviors associated with increased risks offood-borne disease. J. Food Protect. 58: 927-930.

Kobayashi, S. and Ando, S. 2000. Reactive arthritis orReiter’s syndrome and B51-associated seroneagtivespondyloarthropathy. Intern. Med. 39(2): 89.

Kolter, R., Siegele, D.A., and Tormo, A. 1993. The sta-tionary phase of the bacterial life cycle. Annu. Rev. Mi-crobiol. 47: 855-874.

Koo, K. and Jaykus, L. 2000. A modified method todetect PCR products by 5' nuclease activity and anasymmetric fluorogenic probe set. BioTechniques 29:690-694.

Korkeala, H., Stengel, G., Hyytia, E., Vogelsang, B.,Bohl, A., Wihlman, H., Pakkala, P., and Hielm, S.1998. Type E botulism associated with vacuum-pack-aged hot-smoked whitefish. Intl. J. Food Microbiol.43: 1-5.

Kornacki, J.L. and Johnson, J.L. 2001. Enterobacteri-aceae, coliforms, and Escherichia coli as quality and

safety indicators. In ”Compendium of Methods for theMicrobiological Examination of Foods.” ed. F.P.Downes and K. Ito. pp. 69-82. American PublicHealth Association, Washington, D.C.

Korzenik, J.R. and Dieckgraefe, B.K. 2000. Is Crohn’sdisease an immunodeficiency? A hypothesis suggest-ing possible early events in the pathogenesis ofCrohn’s disease. Dig. Dis. Sci. 45(6): 1121-29.

Krishnamachari, K.A.V.R., Bhat, R.V., Nagarajan, V.,and Tilac, T.B.G. 1975. Hepatitis due to aflatoxicosis.Lancet 1: 1061-1063.

Kudva, I. T., C. W. Hunt, C. J. Williams, U. M. Nance,and Hovde, C. J. 1997. Evaluation of dietary influenc-es on Escherichia coli O157:H7 shedding by sheep.Appl. Environ. Microbiol. 63: 3878-3886.

Kulick, I. and Giachino, P. 1997. The alternative sigmafacter sB in Staphylococcus aureus: Regulation of thesignB operon in response to growth phase and heatshock. Arch. Microbiol. 167: 151-159.

Kushner, D.J. 1980. Extreme environments. In “Contem-porary Microbial Ecology,” ed. D.C. Ellwood, J.N.Hedger, M.J. Latham, et al., pp. 29-54. AcademicPress, Inc., London.

LLandini, P., Hajec, L.I., Nguyen, L.H., Burgess, R.R.,

and Volkert, M.R. 1996. The leucine-responsive regu-latory protein (Lrp) acts as a specific repressor for thesS-dependent transcription of the Escherichia coliaidB gene. Molec. Microbiol. 20: 947-955.

Lange, R. and Hengge-Aronis, R. 1991. Identification ofa central regulator of stationary-phase gene expres-sion in Escherichia coli. Molec. Microbiol. 5: 49-59.

Langhendries, J.P., Detry, J., Van Hees J., Lamboray,J.M., Darimont, J., Mozin, M.J., Secretin, M.C., andSenterre, J. 1995. Effect of a fermented infant for-mula containing viable bifidobacteria on the fecal floracomposition and pH of healthy full-term infants. J. Pe-diatr. Gastroenterol. Nutr. 21: 177-181.

Larsen, W.P., Hartzell, T.B., and Diehl, H.S. 1914. Theeffects of high pressure on bacteria. J. Infect. Dis. 22:271-279.

Lawlor, K.A., Pierson, M.D., Hackney, C.R., Claus, J.R.,and Marcy, J.E. 2000. Nonproteolytic Clostridiumbotulinum toxigenesis in cooked turkey stored undermodified atmospheres. J. Food Protect. 63: 1511-1516.

Lawrence, J.G. and Ochman, H. 1998. Molecular ar-chaeology of the Escherichia coli genome. Proc. Natl.Acad. Sci. 95: 9413-7.

Lee, L.A., Gerber, A.R., Lownsay, D.R., Smith, J.D.,Carter, G.P., Puhr, N.D., Puhr, N.D., Parrish, C.M.,Sikes, R.K., Finton, R.J., and Tauxe, R.V. 1990. Yers-inia enterocolitica 0:3 infections in infants and childrenassociated with household preparation of chitterlings.N. Eng. J. Med. 322: 984-7.

Lee, I. S., Lin, J., Hall, H.K., Bearson, B., and Foster,J.W. 1995. The stationary-phase sigma factor ss(RpoS) is required for a sustained acid tolerance re-sponse in virulent Salmonella typhimurium. Molec. Mi-crobiol. 17: 155-167.

Lee, L., Arul, J., Lencki, R., and Castaigne, F. 1996. Areview on modified atmosphere packaging and pres-ervation of fresh fruits and vegetables: Physiologicalbasis and practical aspects. Part 2. Packaging Tech-nol. Sci. 9: 1-17.

Leistner, L. 2000. Use of combined preservative factorsin foods of developing countries. Chpt. 14, In “TheMicrobiological Safety and Quality of Food.” Vol. I. ed.B.M. Lund, T.C. Baird-Parker, and G.W. Gould, pp.294-314. Aspen Publishers, Gaithersburg, Md.

Leung, C-K., Huang, Y-W., and Pansorbo, O.C. 1992.Bacterial pathogens and indicators in catfish andpond environments. J. Food Protect. 55: 424-427.

Levin, E.D., Schmechel, D.E., Burkholder, J.M., Glas-

103EXPERT REPORT

gow, H.B., Deamer-Melia, N., Moser, V.C., and Harry,G.J. 1997. Persistent learning deficits in rats after ex-posure to Pfiesteria piscicida. Environ. Health Per-spect. 105(12): 2-7.

Levine, P., Rose, B., Green, S., Ransom, G., and Hill, W.2001. Pathogen testing of ready-to-eat meat andpoultry products collected at federally inspected es-tablishments in the United States, 1990 to 1999. J.Food Protect. 64: 1188-1193.

Levy, S.B. 1992. The antibiotic paradox. How MiracleDrugs are Destroying the Miracle. Plenum Press, NewYork, N.Y.

Levy, A. 1997. Overview of consumer studies of foodsafety knowledge, attitudes and behavior. Presentedat Changing strategies-changing behavior: What foodsafety communicators need to know. June 12-13,Washington, D.C., U.S. Dept. of Agriculture and Foodand Drug Administration, Washington, D.C. (http://www.fda.gov)

Levy, S.B., Fitzgerald, G.B., and Macone, A.B. 1976.Changes in the intestinal flora of farm personnel afterintroduction of tetracycline-supplemented feed on afarm. N. Eng. J. Med. 295: 583-588.

Leyer, G.J. and Johnson, E.A. 1992. Acid adaptationpromotes survival of Salmonella spp. in cheese. Appl.Environ. Microbiol. 58: 2075-2080.

Leyer, G.J. and Johnson, E.A. 1993. Acid adaptationinduces cross-protection against environmentalstresses in Salmonella typhimurium. Appl. Environ.Microbiol. 59: 1842-1847.

Liao, C. and Sapers, G. 2000. Attachment and growthof Salmonella Chester on apple fruits and in vivo re-sponse of attached bacteria to sanitizer treatments.J. Food Protect. 63: 876-883.

Linton, A.H., Handley, B., Osborne, A.D., Shaw, B.G.,Roberts, T.A., and Hudson, W.R. 1976. Contamina-tion of pig carcasses at two abattoirs by Escherichiacoli with special reference to O-serotypes and antibi-otic resistance. J. Appl. Bacteriol. 41: 89-110.

Loewen, P.C., Hu, B., Strutinsky, J., and Sparling, R.1998. Regulation in the rpoS regulon of Escherichiacoli. Can. J. Microbiol. 44: 707-717.

Losinger, W.C., Wells, S.J., Garber, L.P., Hurd, H.S.,and Thomas, LA. 1995. Management factors relatedto Salmonella shedding by dairy heifers. J. Dairy Sci.78: 2474-2472.

Lou, Y. and Yousef, A.E. 1997. Adaptation to sublethalenvironmental stresses protects listeria monocytoge-nes against lethal preservation factors. Appl. Environ.Microbiol. 63: 1252-1255.

Lund, D. 1989. Food processing: From art to engineer-ing. Food Technol. 43(9): 242-247.

Luo, X. 1994. Food poisoning caused by Fusarium tox-ins, pp. 129-136. International Life Sciences Institute,Chatuchak, Thailand.

