Access EMERGING INFECTIOUS DISEASES at · Atlanta, Georgia, USA Managing Editor Polyxeni Potter, M.A. National Center for Infectious Diseases Centers for Disease Control and Prevention

Access EMERGING INFECTIOUS DISEASES at www.cdc.gov

Liaison RepresentativesAnthony I. Adams, M.D.Chief Medical AdviserCommonwealth Department of Human

Services and HealthCanberra, Australia

David Brandling-Bennett, M.D.Deputy DirectorPan American Health OrganizationWorld Health OrganizationWashington, D.C., USA

Gail Cassell, Ph.D.Liaison to American Society for MicrobiologyUniversity of Alabama at BirminghamBirmingham, Alabama, USA

Thomas M. Gomez, D.V.M., M.S.Staff EpidemiologistU.S. Department of Agriculture Animal and

Plant Health Inspection ServiceRiverdale, Maryland, USA

Richard A. Goodman, M.D., M.P.H.Editor, MMWRCenters for Disease Control and PreventionAtlanta, Georgia, USA

Joseph Losos, M.D.Director GeneralLaboratory Center for Disease ControlOntario, Canada

Gerald L. Mandell, M.D.Liaison to Infectious Diseases Society of

AmericaUniversity of Virginia Medical CenterCharlottesville, Virginia, USA

William J. Martone, M.D.Senior Executive DirectorNational Foundation for Infectious DiseasesBethesda, Maryland, USA

Phillip P. Mortimer, M.D.Director, Virus Reference DivisionCentral Public Health LaboratoryLondon, United Kingdom

Robert Shope, M.D.Professor of ResearchUniversity of Texas Medical BranchGalveston, Texas, USA

Natalya B. Sipachova, M.D., Ph.D.Scientific EditorRussian Republic Information and

Analytic CentreMoscow, Russia

Bonnie Smoak, M.D.U.S. Army Medical Research Unit—KenyaUnit 64109Box 401APO AE 09831-4109

Robert Swanepoel, B.V.Sc., Ph.D.Head, Special Pathogens UnitNational Institute for VirologySandringham 2131, South Africa

Roberto Tapia-Conyer, M.D.Dirección General de EpidemiologíaSecretaría de SaludMéxico

Editors

EditorJoseph E. McDade, Ph.D.National Center for Infectious DiseasesCenters for Disease Control and PreventionAtlanta, Georgia, USA

Perspectives EditorStephen S. Morse, Ph.D.Columbia University School of Public HealthNew York, New York, USA

Synopses EditorPhillip J. Baker, Ph.D.Division of Microbiology and Infectious DiseasesNational Institute of Allergy and Infectious DiseasesNational Institutes of HealthBethesda, Maryland, USA

Dispatches EditorStephen Ostroff, M.D.National Center for Infectious DiseasesCenters for Disease Control and PreventionAtlanta, Georgia, USA

Managing EditorPolyxeni Potter, M.A.National Center for Infectious DiseasesCenters for Disease Control and PreventionAtlanta, Georgia, USA

Editorial and Computer Support

Editing and ProductionBeatrice T. Divine, M.B.A.Teresa M. Hood, M.S.

Editorial AssistanceMaria T. BritoP. Lynne Stockton, V.M.D.

Electronic DistributionCarol Y. Crawford, M.P.A.Rachna Dube, Ph.D.

Retrieve the journal electronically on theWorld-Wide Web (WWW), through file transferprotocol (FTP), or by electronic mail (e-mail).WWW and FTP: Access the journal at http://www.cdc.gov/ncidod/EID/eid.htm, from the CDChome page (http://www.cdc.gov), or download itthrough anonymous FTP at ftp.cdc.gov. Thefiles can be found in the directory pub/EID.LISTSERVer (e-mail lists): To subscribe to a list,send an e-mail to [email protected] with thefollowing in the body of your message: subscribelistname (e.g., subscribe EID-TOC). EID-TOCwill send the table of contents automatically toyour e-mail box. Obtain individual articles onWWW or through FTP, or request copies by e-mail. EID-ASCII will send the the entire journal toyou in ASCII format.

Electronic Access

Emerging Infectious DiseasesEmerging Infectious Diseases is published

four times a year by the National Center forInfectious Diseases, Centers for DiseaseControl and Prevention (CDC), 1600 CliftonRoad, Mailstop C-12, Atllanta, GA 30333, USA.Telephone 404-639-3967, fax 404-639-3075,e-mail [email protected].

The opinions expressed by authorscontributing to this journal do not necessarilyreflect the opinions of CDC or the institutionswith which the authors are affiliated.

All material published in Emerging Infec-tious Diseases is in the public domain and maybe used and reprinted without special permis-sion; proper citation, however, is appreciated.

Use of trade names is for identification onlyand does not imply endorsement by the PublicHealth Service or by the U.S. Department ofHealth and Human Services.

Emerging Infectious Diseases is printed on acidfree paper that meets the requirements of ANSI/NISO 239.48-1992 (Permanence of Paper).

∞

The journal is distributed electronically and in hard copy and is available at no charge .

YES, I would like to receive Emerging Infectious Diseases.

Please print your name andbusiness address in this box andreturn by fax to 404-639-3075 ormail to Journal/CDC/MS C12 1600 Clifton Rd. Atlanta, GA 30333

Indexed in Index Medicus/Medline, Current Contents, and other databases.

Contents

EMERGING INFECTIOUS DISEASESVolume 3 • Number 4 October–December 1997

Special IssueThe National Conference on Emerging Foodborne Pathogens: Implications and Control, March 24-26, 1997, Alexandria, Virginia, USA 415

Infectious Disease as an Evolutionary Paradigm 417

Factors That Influence the Emergence or Reemergenceand Dissemination of Microbial Foodborne Pathogensand Human DiseaseEmerging Foodborne Diseases: An Evolving Public Health Challenge 425Emergence of New Pathogens as a Function of Changes in Host Susceptibility 435Chronic Sequelae of Foodborne Disease 443Public, Animal, and Environmental Health Implications of Aquaculture 453Produce Handling and Processing Practices 459The Impact of Consumer Demands and Trends on Food Processing 467Impact of Changing Consumer Lifestyles on the Emergence/Reemergence of Foodborne Pathogens 471

Identifying and Anticipating New FoodborneMicrobial HazardsHistorical Overview of Key Issues in Food Safety 481Quantitative Risk Assessment: An Emerging Tool for Emerging Foodborne Pathogens 483Communicating Foodborne Disease Risk 489Animal Diseases of Public Health Importance 497Foodborne Disease Control: A Transnational Challenge 503

Controlling Emerging Foodborne Microbial HazardsConsumer Concerns: Motivating to Action 511Identifying and Controlling Emerging Foodborne Pathogens: Research Needs 517Maximizing the Usefulness of Food Microbiology Research 523Epidemiology and Detection as Options for Control of Viral and Parasitic Foodborne Disease 529Quantitative Microbiology: A Basis for Food Safety 541

Strategies for Mobilizing Resources for Rapid Responseto Emerging Foodborne Microbial HazardsStrategies for Rapid Response to Emerging Foodborne Microbial Hazards 551Foodborne Illness: Implications for the Future 555

Cover: J.D. de Heem (1601–ca.1684), Bloemstilleven, Floral Still Life (detail). Reprinted with permission of Mauritshuis, The Hague,The Netherlands.

Cyclospora oocysts in human stool; Dr. MarkEberhard, Centers for Disease Control andPrevention, Atlanta, GA, USA.

Contents

EMERGING INFECTIOUS DISEASESVolume 3 • Number 4 October–December 1997

DispatchesVero Cytotoxin-Producing Escherichia coli O157 Outbreaks in England and Wales, 1995: Phenotypic Methods and Genotypic Subtyping 561Genetic Polymorphism Among Cryptosporidium parvum Isolates: Evidence of Two Distinct Human Transmission Cycles 567

CommentaryIrradiation Pasteurization of Solid Foods: Taking Food Safety to the Next Level 575

LettersNon-O157 Shiga Toxin-Producing Escherichia coli Infections in Europe 578The Taxonomy of Cyclospora 579Reply to W.C. Marquardt 579

News and NotesFoodborne Diseases Active Surveillance Network (FoodNet) 5811998 International Emerging Infectious Disease Conference 583Errata 584

Electron micrograph of Escherichia coliO157:H7; Elizabeth White, Centers for DiseaseControl and Prevention, Atlanta, GA, USA.

Foods from around the world; Dr. Michael Osterholm, Minnesota Department of Health, Minneapolis, MN, USA.

415Vol. 3, No. 4, October–December 1997 Emerging Infectious Diseases

Preface

Infectious diseases transmitted by foods havebecome a major public health concern in recentyears. Response by both the food industry andpublic health and food safety regulatory agenciesto new microbiologic health threats and reemerg-ing pathogens in foods has been primarilyreactive. The multiplicity of factors and complexinteractions involved in the emergence andreemergence of microbial foodborne hazards andthe need for a multifaceted and integratedapproach to protecting the population prompted anational Conference on Emerging FoodbornePathogens: Implications and Control (March 24-26, 1997, Alexandria, Virginia).

The conference, attended by more than 400scientists in basic and applied research,epidemiology, and public health, was organizedto elucidate programs and initiatives that couldbe used to identify and respond appropriatelyand proactively to emerging and reemergingfoodborne disease threats.

Representing the conference organizers wereDr. Alex Malaspina, president of the Interna-tional Life Sciences Institute; Dr. David Satcher,director of the Centers for Disease Control andPrevention; Mr. Thomas Billy, administrator of

*The conference was sponsored by the International Life Sciences Institute (ILSI),ILSI North America Technical Committee on Food Microbiology, the Centers forDisease Control and Prevention, the U.S. Department of Agriculture, and the U.S. Foodand Drug Administration, in cooperation with the Food and Agriculture Organizationof the United Nations and the Pan American Health Organization/World HealthOrganization. Conference grant support was also provided by the National Institute ofDiabetes and Digestive and Kidney Diseases. Additional support was provided throughunrestricted education grants from the American Meat Institute, American Society forMicrobiology, American Veterinary Medical Association, Animal Health Institute,Association of American Veterinary Medical Colleges, Frito-Lay, Inc., Land O’Lakes,Inc., McDonald’s Corporation, National Food Processors Association, The Quaker OatsCompany, Roquette America, Inc., and Ross Products Division, Abbott Laboratories.The members of the Technical Committee on Food Microbiology are Campbell SoupCompany, The Coca-Cola Company, General Mills, Gerber Products Company, H.J.Heinz Company, Kraft Foods, Inc., Lipton, M&M/Mars, Nabisco, Inc., Nestlé USA, Inc.,PepsiCo, Inc., The Pillsbury Company, and The Procter & Gamble Company.

About the National Conference on EmergingFoodborne Pathogens: Implications and Control*

the Food Safety and Inspection Service of theU.S. Department of Agriculture; Dr. Fred Shank,director of the Center for Food Safety andApplied Nutrition in the U.S. Food and DrugAdministration; and Sir George Alleyne, directorof the Pan American Health Organization. Theiropening remarks reflected a strong commitmentto collaboration among different sectors, develop-ment of integrated approaches to food safety,implementation of President Clinton’s foodsafety initiative, and international cooperationin the fight against foodborne disease.

Nobel laureate Dr. Joshua Lederberg deliveredthe keynote address, in which he called for a globalpublic health approach to the threats posed bymicrobial foodborne illness. Dr. Richard Hall’sclosing address summarized the key points of theconference presentations, emphasized the need forconcerted control efforts by the public and privatesectors, and suggested prioritizing foodbornedisease risks according to their probable impact.

Conference organizers hope that the publica-tion of conference presentations and discussionsin this journal will stimulate initiatives toimprove the safety of food and draw much neededattention to foodborne microbial hazards.


Special Issue

Life expectancy in the United States from1900 to the present (Figure 1) shows an overallsteady rise, reflecting improved health conditionsin general, the result of advances in medicalscience, hygiene, personal care, health technolo-gies, and public health administrations. The risedecelerates asymptotically to a near plateau fromthe 1950s to the 1970s, reflecting an epidemic ofcoronary disease, which we do not yet fullyunderstand. Improvements in medical care,attention to life style, or indiscriminate use ofaspirin may all be responsible for the subsequentdecrease in deaths from coronary disease. Up tothe 1940s, the rising curve is jagged, reflectingsporadic infectious disease outbreaks, especiallythe Spanish influenza outbreak of 1918. Whetherthe life expectancy curve continues to risesmoothly or whether it has some jagged declinesdepends on what we do about transmission ofinfectious disease, including foodborne disease.When plotted another way (Figure 2), both theabsolute number of deaths from infectiousdisease and the proportion of total deathsattributable to infectious disease also showsteady amelioration from 1900 almost to thepresent.

The 1918 Spanish influenza pandemic maybe a prototype for future emerging infections.

Infectious Disease asan Evolutionary Paradigm

Joshua LederbergSackler Foundation Scholar, Rockefeller University,

New York, New York, USA

Address for correspondence: Joshua Lederberg, RockefellerUniversity, 1230 York Avenue, New York, NY 10021-6399, USA;fax: 212-327-8651; e-mail: [email protected].

The basic principles of genetics and evolution apply equally to human hosts and toemerging infections, in which foodborne outbreaks play an important and growing role.However, we are dealing with a very complicated coevolutionary process in whichinfectious agent outcomes range from mutual annihilation to mutual integration andresynthesis of a new species. In our race against microbial evolution, new molecularbiology tools will help us study the past; education and a global public health perspectivewill help us deal better with the future.

Figure 1. Life expectancy in the United States, at birth,20th century.

Figure 2. Trends in infectious diseases mortality,1900–1992. Source: CDC, unpub. data.

418Emerging Infectious Diseases Vol. 3, No. 4, October–December 1997

Special Issue

Although minimized as not much more than a badcold, influenza took a terrible toll in 1918,especially on young people (Figure 3). Somewhatolder persons may have been protected byimmunity from prior exposure to related strainsof influenza. The disease, with rapid onset offulminating pneumonic symptoms, killed 20 to 25million persons worldwide. The infectious agentwas not available for study at that time. However,very recently the Armed Forces Institute ofPathology recovered with PCR technologygenetic fragments of the 1918 influenza virus (1).Less than 10% of the entire genome has beenrecovered to date, but recovery of completesequences is likely. Although the target geneshave not yet provided a clue as to why the 1918influenza was so devastating, they demonstratethe enormous potential of today’s molecularbiology tools.

These tools will enable us to better studypaleovirology and paleomicrobiology. We areaccustomed to stereotyping historical diseaseoutbreaks as if we really knew what they were,but we really know very little detail about theirgenetic features. For example, we talk about thegreat historic plagues as if they indeed wereYersinia or cholera or malaria. We should lookforward to finding out about the 14th centuryblack death, if it was indeed Yersinia pestis.Although clinically unmistakable, that is not tosay it was caused by the identical genotype ofpresent Yersinia strains.

We need to look ahead as well as back. In thiscentury, emerging and reemerging infections

have stimulated flurries of interest, but ingeneral we have been complacent aboutinfectious diseases ever since the introduction ofantibiotics. The effect of antibiotics on acuteinfections and tuberculosis as well as the effect ofpolio vaccination led to a national, almostworldwide, redirection of attention to chronic andconstitutional diseases. However, the HIVpandemic in the early 1980s caught us off guard,reminding us that there are many moreinfectious agents in the world. It is fortuitous thatretroviruses had already been studied from theperspective of cancer etiology; otherwise, wewould have had no scientific platform whatsoeverfor coping with HIV and AIDS.

The Committee on International ScienceEngineering and Technology provided an inter-agency review setting out a policy framework forthe United States’ global response to infectiousdisease (Table 1). The policy provides aworldwide mantle for surveillance and monitor-ing, remedial measures, development of newdrugs, vaccines, and treatment modalities. Theglobal outlook is necessary, even if for purelyselfish reasons, because to infectious agents theworld is indivisible, with no national boundaries.Our thinking has been impoverished in terms ofbudget allocations for dealing with health on aninternational basis.

We are engaged in a type of race, enmeshingour ecologic circumstances with evolutionarychanges in our predatory competitors. To ouradvantage, we have wonderful new technology;we have rising life expectancy curves. To ourdisadvantage, we have crowding; we have social,political, economic, and hygienic stratification.We have crowded together a hotbed ofopportunity for infectious agents to spread over asignificant part of the population. Affluent andmobile people are ready, willing, and able to carryafflictions all over the world within 24 hours’notice. This condensation, stratification, andmobility is unique, defining us as a very differentspecies from what we were 100 years ago. We areenabled by a different set of technologies. Butdespite many potential defenses—vaccines,antibiotics, diagnostic tools—we are intrinsicallymore vulnerable than before, at least in terms ofpandemic and communicable diseases.

We could imaginably adapt in a Darwinianfashion, but the odds are stacked against us. Wecannot compete with microorganisms whosepopulations are measured in exponents of 1012,

Figure 3. Pneumonia and influenza mortality, by age,in certain epidemic years. (Reprinted with permissionof W. Paul Glezen and Epidemiologic Reviews. Emerg-ing Infections: Pandemic Influenza. Epi Rev 1996;18:66).


Special Issue

1014, 1016 over periods of days. Darwinian naturalselection has led to the evolution of our speciesbut at a terrible cost. If we were to rely strictly onbiologic selection to respond to the selectivefactors of infectious disease, the populationwould fluctuate from billions down to perhapsmillions before slowly rising again. Therefore,our evolutionary capability may be dismissed asalmost totally inconsequential. In the raceagainst microbial genes, our best weapon is ourwits, not natural selection on our genes.

New mechanisms of genetic plasticity of onemicrobe species or another are uncovered almostdaily. Spontaneous mutation is just the begin-ning. We are also dealing with very largepopulations, living in a sea of mutagenicinfluences (e.g., sunlight). Haploid microbes canimmediately express their genetic variations.They have a wide range of repair mechanisms,themselves subject to genetic control. Some

strains are highly mutable by not repairing theirDNA; others are relatively more stable. They areextraordinarily flexible in responding to environ-mental stresses (e.g., pathogens’ responses toantibodies, saprophytes’ responses to newenvironments). Mechanisms proliferate wherebybacteria and viruses exchange genetic materialquite promiscuously. Plasmids now spreadthroughout the microbial world (3). They cancross the boundaries of yeast and bacteria.Lateral transfer is very important in theevolution of microorganisms. Their pathogenic-ity, their toxicity, their antibiotic resistance donot rely exclusively on evolution within a singleclonal proliferation.

We have a very powerful theoretical basiswhereby the application of selective pressure(e.g., antibiotics in food animals) will result indrug resistance carried by plasmids, or patho-gens attacking humans. It is not easy to get direct

Table 1. Examples of pathogenic microbes and infectious diseases recognized since 1973 (2)

Year Microbe Type Disease1973 Rotavirus Virus Major cause of infantile diarrhea worldwide1975 Parvovirus B19 Virus Aplastic crisis in chronic hemolytic anemia1976 Cryptosporidium Parasite Acute and chronic diarrhea parvum1977 Ebola virus Virus Ebola hemorrhagic fever1977 Legionella Bacteria Legionnaires’ disease pneumophila1977 Hantaan virus Virus Hemorrhagic fever with renal syndrome (HRFS)1977 Campylobacter jejuni Bacteria Enteric pathogens distributed globally1980 Human T-lymphotropic virus I Virus T-cell lymphoma-leukemia

(HTLV-1)1981 Toxic producing strains of Bacteria Toxic shock syndrome(tampon use)

Staphylococcus aureus1982 Escherichia coli O157:H7 Bacteria Hemorrhagic colitis; hemolytic uremic syndrome1982 HTLV-II Virus Hairy cell leukemia1982 Borrelia burgdorferi Bacteria Lyme disease1983 Human immunodeficiency Virus Acquired immunodeficiency syndrome (AIDS)

virus(HIV)1983 Helicobacter pylori Bacteria Peptic ulcer disease1985 Enterocytozoon bieneusi Parasite Persistent diarrhea1986 Cyclospora cayetanensis Parasite Persistent diarrhea1988 Human herpes-virus-6 (HHV-6) Virus Roseola subitum1988 Hepatitis E Virus Enterically transmitted non-A, non-B hepatitis1989 Ehrlichia chafeensis Bacteria Human ehrlichiosis1989 Hepatitis C Virus Parenterally transmitted non-A, non-B liver infection1991 Guanarito virus Virus Venezuelan hemorrhagic fever1991 Encephalitozoon hellem Parasite Conjunctivitis, disseminated disease1991 New species of Babesia Parasite Atypical babesiosis1992 Vibrio cholerae O139 Bacteria New strain associated with epidemic cholera1992 Bartonella henselae Bacteria Cat-scratch disease; bacillary angiomatosis1993 Sin Nombre virus Virus Adult respiratory distress syndrome1993 Encephalitozoon cuniculi Parasite Disseminated disease1994 Sabia virus Virus Brazilian hemorrhagic fever1995 HHV-8 Virus Associated with Kaposi sarcoma in AIDS patients


Special Issue

and immediate epidemiologic evidence, but thefoundations for these phenomena exist and mustbe taken into account in the development ofpolicies. We have barely begun to study theresponses of microorganisms under stress,although we have examples where root mecha-nisms of adaptive mutability are themselvesresponses to stress. In recent experiments,bacterial restriction systems are more permissiveof the introduction of foreign DNA, possiblyletting down their guard in response to “mutateor die” circumstances. This does not reflectbacterial intelligence—that they know exactlywhat mutations they should undergo in responseto environmental situations. Their intrinsicmutability and capacity to exchange geneticinformation without knowing what it is going tobe is not a constant; it is certainly under geneticcontrol and in some circumstances varies withthe stress under which the microbes are placed.

Evolution is more or less proportionate to thedegree of genetic divergence among the differentbranches of the three-tiered tree of life, with thearchaeal branch, the eubacterial branch, and theeukaryotes (Figure 4). The tree illustrates thesmall territory occupied by humans in the overallworld of biodiversity. It shows mitochondria right

next to Escherichia coli. Bacterial invasion of aprimitive eukaryote 2-1/2 to 3 billion years ago,synchronized with the development of primitivegreen oxygen-generating plants, conferred aselective advantage to complexes that could useoxygen in respiration. Our ancestors were onceinvaded by an oxidative-capable bacterium thatwe now call a mitochondrium and that is presentin every cell of every body and almost everyspecies of eukaryote. We did not evolve in amonotonous treelike development; we are alsothe resynthesis of components of geneticdevelopment that diverged as far as the bacteriaand were reincorporated into the mitochondrialpart of our overall genome. Another example oflateral transfer is the symbiosis that resultedfrom chloroplast invasion of green plants.

The outcome of encounters between mutuallyantagonistic organisms is intrinsically unpredict-able. The 1918 influenza outbreak killed halfpercent of the human population; but because theconsequences were to either kill the host or leavethe host immune, the virus died out totally,leaving no trace in our genomes, as far as weknow. Historic serology on survivors has foundmemory cells and antibodies against H1N1, theserotype of the resurrected 1918 virus. Unlike theinfluenza virus, which left no known geneticimprint, 400 to 500 retroviruses are integratedinto our human genome. The full phylogeny ofthese encounters is unknown, but many of theseviruses may precede the separation of homosapiens from the rest of the hominid line.

Infectious agent outcomes range from mutualannihilation to mutual integration and resynthe-sis of a new species. Much has been made of thefact that zoonoses are often more lethal tohumans than to their original host, but thisphenomenon cannot necessarily be generalized.Most zoonoses do not affect humans adversely.Some are equally capable in a new host. We tendto pay most attention, however, to those, such asyellow fever, for which we have not genetically orserologically adapted and which cause severedisease.

Canine distemper provides an example of aquasihereditary adaptation. In the Serengeti, thedisease migrated from village dogs to jackals,which shared prey and had contact with lions.About one-fourth of the preserve’s 4,000 lionsdied of canine distemper (4) but the survivors areimmune and will pass immunoglobulin, to theiroffspring. The cubs’ maternal immunity will

Figure 4. The three-domain tree of life based on small-subunit rRNA sequences. Reprinted with permission ofNorman R. Pace and ASM News. ASM News1996;62(9):464.


Special Issue

likely mitigate infection and permit a newequilibrium, not because of genetic adaptationbut because of the preimmunized host. This isalso the most plausible explanation for howsavage the polio virus has been as a paralyticinfection of young people. It may also apply tohepatitis, where cleaner is not always better if itmeans we do not have the “street smarts” torespond to new infectious challenges. Thesenongenetic adaptations between parasite andhost complicate our outcome expectations.

Short-term shifts in equilibrium can giveferocious but temporary advantages to a virus.Long-term outcomes are most stable when theyinvolve some degree of mutual accommodation,with both surviving longer. New short-termdeviants, however, can disrupt this equilibrium.The final outcome of the HIV pandemic cannot bepredicted. More strains with longer latency maybe taking over, mitigating the disease. However,deviant strains could counteract this effect byovercoming immunity and rapidly proliferating,with earlier and more lethal consequences.

We should also consider somatic evolution, aDarwinian process that occurs with every infection.In the clonal selection model of immunogenesis (5),an apparently random production of immunoglobu-lin variants, both by reassortment of parts and bylocalized mutagenesis, gives rise to candidateantibodies, which then proliferate in response tomatching epitopes. We do not understand thedetails of how a given epitope enhances stepwiseimprovements in affinity and productivity ofantibodies at various stages. The process may bemore complicated than we realize; so mayDarwinian evolution.

Despite the prior arguments against relyingon host or genotype evolution as a response toinfection, historically we have done so and nowhave “scars of experience.” A notable example ismalaria, wherein the Duffy mutation againstPlasmodium vivax is the only host defense withno deleterious consequences. The thalassemias,G6PD deficiency, and hemoglobin S are allhemopoietic modifications that thwart theplasmodia; but in homozygotes, they themselvescause disease. In the evolution of our species, forevery child spared an early death because ahemoglobin S mutation impeded Plasmodiumdevelopment, another will succumb to sickle celldisease unless we can intervene. Specificremedies do not exist. Although somatic genetherapy is an interesting possibility, one that will

Table 2. The origin of viruses

Viruses are genomic fragments that can replicateonly in the context of an intact living cell. They cannottherefore be primitive antecedents of cells.

Within a given species, viruses may have emergedas genetic fragments or reduced versions fromchromosomes, plasmids, or RNA of

1) the host or related species2) distant species3) larger parasites of the same or different hosts4) further evolution and genetic interchange among existing virusesOnce established, they may then cycle back into

the genome of the host as an integrated episome; therethey may have genetic functions or in principle mightreemerge as new viruses.

These cycles have some substantiation in theworld of bacterial viruses; but we have no clear data onthe provenience of plant or animal viruses.

probably progress in the next 20 years, it isparadoxical that we know more about hemoglo-bin S than any other molecular disease. Theentire concept of genetic determination of proteinstructure has been based on these earlyobservations, yet we are still searching with limitedsuccess for ways to put it to therapeutic use.

Biotechnology may enable other forms ofgenetic intervention through which homosapiens could conceivably bypass natural selec-tion and random variation. In the absence ofalternatives, we might speculate about thesekinds of “aversive therapies” as a last resort tosave our species.

The ultimate origin of life is still the subject ofmany theories, as is the origin of viruses (Table2). Each virus is different. We know nothing ofvirus phylogenies and cannot even substantiatethe distinctions of the several hundred catego-ries. We do not know their origin, only that theyinteract with host genomes in many ways.Particles could come out of any genome, becomefree-living (i.e., independent, autonomouslyreplicating units in host cells), reenter a hostgenome as retroviruses and possibly others do,and repeat the cycle dozens of times. But no onecan give a single example or claim to havesignificant knowledge of how any particular virusevolved, thus presenting a scientific challenge forthe next 20 or 30 years.