MMacKenzie, W.R., Hoxie, N.J., Proctor, M.E., Gradus,

M.S., Blair, K.A., Petersen, D.E., Kazmierczak, J.J.,Fox, K.R., Addiss, D.G., Rose, J.B., and Davis, J.P.1994. A massive outbreak in Milwaukee ofCryptosporidium infection transmitted through thepublic water supply. N. Eng. J. Med. 331: 161-167.

Mackey, B.M. 2000. Injured Bacteria. In “The Microbio-logical Safety and Quality of Foods.” ed. B.M. Lund,T.C. Baird-Parker and G.W. Gould. Aspen Publishers,Gaithersburg, Md.

Mahon, B.E., Ponka, A., Hall, W., Komatsu, K., Dietrich,S.E., Siitonen, A., Cage, G., Hayes, P.S.,Lambert-Fair,M.A., Bean, N.H., Griffin, P.M., and Slutsker, L.1997. An international outbreak of Salmonella infec-tions caused by alfalfa sprouts grown from contami-nated seeds. J. Infect. Dis. 175: 876-82.

Malinverni, R., Glauser, M.P., Bille, J., and Rocourt, J.1986. Unusual clinical features of an epidemic of list-

eriosis associated with a particular phage type. Eur.J. Clin. Microbiol. 5: 169-171.

Mallett, J.C., Beghian, L.E., Metcalf, T.G., and Kaylor,J.D. 1991. Potential of irradiation technology for im-proved shellfish sanitation. J. Food Safety 11: 231-245.

Manchester, A. and Clauson, A. 1995. 1994 spendingfor food away from home outpaces food at home.Food Review 18: 12-5.

Mandell, D., Bennett, G.L., and Dolin, J.E. 1995.“Bennett’s Principles and Practice of Infectious Dis-eases.” 4th Ed. W.B. Saunders and Co., New York,N.Y.

Manger, R.L., Leja, L.S., Lee, S.Y., Hungerford, J.M.,Hokama, Y., Dickey, R.W., Granade, H.R., Lewis, R.,Yasumoto, T., and Wekell, M.M. 1995. Detection ofsodium channel toxins: Directed cytotoxicity assaysof purified ciguatoxins, brevetoxins, saxitoxins, andseafood extracts. J. AOAC Intl. 78(2): 521-527.

Marks, P.J., Vipond, I.B., Carlisle, D., Deakin, D., Fey,R.E., and Caul, E.O. 2000. Evidence for airbornetransmission of Norwalk-like virus (NLV) in a hotelrestaurant. Epidemiol. Infect. 124: 481-487.

Marron, L., Emerson, N., Gahan, C.G., and Hill, C.1997. A mutant of Listeria monocytogenes L028unable to induce an acid tolerance response displaysdiminished virulence in a murine model. Appl. Envi-ron. Microbiol. 63(12): 4945-4947.

Marth, E.H. 1998. Extended shelf life refrigeratedfoods: Microbiological quality and safety. A ScientificStatus Summary by the Institute of Food Technolo-gists Expert Panel on Food Safety and Nutrition. Chi-cago, Ill. Food Technol. 52(2): 57-62.

Masson, P. 1992. Protein denaturation of proteins. In“High Pressure and Biotechnology,” ed. C. Balny, R.Hayashi, and P. Masson. pp. 89-99. John Libby Eu-rotext Ltd., London.

Matin, A., Auger, E.A., Blum, P.H., and Schultz, J.E.1989. Genetic basis of starvation survival in nondif-ferentiating bacteria. Ann. Rev. Microbiol. 43: 293-316.

Mattick, K.L., Jorgensen, F., Legan, J.D., Lappin-Scott, H.M., and Humphrey, T.J. 2000. Habituationof Salmonella spp. at reduced water activity and itseffect on heat tolerance. Appl. Environ. Microbiol.66: 4921-4925.

Maule, A. 2000. Survival of verocytotoxigenic Escheri-chia coli O157: H7 in soil, water and on surfaces. J.Appl. Microbiol. 88 (Suppl): 71S-78S.

Mazotta, A.S. 1999. Problems and solutions to devel-opment of pathogen resistance to thermal process-ing. ISSI Meeting.

McDaniel, T.K., Jarvis, K.G., Donnenberg, M.S., andKaper, J.B. 1995. A genetic locus of enterocyte ef-facement conserved among diverse enterobacterialpathogens. Proc. Natl. Acad. Sci. 92: 1664-8.

McDermott, M.R. and Bienestock, J. 1979. Evidencefor a common mucosal immunologic system I. Mi-gration of B immunoblasts into intestinal, respiratoryand genital tissues. J. Immunol. 122: 1892-1898.

McDougald, D., Rice, S.A., Weichart, D., and Kjelle-berg, S. 1998. Nonculturability: Adaptation or debil-itation? FEMS Microbiol. Ecol. 25: 1-9.

McFeters, G.A., Yu, F.P., Pyle, B.H., and Stewart, P.S.1995. Physiological assessment of bacteria usingfluorochromes. J. Microbiol. Meth. 21: 1-13.

McGhee, J.R., Mestecky, J., Dertzbaugh, M.T., Eld-ridge, J.H., Hirasawa, M., and Kiyono, H. 1992. Themucosal immune system: From fundamental con-cepts to vaccine development. Vaccine 10 (2): 75-88.

McGlynn, K.A., Rosvold, E.A., Lustbader, E.D., Hu, Y.,Clapper, M.L., Zhouo, T., Wild, C.P., Xia, X.L., Baf-foe-Bonnie, A., Ofori Adjei, D., Chen, G.-C., London,W.T., Shen, F.-M., and Buetow, K.W. 1995. Sus-

ceptibility to hepatocellular carcinoma is associatedwith genetic variation in the enzymatic detoxification ofaflatoxin B1. Proc. Natl. Acad. Sci. 92: 2384-2387.

McKay, A.M. 1992. Viable but non-culturable forms ofpotentially pathogenic bacteria in water. Lett. Appl.Microbiol. 14: 129-135.

McKillip, J.L. and Drake, M.A. 2000. Molecular bea-con-PCR detection of Escherichia coli 0157:H7 inmilk. J. Food Protect. 63: 855-859.

McKillip, J.L., Jaykus, L.A., and Drake, M.A. 1998.rRNA stability in heat-killed and UV-irradiated entero-toxigenic Staphylococcus aureus and Escherichia coliO157:H7. Appl. Environ. Microbiol. 64: 4264-4268.

McMullen, M., Jones, R., and Gallenberg, D. 1997.Scab of wheat and barley: A re-emerging disease ofdevasting impact. Plant Dis. 81(12): 1340-1348.

Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee,J.S., Shapiro, C., Griffin, P.M., and Tauxe, R.V. 1999.Food-related illness and death in the United States.Emerg. Infect. Dis. 5: 607-625.

Mecsas, J., Rouviere, P.E., Erickson, J.W., Donohue,T.J., and Gross, C.A. 1993. The activity of s E, anEschericia coli heat-inducible s-factor, is modulated byexpression of outer membrane proteins. Genes Dev.7: 2618-2628.

Meer, R.R. and Misner, S.L. 2000. Food safety knowl-edge and behavior of expanded food and nutritionprogram participants in Arizona. J. Food Protect. 63:1725-1731.

Mermelstein, N.H. 2001. Sanitizing Meat. Food Technol.55(3): 64-68.

Metchnikoff. E. 1908. “The prolongation of life.” 1st Ed.Putnams Sons, New York,l N.Y.

Meyer, R.S., Cooper, K.L., Knorr, D., and Lelieveld,H.L.M. 2000. High-pressure strerilization of foods.Food Technol. 54(11): 67-68, 70, 72.

Midgley, J. and Desmarchelier, P. 2001. Pre-slaughterhandling of cattle and Shiga toxin-producing Escheri-chia coli (STEC). Lett. Appl. Microbiol. 32: 307-311.

Miller, L.G. and Kaspar, C.W. 1993. Escherichia coliO157:H7 acid tolerance and survival in apple cider. J.Food Protect. 57: 460-464 [errata 57: 645].

Mitsuoka, T. 1989. “Microbes in the Intestine,” pp. 24-25. Yakult Honsha Co., Ltd., Tokyo.

Mitsuoka, T. 1990. “A Profile of Intestinal Bacteria.”Yakult Honsha Co., Ltd., Tokyo.

Miwa, N., Nishina, T., Kubo, S., Atsumi, M., and Honda,H. 1998. Amount of enterotoxigenic Clostridium per-fringens in meat detected by nested PCR. Intl. J.Food Microbiol. 42: 195-200.

Moe, C.L. 1996. Waterborne transmission of infectiousagents. In “Manual of Environmental Microbiology.”ed. C.J. Hurst, G.R. Knudsen, M.J. McInerney, L.D.Stenzenbach, and M.V. Walter. ASM Press, Washing-ton, D.C.

Moe, C.L., Rhodes, D., Pusek, S., Tseng, F., Heizer, W.,Kapoor, C., Gilliam, B., Harb, M., Stewart, P., Miller,S., Sobsey, M., Herrmann, J., Blacklow, N., and Cal-deron, R. 1998. Determination of Norwalk virusdose-response in human volunteers. Poster # C-384, In “Abstracts of the 98th Annual Meeting of theAmerican Society for Microbiology,” Atlanta, Ga.

Moe, C.L., Sobsey, M.D., Stewart, P.W., and Crawford-Brown, D. 1999. Estimating the risk of human calicivi-rus infection from drinking water. Abstr. P4-6. In “Ab-stract Book of the International Workshop on HumanCaliciviruses,” Atlanta, Ga.

Molbak, K., Baggesen, D.L., Aarestrup, F.M., Ebbesen,J.M., Engberg, J., Frydendahl, K., Gerner-Smidt, P.,Petersen, A.M., and Wegener, H.C. 1999. An out-break of multidrug resistant, quinolone-resistant Sal-monella enterica serotype Typhimurium DT104. N.Eng. J. Med. 341: 1420-1425.

Moleyar, V. and Narasimham, P. 1994. Modified atmo-sphere packaging of vegetables: An appraisal. J.


Food Sci. Technol. 31: 267-278.Molin, G. 2000. Modified Atmospheres. In “The Micro-

biological Quality and Safety of Food,” ed. B.M. Lund,T. C. Baird-Parker, and G.W. Gould. pp. 214-234.Aspen Publishers Inc., Gaithersburg, Md.

Monk, J.D., Beuchat, L.R., and Doyle, M.P. 1995. Irra-diation inactivation of food-borne microorganisms. J.Food Protect. 58: 197-208.

Morita, M.T., Kanemori, M., Yanagi, H., and Yura, T.2000. Dynamic interplay between antagonistic path-ways controlling the sigma 32 level in Escherichia coli.Proc. Natl. Acad. Sci. 97(11): 5860-5865.

Mossel, D.A., Corry, J.E.L., Struijk, C.B., and Baird,R.M. 1995. Essentials of the microbiology of foods:a textbook for advanced studies. John Wiley andSons. Chichester, England.

Mulvey, M.R., Switala, J., Borys, A., and Loewen, P.C.1990. Regulation of transcription of katE and katF inEscherichia coli. J. Bacteriol. 172: 6713-6720.

Murray, C.J.L. and Lopez, A.D. 1996. The Global Bur-den of Disease: A Comprehensive Assessment ofMortality and Disability from Diseases, Injuries, andRisk Factors in 1990 and Projected to 2020. Vol. I.Harvard Sch. of Public Health, Harvard UniversityPress, Cambridge, Mass.

Murray, C.J.L. and Lopez, A.D. 2001. The Global Epi-demiology of Infectious Diseases. In “Global Burden ofDisease and Injury.” Harvard Univ. Press. Cambridge,Mass.

Musser, J.M. 1996. Molecular population genetic anal-ysis of emerged bacterial pathogens: Selected in-sights. Emerg. Infect. Dis. 2: 1-17.

NNACMCF. 1992. Hazard analysis and critical control

point system. National Advisory Committee on Micro-biological Criteria for Foods. Intl. J. Food Microbiol.16: 1-25.

NACMCF. 1998. Hazard analysis and critical controlpoint principles and application guidelines. NationalAdvisory Committee on Microbiological Criteria forFoods. J. Food Protect. 61: 1246-1259.

NACMCF. 1999. Microbiological safety evaluations andrecommendations on sprouted seeds. National Advi-sory Committee on Microbiology Criteria for Food.Int. J. Food Microbiol. 52: 123-153.

Naniche, D. and Oldstone, M.B. 2000. Generalized im-munosuppression: How viruses undermine the im-mune response. Cell. Mol. Life Sci. 57(10): 1399-1407.

Narayan, K.G. 1966. Studies on clostridia incidence inbeef cattle. Acta Veterinaria Acad. Sci. Hung. 16: 65-72.

NAS. 1991. “Seafood Safety.” Committee on Evaluationof the Safety of Fishery Products, Food and NutritionBoard, Institute of Medicine. National Academies.Washington, D.C. National Academy Press. Washing-ton, D.C.

Nataro, J.P. and Kaper, J.B. 1998. Diarrheagenic Es-cherichia coli. Clin. Microbiol. Rev. 11: 142-201.

Navin, T.R. and Juranek, D.D. 1984. Cryptosporidiosis:Clinical, epidemiologic, and parasitologic review. Rev.Infect. Dis. 6: 313-27.

Newby, T.J. 1984. Protective immune responses in theintestinal tract. In “Local Immune Responses of theGut,” ed. T.J. Newby and C.R. Stokes. pp. 143-198.CRC Press, Boca Raton, Fla.

Ngindu, A., Kenya, P.R., Ocheng, D.M., Omonde, T.N.,Ngare, W., Gatei, D., Johnson, B.K., Ngira, J.A.,Nandwa, H., Jansen, A.J., Kaviti, J.N., and Siongok,T.A. 1982. Outbreak of acute hepatitis caused byaflatoxin poisoning in Kenya. Lancet 1: 1346-1348.

Nguyen-the, C. and Carlin, F. 2000. Fresh and pro-cessed vegetables. In “The Microbiological Safety andQuality of Food.” ed. B.M. Lund, T.C. Baird-Parker,

and G.W. Gould. pp. 620-684. Aspen Publishers, Inc.Gaithersburg, Md.

NIH/DMID. 1998. Crohn’s disease: Is there a microbialetiology? Division of Microbiology and Infectious Dis-eases. National Institute of Allergy and Infectious Dis-eases/National Institutes of Health, Bethesda, Md.

NLMB. 1994. Solving the E. coli O157:H7 problem: Ablueprint for industry action. Final Report of the BlueRibbon Task Force. National Live Stock and MeatBoard, National Cattlemens Beef Association, Centin-nial, Colo.

NMFS. 2000. Fisheries of the U.S. Current FisheriesStatistics. 1999. U.S. Department of Commerce,NOAA, National Marine Fisheries Service. SilverSpring, Md.

Noga, E.J. 1997. Ichthyotoxicity of Pfiesteria piscicida.NIEHS Workshop on Hazardous Marine/FreshwaterMicrobes and Toxins, North Carolina BiotechnologyCenter, Research Triangle Park, Raleigh, N.C.

Noga, E.J. 2000. Culturing of Pfiesteria and productionof Pfiesteria toxin(s). North Carolina State UniversityResearch Triangle Park, Raleigh, N.C.

Nogva, H.K., Bergh, A., Holck, A., and Rudi, K. 2000.Application of the 5'-nuclease PCR assay in evaluationand development of methods for quantitative detec-tion of Campylobacter jejuni. Appl. Environ. Microbi-ol. 66: 4029-4036.

Norton, D.-M. and Batt, C.A. 1999. Detection of viableListeria monocytogenes with a 5' nuclease PCR as-say. Appl. Environ. Microbiol. 65: 2122-2127.

Norton, D.M., McCamey, M.A., Gall, K.L., Scarlett,J.M., Boor, K.J., and Wiedmann, M. 2001. Molecu-lar studies on the ecology of Listeria monocytogenesin the smoked fish processing industry. Appl. Environ.Microbiol. 67: 198-205.

NRC. 1989. “Improving Risk Communication.” Commit-tee on Risk Perception and Communication. NationalResearch Council. National Academy Press, Washing-ton, D.C.

NRC. 1996. “Understanding Risk: Informing Decisionsin a Democratic Society.” Committee on Risk Charac-terization. Commission on Behavioral and Social Sci-ence and Education. National Research Council. Na-tional Academy Press, Washington, D.C.

Nurmi, E. and Rantala, M. 1973. New aspects of Sal-monella infection in broiler production. Nature 241:210-211.

Nurmi, E., Nuotio, L., and Schneitz, C. 1992. The com-petitive exclusion concept: Development and Future.Int. J. Food Microbiol. 15: 237-240.