We are dealing with more than just predationand competition. We are dealing with a very


Special Issue

Table 3. Genetic evolution

Microbes (bacteria, viruses, fungi, protozoa)Rapid and incessantHuge population sizes 1014+ and generation times in minutes vs. years

Intraclonal processDNA replication—may be error-prone—in sea of mutagens sunlight; unshielded chemicals, incl.natural productsRNA replication—intrinsically unedited, >10-3

swarm speciesHaploid: immediate manifestation, but partial recessives not accumulated contra multicopy plasmidsAmplificationSite-directed inversions and transpositions: phase variation?? Other specifically evolved mechanisms: genome quadrant duplication; silencing

Interclonal processPromiscuous recombination—not all mechanisms are knownConjugation—dozens of speciesViral transduction and lysogenic integration: universal

Classical: phage-borne toxins in C. diphtheriae

Plasmid interchange (by any of above) and integration

Toxins of B. anthracisPasteur: heat attenuation: plasmid loss; chemically induced

RNA viral reassortment; ?? and recombination?Transgressive—across all boundaries

Artificial gene splicingBacteria and viruses have picked up host genes (antigenic masking?)Interkingdom: P. tumefaciens and plants, E. coli and yeastVegetable and mineral! oligonucleotides and yeast.

Host-parasite coevolutionCoadaptation to mutualism or accentuation of virulence?Jury is still out (May and Anderson). Many zoonotic convergences.Probably divergent phenomena, with short- term flareups and Pyrrhic victories, atop long-term trend to coadaptation.

complicated coevolutionary process, involvingmerger, union, bifurcation, and reemergence ofnew species (Table 3). Divergent phenomena canoccur in any binary association, with unpredict-able outcomes. We have hundreds of retrovirusesin our genome and no knowledge of how they gotthere. As to HIV, we have no evidence as yet thatit has ever entered anyone’s germ line genome:we really do not know whether it ever entersgerm cells. The outcomes of even that interactioncould be much more complicated than the purelyparasite/host relationships we are accustomed to.

Innovative technologies for dealing withmicrobial threats have the potential for fascinat-ing therapeutic opportunities (Table 4). Some,like bacteriophage, have been set aside aslaboratory curiosities. Nothing is more excitingthan unraveling the details of pathogenesis.Having the full genomes of half a dozen parasiticorganisms opens up new opportunities fortherapeutic invention in ways that we could nothave dreamed of even 5 years ago, which will leadto many more technologies. In food microbiology,we should keep in mind the probiotic as well asthe adversarial and pathogenetic opportunitiesin our alimentary tracts.

The Committee on International ScienceEngineering and Technology report (2) providessome recommendations (Table 5). We need aglobal perspective. We need to invest in publichealth, especially food microbiology, not justmedical care, in dealing with disease. It isimportant to prevent foodborne disease throughsensible monitoring, standards of cleanliness,and consumer and food-handler education andnot just care for its victims.

Today we emphasize individual rights overcommunity needs more than we did 50 to 75 yearsago. Restraining the rights and freedoms ofindividuals is a far greater sin than allowing theinfection of others. The restraints placed onTyphoid Mary might not be acceptable today,when some would prefer to give her unlimitedrein to infect others, with litigation their onlyrecourse. In the triumph of individual rights, thepublic health perspective has had an uphillstruggle in recent pandemics.

Education, however, is a universally acceptedcountermeasure, especially important infoodborne diseases. Food safety programs shouldmore specifically target food handlers, examiningtheir hands to determine if they are carriers, toensure they are complying with basic sanitation.


Special Issue

We typically do this only after an outbreak.Perhaps we should have further debate on thesocial context for constraints and persuasion tocontain the spread of infectious agents.

References 1. Taubenberger JK, Reid AH, Frafft AE, Bijwaard KE,

Fanning TG. Initial genetic characterization of the1918 “Spanish” influenza virus. Science 1997;275:1793-6.

2. NSTC-CISET Working Group on Emerging andReemerging Infectious Diseases. Infectious disease—aglobal health threat. Washington (DC): The Group;1996.

3. Lederberg J. Plasmid (1952–1997). Plasmid. In press1997.

4. Roelke-Parker ME, Munson L, Packer C, Kock R,Cleaveland S, Carpenter M, et al. A canine distempervirus epidemic in Serengeti lions (Panthera leo).Nature 1996;379:441-5.

5. Lederberg J. The ontogeny of the clonal selectiontheory of antibody formation: reflections on Darwinand Ehrlich. Ann N Y Acad Sci 1988;546:175-87.

Table 4. Technologies to address microbial threats

Antibacterial chemotherapyPotentially unlimited capability; bacterial metabolism and genetic structure notably different from human genome sequencing pointing to bacterial vulnerabilitiesEconomic-structural factors—public expectation for unachievable bargains in safety assurance, cost of development, and ultimate pricingDilemmas of regulation of (ab)useResurgent interest in bacteriophage and other biologically oriented approaches

Antiviral chemotherapyMuch more difficult program, inherentlyGross underinvestmentNew approaches: antisense, ribozymes, targeted D/RNA cleaversProblematics of sequence-selective targets

VaccinesGross underinvestment; other structural problems as aboveLiability/indemnificationVaccination as service to the herdNew approaches: hot biotechnology is coming along especially live attenuated: but testing dilemmasSafety issues about use of human cells lines; adjuvants

Immunoglobulins and their progenyPhage display and diversification: biosynthetic antibodyPassive immunization for therapy

Biologic response modifiersNew world of interleukins, cell growth factors so far just scratching surfaceInteraction with pathogenesisIntersection with somatic gene therapy

Technologies for diagnosis and monitoringEtiologic agents and controlHost polymorphisms and sensitivities

Homely technologies neededSimple, effective face-masksPalatable water-disinfectantsHome-use diagnostics of contamination

Table 5. CISET* recommendations for addressing globalinfectious disease threats1. Concerted global and domestic surveillance and

diagnosis of disease outbreaks and endemicoccurrence. This must entail the installation ofsophisticated laboratory capabilities at manycenters now lacking them.

2. Vector management and monitoring andenforcement of safe water and food supplies; andpersonal hygiene (e.g., Operation Clean Hands).

3. Public and professional education.4. Scientific research on causes of disease, pathogenic

mechanisms, bodily defenses, vaccines, andantibiotics.

5. Cultivation of the technical fruits of such research,with the full involvement of the pharmaceuticalindustry and a public understanding of theregulatory and incentive structures needed tooptimize the outcomes.

*Committee on International Science, Engineering andTechnology Policy of the National Science and TechnologyCouncil.


Special Issue

Every year, in the United States foodborneinfections cause millions of illnesses andthousands of deaths; most infections go undiag-nosed and unreported. As the epidemiology offoodborne infections evolves, old scenarios andsolutions need to be updated. This articlereviews main trends in the evolution offoodborne disease epidemiology and their effecton surveillance and prevention activities.

Preventing foodborne disease is a multifac-eted process, without simple and universalsolutions. For most foodborne pathogens, novaccines are available. Consumer educationabout basic principles of food safety, an importantcomponent of prevention, by itself is insufficient.Food reaches the consumer through long chainsof industrial production, in which manyopportunities for contamination exist. Thegeneral strategy of prevention is to understandthe mechanisms by which contamination anddisease transmission can occur well enough tointerrupt them. An outbreak investigation orepidemiologic study should go beyond identifying

a suspected food and pulling it from the shelf todefining the chain of events that allowedcontamination with an organism in large enoughnumbers to cause illness. We learn from theinvestigation what went wrong, in order to devisestrategies to prevent similar events in the future.Although outbreaks make the news, mostfoodborne infections occur as individual orsporadic cases. Therefore, the sources of sporadiccases must also be investigated and understood.

Emerging Foodborne PathogensSubstantial progress has been made in

preventing foodborne diseases. For example,typhoid fever, extremely common at thebeginning of the 20th century, is now almostforgotten in the United States. It was conqueredin the preantibiotic era by disinfection of drinkingwater, sewage treatment, milk sanitation andpasteurization, and shellfish bed sanitation(Figure 1). Similarly, cholera, bovine tuberculo-sis, and trichinosis have also been controlled inthe United States. However, new foodbornepathogens have emerged. Among the first ofthese were infections caused by nontyphoidstrains of Salmonella, which have increaseddecade by decade since World War II (Figure 1).

Emerging Foodborne Diseases: AnEvolving Public Health Challenge

Robert V. TauxeCenters for Disease Control and Prevention, Atlanta, Georgia, USA

Address for correspondence: Robert V. Tauxe, Centers forDisease Control and Prevention, 1600 Clifton Road, MS A38,Atlanta, GA 30333 USA; fax: 404-639-2205.

The epidemiology of foodborne disease is changing. New pathogens haveemerged, and some have spread worldwide. Many, including Salmonella, Escherichiacoli O157:H7, Campylobacter, and Yersinia enterocolitica, have reservoirs in healthyfood animals, from which they spread to an increasing variety of foods. These pathogenscause millions of cases of sporadic illness and chronic complications, as well as largeand challenging outbreaks over many states and nations. Improved surveillance thatcombines rapid subtyping methods, cluster identification, and collaborativeepidemiologic investigation can identify and halt large, dispersed outbreaks. Outbreakinvestigations and case-control studies of sporadic cases can identify sources ofinfection and guide the development of specific prevention strategies. Betterunderstanding of how pathogens persist in animal reservoirs is also critical to successfullong-term prevention. In the past, the central challenge of foodborne disease lay inpreventing the contamination of human food with sewage or animal manure. In thefuture, prevention of foodborne disease will increasingly depend on controllingcontamination of feed and water consumed by the animals themselves.


Special Issue

In the last 20 years, other infectious agents havebeen either newly described or newly associatedwith foodborne transmission (Table 1). Vibriovulnificus, Escherichia coli O157:H7, andCyclospora cayetanensis are examples of newlydescribed pathogens that often are foodborne. V.vulnificus was identified in the bloodstream ofpersons with underlying liver disease who hadfulminant infections after eating raw oysters orbeing exposed to seawater; this organism lives inthe sea and can be a natural summertimecommensal organism in shellfish (1). E. coliO157:H7 was first identified as a pathogen in1982 in an outbreak of bloody diarrhea traced tohamburgers from a fast-food chain (2); it wassubsequently shown to have a reservoir inhealthy cattle (3). Cyclospora, known previouslyas a cyanobacterialike organism, received itscurrent taxonomic designation in 1992 andemerged as a foodborne pathogen in outbreakstraced to imported Guatemalan raspberries in1996 (4,5). The similarity of Cyclospora toEimeria coccidian pathogens of birds suggestsan avian reservoir (4,5).

Some known pathogens have only recentlybeen shown to be predominantly foodborne. Forexample, Listeria monocytogenes was long knownas a cause of meningitis and other invasiveinfections in immunocompromised hosts. Howthese hosts became infected remained unknownuntil a series of investigations identified food asthe most common source (6). Similarly, Campy-lobacter jejuni was known as a rare opportunisticbloodstream infection until veterinary diagnosticmethods used on specimens from humans showedit was a common cause of diarrheal illness (7).

Subsequent epidemiologic investigations im-plicated poultry and raw milk as the mostcommon sources of sporadic cases and out-breaks, respectively (8). Yersinia enterocolitica,rare in the United States but a common cause ofdiarrheal illness and pseudoappendicitis innorthern Europe and elsewhere, is now knownto be most frequently associated withundercooked pork (9).

These foodborne pathogens share a numberof characteristics. Virtually all have an animalreservoir from which they spread to humans; thatis, they are foodborne zoonoses. In markedcontrast to many established zoonoses, these newzoonoses do not often cause illness in the infectedhost animal. The chicken with lifelong ovarianinfection with Salmonella serotype Enteritidis,the calf carrying E. coli O157:H7, and the oystercarrying Norwalk virus or V. vulnificus appearhealthy; therefore, public health concerns mustnow include apparently healthy animals. Limitedexisting research on how animals acquire andtransmit emerging pathogens among themselvesoften implicates contaminated fodder and water;therefore, public health concerns must nowinclude the safety of what food animalsthemselves eat and drink.

For reasons that remain unclear, thesepathogens can rapidly spread globally. Forexample, Y. enterocolitica spread globally amongpigs in the 1970s (10); Salmonella serotypeEnteritidis appeared simultaneously around theworld in the 1980s (11); and SalmonellaTyphimurium Definitive Type (DT) 104 is nowappearing in North America, Europe, and

Figure 1. Reported incidence of typhoid fever andnontyphoidal salmonellosis in the United States,1920–1995.

Table 1. New pathogens that are foodborne andpathogens newly recognized as predominantlyfoodborne in the United States in the last 20 yearsCampylobacter jejuniCampylobacter fetus ssp. fetusCryptosporidium cayetanensisEscherichia coli O157:H7 and related E. coli (e.g., O111:NM, O104:H21)Listeria monocytogenesNorwalk-like virusesNitzschia pungens (cause of amnesic shellfish poisoning)Salmonella EnteritidisSalmonella Typhimurium DT 104Vibrio cholerae O1Vibrio vulnificusVibrio parahaemolyticusYersinia enterocolitica

1920 1940 1960 1980

0

10

20

30

40

50

Years

Incidence per 100,000 population

Typhoid fever

Nontyphoid Salmonellosis


Special Issue

perhaps elsewhere (12); therefore, publichealth concerns must now include eventshappening around the world, as harbingers ofwhat may appear here.

Many emerging zoonotic pathogens arebecoming increasingly resistant to antimicrobialagents, largely because of the widespread use ofantibiotics in the animal reservoir. For example,Campylobacter isolated from human patients inEurope is now increasingly resistant tofluoroquinolones, after these agents wereintroduced for use in animals (13). Salmonellaehave become increasing resistant to a variety ofantimicrobial agents in the United States (14);therefore, public health concerns must includethe patterns of antimicrobial use in agriculture aswell as in human medicine.