OObiri-Dnaso, K., Paul, N., and Jones, K. 2001. The ef-

fects of UVB and temperature on the survival of natu-ral populations and pure cultures of Campylobacterjejuni, Camp. coli, Camp. lari and urease-positive ther-mophilic campylobacters (UPTC) in surface waters. J.Appl. Microbiol. 90: 256-267.

Oh, M.K. and Liao, J.C. 2000. Gene expression profil-ing by DNA microarrays and metabolic fluxes in Es-cherichia coli. Biotechnol. Prog. 16: 278-286.

Ojeniyi, B., Christensen, J., and Bisgaard, M. 2000.Comparative investigations of Listeria monocytogenesisolated from a turkey processing plant, turkey prod-ucts, and from human cases of listeriosis in Denmark.Epidemiol. Infect. 125(2): 303-308.

Okamura, N., Nakaya, R., Yokota, H., Yanai, N., andKawashima, T. 1986. Interaction of Shigella with bifi-dobacteria. Bifidobacteria Microflora 5(1): 51-55.

Olive, D.M. and Bean, P. 1999. Principles and applica-tions of methods for DNA-based typing of microbialorganisms. J. Clin. Microbiol. 37(6): 1661-1669.

Oliver, J.D. and Wanucha, D. 1989. Survival of Vibriovulnificus at reduced temperatures and elevated nutri-ent. J. Food Safety 10: 79-86.

Olsen, G.J., Overbeek, R., Larsen, N., Marsh, T.L., Mc-Caughey, M.J., Maciukenas, M.A., Kuan, W.M.,Macke, T.J., Xing, Y., and Woese, C.R. 1992. The ri-bosomal database project. Nucleic Acids Res. 11:2199-200.

Olsen, S.J., MacKinon, L.C., Goulding, J.S., Bean, N.H.,and Slutsker, L. 2000. Surveillance for foodborne-disease outbreaks-United States, 1993-1997.MMWR 49 (SS-1): 1-51.

Olson, D.G. 1998. Irradiation of food. A Scientific StatusSummary published by Institute of Food TechnologistsExpert Panel on Food Safety and Nutrition, Chicago,Ill. Food Technol. 52(1): 56-62.

O’Mahony, M., Mitchell, E., Gilbert R.J., Hutchinson,D.N., Begg, N.T., Rodhouse, J.C., and Morris, J.E.1990. An outbreak of foodborne botulism associatedwith contaminated hazelnut yoghurt. Epidemiol. Infect.104(3): 389-395.

O’Neal, C.R., Gabriel, W.M., Turk, A., Libby, S.J., Fang,F., and Spector, M.P. 1994. RpoS is necessary forboth the positive and negative regulation of starvationsurvival genes during phosphate, carbon, and nitrogenstarvation in Salmonella typhimurium. J. Bacteriol.176: 4610-4616.

Orr, R.V. and Beuchat, L.R. 2000. Efficacy of disinfec-tants in killing spores of Alicyclobacillus acidoterrestisand performance of media for supporting colony de-velopment of survivors. J. Food Protect. 63: 1117-1122.

Ortega, Y.R., Sterling, C.R., Gilman, R.H., Cama, V.A.,Diaz, F. 1993. Cyclospora species-A new protozoanpathogen of humans. N. Eng. J. Med. 328: 1308-1312.

Owen, R.L. and Ermak, T.H. 1990. Structural special-izations for antigen uptake and processing in the di-gestive tract. Springer Semin. Immunopathol. 12:139-152.

PPaci, E. 1995. Exploring new tourism marketing oppor-

tunities around the world. Proc. of the 11th GeneralAssembly of the World Tourism Organization, Oct. 15.Cairo.

Pagan, R., Jordan, S., Benito, A., and Mackey, B.2001. Enhanced acid sensitivity of pressure-damagedEscherichia coli 0157 cells. Appl. Environ. Microbiol.67: 1983-1985.

Parashar, U.D., Dow, L., Fankhauser, R.L., Humphrey,C.D., Miller, J., Ando, T., Williams, K.S., Eddy, C.R.,Noel, J.S., Ingram, T., Bresee, J.S., Monroe, S.S.,and Glass, R.I. 1998. An outbreak of viral gastroen-teritis associated with consumption of sandwiches:Implications for the control of transmission by foodhandlers. Epidemiol. and Infect. 121: 615-621.

Parker, C.T. and Thomas, D. 2000. Reiter’s syndromeand reactive arthritis. J. Am. Osteopath. Assn. 100(2):101-104.

Patterson, M.F. and Kilpatrick, D.J. 1998. The com-bined effect of high hydrostatic pressure and mildheat on inactivation of pathgens in milk and poultry. J.Food Protect. 61: 432-436.

Patterson, M.F., Quinn, M., Simpson, R., and Gilmour,A. 1995. Sensitivity of vegetative pathogens to highhydrostatic pressure treatment in phosphate-bufferedsaline and foods. J. Food Protect. 58: 524-529.

Patterson, W., Haswell, P., Fryers, P.T., and Green, J.1997. Outbreak of small round structured virus gas-troenteritis arose after kitchen assistant vomited.Commun. Dis. Rep. 7(7): R101-3.

Paul, K.Y. and Morita, R.Y. 1971. Effects of hydrostaticpressure and temperature on the uptake and respira-tion of amino acids by a facultatively psychrophilicmarine bacterium. J. Bacteriol. 108: 835.

Pauli, G.H. and Tarantino, L.M. 1995. FDA regulatoryaspects of food irradiation. J. Food Protect. 58:

105EXPERT REPORT

209-212.Peck, M.W. and Fernandez, P.S. 1995. Effect of

Iysozyme concentration, heating at 90°C, and thenicubation at chilled temperatures on growth fromspores of non-proteolytic Clostridium botulinum. Let.Appl. Microbiol. 21: 50-54.

Penner, K., Kramer, C., and Frantz, G. 1985. Consumerfood safety perceptions. MF774. Kansas State Univ.Ext. Service, Manhattan, Kansas.

Perna, N.T., Plunkett, G., Burland, V., Mau, B., Glasner,J.D., Rose, J.D., Mayhew, G.F., Evans, P.S., Gregor,J., Kirkpatrick, H.A., Posfai, G., Hackett, J., Klink, S.,Boutin, A., Shao, Y., Miller, L., Grotbeck, E.J., Davis,N.W., Lim, A., Dimalanta, E.T., Potamousis, K.D.,Apodaca, J., Anantharaman, T.S., Lin, J., Yen, G.,Schwartz, D.C., Welch, R.A., and Blattner, F.R. 2001.Genome sequence of enterohaemorrhagic Escherichiacoli O157:H7. Nature 409: 529-533.

Pestka, J.J. 1993. Food, diet and gastrointestinal im-mune function. Adv. Food Nutr. Res. 37: 1-66.

Pestka, J.J. and Casale, W.L. 1990. Naturally occurringfungal toxins. Adv. Environ. Sci. Technol. 23: 613-638.

Pestka, J.J. and Witt, M.F. 1985. An overview of im-mune function. Food Technol. 39(2): 83-90.

Pestka, J.J., Abouzied, M.M., and Sutikno, M. 1995.Immunologically-based assays for mycotoxin detec-tion. Food Technol. 49(2): 120-128.

Pflug, I.J. and Gould, G.W. 2000. Heat treatment. Chpt.3, In “The Microbiological Safety and Quality ofFood.” Vol. I. ed. B.M. Lund, T.C. Baird-Parker, andG.W. Gould. pp. 36-64. Aspen Publishers Inc., Gaith-ersburg, Md.

Phebus, R.K., Nutsch, A.L., Schafer, D.E., Wilson, R.C.,Riemann, M.J., Leising, J.D., Kastner, C.L., Wolf,J.R., and Prasia, R.K. 1997. Comparison of steampasteurization and other methods for reduction ofpathogens on surfaces of freshly slaughtered beef. J.Food Protect. 59: 751-756.

Phillips, C.A. 1996. Review: Modified atmosphere pack-aging and its effects on the microbiological qualityand safety of produce. Intl. J. Food Sci. Technol.31: 463-479.

Pitlik, S.D., Fainstein, V., Garza, D., Guarda, L., Bolivar,R., Rios, A., Hopfer, R.L., Mansell, P.A. 1983. Hu-man cryptosporidiosis: Spectrum of disease. Reportof six cases and review of the literature. Arch. Intern.Med. 143: 2269-75.

Pitt, J.I. 2000. Toxigenic fungi and mycotoxins. Br.Med. Bull. 56(1): 184-92.

Plumley, F.G. 1997. Marine algal toxins: Biochemistry,genetics, and molecular biology. Limnol. Oceanog.42(5): 1252-1264.

Porter, J., Edwards, C., and Pickup, R.W. 1995. Rapidassessment of physiological status in Escherichia coliusing fluorescent probes. J. App. Bacteriol. 79: 399-408.

Potter, M.E. and Tauxe, R.V. 1997. Epidemiology offoodborne diseases: Tools and Applications. WorldHealth Statistics Quarterly 50: 24-29.

Price, S. B., Cheng, C. -M., Kaspar, C.W., Wright, J.C.,DeGraves, F., Penfound, T.A., Castanie-Cornet, M.-P.,and Foster, J.W. 2000. Role of rpoS in acid resis-tance and fecal shedding of Escherichia coli O157:H7. Appl. Environ. Microbiol. 66: 632-637.

Proctor, M.E., Brosch, R., Mellen, J.W., Garrett, L.A.,Kaspar, C.W., and Luchansky, J.B. 1995. Use ofpulsed field gel electrophoresis to link sporadic casesof invasive listeriosis with recalled chocolate milk.Appl. Environ. Microbiol. 61(8): 3177-3179.

Pruimboom-Brees, I.M., Morgan, T.W., Ackermann,M.R., Nystrom, E.D., Samuel, J.E., Cornick, N.A.,and Moon, H.W. 2000. Cattle lack vascular receptorsfor Escherichia coli O157: H7 Shiga toxins. Proc.Natl. Acad. Sci. 97: 10325-10329.

RRabsch, W., Hargis, B.M., Tsolis, R.M., Kingsley, R.A.,

Hinz, K.H., Tschape, H., and Baumler, A.J. 2000.Competitive exclusion of Salmonella enteritidis by Sal-monella gallinarum in poultry. Emerg. Infect. Dis. 6:443-448.

Rahyaman, S.M., Chira, P., and Koff, R.S. 1994. Idio-pathic autoimmune chronic hepatitis triggered byHepatitis A. Am. J. Gastroenterol. 89(1): 106-8.

Ralston, K., Starket, Y., Brent, P., and Riggins, T. 2000.Awareness of risks changing how hamburgers arecooked. Food Rev. May-Aug. pp. 44-50.

Ramsaran, H., Chen, J., Brunke, B., Hill, A., and Grif-fiths, M.W. 1998. Survival of bioluminescent Listeriamonocytogenes and Escherichia coli O157: H7 insoft cheeses. J. Dairy Sci. 81: 1810-1817.

Rangarajan, A., Pritts, M., Reiners, S., and Pedersen, L.2000. Reduce microbial contamination with goodagricultural practices. Cornell University, Ithaca, N.Y.(http://www.hort.cornell.edu/veggie/foodsafety/foodsafety.htm.).

Rashid, H.O., Ito, H., and Ishigaki, I. 1992. Distributionof pathogenic vibrios and other bacteria in importedfrozen shrimps and their decontamination by gamma-irradiation. World J. Microbiol. Biotechnol. 8: 494-499.

Rasmussen, O.F., Skouboe, P., Dons, L., Rossen, L.,and Olsen, J.E. 1995. Listeria monocytogenes existsin at least three evolutionary lines: Evidence fromflagellin, invasive associated protein, and listeriolysin Ogenes. Microbiol. 141: 2053-2061.

Ray, B. 1979. Methods to detect stressed microorgan-isms. J. Food Protect. 42: 346-355.

Ray, B. 1989. “Injured Index and Pathogenic Bacteria:Occurrence and Detection in Foods, Water andFeeds.” CRC Press, Inc., Boca Raton, Fla.

Reagan, J.O., Acuff, G.R., Buege, D.R., Buyck, M.J.,Dickson, J.S., Kastner, C.L., Marsden, J.L., Morgan,J.B., Nickelson, R., Smith, G.C., and Sofos, J.N.1996. Trimming and washing of beef carcasses as amethod of improving the microbiological quality ofmeat. J. Food Protect. 59: 751-756.

Reid, S.D., Herbelin, C.J., Bumbaugh, A.C., Selander,R.K., and Whittam, T.S. 2000. Parallel evolution ofvirulence in pathogenic Escherichia coli. Nature.406(6791): 64-67.

Reilly, P.J.A., Twiddy, D.R., and Fuchs, R.S. 1992. Re-view of the occurrence of salmonella in cultured tropi-cal shrimp. FAO. Fisheries Cirular No. 851. Food andAgriculture Organization of the United Nations, Rome,Italy.

Reis, A.A., Zaza, S., Langkop, C., Tauxe, R.V., andBlake, P.A. 1990. A multistate outbreak of Salmonellachester linked to imported cantaloupe (Abstr.). Proc.of the 3Oth Conference on Antimicrobial Agents andChemotherapy, Oct. 21-24, Atlanta, Ga. p. 38. Amer-ican Society for Microbiology, Washington, D.C.

Rice, E.W. and Johnson, C.H. 2000. Short communica-tion: Survival of Escherichia coli O157:H7 in dairycattle drinking water. J. Dairy Sci. 83: 2021-2023.

Richter, H., Timm, M., Lehmann, S., Lochotzke, H.-M.,Perlberg, K.-W., Feier, U., and Protz, D. 2000. Influ-ence of different feed rations on E. coli and VTEC inthe faeces of cattle. (Abstr. 150). Kyoto, Japan.(http://www.iss.it/laboratori/leb/vtec/abstractsP2.htm).

Robbe-Saule, V., Schaeffer, F., Kowarz, L., and Norel, F.1997. Relationships between H-NS, ss, SpvR andgrowth phase in the control of spvR, the regulatorygene of the Salmonella typhimurium plasmid virulenceoperon. Mol. Gen. Genet. 256: 333-347.

Roberts, C.M. and Hoover, D.G. 1996. Sensitivity ofBacillus coagulans spores to combinations of high hy-drostatic pressure, heat, acidity and nisin. J. App.Bacteriol. 81: 363-368.

Roberts, R.A. 1976. Microbiological guidelines formeat. Proc. 22nd Meeting of European Meat Re-search Workers, pp. LO: 1 – LO: 4. Malmo, Sweden.

Robertson, R.E. 1999. Fresh Produce: Potential Impli-cations of Country-of-Origin Labeling. Before the Sen-ate Committee on Agriculture, Nutrition, and Forestry.May 26. Testimony GAO/T-RCED-99-200.

Rodrigo, D., Martinez, A., Harte, F., Barbosa-Canovas,G.V., and Rodrigo, M. 2001. Study of inactivation ofLactobacillus plantarum in orange-carrot juice bymeans of pulsed electric fields: Comparison of inacti-vation kinetics models. J. Food Protect. 64(2): 259-263.

Rouquette, C., Ripio, M.T., Pellegrini, E., Bolla, J.M.,Tascon, R.I., Vazquez-Boland, J.A., and Berche, P.1996. Identification of a ClpC ATPase required forstress tolerance and in vivo survival of Listeria mono-cytogenes. Mol. Microbiol. 21: 977-987.

Rowbury, R.J. 1995. An assessment of environmentalfactors influencing acid tolerance and sensitivity in Es-cherichia coli, Salmonella spp. and other enterobacte-ria. Lett. Appl. Microbiol. 20: 333-337.

Russell, M.W., Lue, C., Van Den Wall Bake, A.W.L.,Moldoveanu, Z., and Mestecky, J. 1991. Molecularheterogeneity of human IgA antibodies during an im-mune response. Clin. Exp. Immunol. 87: 1-6.

Russell, N.J., Evans, R.I., ter Steeg, P.F., Hellemons, J.,Verheul, A., and Abee, T. 1995. Membranes as a tar-get for stress adaptation. Intl. J. Food Microbiol. 28:255-261.

SSaavedra, J.M., Bauman, N.A., Oung, I., Perman, J.A.,

and Yolken, R.H. 1994. Feeding of Bifidobacteriumbifidum and Streptococcus thermophilus to infants inhospital for prevention of diarrhoea and shedding ofrotavirus. Lancet 344: 1046-1049.

Sagliocca, L., Amoroso, P., Stroffolini, T., Adamo, B.,Tosti, M.E., Lettieri, G., Esposito, C., Buoocore, S.,Pierri, P., and Mele, A. 1999. Efficacy of hepatitis Avaccine in prevention of secondary hepatitis infection:A randomised trial. Lancet 353(9159): 1136-1139.