The foods contaminated with emergingpathogens usually look, smell, and taste normal,and the pathogen often survives traditionalpreparation techniques: E. coli O157:H7 in meatcan survive the gentle heating that a rarehamburger gets (15); Salmonella Enteritidis ineggs survives in an omelette (16); and Norwalkvirus in oysters survives gentle steaming (17).Following standard and traditional recipes cancause illness and outbreaks. Contamination withthe new foodborne zoonoses eludes traditionalfood inspection, which relies on visual identifica-tion of foodborne hazards. These pathogensdemand new control strategies, which wouldminimize the likelihood of contamination in thefirst place. The rate at which new pathogens havebeen identified suggests that many more remainto be discovered. Many of the foodborne infectionsof the future are likely to arise from the animalreservoirs from which we draw our food supply.

Once a new foodborne disease is identified, anumber of critical questions need to be answeredto develop a rational approach to prevention:What is the nature of the disease? What is thenature of the pathogen? What are simple waysto easily identify the pathogen and diagnosethe disease? What is the incidence of theinfection? How can the disease be treated?Which foods transmit the infection? How doesthe pathogen get into the food, and how welldoes it persist there? Is there is an animalreservoir? How do the animals themselvesbecome infected? How can the disease beprevented? Does the prevention strategy work?

The answers to these questions do not comerapidly. Knowledge accumulates gradually, as a

result of detailed scientific investigations, oftenconducted during outbreaks (18). After 15 yearsof research, we know a great deal about infectionswith E. coli O157:H7, but we still do not knowhow best to treat the infection, nor how the cattle(the principal source of infection for humans)themselves become infected. Better slaughterprocedures and pasteurization of milk are usefulcontrol strategies for this pathogen in meat andmilk, as irradiation of meat may be in the future.More needs to be learned: for example, it remainsunclear how best to prevent this organism fromcontaminating lettuce or apple juice. For morerecently identified agents, even less is known.

New Food Vehicles of TransmissionAlong with new pathogens, an array of new

food vehicles of transmission have been impli-cated in recent years. Traditionally, the foodimplicated in a foodborne outbreak wasundercooked meat, poultry or seafood, orunpasteurized milk. Now, additional foodspreviously thought safe are considered hazard-ous. For example, for centuries, the internalcontents of an egg were presumed safe to eat raw.However, epidemic Salmonella Enteritidis infec-tion among egg-laying flocks indicates that intacteggs may have internal contamination with thisSalmonella serotype. Many outbreaks are causedby contaminated shell eggs, including eggs usedin such traditional recipes as eggnog and Caesarsalad, lightly cooked eggs in omelettes andFrench toast, and even foods one would presumethoroughly cooked, such as lasagna andmeringue pie (19,20). E. coli O157:H7 has causedillness through an ever-broadening spectrum offoods, beyond the beef and raw milk that aredirectly related to the bovine reservoir. In 1992,an outbreak caused by apple cider showed thatthis organism could be transmitted through afood with a pH level of less than 4.0, possibly aftercontact of fresh produce with manure (21). Arecent outbreak traced to venison jerky suggestsa wild deer reservoir, so both cattle and feral deermanure are of concern (22). Imported raspberriescontaminated with Cyclospora caused an epi-demic in the United States in 1996, possiblybecause contaminated surface water was used tospray the berries with fungicide before harvest(5). Norwalk-like viruses, which appear to have ahuman reservoir, have contaminated oystersharvested from pristine waters by oyster catcherswho did not use toilets with holding tanks on


Special Issue

their boats and were themselves the likelysource of the virus (23).

The new food vehicles of disease shareseveral features. Contamination typically occursearly in the production process, rather than justbefore consumption. Because of consumer demandand the global food market, ingredients from manycountries may be combined in a single dish, whichmakes the specific source of contamination difficultto trace. These foods have fewer barriers tomicrobial growth, such as salt, sugar, orpreservatives; therefore, simple transgressions canmake the food unsafe. Because the food has a shortshelf life, it may often be gone by the time theoutbreak is recognized; therefore, efforts to preventcontamination at the source are very important.

An increasing, though still limited, propor-tion of reported foodborne outbreaks are beingtraced to fresh produce (24). A series of outbreaksrecently investigated by the Centers for DiseaseControl and Prevention (CDC) has linked avariety of pathogens to fresh fruits andvegetables harvested in the United States andelsewhere (Table 2). The investigations haveoften been triggered by detection of more casesthan expected of a rare serotype of Salmonella orShigella or by diagnosis of a rare infection likecyclosporiasis. Outbreaks caused by commonserotypes are more likely to be missed. Variouspossible points of contamination have beenidentified during these investigations, includingcontamination during production and harvest,initial processing and packing, distribution, andfinal processing (Table 3). For example, fresh orinadequately composted manure is used some-times, although E. coli O157:H7 has been shownto survive for up to 70 days in bovine feces (25).Untreated or contaminated water seems to be aparticularly likely source of contamination.Water used for spraying, washing, and maintain-ing the appearance of produce must bemicrobiologically safe. After two large outbreaksof salmonellosis were traced to importedcantaloupe, the melon industry considered a“Melon Safety Plan,” focusing particularly on thechlorination of water used to wash melons and tomake ice for shipping them. Although the extentto which the plan was implemented is unknown,no further large outbreaks have occurred. Aftertwo large outbreaks of salmonellosis were tracedto a single tomato packer in the Southeast, anautomated chlorination system was developed forthe packing plant wash tank. Because tomatoes

absorb water (and associated bacteria) if washedin water colder than they are, particularattention was also focused on the temperature ofthe water bath (26,27). No further outbreakshave been linked to southeastern tomatoes.Similar attention is warranted for water used torinse lettuce heads in packing sheds and to crispthem in grocery stores as well as for water used inprocessing other fresh produce.

A New Outbreak ScenarioBecause of changes in the way food is

produced and distributed, a new kind of outbreakhas appeared. The traditional foodborne out-break scenario often follows a church supper,family picnic, wedding reception, or other social

Table 2. Foodborne outbreaks traced to fresh produce,1990–1996

Cases StatesYr. Pathogen Vehicle (No.) (No.) Source'90 S. Chester Cantaloupe 245 30 C.A.a

'90 S. Javiana Tomatoes 174 4 U.S.b

'90 Hepatitis A Strawberries 18 2 U.S.'91 S. Poona Cantaloupe >400 23 U.S./

C.A.'93 E. coli O157:H7 Apple cider 23 1 U.S.'93 S. Montevideo Tomatoes 84 3 U.S.'94 Shigella flexneri Scallions 72 2 C.A.'95 S. Stanley Alfalfa 242 17 N.K.c

sprouts'95 S. Hartford Orange juice 63 21 U.S.'95 E. coli O157:H7 Leaf lettuce 70 1 U.S.'96 E. coli O157:H7 Leaf lettuce 49 2 U.S.'96 Cyclospora Raspberries 978 20 C.A.'96 E. coli O157:H7 Apple juice 71 3 U.S.aCentral AmericabUnited StatescSource not known

Table 3. Events and potential contamination sourcesduring produce processingEvent Contamination sourcesProduction and harvest

Growing, picking, Irrigation water, manure,bundling lack of field sanitation

Initial processingWashing, waxing, Wash water, handlingsorting, boxing

DistributionTrucking Ice, dirty trucks

Final processingSlicing, squeezing, Wash water, handling,shredding, peeling cross-contamination


Special Issue

event. This scenario involves an acute and highlylocal outbreak, with a high inoculum dose and ahigh attack rate. The outbreak is typicallyimmediately apparent to those in the local group,who promptly involve medical and publichealth authorities. The investigation identifiesa food-handling error in a small kitchen thatoccurs shortly before consumption. The solu-tion is also local. Such outbreaks still occur, andhandling them remains an important functionof a local health department.

However, diffuse and widespread outbreaks,involving many counties, states, and evennations (28), are identified more frequently andfollow an entirely different scenario. The newscenario is the result of low-level contaminationof a widely distributed commercial food product.In most jurisdictions, the increase in cases maybe inapparent against the background illness.The outbreak is detected only because of afortuitous concentration of cases in one location,because the pathogen causing the outbreak isunusual, or because laboratory-based subtypingof strains collected over a wide area identifies adiffuse surge in one subtype. In such outbreaks,investigation can require coordinated efforts of alarge team to clarify the extent of the outbreak,implicate a specific food, and determine thesource of contamination. Often, no obviousterminal food-handling error is found. Instead,contamination is the result of an event in theindustrial chain of food production. Investigat-ing, controlling, and preventing such outbreakscan have industrywide implications.

These diffuse outbreaks can be caused by avariety of foods. Because fresh produce is usuallywidely distributed, most of the produce-relatedoutbreaks listed in Table 2 were multistateevents. Some of the largest outbreaks affectedmost states at once. For example, a recentoutbreak of Salmonella Enteritidis infectionscaused by a nationally distributed brand of icecream affected the entire nation (29). Although itcaused an estimated 250,000 illnesses, it wasdetected only when vigorous routine surveillanceidentified a surge in reported infections with S.Enteritidis in one area of southern Minnesota.The consumers affected did not make food-handling errors with their ice cream, so foodsafety instruction could not have prevented thisoutbreak. The ice cream premix was transportedafter pasteurization to the ice cream factory intanker trucks that had been used to haul raw

eggs. The huge epidemic was the result of abasic failure on an industrial scale to separatethe raw from the cooked.

S. Enteritidis infections also illustrate whysurveillance and investigation of sporadic casesare needed. A diffuse increase in sporadic casescan occur well before a local or large outbreakfocuses attention on the emergence of a pathogen.The isolation rate for S. Enteritidis began toincrease sharply in the New England region in1978 (Figure 2); all cases were sporadic. In 1982,an outbreak in a New England nursing home wastraced to eggs from a local supplier. However, theegg connection was not really appreciated until1986, when a large multistate outbreak of S.Enteritidis infections was traced to stuffed pastamade with raw eggs and labeled “fully cooked.”This outbreak, affecting an estimated 3,000persons in seven states, led to the documentationthat S. Enteritidis was present on egg-layingfarms and to the subsequent demonstration thatboth outbreaks and sporadic cases of infectionswere associated with shell eggs (19,30). Sincethen, Enteritidis has become the most commonserotype of Salmonella isolated in the UnitedStates, accounting for 25% of all Salmonellareported in the country and causing outbreakscoast to coast. Eggs remain the dominant sourceof these infections, causing large outbreaks whenthey are pooled and undercooked and individualsporadic cases among consumers who eatindividual eggs (20,31). Perhaps focused investi-gation and control measures taken when thelocalized increase in sporadic Salmonella caseswas just beginning might have prevented thesubsequent spread.

Figure 2. Salmonella Enteritidis isolation rates fromhumans by region, United States, 1970–1996.

1970 72 74 76 78 80 82 84 86 88 90 92 94 960

2

4

6

8

10

12

Year

Total New England Mid Atlantic Pacific Other


Special Issue

Changing Surveillance StrategiesIn the United States, surveillance for

diseases of major public health importance hasbeen conducted for many years. The legalframework for surveillance resides in the statepublic health epidemiology offices, which sharedata with CDC. The first surveillance systemsdepended on physician or coroner notification ofspecific diseases and conditions, with reportsgoing first to the local health department, then tostate and federal offices. Now electronic, thisform of surveillance is still used for many specificconditions (32). In 1962, a second channel wasdeveloped specifically for Salmonella, to takeadvantage of the added public health informationprovided by subtyping the strains of bacteria (33).Clinical laboratories that isolated Salmonellafrom humans were requested or required to sendthe strains to the state public health laboratoryfor serotyping. Although knowing the serotype isusually of little benefit to the individual patient,it has been critical to protecting and improvingthe health of the public at large. Serotypingallows cases that might otherwise appearunrelated to be included in an investigationbecause they are of the same serotype. Moreover,infections that are close in time and space to anoutbreak but are caused by nonoutbreakserotypes and are probably unrelated can bediscounted. Results of serotyping are now sentelectronically from public health laboratories andcan be rapidly analyzed and summarized.Salmonella serotyping was the first subtype-basedsurveillance system and is a model for similarsystems (34). Yet another source of surveillancedata involves summary reports of foodbornedisease outbreak investigations from local and statehealth departments (35). About 400 such outbreaksare reported annually, by a system that remainspaper-based, labor-intensive, and slow.

Existing surveillance systems provide alimited and relatively inexpensive net for tracinglarge-scale trends in foodborne diseases undersurveillance and for detecting outbreaks ofestablished pathogens in the United States.However, they are less sensitive to diffuseoutbreaks of common pathogens, provide littledetail on sporadic cases, and are not easy toextend to emerging pathogens. In the future,changes in health delivery may impinge on theway that diagnoses are made and reported,leading to artifactual changes in reporteddisease incidence.

Therefore, CDC, in collaboration with statehealth departments and federal food regulatoryagencies, is enhancing national surveillance forfoodborne diseases in several ways. First, the roleof subtyping in public health laboratories is beingexpanded to encompass new molecular subtypingmethods. Beginning in 1997, a national subtypingnetwork for E. coli O157:H7 of participating statepublic health department laboratories and CDCwill use a single standardized laboratory protocolto subtype strains of this important pathogen.The standard method, pulsed-field gel electro-phoresis, can be easily adapted to other bacterialpathogens. In this network, each participatinglaboratory will be able to routinely compare thegenetic gel patterns of strains of E. coli O157:H7with the patterns in a national pattern bank. Thiswill enable rapid detection of clusters of relatedcases within the state and will focus investiga-tive resources on the cases most likely to belinked. It will also enable related casesscattered across several states to be linked sothat a common source can be sought.

Another surveillance strategy, now imple-mented, is active surveillance in sentinelpopulations. Since January 1996, at five U.S.sentinel sites, additional surveillance resourcesmake it possible to contact laboratories directlyfor regular reporting of bacterial infections likely tobe foodborne (36; Figure 3). In addition, surveys ofthe population, physicians, and laboratoriesmeasure the proportion of diarrheal diseases thatare undiagnosed and unreported so that the truedisease incidence can be estimated. Thissurveillance, known as FoodNet, is the platformon which more detailed investigations, includingcase-control studies of sporadic cases of commonfoodborne infections, are being conducted.

Figure 3. Incidence of three infections in FoodNetsurveillance areas, 1996.