Salman, M.D., Frank, G.R., MacVean, D.W., Reif, J.S.,Collins, J.K., and Jones, R. 1988. Validation of dis-ease diagnoses reported to the National Animal HealthMonitoring System from a large Colorado beef feed-lot. J. Am. Vet. Med. Assn. 192: 1069-1073.

Sandman, P.M. 1987. Risk communication: Facingpublic outrage. EPA Journal 13(9): 21-22.

Sandman, P.M. 1997. Persuading the public to takeserious food risks seriously and trivial food risks trivial-ly. Presented at “Changing strategies-changing behav-ior: What food safety communicators need to know.”June 12-13, Washington, D.C., U.S. Dept. of Agricul-ture and Food and Drug Administration, Washington,D.C. (http://www.fda.gov).

Sargeant, J.M., Gillespie, J.R., Oberst, R.D., Phebus,R.K., Hyatt, D.R., Bohra, L.K., and Galland, J.C.2000. Results of a longitudinal study of the preva-lence of Escherichia coli O157:H7 on cow-calf farms.Am. J. Vet. Res. 61: 1375-1379.

Savitz, D. 1998. Final Report: Medical evaluation of es-tuary-exposed persons in North Carolina, November,1997. Department of Epidemiology, University ofNorth Carolina,. Chapel Hill.

Schuchat, A., Deaver, K.A., Wenger, J.D., Plikaytis,B.D., Mascola, L., Pinner, R.W., Reingold, A.L.,Broome, C.V., and the Listeria Study Group. 1992.Role of Foods in Sporadic Listeriosis. J. Am. Med.Assn. 267(15): 2041-2045.

Schwartz, I. 2000. Microbial genomics: From sequenceto function. Emerg. Infect. Dis. 6: 493-495.

Seidegard, J. and Ekstrom, G. 1997. The role of huy-man glutathione transferases and epoxide hydrolasesin the metabolism of xenobiotics. Environ. Health Per-


spect. 105: 791-799.Shadbolt, C., Ross, T., and McMeekin, T.A. 2001. Dif-

ferentiation of the effects of lethal pH and water activ-ity: Food safety implications. Lett. Appl. Microbiol. 32:99-102.

Shane, S.M. 2000. Campylobacter infection of com-mercial poultry. Rev. Sci. Tech. 19: 376-395.

Sharma,V.K. and Carlson, S.A. 2000. Simultaneous de-tection of Salmonella strains and Escherichia coliO157: H7 with fluorogenic PCR and single-enrich-ment-broth culture. Appl. Environ. Microbiol. 66:5472-5476.

Sharpe, A.N. 1997. Separation and Concentration ofPathogens from Foods. In “Food MicrobiologicalAnalysis: New Technologies.” ed. M.L. Tortorello andS.M. Gendel. Marcel Dekker. New York, N.Y.

Shearer, A.E.H., Dunne, C.P., Sikes, A., and Hoover,D.G. 2000. Bacterial spore inhibition and inactivationin foods by pressure, chemical preservatives and mildheat. J. Food Protect. 63: 1503-1510.

Shere, J.A., Bartlett, K.J., and Kaspar, C.W. 1998.Longitudinal study of Escherichia coli O157:H7 dis-semination on four dairy farms in Wisconsin. Appl. En-viron. Microbiol. 64: 1390-1399.

Sheridan, J.J. and McDowell, D.A. 1998. Factors af-fecting the emergence of pathogens on foods. MeatSci. 49(Suppl. 1): 151-167.

Sheridan, G.E.C., Masters, C.I., Shallcross, J.A., andMackey, B.M. 1998. Detection of mRNA by reversetranscription PCR as an indicator of viability in Escher-ichia coli cells. Appl. Environ. Microbiol. 64: 1313-1318.

Shoenfeld, Y., George, J., and Peter, J.B. 1996. Guil-lain-Barré as an autoimmune disease. Intl. Arch. Aller-gy Immunol. 109: 318-26.

Shumway, S.E. 1989. Toxic algae. World Aquacult. 20:65-73.

Simonsen, G.S., Haaheim, H., Dahl, K.H., Kruse, H.,Lovseth, A., Olsvik, O., Sundsfjord, A. 1998. Trans-mission of VanA-type vancomycin-resistant enterococ-ci and vanA resistance elements between chicken andhumans at avoparcin-exposed farms. Microb. DrugResist. 4: 313-8.

Simpkins, S.A., Chan, A.B., Hays, J., Popping, B., andCook, N. 2000. An RNA transcription-based amplifi-cation technique (NASBA) for the detection of viableSalmonella enterica. Lett. Appl. Microbiol. 30: 75-79.

Siragusa, G.R. 1996. The effectiveness of carcass de-contamination systems for controlling the presence ofpathogens on the surfaces of meat animal carcasses.In “HACCP: An Integrated Approach to Assuring theMicrobiological Safety of Meat and Poultry.” ed. J.L.Sheridan, R.L. Buchanan, and T.J. Montville. Foodand Nutrition Press, Trumbull, Conn.

Siragusa, G.R., Nawotka, K., Spilman, S.D., Contag,P.R., and Contag, C.H. 1999. Real-time monitoringof Escherichia coli O157: H7 adherence to beef car-cass surface tissues with a bioluminescent reporter.Appl. Environ. Microbiol. 65: 1738-1745.

Smelt, J.P.P.M., Rijke, A.G.F., and Hayhurst, A. 1994.Possible mechanism of high pressure inactivation ofmicroorganisms. High Pressure Res. 12: 199-203.

Smith, J.A. 1995. Arthritis, guillain-barre’ syndrome, andother sequelae of Campylobacter jejuni enteritis. J.Food Protect. 58: 1153-1170.

Smith, S.A., Noga, E.J., and Bullis, R.A. 1988. Mortali-ty in Tilapia aurea due to a toxic dinoflagellate bloom.(Abstr.) In Proc. of the 3rd Internat. Colloq. Pathol.Marine Aquacul. pp. 167-168. Gloucester Point, Va.

Smith, K.E., Besser, J.M., Hedberg, C.W., Leano, F.T.,Bender, J.B., Wicklund, J.H., Johnson, B.P., Moore,K.A., and Osterholm, M.T. 1999. Quinolone-resistantCampylobacter jejuni infections in Minnesota,1992-1998. Investigation Team. N. Eng. J. Med.340: 1525-32.

Smoot, L.M. and Pierson, M.D. 1998. Effect of envi-

ronmental stress on the ability of Listeria monocyto-genes Scott A to attach to food contact surfaces. J.Food Protect. 61: 1293-1298.

Soderlund, D. et al. 2000. Intl. Conf. Emerg. Infect. Dis.(Abstr. 62) p. 148.

Sorvillo, F.J., Lieb, L.E., Kerndt, P.R., and Ash, L.R.1994. Epidemiology of cryptosporidiosis among per-sons with acquired immunodeficiency syndrome inLos Angeles County. Am. J. Trop. Med. Hyg. 51(3):326-331.

Spencer, J.L. and Garcia, M.M. 1995. Resistance ofchicks and poults fed vermicompost to caecal coloni-zation by Salmonella. Avian Pathol. 24: 157-170.

Sperandio, V., Mellies, J.L., Nguyen, W., Shin, S., andKaper, J.B. 1999. Quorum sensing controls the ex-pression of the type III secretion gene transcriptionand protein secretion in enterohemorrhagic and en-teropathogenic Escherichia coli. Proc. Nat. Acad. Sci.96(26): 15196-15201.

Spratt, B.G. 1999. Multilocus sequence typing: Molecu-lar typing of bacterial pathogens in an era of rapidDNA sequencing and the internet. Curr. Opin. Microbi-ol. 2(3): 312-316.

St. Louis, M.E., Morse, D.L., Potter, M.E., Demelfig,T.M., Guzewich, J.J., Tauxe, R.V., Blake, P.A., and theSalmonella enteritidis Working Group. 1988. Theemergence of grade A eggs as a major source ofSalmonella enteritidis infections: New implications forthe control of salmonellosis. J. Am. Med. Assn. 259:2103-2107.

Stehr-Green, J.K. and Schantz, P.M. 1986. Trichinosisin Southeast Asian refugees in the United States. Am.J. Publ. Health 76: 1238-39.

Steidinger, K.A. and Penta, H.L.M. 1999. “Harmful Mi-croalgae and Associated Public Health Risks in theGulf of Mexico”. Marine Research Institute, St. Peters-burg, Fla.

Stewart, G.S.A.B., Aldsworth, T.G., Sharman, R.L., Gib-son, P.T., and Dodd, C.E.R. 1997. Bioluminescence:Lux as an enabling tool for the microbiological analy-sis of food. Chpt. 13 in “Food Microbiological Analy-sis: New Technologies,” ed. M. L. Tortorello and S.M.Gendel, pp. 265–288. Marcel Dekker, New York,N.Y.

Stokes, C.R. 1984. Induction and control of intestinalimmune responses. In “Local Immune Responses inthe Gut.” ed. T.J. Newby and C.R. Stokes. CRCPress, Boca Raton, Fla.

Sulaiman, I.M., Xiao, L., Yang, C., Escalante, L., Moore,A., Beard, C.B., Arrowood, M.J., and Lal, A.A.1998. Differentiating human from animal isolates ofCryptosporidium parvum. Emerg. Infect. Dis. 4: 681-685.

Surette, M.G. and Bassler, B.L. 1998. Quorum sensingin Escherichia coli and Salmonella typhimurium. Proc.Nat. Acad. Sci. 95(12): 7046-7050.

Swaminathan, B. and Feng, P. 1994. Rapid detection offood-borne pathogenic bacteria. Ann. Rev. Microbiol.48: 401-426.

Swaminathan, B., Barrett. T.J., Hunter, S.B., Tauxe, R.V.,and the CDC PulseNet Task Force. 2001. PulseNet:The molecular subtyping network for foodborne bac-terial disease surveillance, United States. Emerg. In-fect. Dis. 7(3): 382-389.

Swift and Company. 1973. Carcass chilling process.U.S. Pat. 1,380,891.

TTacket, C.O., Brenner, F., and Blake, P.A. 1984. Clinical

features of an epidemiological study of Vibrio vulnifi-cus infections. J. Infect. Dis. 149: 558-561.

Tai, J-H. and Pestka, J.J. 1988. Synergistic interactionbetween the trichothecene T-2 toxin and Salmonellatyphimurium lipopolysaccharide in C3H/HeN andC3H/HeJ mice. Toxicol. Lett. 44: 191-200.

Takahashi, I. and Kiyono, H. 1999. Gut as the largestimmunologic tissue. J. Parenter. Enteral. Nutr. 23(5Supppl): 7S-12S.

Tao, H., Bausch, C., Richmond, C., Blattner, F.R., andConway, T. 1999. Functional genomics: Expressionanalysis of Escherichia coli growing on minimal andrich media. J. Bacteriol. 181: 6425-6440.

Taormina, P.J., Beuchat, L.R., and Slutsker, L. 1999.Infections associated with eating seed sprouts: An in-ternational concern. Emerg. Infect. Dis. 5: 626-634.

Taunk, J., Roberts, A.I., and Ebert, E.C. 1992. Sponta-neous cytotoxicity of human intraepithelial lympho-cytes against epithelial cell tumors. Gastroenterol.102: 69-75.

Tauxe, R.V. 1997. Emerging Foodborne Diseases: AnEvolving Public Health Challenge. Emerg. Infect. Dis.3(4): 425-434.

Tauxe, R., Kruse, H., Hedberg, C., Potter, M., Madden,J., and Wachsmuth, K. 1997. Microbial hazards andemerging issues associated with produce. A prelimi-nary report to the National Advisory Committee OnMicrobiological Criteria For Foods. J. Food Protect.60: 1400-1408.

Tenter, A.M., Heckeroth, A.R., and Weiss, L.M. 2000.Toxoplasma gondi: From animals to humans. Intl. J.Parasitol. 30(12-13): 1217-58.

Thacker, S.B. and Berkelman, R.L. 1988. Public HealthSurveillance in the United States. Epidemiol. Rev.10(2): 187-203.

Thompson B., Demark-Wahnefried, W., Taylor, G., Mc-Clelland, J.W., Stables, G., Havas, S., Feng, Z., Topor,M., Heimendinger, J., Reynolds, K.D., and Cohen, N.1999. Baseline fruit and vegetable intake amongadults in seven 5 a day study centers located in di-verse geographic areas. J. Am. Diet. Assn. 99: 1241-1248.

Threlfall, E.J., Frost, J.A., Ward, L.R., and Rowe, B.1995. Epidemic in cattle of S. typhimurium DT104with chromosomally-integrated multiple drug resis-tance. Vet. Rec. 134: 577.

Threlfall, E.J., Ward. L.R., and Rowe, B. 1997. Increas-ing incidence of resistance to trimethoprim and cipro-floxacin in epidemic Salmonella typhimurium DT104 inEngland and Wales. Eurosurveill. 2: 81-84.

Thorne, G.M. 1986. Gastrointestinal infections: Dietaryinteractions. J. Am. Coll. Nutr. 5: 487-499.

Timbo, B., Headrick, M., Altekruse, S.F., and Klontz,K.C. 1995. Raw shellfish consumption in California:The 1992 California behavioral risk factor survey. Am.J. Prev. Med. 11: 214-6.

Tompkin, R.B. 1983. Indicator organisms in meat andpoultry products. Food Technol. 37(6): 107-110.

Tompkins, L.S. 1992. The use of molecular methods ininfectious diseases. N. Engl. J. Med. 327(18): 1290-1297.

Toro, A.M. Resto, Gonzalez, N., Dufour, A., and Toran-zos, G.A. 1995. Oysters as possible vectors for Giar-dia and Cryptosporidium. (Abstr.) p. 405. 95th Gen-eral Meeting of the American Society of Microbiology,May 21-25, Washington, D.C. American Society forMicrobiology, Washington, D.C.

UUeno, Y. 1974. Citreoviridin from Penicillium citreoviride

Biourge. In “Mycotoxins.” ed. I.F.H. Purchase. pp.283-302. Elsevier, New York, N.Y.

Uljas, H.E., Schaffner, D.W., Duffy, S., Zhao, L.H., andIngham, S.C. 2001. Modeling of combined process-ing steps for reducing Escherichia coli O157: H7populations in apple cider. Appl. Environ. Microbiol.67(1): 133-141.

USDA/AMS. 2000. National Organic Program; FinalRule. 56FR 80548. U.S. Dept. of Agriculture/Agricul-tural Marketing Service, Washington, D.C.

USDA/FSIS. 1996. Nationwide federal plant ground

107EXPERT REPORT

beef microbiological survey. U.S. Dept. of Agriculture/Food Safety and Inspection Service. Washington, D.C.

USDA/FSIS. 1998a. Salmonella Enteritidis risk assess-ment: Shell eggs and egg products. Final Report.Prepared for the Food Safety and Inspection Serviceby the Salmonella Enteritidis Risk Assessment Team.U.S. Dept. Agriculture/Food Safety and InspectionService, Washington, D.C. (http://www.fsis.usda.gov/OPHS/risk/index.htm).

USDA/FSIS. 1998b. Risk assessment for E. coliO157:H7 in beef and ground beef. U.S. Dept. of Ag-riculture/Food Safety and Inspection Service, Wash-ington, D.C.

USDA/FSIS. 2000. Progress report on Salmonella test-ing of raw meat and poultry products, 1998-2000.U.S. Dept. of Agriculture/Food Safety and InspectionService, Washington, D.C. (http://www.fsis.usda.gov).

USDA/FSIS. 2001. FSIS Guidance for Water, Ice, andSolution Reuse. U.S. Dept. of Agriculture/Food Safetyand Inspection Service. Washington, D.C. (http://www.fsis.usda.gov/oppde/rdad/frpubs/san-update1.pdf).

USDA/FSIS. 2002. Microbiological results of rawground beef products analyzed for Escherichia coliO157:H7: Calendar year 2001. U.S. Dept. of Agricul-ture/Food Safety and Inspection Service. Washington,D.C. (http://www.fsis.usda.gov/OPHS/ecoltest/ecpositives2001.htm).

USDA/NASS. 2000. Meat animals production, disposi-tion, and income, 1999 summary. U.S. Dept. of Agri-culture/ National Agriculture Statistics Service. Wash-ington, D.C.

USDA/NASS. 2001. Fruit and vegetable agriculturalpractices –1999, June. U.S. Department of Agricul-ture/National Agricultural Statistics Service, Washing-ton, D.C. (http://www.usda.gov/nass/pubs/rpts106.htm).

USPSTF. 1997. Guide to Clinical Preventive Services,2nd Ed. pp. 625-642. U.S. Preventive Services TaskForce. International Medical Publishing. Washington,D.C.