California Connecticut Georgia Minnesota Oregon0

10

20

30

40

50

60

70

Cases per 100,000 population

Campylobacter Salmonella Shigella


Special Issue

Yet another new surveillance initiative is theroutine monitoring of antimicrobial resistanceamong a sample of Salmonella and E. coliO157:H7 bacteria isolated from humans (37). Anew cluster detection algorithm is beingapplied routinely to surveillance data forSalmonella at the national level, making itpossible to detect and flag possible outbreaks assoon as the data are reported (38). Implementa-tion of such algorithms for other infections andat the state level will further increase theusefulness of routine surveillance.

Further enhancements are possible as activesurveillance through FoodNet is extended to awider spectrum of infections, including foodborneparasitic and viral infections. In 1997, activesurveillance for Cyclospora began in FoodNet,which quickly resulted in the detection of adiffuse outbreak among persons who had been ona Caribbean cruise ship that made stops inMexico and Central America (CDC, unpub. data).Application of standardized molecular subtypingmethods to other foodborne pathogens will providea more sensitive warning system for diffuseoutbreaks of a variety of pathogens. To handleoutbreaks in areas not covered by FoodNet,standard surveillance and investigative capacitiesin state health department epidemiology officesand laboratories should be strengthened. Inaddition, enhanced international consultation willbe critical to better detect and investigateinternational or global outbreaks (28).

Implications of the New OutbreakScenario for Public Health Activities

Our public health infrastructure is tiered,both in surveillance responsibilities and inresponse to emergency situations (39). At thelocal level, the county or city health department,first developed in response to epidemic choleraand other challenges in the 19th century, isresponsible for most basic surveillance, investi-gation, and prevention activities. At the statelevel, epidemiologists, public health laboratorians,sanitarians, and educators conduct statewidesurveillance and prevention activities andconsult with and support local authorities. At thenational level, CDC is the primary risk-assessment agency for public health hazards andconducts the primary national surveillance aswell as epidemic response in support of statehealth departments. The Food and DrugAdministration, Department of Agriculture, and

Environmental Protection Agency are theprimary regulatory agencies, charged withspecific responsibilities regarding the nation’sfood and water supplies that interlock and are notalways predictable. The Food and Drug Adminis-tration regulates low-acid canned foods, im-ported foods, pasteurized milk, many seafoods,rabbits raised for meat, and food and waterprovided on aircraft and trains. The Departmentof Agriculture regulates meat and poultry,including primary slaughter and further process-ing, and pasteurized eggs; investigates animaland plant diseases; and maintains the countyextension outreach program. Shell eggs do nothave a clear regulatory home, as the Departmentof Agriculture regulates the grading of shell eggsfor quality, but the Food and Drug Administra-tion, since 1995, has responsibility for themicrobiologic safety of shell eggs.

The new outbreak scenario has severalimplications for the practice of public health,starting at the local level. One is that whendiffuse outbreaks are detected, a local healthdepartment may need to investigate a few casesthat are part of a larger outbreak despite theirapparently small local impact. Second, anapparently local outbreak may herald the firstrecognized manifestation of a national or eveninternational event.

When a diffuse outbreak of a potentiallyfoodborne pathogen is detected, rapid investiga-tion is needed to determine whether the outbreakis foodborne, and if possible, identify a specificfood vehicle. These investigations, which typi-cally include case-control studies, may need to beconducted in several locations at once. While allcases or all affected states may not need to beincluded in such an investigation, combiningcases from several locations in one investigationand repeating the investigation in more than onelocation can be helpful. For example, in a recentinternational outbreak of Salmonella Stanleyinfections traced to alfalfa sprouts, concentra-tions of cases in Arizona, Michigan, and Finlandled to case-control studies in each location, eachof which linked illness to eating sprouts grownfrom the same batch of alfalfa seeds. This provedthat the seeds were contaminated at the source(40). Parallel investigations can also lead to newtwists. In the large West Coast outbreak of E. coliO157:H7 infections in 1993, a parallel investiga-tion conducted in Nevada identified a type ofhamburger other than the one implicated in the


Special Issue

initial case-control investigation in Washington,leading to a broader recall and a more completeinvestigation of the circumstances of contamina-tion (15,41). Because well-conducted investiga-tions may lead to major product recalls,industrial review, and overhaul, and eveninternational embargoes, it is essential that theybe of the highest scientific quality.

Foodborne outbreaks are investigated for twomain reasons. The first is to identify and controlan ongoing source by emergency action: productrecall, restaurant closure, or other temporary butdefinitive solutions. The second reason is to learnhow to prevent future similar outbreaks fromoccurring. In the long run this second purposewill have an even greater impact on public healththan simply identifying and halting theoutbreaks. Because all the answers are notavailable and existing regulations may not besufficient to prevent outbreaks, the scientificinvestigation often requires a careful evaluationof the chain of production. This traceback is anintegral part of the outbreak investigation. It isnot a search for regulatory violations, but ratheran effort to determine where and howcontamination occurred. Often, the contamina-tion scenario reveals that a critical point has beenlost. Therefore, epidemiologists must participatein traceback investigations.

Intervention during outbreaks often dependson having enough good epidemiologic data to actwith confidence, without waiting for a definitivelaboratory test, particularly if potentially lethalillnesses are involved. For example, if fivepersons with classic clinical botulism ate at thesame restaurant the preceding day (but havenothing else apparent in common), prudencedictates closing the restaurant quickly while theoutbreak is sorted out—that is, before a specificfood is identified or confirmatory cultures aremade, which may take several days or evenweeks. Good epidemiologic data, includingevidence of a clear statistical association with aspecific exposure, biologic plausibility of theillness syndrome, the potential hazard of thatfood, and the logical consistency of distribution ofthe suspect food and cases are essential.

The role of the regulatory agency laboratoryis also affected by the new scenario. Because ofthe short shelf life and broad distribution of manyof the new foods responsible for infection, by thetime the outbreak is recognized and investigatedthe relevant food may no longer be available for

culture. Because contamination may be re-stricted to a single production lot, blind samplingof similar foods that does not include theimplicated lot can give a false sense of security.Good epidemiologic information pointing tocontamination of a specific food or production lotshould guide the microbiologic sampling and theinterpretation of the results. Available methodsmay be insufficient to detect low-level contamina-tion, even of well-established pathogens.

New Approaches to the Prevention ofFoodborne Disease

Meeting the complex challenge of foodbornedisease prevention will require the collaborationof regulatory agencies and industry to make foodsafely and keep it safe throughout the industrialchain of production. Prevention can be “built in”to the industry by identifying and controlling thekey points—from field, farm, or fishing ground tothe dinner table—at which contamination caneither occur or be eliminated. The generalstrategy known as Hazard Analysis and CriticalControl Points (HACCP) replaces the strategy offinal product inspection. Some simple controlstrategies are self-evident, once the reality ofmicrobial contamination is recognized. Forexample, shipping fruit from Central Americawith clean ice or in closed refrigerator trucks,rather than with ice made from untreated riverwater, is common sense. Similarly, requiringoyster harvesters to use toilets with holdingtanks on their oyster boats is an obvious way toreduce fecal contamination of shallow oysterbeds. Pasteurization provides the extra barrierthat will prevent E. coli O157:H7 and otherpathogens from contaminating a large batch offreshly squeezed juice.

For many foodborne diseases, multiplechoices for prevention are available, and the bestanswer may be to apply several steps simulta-neously. For E. coli O157:H7 infections related tothe cattle reservoir, pasteurizing milk andcooking meat thoroughly provide an importantmeasure of protection but are insufficient bythemselves. Options for better control includecontinued improvements in slaughter planthygiene and control measures under HACCP,developing additives to cattle feed that alter themicrobial growth either in the feed or in thebovine rumen to make cows less hospitable hostsfor E. coli O157, immunizing or otherwiseprotecting the cows so that they do not become


Special Issue

infected in the first place, and irradiating beefafter slaughter. For C. jejuni infections related tothe poultry reservoir, future control options mayinclude modification of the slaughter process toreduce contamination of chicken carcasses by bileor by water baths, freezing chicken carcasses toreduce Campylobacter counts, chlorinating thewater that chickens drink to prevent them fromgetting infected, vaccinating chickens, andirradiating poultry carcasses after slaughter.

Outbreaks are often fertile sources of newresearch questions. Translating these questionsinto research agendas is an important part of theoverall prevention effort. Applied research isneeded to improve strategies of subtyping andsurveillance. Veterinary and agricultural re-search on the farm is needed to answer thequestions about whether and how a pathogensuch as E. coli O157:H7 persists in the bovinereservoir, to establish the size and dynamics of areservoir for this organism in wild deer, and tolook at potential routes of contaminationconnecting animal manure and lettuce fields.More research is needed regarding foods definedas sources in large outbreaks to develop bettercontrol strategies and better barriers tocontamination and microbial growth and tounderstand the behavior of new pathogens inspecific foods. Research is also needed to improvethe diagnosis, clinical management, and treat-ment of severe foodborne infections and toimprove our understanding of the pathogenesisof new and emerging pathogens. To assess andevaluate potential prevention strategies, appliedresearch is needed into the costs and potentialbenefits of each or of combinations.

To prepare for the 21st century, we willenhance our public health food safety infra-structure by adding new surveillance andsubtyping strategies and strengthening theability of public health practitioners toinvestigate and respond quickly. We need toencourage the prudent use of antibiotics inanimal and human medicine to limit antimicro-bial resistance. We need to continue basic andapplied research into the microbes that causefoodborne disease and into the mechanisms bywhich they contaminate our foods and causeoutbreaks and sporadic cases. Better under-standing of foodborne pathogens is thefoundation for new approaches to diseaseprevention and control.

References 1. Blake PA, Merson MH, Weaver RE, Hollis DG,

Heublein PC. Disease caused by a marine vibrio:clinical characteristics and epidemiology. N Engl JMed 1979;300:1-5.

2. Riley LW, Remis RS, Helgerson SD, McGee HB, WellsJG, Davis BR, et al. Hemorrhagic colitis associatedwith a rare Escherichia coli serotype. N Engl J Med1983;308:681.

3. Martin ML, Shipman LD, Wells JG, Potter ME, HedbergK, Wachsmuth IK, et al. Isolation of Escherichia coliO157:H7 from dairy cattle associated with two cases ofhemolytic uremic syndrome. Lancet 1986;2:1043.

4. Ortega YR, Sterling CR, Gilman RH, Cama VA, Diaz F.Cyclospora species—a new protozoan pathogen ofhumans. N Engl J Med 1993;328:1308-12.

5. Herwaldt BL, Ackers M-L, the Cyclospora WorkingGroup. International outbreak of cyclosporiasisassociated with imported raspberries. N Engl J Med. Inpress 1997.

6. Jackson LA, Wenger JD. Listeriosis: a foodbornedisease. Infections in Medicine 1993;10:61-6.

7. Dekeyser PJ, Gossin-Detrain M, Butzler JP, Sternon J.Acute enteritis due to related Vibrio; first positive stoolcultures. J Infect Dis 1972;125:390-2.

8. Tauxe RV. Epidemiology of Campylobacter jejuniinfections in the United States and other industrializednations. In: Nachamkin I, Blaser MJ, Tompkins L,editors. Campylobacter jejuni: current status andfuture trends, eds. Washington (DC): American Societyof Microbiology, 1992. pp 9-19.

9. Tauxe RV, Vandepitte J, Wauters G, Martin SM,Goosens V, DeMol P, et al. Yersinia enterocoliticainfections and pork: the missing link. Lancet1987;1:1129-32.

10. World Health Organization. Worldwide spread ofinfections with Yersinia enterocolitica. WHO Chronicle1976;30:494-6.

11. Rodrigue DC, Tauxe RV, Rowe B. Internationalincrease in Salmonella enteritidis: a new pandemic?Epidemiol Infect 1990;105:21-7.

12. Centers for Disease Control and Prevention. Multidrug-resistant Salmonella serotype Typhimurium—UnitedStates, 1996. MMWR Morb Mortal Wkly Rep1997;46:308-10.

13. Endt HP, Ruijs GJ, van Klingeren B, Jansen WH, vander Reyden T, Mouton RP. Quinolone resistance inCampylobacter isolated from man and poultryfollowing the introduction of fluoroquinolones inveterinary medicine. J Antimicrob Chemother1991;27:199-208.

14. Lee LA, Puhr ND, Maloney K, Bean NH, Tauxe RV.Increase in antimicrobial-resistant Salmonellainfections in the United States, 1989-1990. J Infect Dis1994;170:128-34.

15. Cieslak PR, Noble SJ, Maxson DJ, Empey LC,Ravenholt O, Legarza G, et al. Hamburger-associatedEscherichia coli O157:H7 in Las Vegas: a hiddenepidemic. Am J Public Health 1997;87:176-80.

16. Humphrey TJ, Greenwood M, Gilbert RJ, Rowe B,Chapman PA. The survival of salmonellas in shell eggscooked under simulated domestic conditions. EpidemiolInfect 1989;103:35-45.


Special Issue

17. Kirkland KB, Meriwether RA, Leiss JK, MacKenzieWR. Steaming oysters does not prevent Norwalk-likegastroenteritis. Public Health Rep 1996;111:527-30.

18. Holmberg SD, Feldman RA. New and newer entericpathogens: stages in our knowledge. Am J PublicHealth 1984;74:205-7.

19. St. Louis ME, Morse DL, Potter ME, DeMelfi TM,Guzewich JJ, Tauxe RV, et al. The emergence of GradeA eggs as a major source of Salmonella enteritidisinfections: implications for the control of salmonellosis.JAMA 1988;259:2103-7.

20. Mishu B, J Koehler, Lee LA, Rodrigue D, Brenner FH,Blake P, Tauxe RV. Outbreaks of Salmonellaenteritidis infections in the United States, 1985-1991. JInfect Dis 1994;169:547-52.

21. Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ,Wells JG, Griffin PM. An outbreak of diarrhea andhemolytic uremic syndrome from Escherichia coliO157:H7 in fresh-pressed apple cider. JAMA1993;269:2217-20.

22. Keene WE, Sazie E, Kok J, Rice DH, Hancock DD,Balan VK, et al. An outbreak of Escherichia coliO157:H7 infections traced to jerky made from deermeat. JAMA 1997;277:1229-31.

23. Kohn MA, Farley TA, Ando T, Curtis M, Wilson SA, JinQ, et al. An outbreak of Norwalk virus gastroenteritisassociated with eating raw oysters: implications formaintaining safe oyster beds. JAMA 1995;273:466-71.

24. Tauxe R, Kruse H, Hedberg C, Potter M, Madden J,Wachsmuth K. Microbial hazards and emerging issuesassociated with produce; a preliminary report to theNational Advisory Committee on Microbiologic Criteriafor Foods. Journal of Food Protection. In press 1997.

25. Wang G, Zhao T, Doyle MP. Fate of enterohemorrhagicEscherichia coli O157:H7 in bovine feces. Appl EnvironMicrobiol 1996;62:2567-70.

26. Zhuang R-Y, Beuchat LR, Angulo FJ. Fate ofSalmonella montevideo on and in raw tomatoes asaffected by temperature and treatment with chlorine.Appl Environ Microbiol 1995;61:2127-31.

27. Rushing JW, Angulo FJ, Beuchat LR. Implementationof a HACCP program in a commercial fresh-markettomato packinghouse: a model for the industry. Dairy,Food and Environmental Sanitation 1996;16:549-53.

28. Tauxe RV, Hughes JM. International investigations ofoutbreaks of foodborne disease: public health respondsto the globalization of food. BMJ 1996;313:1093-4.

29. Hennessey TW, Hedberg CW, Slutsker L, White KE,Besser-Wiek JM, Moen ME, et al. A national outbreakof Salmonella enteritidis infections from ice cream. NEngl J Med 1996;334:1281-6.

30. Passarro DJ, Reporter R, Mascola L, Kilman L,Malcolm GB, Rolka H, et al. Epidemic SalmonellaEnteritidis infection in Los Angeles County, California:the predominance of phage type 4. West J Med1996;165:126-30.

31. Centers for Disease Control and Prevention. Outbreaks ofSalmonella serotype Enteritidis infection associated withconsumption of raw shell eggs—United States, 1994-1995. MMWR Morb Mortal Wkly Rep 1996;45:737-42.

32. Centers for Disease Control and Prevention. Summaryof notifiable diseases, United States, 1995. MMWRMorb Mortal Wkly Rep 1995;44(53).

33. Centers for Disease Control and Prevention.Proceedings of a national conference on salmonellosis,March 11-13, 1964. U.S. Public Health ServicePublication No 1262. Washington (DC): U.S.Government Printing Office; 1965.

34. Bean NH, Morris SM, Bradford H. PHLIS: anelectronic system for reporting public health data fromremote sites. Am J Public Health 1992;82:1273-6.

35. Bean NH, Goulding JS, Lao C, Angulo FJ. Surveillancefor foodborne-disease outbreaks—United States, 1988-1992. CDC Surveillance Summaries, October 25, 1996.MMWR Morb Mortal Wkly Rep 1996;45(SS-5).

36. Centers for Disease Control and Prevention. FoodborneDiseases Active Surveillance Network, 1996. MMWRMorb Mortal Wkly Rep 1997;46:258-61.

37. Centers for Disease Control and Prevention.Establishment of a national surveillance program forantimicrobial resistance in Salmonella. MMWR MorbMortal Wkly Rep 1996;45:110-1.

38. Hutwagner LC, Maloney EK, Bean NH, Slutsker L,Martin SM. Using laboratory-based surveillance datafor prevention: an algorithm for detecting Salmonellaoutbreaks. Emerg Infect Dis 1997;3:395-400.

39. Meriwether RA. Blueprint for a national public healthsurveillance system for the 21st Century. Journal ofPublic Health Management and Practice 1996;2:16-23.

40. Mahon BE, Pönkä A, Hall WN, Komatsu K, DietrichSE, Siitonen A, et al. An international outbreak ofSalmonella infections caused by alfalfa sprouts grownfrom contaminated seed. J Infect Dis 1997;175:876-82.

41. Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC,Tarr PI, et al. A multistate outbreak of Escherichia coliO157:H7-associated bloody diarrhea and hemolyticuremic syndrome from hamburgers: the Washingtonexperience. JAMA 1994;272:1349-53.


Special Issue

Occurrence of disease is a function of severalmajor variables: the virulence of the microorgan-ism (i.e., its possession of factors that allow it tocause illness), its mode of transmission (i.e., howit gets to the host), and host susceptibility (i.e.,how well the host can defend itself against themicroorganism). Increased susceptibility toinfection may be measured in terms of infectiousdose (the number of microorganisms it takes tocause illness) and of the ability of the host to limitspread of the microorganism (e.g., the ability tolimit spread of microorganisms from the intestinaltract to the bloodstream). These same variablesapply to emerging pathogens. Microorganisms mayemerge as pathogens because they have developednew virulence genes or resistance to standardtherapeutic methods. Emergence may be relatedto changes in transmission pathways, whichpermit a known pathogen to move into new,previously unexposed populations. Finally, in-creases in the number of persons susceptible to aspecific microorganism may result in its emergenceas an important public health problem; at the sametime, when attention is focused on populations

with increased susceptibility to infection, organ-isms that might not otherwise be recognized aspathogens may be identified.

Many factors influence the susceptibility ofpopulations to infection, including increases indiseases that cause immunosuppression, in-creased use of immunosuppressive agents, agingof the population, and malnutrition. In consider-ing these categories, it should be recognized that“host susceptibility” is not a single entity. Changesin host susceptibility may be due to variousmechanisms, with each mechanism having agreater or lesser impact on the ability of the hostto defend itself against infection with specificpathogens or classes of pathogens. These are verycomplex biologic systems; nonetheless, the generalcategories outlined below may be of value inidentifying groups or populations for further study.

Increases in Diseases That CauseImmunosuppression

Hereditary diseases associated with immu-nosuppression are present in a small butrelatively constant proportion of the population.The most common of these diseases, a selectiveimmunoglobulin A deficiency, has been found inas many as 0.3% of some blood donor populations(1) and may be associated with recurrentdiarrhea; infections with giardia, in particular,

Emergence of New Pathogensas a Function of Changes

in Host Susceptibility

J. Glenn Morris, Jr.* and Morris Potter†*University of Maryland School of Medicine, Baltimore, Maryland, USA; and

†Centers for Disease Control and Prevention, Atlanta, Georgia, USA

Address for correspondence: Glenn Morris, 934 MSTF,University of Maryland School of Medicine, 10 S. Pine St.,Baltimore, MD, 21201, USA; fax: 410-706-4581; e-mail:[email protected].

A pathogen may emerge as an important public health problem because of changesin itself or its transmission pathways. Alternatively, a microorganism may emerge as apathogen or acquire new public health importance because of changes in hostsusceptibility to infection. Factors influencing host susceptibility within the population asa whole include increases in the number of immunocompromised patients; increaseduse of immunosuppressive agents, particularly among persons receiving cancerchemotherapy or undergoing organ transplantation; aging of the population; andmalnutrition. In considering the emergence of foodborne pathogens and designinginterventions to limit their spread, the susceptibility of these population subgroups tospecific infections should be taken into account.


Special Issue

have been noted to be more common among suchimmunocompromised patients.

In contrast to patients with hereditaryimmunodeficiencies, the population of patientswith acquired immunodeficiencies is rapidlyincreasing. As of June 1996 in the United States,an estimated 223,000 persons ≥ 13 years of agehad AIDS (Figure 1); this represents increases of10% and 65% since mid-1995 and January 1993,respectively. During 1995, human immunodefi-ciency virus (HIV) infection remained the leadingcause of death among persons 25 to 44 years ofage (Figure 2), accounting for 19% of deaths fromall causes in this age group (2).

Persons with AIDS show a clear increase insusceptibility to infection with Salmonellaspecies. Data suggest that risk for nontyphiSalmonella infections is increased 20- to 100-foldamong AIDS patients (3-7). Among personsinfected with Salmonella, AIDS results in aseveralfold increase in the risk for septicemia(3,5,6); AIDS also results in increases ininfections at other extraintestinal sites, compati-ble with an overall increase in risk fordissemination of the organism. This increase inrisk is reflected in increases in the percentage oftotal Salmonella isolated from blood. Forexample, for persons ages 25 to 49 years in stateswith high AIDS incidence, the percentage ofSalmonella isolates from blood increased from2.3% in 1978 to 17.8% in 1987 among men and

from 3.1% to 8.1% among women; in contrast, nochanges in blood-isolate percentages occurred foreither sex in states with low AIDS incidence (8).These latter studies further suggest that serotype isan important risk determinant, with increases inbacteremia in states with high AIDS incidenceassociated primarily with infections due toSalmonella Enteritidis and Salmonella Typh-imurium. While these data are for nontyphoidalSalmonella, studies outside the United Statessuggest that AIDS patients have a similarincrease in risk for infection with Salmonellatyphi in areas endemic for typhoid fever. Forexample, in Lima, Peru, the risk for typhoidincreased 25-fold in HIV-infected persons 15 to 35years of age (9); HIV infection also appeared toinfluence the clinical presentation of typhoidfever, with severe diarrhea and gastrointestinalsymptoms seen more often than expected.

While the above data suggest a continuingclimb in salmonellosis and, in particular,Salmonella bacteremia in conjunction with anincreasing AIDS prevalence, anecdotal data fromAIDS clinicians do not support this view. Thestandard of care for AIDS patients now includesroutine prophylactic therapy with trimethoprim/sulfamethoxazole to prevent Pneumocystis cariniipneumonia. Trimethoprim/sulfamethoxazole isalso one of the first-line therapies for salmonello-sis; widespread prophylactic use of this drug inthe AIDS population may have reduced the

Figure 1. Cases among persons aged > 13 years,adjusted for delays in reporting, by quarter year,United States, 1988–June 1996. (Points representquarterly prevalence; the line represents “smoothed”prevalence. Estimates are not adjusted for incompletereporting of diagnosed AIDS cases or AIDS deaths.From the Centers for Disease Control and Prevention.Update: Trends in AIDS incidence, deaths, andprevalence—United States, 1996. MMWR MorbMortal Wkly Rep 1997;46:165-73).

Figure 2. Death rates per 100,000 population for leadingcauses of death among persons ages 25 to 44 years, byyear, United States, 1982–1995. (Based on underlyingcauses of death reported on death certificates, using finaldata for 1982–1994 and preliminary data for 1995. Fromthe Centers for Disease Control and Prevention. Update:Trends in AIDS incidence, deaths, and prevalence,United States, 1996. MMWR Morb Mortal Wkly Rep1997;46:165-73).

1982 1984 1986 1988 1990 1992 1994

Year

0

10

20

30

40

HIV Infection

Heart Disease

Cancer

Stroke

Unintentional

Injury

Diabetes

Suicide

Liver Disease

Homicide


Special Issue

incidence of serious Salmonella infections,although this protective effect could bediminished in the face of increasing resistanceto this antimicrobial agent among clinicalisolates of Salmonella (10).

Fewer data are available on susceptibility ofAIDS patients to other acute bacterial foodborneinfections. While the initial impression was thatthe risk for Campylobacter jejuni infections wasnot higher among AIDS patients, a 35-foldincrease in the Campylobacter case rate amongpersons with AIDS was noted in one study fromLos Angeles (11). Other data indicate that HIV-positive patients can contract persistent C. jejuniinfections, with chronic diarrhea, fever, and fecalleukocytes (12). In studies in the San Franciscoarea, AIDS patients were estimated to have a280-fold increase in incidence of listeriosis, ascompared with the general population (13). Of 98nonpregnant adults with invasive Listeriainfection identified between November 1988 andDecember 1990 in selected counties in California,Tennessee, Georgia, and Oklahoma, 20% wereHIV-positive (13).

Before AIDS, Toxoplasma gondii was ofconcern primarily because of the risk forcongenital infection in infants of mothers whohad acute illness during pregnancy. T. gondii isnow the leading cause of space-occupying craniallesions in persons with AIDS (14,15); data fromthe 1980s suggest that 5% to 10% of AIDSpatients get toxoplasmic encephalitis (16). In anestimated 50% of cases, Toxoplasma is transmit-ted by food (17). In this context, Toxoplasma mustbe regarded as an important emerging pathogenin this patient population.

The AIDS epidemic has also drawn attentionto microorganisms not previously recognized aspathogens. Perhaps the most important of theseis Cryptosporidium. In early investigations ofAIDS-associated diarrhea, it rapidly becameapparent that most patients were not infectedwith traditional enteric pathogens. Many of thesepatients were infected with Cryptosporidium; anestimated 10% to 20% of cases of AIDS-associateddiarrhea are due to this microorganism (18). Insubsequent years, Cryptosporidium was alsorecognized as the cause of intestinal infections inhealthy hosts (19); and, most recently, it hasbeen recognized as a major waterbornepathogen (20). Isospora belli has also beenimplicated as a cause of diarrhea in AIDSpatients, as have the microsporidia (18). More

recently, enteroaggregative Escherichia coli andnonpathogenic bioserovars of Yersiniaenterocolitica were associated with diarrhealdisease in AIDS cases (21,22). Further work willdetermine if these latter agents are indeedemerging pathogens in this patient population.

Increased Use of ImmunosuppressiveAgents

Advances in medical treatment have resultedin increasing numbers of immunosuppressedpatients (including patients undergoing organtransplantation or cancer chemotherapy) andpatients with serious underlying chronic dis-eases; these patients, too, may be at increasedrisk for infection with microorganisms that mightotherwise not be recognized or associated withserious illness. The number of new cancer caseshas steadily increased over the past 20 years. Forwhite males, cancers at all sites have increasedfrom an estimated 364 new cases/100,000population in 1973 to 462 new cases/100,000population in 1994; white females show anincrease from 295 new cases/100,000 populationin 1973 to 347 new cases/100,000 population in1994 (23). Patients are also surviving longer. Forwhite males, the 5-year relative cancer survivalrate was 56.6% in 1986 to 1993, compared with42% in 1974 to 1976; for white females, the 1986to 1993 rate was 62.3%, compared with 57.6% in1974 to 1976. While specific data are lacking,these increases appear to have been accompaniedby increased use of chemotherapeutic regimensand regimens that may have more toxicity thanthose used in the past. The number of solid organtransplants has also increased substantively(Figure 3). In particular, more complex proce-dures such as liver, heart, and lung transplantshave increased. This, in turn, has resulted in largernumbers of chronically immunocompromisedpersons in the general population.