U.S. Senate Agriculture Committee. 1998. Animal WastePollution in America: An Emerging National Problem.U.S. Congress, Democratic Staff, Washington, D.C.

VVan Beneden, C.A, Keene, W.E., Strang, R.A., Werker,

D.H., King, A.S., Mahon, B., Hedberg, K., Bell, A.,Kelly, M.T., Balan, V.K., MacKenzie, W.R., and Flem-ing, D. 1999. Multinational outbreak of Salmonellaenterica serotype. Newport infections due to contami-nated alfalfa sprouts. J. Am. Med. Assn. 281(2):158-162.

Van den Braak, N., van Belkum, A., van Keulen, M.,Vliegenthart, J., Verbrugh, H.A., and Endtz, H.P.1998. Molecular characterization ofvancomycin-resistant enterococci from hospitalizedpatients and poultry products in The Netherlands. J.Clin. Microbiol. 36: 1927-32.

Van der Poel, W.H.M., Vinje, J., van der Heide, R., Her-rere, M.-I., Vivo, A., and Koopmans, M.P.G. 2000.Norwalk-like calicivirus genes in farm animals. Emerg.Infect. Dis. 6: 36-41.

Van Embden, J.D., Cave, M.D., Crawford, J.T., Dale,J.W., Eisenach, K.D., Gicquel, B., Hermans, P., Mar-tin, C., McAdam, R., Shinnick, T.M., and Small, P.M.

1993. Strain identification of Mycobacterium tubercu-losis by DNA fingerprinting: Recommendations for astandardized methodology. J. Clin. Microbiol. 31(2):406-409.

Versalovic, J., Woods, C.R., Jr., Georghiou, P.R., Hamill,R.J., and Lupski, J.R. 1993. DNA based identifica-tion and epidemiologic typing of bacterial pathogens.Arch. Pathol. Lab Med. 117(11): 1088-1098.

WWaldor, M.K. and Mekalanos, J.J. 1996. Lysogenic

conversion by a filamentous phage encoding choleratoxin. Science 272: 1910-1914.

Walker, W.A. 1987. Pathophysiology of intestinal uptakeand absorption of antigens in food allergy. Annals Al-lergy 59: 7-16.

Wall, P.G., Morgan, D., Lamden, K., Ryan, M., Griffin,M., Threlfall, E.J., and Ward, L.R. 1994. A case con-trol study of infection with an epidemic strain of multi-resistant Salmonella Typhimurium DT104 in Englandand Wales. CDR Rev. 4: R130-R135.

Wallace, B.J., Guzewich, J.J., Cambridge, M., Al-tekrause, S., and Morse, D.L. 1999. Seafood-associ-ated disease outbreaks in New York, 1980-1994.Am. J. Prev. Med. 17: 48-54.

Wang, Q.P. and Kaguni, J.M. 1989. A novel sigma fac-tor is involved in expression of the rpoH gene of Es-cherichia coli. J. Bacteriol. 171: 4248-4253.

Weijer, C. 2000. The future of research into rotavirusvaccine: Benefits of vaccine may outweigh risks forchildren in developing countries. Brit. Med. J.321(7260): 525-526.

Wells, C.L., Maddaus, M.A., Simmons, R.L. 1988. Pro-posed mechanisms for the translocation of intestinalbacteria. Rev. Infect. Dis. 10: 958-979.

WHO. 1999. Food safety issues associated with prod-ucts from aquaculture. Report of a Joint FAO/NACA/WHO Study Group. WHO Technical Report Series883. World Health Organization, Geneva.

WHO. 2002. Use of antimicrobials outside human medi-cine. World Health Organization, Geneva. AccessedJan. 29,.2002. (http://www.who.int/emc/diseases/zoo/antimicrobial.html).

Wiedmann, M., Bruce, J.L., Keating, C., Johnson, A.E.,McDonough, P.L., and Batt, C.A. 1997. Ribotypesand virulence gene polymorphisms suggest three dis-tinct Listeria monocytogenes lineages with differencesin pathogenic potential. Infect. Immun. 65: 2707-2716.

Williamson, D.M. 1991. Home food preparation prac-tice: Results of a national consumer survey. M.S. the-sis. Cornell Univ., Ithaca, N.Y.

Witte, W. 1998. Medical consequences of antibiotic usein agriculture. Science 279: 996-997.

Woese, C.R., Kandler, O., and Wheelis, M.L. 1990. To-wards a natural system of organisms: Proposal forthe domains Archaea, Bacteria, and Eucarya. Proc.Natl. Acad. Sci. 87(12): 4576-4579.

Wong, A.C. 1998. Biofilms in food processing environ-ments. J. Dairy Sci. 81: 2765-2770.

Wood, R.C., Hedberg, C., White, K., MacDonald, K., Os-terholm, M. 1991. A multi-state outbreak of Salmonellajaviana infections associated with raw tomatoes. (Abstr.)Proc. of the 40th Annual Conference of the EpidemicIntelligence Service, p. 69. April 8-12. Centers for Dis-ease Control and Prevention, Washington, D.C.

Wright, A.C., Simpson, L.M., and Oliver, J.D. 1981.Role of iron in the pathogenesis of Vibrio vulnificus in-fections. Infect. Immun. 34: 503-507.

Wu, S., de Lancastre, H., and Tomasz, A. 1996. Sig-ma-B, a putative operon encoding alternative sigmafactor of Staphylococcus aureus RNA polymerase:Molecular cloning and DNA sequencing. J. Bacteriol.178: 6036-6042.

XXu, H.S., Robert, N., Singleton, F.L., Attwell, R.W.,

Grimes, D.J. and Colwell, R.R. 1982. Survival and vi-ability of non-culturable Escherichia coli and Vibriocholerae in the estuarine and marine environment.Microb. Ecol. 8: 313-323.

YYang, S., Leff, M.G., McTague, D., Horvath, K.A., Jack-

son-Thompson, J., Murayi, T. Boeselager, G.K.,Melnik, T.A., Gildemaster, M.C., Ridings, D.L., Altek-ruse, S.F., and Angulo, F.J. 1998. Multistate surveil-lance for food-handling, preparation, and consump-tion behaviors associated with foodborne diseases:1995 and 1996 BRFSS food-safety questions. CDCSurveill. Summ. 47(SS-4): 33-54.

Yang, S., Angulo, F.J., and Altekruse, S.F. 2000. Evalua-tion of safe food-handling instructions on raw meatand poultry products. J. Food Protect. 63: 1321-1325.

ZZagaevskii, I.S. 1973. Beef contamination by Clostridi-

um perfringens. Veterina. 1: 101-103.Zagory, D. 1995. Principles and practice of modified

atmosphere packaging of horticultural commodities.In “Principles of Modified Atmosphere and Sous VideProduct Packaging.” ed. J. M. Farber and K. LDodds. pp. 175-204.Technomic Publishing Co. Inc.,Lancaster, Pa.

Zambrano, M. M. and Kolter, R. 1996. GASPing for lifein stationary phase. Cell 86: 181-184.

Zhou, H-R., Harkema,J.R., Yan, D., and Pestka, J.J.1999. Amplified proinflammatory cytokine expressionand toxicity in mice coexposed to lipopolysaccharideand the trichothecene vomitoxin (deoxynivalenol). J.Toxicol. Environ. Health 56: 115-136.

Zhou, H-R., Harkema, J.R., Hotchkiss, J.A., Yan, D.,Roth, R.A, and Pestka J.J. 2000. Lipopolysaccharideand the trichothecene vomitoxin (deoxynivalenol) syn-ergistically induce apoptosis in murine lymphoid or-gans. Toxicol. Sci. 53(2): 253-263.

Zook, C.D., Parish, M.E., Braddock, R.J., and Balaban,M.O. 1999. High pressure inactivation kinetics ofSaccharomyces cerevisiae ascospores in orange andapple juice. J. Food Sci. 64(3): 533-535.

Zottola, E.A. 1994. Microbial attachment and biofilmformation: A new problem for the food industry? AScientific Status Summary published by the Institute ofFood Technologists’ Expert Panel on Food Safety andNutrition, Chicago, Ill. Food Technol. 48: 107-114.

Zottola, E.A. 1997. Special techniques for studying mi-crobial biofilms in food systems. Chpt. 16 In “FoodMicrobiological Analysis: New Technologies,” ed. M. L.Tortorello and S. M.Gendel, pp. 315-346. MarcelDekker, New York, N.Y.

ift expert report on emerging microbiological food safety issues

Documents