Aside from the direct immunosuppressiveeffect of the agents administered to thesepatients, other associated factors may contributeto susceptibility to infection. Many, if not most,patients receiving chemotherapy or immunosup-pressive agents are also treated with antimicro-bial drugs, which can have profound effects on thebacterial flora of the intestinal tract. Thesedisturbances of gut microbial ecology maypredispose to colonization and infection withother microorganisms, some of which may haveincreased virulence. Many chemotherapeutic


Special Issue

agents have direct toxicity for the gut mucosa; theresulting mucositis increases the susceptibility ofthese patients to bloodstream infections withwhatever microorganisms are present in gut. Theconcentration of these highly susceptible patientson certain wards or units in a hospital may alsoincrease the risk for nosocomial transfer ofspecific microorganisms. For example, vancomy-cin-resistant enterococci (VRE) (which, at least inEurope, have been associated with food [24]) haveemerged as a substantive problem in cancercenters and transplant units (25-27). Personswho are immunosuppressed or have seriousunderlying illness are much more susceptible tocolonization and infection with the organism. Insome oncology and transplant units, more than10% of patients are colonized or infected withVRE (26,27), providing well-documented oppor-tunities for transfer of the organism to otherimmunosuppressed hosts in the same unit.

Cancer patients who have just undergonechemotherapy are often profoundly neutropenic.In this setting, especially in conjunction with themucositis mentioned above, virtually any micro-organism in the intestinal tract can enter thebloodstream and cause potentially fatal illness.For example, it has been found that raw producein salads may be an important route by whichpatients acquire Pseudomonas (28); as a result,severely neutropenic patients are generallyrestricted to cooked food. Salmonella infectionsare reported among cancer patients, although therelative risk for infection in this population is notwell characterized. Patients with neoplastic

disease do appear to have a substantivelyincreased risk for Salmonella septicemia, with35% patients in one series having septicemia (29),versus fewer than 1% in healthy hosts. Cancerpatients appear to be at increased risk forinvasive Listeria infections (13). Toxoplasmosistends to be of particular concern amongtransplant patients (30,31).

Aging of the PopulationThe absolute number of the elderly in the

United States is rapidly increasing, as is theproportion of the U.S. population they comprise.In 1950, 3.8 million persons (2.6% of thepopulation) were over the age of 74, as opposed to14.7 million (5.6% of the population) in 1995 (23;Figures 4 and 5). The elderly appear to be at aclearly increased risk for death from foodborneand diarrheal disease. Between 1979 and 1987,28,538 persons in the United State had diarrheaas an immediate or underlying cause of death;51% of these persons were more than 74 years ofage, 27% were adults age 55 to 74, and 11% werechildren under the age of 5 (32).

The increased susceptibility of the elderly toinfection and death may be due to a number offactors. Aging results in senescence of the gut-associated lymphoid tissue, increasing suscepti-bility to infection. Aging may also result in adecrease in gastric acid secretion: in one study,stimulated acid output was reduced approxi-mately 30% in persons aged 65 to 98, with a 40%reduction in pepsin output (33). As a low pH of the

Figure 3. Number of organ transplants, by year and bysite, 1988–1996 (data obtained from the UnitedNetwork for Organ Sharing).Note: Heart-lung transplants too few to be visible.

Figure 4. Number of persons >74 years of age, U.S.population, for selected years, 1950–1990. From theNational Center for Health Statistics. Health, UnitedStates, 1996–97 and Injury Chartbook. Hyattsville,Maryland, 1997.


Special Issue

stomach is a major barrier to entry of entericpathogens, reductions in gastric acidity canclearly increase the susceptibility to infection.Social factors may also influence susceptibility.For the elderly, residence in a long-term carefacility was a major independent risk factor fordiarrheal death (32). While a number of factorscontribute to the increased risk, the communalliving environment, combined with problems offecal incontinence, create an environment inwhich enteric and foodborne pathogens areeasily spread (32,34).

Incidence of salmonellosis and Campylobacterdiarrhea appears to be higher among the elderly(35,36). More striking, however, is the increase infrequency of Salmonella bacteremia as comparedwith isolations from other sites: Salmonellainfections in the elderly are more likely to causebacteremia (37; Figure 6), which, in turn,substantively increases the risk for death. Forexample, in a recent nursing home outbreak inMaryland, 50 (35%) of 141 residents became ill,seven had bacteremia, nine were hospitalized(with a median length of hospitalization of 22days), and four died (38). E. coli O157:H7 is also acommon pathogen in nursing homes and amongthe elderly. In one reported nursing home outbreak,55 of 169 residents were affected; overall, 19(35%) of the affected residents died (39).

MalnutritionThe above discussions have focused on issues

most relevant to the United States and otherindustrialized countries. However, on a globalscale, probably the leading cause of increasedhost susceptibility to infection is malnutrition.While accurate data on the prevalence ofmalnutrition are difficult to obtain, problems areaccentuated in developing countries, in areas ofpolitical unrest, and among marginalized popula-tions in the United States and other affluentnations. In Mexico, according to a probabilisticsurvey in 1990, 42.3% of children under 5 years ofage had some degree of malnutrition (40).

Malnutrition increases host susceptibilitythrough a number of mechanisms. It weakensepithelial integrity and may have a profoundeffect on cell-mediated immunity, with functionaldeficiencies in immunoglobulins and defects inphagocytosis. Malnutrition also may initiate a“vicious cycle” of infection predisposing tomalnutrition and growth faltering, which in turnmay lead to an increased risk for further infection(40,41). In studies in Bangladesh, malnourishedand well-nourished children had the same numberof infections with diarrheal pathogens such asenterotoxigenic E. coli; however, diarrhea inmalnourished children was of longer durationand had greater potential long-term nutritionalconsequences (42). Overall, malnutrition appears

Figure 5. Percentage of U.S. population >74 years of age,for selected years, 1950–1990. From National Center forHealth Statistics. Health, United States, 1996–97 andInjury Chartbook. Hyattsville, Maryland, 1997.

0

1

2

3

4

5

6

7

8

9

0-2

3-5

6-8

9-11

1 2 3 4 5-9

10-14

15-19

20-29

30-39

40-49

50-59

60-69

70-79

80+

Months Years

Age

All serotypes

Excluding S. typhi and S. paratyphi A

Figure 6. Ratio of blood isolates to total isolates ofSalmonella by age group of the person from whom theisolate was obtained as reported to the Centers forDisease Control, Atlanta, GA, in 1968–79. FromBlaser MJ, Feldman RA. Salmonella bacteremia:reports to the Centers for Disease Control andPrevention, 1968–1979. J Infect Dis 1981;143:743-6.


Special Issue

to result in a 30-fold increase in the risk fordiarrhea-associated death (40).

ConclusionsHost susceptibility (and changes in the

susceptibility to infection of groups within thegeneral population) is a critical variable inassessing the public health effects and under-standing the emergence and spread of pathogenicmicroorganisms. Surveillance within popula-tions with increased susceptibility to infectionmay allow identification of new pathogens beforethey are recognized within the general popula-tion. Studies designed to identify the reasons forthe increased susceptibility of a specificpopulation to a specific agent may reveal how amicroorganism is able (or not able) to breachnormal host defense mechanisms. Finally, from apublic health standpoint, risk managementstrategies for emergent foodborne pathogensmust clearly identify and focus on populationswith increased susceptibility to infection.

References 1. Buckley RH. Primary immunodeficiency diseases. In:

Bennett JC, Plum F, editors. Cecil textbook of medicine.Philadelphia: W.B. Saunders Company; 1996.p.1401-8.

2. Centers for Disease Control and Prevention. Update:trends in AIDS incidence, deaths, and prevalence—United States, 1996. MMWR Morb Mortal Wkly Rep1997;46:165-73.

3. Celum CL, Chaisson RE, Rutherford GW, Barnhart JL,Echenberg DF. Incidence of Salmonellosis in patientswith AIDS. J Infect Dis 1987;156:996-1002.

4. Smith PD, Macher AM, Bookman MA, Boccia RV, SteisRG, Gill V, et al. Salmonella typhimurium enteritisand bacteremia in the acquired immunodeficiencysyndrome. Ann Intern Med 1985;102:207-9.

5. Profeta S, Forrester C, Eng RHK, Liu R, Johnson E,Palinkas R, Smith SM. Salmonella infections inpatients with acquired immunodeficiency syndrome.Arch Intern Med 1985;145:670-2.

6. Gruenewald R, Blum S, Chan J. Relationship betweenhuman immunodeficiency virus infection andSalmonellosis in 20- to 59-year-old residents of NewYork City. Clin Infect Dis 1994;18:358-63.

7. Angulo FJ, Swerdlow DL. Bacterial enteric infectionsin persons infected with human immunodeficiencyvirus. Clin Infect Dis 1995;21:S84–93.

8. Levine WC, Buehler JW, Bean NH, Tauxe RV.Epidemiology of nontyphoidal Salmonella bacteremiaduring the human immunodeficiency virus epidemic. JInfect Dis 1991;164:81-7.

9. Gotuzzo E, Frisancho O, Sanchez J, Liendo G, Carrillo C,Black RE, Morris JG Jr. Association between the acquiredimmunodeficiency syndrome and infection with Salmonellatyphi or Salmonella paratyphi in an endemic typhoid area.Arch Intern Med 1991;151:381-2.

10. Lee LA, Puhr ND, Maloney EK, Bean NH, Tauxe RV.Increase in antimicrobial-resistant Salmonellainfections in the United States, 1989–1990. J Infect Dis1994;170:128–34.

11. Sorvillo FJ, Lieb LE, Waterman SH. Incidence ofcampylobacteriosis among patients with AIDS in LosAngeles County. J Acquir Immune Defic Syndr HumRetrovirol 1991;4:598-602.

12. Perlman DM, Ampel NM, Schifman RB, Cohn DL,Patton CM, Aguirre ML, et al. Persistent Campylobacterjejuni infections in patients infected with humanimmunodeficiency virus (HIV). Ann Intern Med1988;108:540-6.

13. Schuchat A, Deaver KA, Wenger JD, Plikaytis BD,Mascola L, Pinner RW, et al. Role of foods in sporadiclisteriosis. 1. Case-control study of dietary risk factors.JAMA 1992;267:2041-50.

14. Luft BJ, Brooks RG, Conley FK, McCabe RE,Remington JS. Toxoplasmid encephalitis in patientswith acquired immune deficiency syndrome. JAMA1984;252:913-7.

15. Porter SB, Sande MA. Toxoplasmosis of the centralnervous system in the acquired immunodeficiencysyndrome. N Engl J Med 1992;327:1643-8.

16. McCabe RE, Remington JS. Toxoplasma gondii. In:Mandell GL, Douglas RG, Bennett JE, editors.Principles and practice of infectious diseases. 3rd ed.New York: Churchill Livingston; 1990. p. 2090-103.

17. United States Department of Agriculture. Pathogenreduction; hazard analysis and critical control point(HACCP) systems; proposed rule. Washington (DC):Federal Register 1995;60:6774-889.

18. Smith PD, Quinn TC, Strober W, Janoff EN, Masur H.Gastrointestinal infections in AIDS. Ann Intern Med1992;116:63-77.

19. Wolfson JS, Richter JM, Waldron MA, Weber DJ,McCarthy DM, Hopkins CC. Cryptosporidiosis inimmunocompetent patients. N Engl J Med1985;312:1278-82.

20. Centers for Disease Control and Prevention. Assessingthe public health threat associated with waterbornecryptosporidiosis: report of a workshop. MMWR MorbMortal Wkly Rep 1995;44(RR-6):19.

21. Mayer HB, Acheson D, Wanke CA. EnteroaggregativeEscherichia coli are a potential cause of persistentdiarrhea in adult HIV patients in the United States. In:Program and abstracts of the 31st US-Japan Choleraand Related Diarrheal Diseases Conference; 1995 Dec1-3; Kiawah Island, South Carolina.

22. Saillour M, de Truchis P, Risbourg M, Nordman P,Nauciel C, Perronne C. Yersinia enterocoliticagastroenteritis in HIV infected patients. In: Abstractsof the 35th Interscience Conference on AntimicrobialAgents and Chemotherapy; 1995 Sept 17-20; SanFrancisco, California.

23. National Center for Health Statistics. Health, UnitedStates, 1996-97 and Injury Chartbook. Hyattsville(MD): The Center; 1997.

24. Bates J, Jordens JZ, Griffiths DT. Farm animals asa putative reservoir for vancomycin-resistantenterococcal infection in man. J AntimicrobChemother 1994;34:507-16.


Special Issue

25. Morris JG Jr, Shay DK, Hebden JN, McCarter RJ,Perdue BE, Jarvis W, et al. Enterococci resistant tomultiple antimicrobial agents, including vancomycin:establishment of endemicity in a university medicalcenter. Ann Intern Med 1995;123:250-9.

26. Papanicolaou GA, Meyers BR, Meyers J, MendelsonMH, Lou W, Emre S, et al. Nosocomial infections withvancomycin-resistant Enterococcus faecium in livertransplant recipients: risk factors for acquisition andmortality. Clin Infect Dis 1996;23:760-6.

27. Montecalvo MA, Horowitz H, Gedris C, Carbonaro C,Tenover FC, Issah A, et al. Outbreak of vancomycin-,ampicillin-, and aminoglycoside-resistant Enterococcusfaecium bacteremia in an adult oncology unit.Antimicrob Agents Chemother 1994;38:1363-7.

28. Remington JS, Schimpff SC. Please don’t eat thesalads. N Engl J Med 1981;304:433-5.

29. Wolfe MS, Armstrong D, Louria DB, Blevins A.Salmonellosis in patients with neoplastic disease. ArchIntern Med 1971;128:546-54.

30. Luft BJ, Naot Y, Araujo FG, Stinson EB, RemingtonJS. Primary and reactivated toxoplasma infection inpatients with cardiac transplants. Ann Intern Med1983;99:27-31.

31. Hofflin JM, Potasman I, Baldwin JC, Oyer PE, StinsonEB, Remington JS. Infectious complications in hearttransplant recipients receiving cyclosporine andcorticosteroids. Ann Intern Med 1987;106:209-16.

32. Lew JF, Glass RI, Gangarosa RE, Cohen IP, Bern C,Moe CL. Diarrheal deaths in the United States, 1979-87. JAMA 1991;265:3280-4.

33. Feldman M, Cryer B, McArthur KE, Huet BA, Lee E.Effects of aging and gastritis on gastric acid and pepsinsecretion in humans: a prospective study.Gastroenterology 1996;110:1043-52.

34. Bennett RG. Diarrhea among residents of long-termcare facilities. Infect Control Hosp Epidemiol1993;14:397-404.

35. Hargrett-Bean N, Pavia AT, Tauxe RV. Salmonellaisolates from humans in the United States, 1982-1986.MMWR Morb Mortal Wkly Rep 1988;30(SS-2):25-31.

36. Tauxe RV, Hargrett-Bean N, Patton CM,Wachsmuth K. Campylobacter isolates in theUnited States, 1982-1986. MMWR Morb MortalWkly Rep 1988;37(SS-2):1-13.

37. Blaser MJ, Feldman RA. Salmonella bacteremia:reports to the Centers for Disease Control, 1968-1979.J Infect Dis 1981;143:743-6.

38. Taylor JL, Dwyer DM, Groves C, Bailowitz A,Tilghman D, Kim V, et al. Simultaneous outbreak ofSalmonella enteritidis and Salmonella schwarzengrundin a nursing home: association of S. enteritidis withbacteremia and hospitalization. J Infect Dis1993;167:781-2.

39. Carter AO, Borczyk AA, Carlson JAK, Harvey B, HogkinJC, Karmali MA, et al. A severe outbreak of Escherichiacoli O157:H7-associated hemorrhagic colitis in a nursinghome. N Engl J Med 1987;317:1496-500.

40. Santos JI. Nutrition, infection, and immunocompetence.Inf Dis Clin North Am 1994;8:243-67.

41. Black RE, Brown KH, Becker S. Effects of diarrheaassociated with specific enteropathogens on thegrowth of children in rural Bangladesh. Pediatrics1984;73:799-805.

42. Black RE, Brown KH, Becker S. Malnutrition is adetermining factor in diarrheal duration, but notincidence, among young children in a longitudinal study inrural Bangladesh. Am J Clin Nutr 1984;37:87-94.


Special Issue

The term foodborne disease encompasses avariety of clinical and etiologic conditions anddescribes a subset of enteric disease (1-4), whichin the United States ranks second in prevalence torespiratory disease (2). In foodborne disease, thefood may act as a vehicle for the transmission ofactively growing microorganisms or products ofmetabolism (toxins), or it may have a passive roleas a vehicle for the transmission of nonreplicatingbacteria, viruses or protozoa, or stable biologictoxins. In most cases, the clinical conditionsusually associated with foodborne disease areacute: diarrhea, vomiting, or other gastrointesti-nal manifestations such as dysentery. However,other pathophysiologic responses may occurindependently or accompany acute-phase re-sponses (1-4). A number of chronic sequelae mayresult from foodborne infections, includingankylosing spondylitis, arthropathies, renaldisease, cardiac and neurologic disorders, andnutritional and other malabsorptive disorders(incapacitating diarrhea). The evidence thatmicroorganisms or their products are eitherdirectly or indirectly associated with these long-term sequelae ranges from convincing tocircumstantial (1-4). The reason for this disparityis that, except in rare circumstances, chroniccomplications are unlikely to be identified orepidemiologically linked to a foodborne causebecause these data are not systematicallycollected. Moreover, host symptoms induced by aspecific pathogen or product of a pathogen areoften wide-ranging and overlapping and thereforedifficult to link temporally to a specific incident.These impediments manifest themselves because

the problems associated with chronic disease canresult from an infection without overt illness.Alternatively, the chronic sequelae may beunrelated to the acute illness and may occur evenif the immune system successfully eliminates theprimary infection; therefore, activation of theimmune system may initiate the chroniccondition as a result of an autoimmune response(2-4). The variability of the human response—from overt illness to chronic carrier status—isperhaps the most confounding issue.

Cost of Chronic SequelaeAs the incidence of foodborne disease

increases, the incidence of chronic sequelaemay also rise. Several authors have estimatedthat chronic sequelae may occur in 2% to 3% offoodborne disease cases and suggest that thelong-term consequences to human health andthe economy may be more detrimental than theacute disease. An estimated 80 million cases offoodborne disease occur annually in the UnitedStates, which suggests significant morbidityfigures and costs to society in the billions ofdollars per year (2,4).

Infection: The Microbe/Toxinversus the Host

Several microbial pathogens are highlyadapted to parasitization, exhibiting environmen-tally responsive and adaptive traits that allowattachment, invasion, and replication in the host(2,4). Microbial pathogenicity can be viewed solelyfrom the perspective of the microbe; however, thiswould be not only unidimensional, but also wrong(2,4). A major selective force that regulates thephenotype of an infecting microbial pathogenpopulation is the host’s immune system, which isalso highly adaptive, especially in discriminating

Chronic Sequelae of Foodborne Disease

James A. LindsayUniversity of Florida, Gainesville, Florida, USA

Address for correspondence: James A. Lindsay, Food Scienceand Human Nutrition Department, University of Florida,Gainesville, FL 32611 USA; fax 352-392-9467; e-mail:[email protected].

In the past decade, the complexity of foodborne pathogens, as well as theiradaptability and ability to cause acute illness, and in some cases chronic (secondary)complications, have been newly appreciated. This overview examines long-termconsequences of foodborne infections and intoxications to emphasize the need formore research and education.


Special Issue

self and nonself antigens (2,4). When the host-parasite relationship is examined holistically,mechanisms that successful pathogens haveapparently evolved to elude the immune systeminclude antigenic heterogeneity or variation;sequestration, either intracellularly or in certainspecific host sites; molecular mimicry, througheither imitation (cross-reaction) or adsorption ofhost protein; and direct immune stimulation and/or suppression (2-5).

Rheumatoid DiseaseSeveral bacteria, including salmonellae,

induce septic arthritis by hematogenous spread tothe synovial space, causing inflammation. Viableorganisms are recoverable from synovial fluid,and treatment usually involves antibiotic therapy.Prognosis depends on host factors and virulence ofthe organism; either complete resolution orpermanent joint damage can occur (1-4,6,7).

Yersinia enterocolitica, Y. pseudotuberculo-sis, Shigella flexneri, Sh. dysenteriae, Salmonellaspp., Campylobacter jejuni, and Escherichia coliinitiate aseptic or reactive arthritis, an acute,nonpurulent joint inflammation following infec-tion elsewhere in the body, for example the bowel.Klebsiella pneumoniae has been implicated,although it appears now that the bacterium isconnected with fecal carriage by ankylosingspondylitis probands (4). Although a distinctclinical disease, reactive arthritis also occurs inthe Reiter syndrome triad with conjunctivitis anduveitis. A subset of patients with symptoms ofreactive arthritis and Reiter syndrome getankylosing spondylitis, a rheumatoid inflamma-tion of synovial joints and entheses within anddistal to the spine (8).

The relative risk of developing theseseronegative spondyloarthropathies after a gram-negative enterobacterial infection is high forpersons positive for the major histocompatibilityclass (MHC) antigen B27 and the cross-reactingMHC B7 group. Indeed, persons who are humanleukocyte antigen (HLA)-B27 positive have an 18-fold greater risk for reactive arthritis, a 37-foldgreater risk for Reiter syndrome, and up to a 126-fold greater risk for ankylosing spondylitis thanpersons who are HLA-B27 negative and have thesame enteric infections. Other genes that may berelated or act in concert appear to determinewhich disease is acquired (2,5,7,9). These chroniccomplications are related to a geneticallydetermined host risk factor in combination with

an environmental trigger. No cause-and-effectrelationship of enteric pathogens in ankylosingspondylitis has been established (4); however, alow but consistent incidence (0.2% to 2.4%) ofreactive arthritis occurs after outbreaks of S.Typhimurium, Sh. flexneri, and C. jejuni.Biotypes and phage types of Y. enterocolitica O:3and O:9, endemic to Scandinavia, are either highlyarthritogenic or affect a more genetically predis-posed population with persistent and debilitatingsymptoms that may last for years (2,3).

The sharing of antigenic determinants by amicrobe and its host is a frequent naturaloccurrence, and bacterial antigens from thepathogens that directly cross-react with MHCB27 have been demonstrated (6,9). Additionally,the plasmid-mediated synthesis of bacterial B27“modifying factor,” a protein that binds to andsubsequently alters the conformation of B27, hasbeen reported (10). In both of these models,immune recognition of the foreign antigen leads toan autoimmune anti-B27 response. Alternatively,B27 may act nonimmunologically because dissemi-nation of bacterial antigens to infected jointsstimulates a local T-cell inflammatory response.Here, B27 may act as a receptor for bacteria orantigens thereof, facilitating invasiveness frommucosal surfaces in the gut (9). Indeed, transfectedB27 on the surface of mouse L cells reportedly canalter bacterial invasion capability (11).

Despite the strong familial association relatedto the MHC B27 gene, B27-negative persons areknown to become ill, albeit less often, but withapparently equal severity, as shown by anepidemiologic investigation of rheumatoid arthri-tis following the 1985 S. Typhimurium gastroen-teritis outbreak due to contaminated milk (4).

Autoimmune Thyroid DiseaseGraves disease is an autoimmune disease

mediated by autoantibodies to the thyrotropinreceptor (12,13). The first indication that thedisease may be linked to infection was findingantibody titers to Y. enterocolitica serotype O:3suggestive of molecular mimicry in a majority ofpatients with Graves disease. Several studieshave shown that two low molecular weightenvelope proteins of Yersinia contain epitopescross-reactive with the thyrotropin receptor. Theseproteins are chromosomally encoded, exposed tothe surface of the bacterium, and produced byboth virulent and avirulent strains of Yersinia(Y. pestis, Y. pseudotuberculosis, Y. enterocolitica


Special Issue

VW+ and WV-). In addition to autoantibody, asuppressor cell dysfunction may be involved inGraves disease (12,13). Severe hypothyroidismmay also result from chronic intestinal giardiasisdue to infection by Giardia lamblia; treatmentwith metronidazole can result in completeelimination of the parasite and recovery of regularintestinal thyroid hormone absorption (14).

Inflammatory Bowel DiseaseInflammatory bowel disease is the collective

term for Crohn disease and ulcerative colitis.While both infections are chronic inflammatorydiseases with histologic infiltrates of macroph-ages and lymphocytes and a prolonged clinicalcourse, the primary clinical and pathologic effectsare gastrointestinal. The infections can bedifficult to differentiate because the symptomsare often similar (15). The acute clinicalcharacteristics are diarrhea, abdominal pain, fever,and weight loss; and the acute pathologic featuresinclude a constant flux of neutrophils into inflamedmucosa, eventually penetrating the epithelium intothe intestinal lumen. The chronic spontaneouslyrelapsing disorder exhibits many of the symptomsof the acute state; however, this phase has anaverage symptom duration of 3.2 years beforecorrect diagnosis. Abdominal abscesses are acommon and dangerous complication of Crohndisease, while in ulcerative colitis, abdominalperforations may lead to peritonitis. Crohn diseaseinvolves the ileum or colon (anaerobes areimportant), while ulcerative colitis appearsrestricted to the colon (aerobes are important).Nationality and familial associations suggest agenetic predisposition for the disease (4,15).

Although the cause of inflammatory boweldisease and the mechanism(s) for spontaneousexacerbations and remissions are unknown, muchresearch has focused on transmissible agents,including foodborne pathogens. An associationbetween bacterial L-forms and inflammatorybowel disease has been sporadically reported,with isolation of Pseudomonas, Mycobacterium,Enterococcus fecalis, and E. coli from affectedtissue but not from appropriate controls. There isconsiderable debate as to whether L-forms arepathogenic in humans or persist in affected tissue.

Mycobacterium paratuberculosis, the caus-ative agent of Johne disease in ruminants, may beassociated with Crohn disease through theproduction of L-forms of the bacterium. Subclini-cally infected cows shed M. paratuberculosis, and

the organism has been identified in pasteurizedmilk by polymerase chain reaction specific for theM. paratuberculosis insertion sequence IS900.The pathogen model suggests that a susceptiblehuman neonate first contracts the organism afteringesting commercial dairy products. Thisinvokes an antigen-poor (lacking a cell wall) L-form that grows slowly and persists in the laminapropria, stimulating a chronic low-grade inflam-mation. The immune response increases inseverity over years without bacterial replication,ultimately producing the pathologic features ofCrohn disease (15,16). Another model proposes anautoimmune phenomenon mediated by alter-ations in inflammatory cytokine profiles, possiblyas a result of infection (4).

Recent immunocytochemical techniquesdemonstrated antigens to Listeria monocytogenes,E. coli, and Streptococcus spp. in Crohn diseasetissues. Macrophages and giant cellsimmunolabelled for antigen specific to theseorganisms were found beneath ulcers, aroundabscesses, along fissures, within the laminapropria, in granulomas, and in germinal centers ofmesenteric lymph nodes (17).

Superantigens and AutoimmunityIn contrast to conventional antigens,

superantigens interact with the variable side ofthe Vß chain of the T-cell receptor by recognizingelements shared by a subset of T cells. Dependingon the type of interaction, recognition can havedifferent consequences, including proliferationand expansion, suppression (clonal deletion), or,alternatively, induction of prolonged unrespon-siveness (anergy) or cell death (apoptosis) (18-21).Superantigens from several foodborne bacteria(e.g., Staphylococcus, Streptococcus, Yersinia,and Clostridium) have been isolated andcharacterized. Many are thought to be associatedwith several autoimmune disorders, for example,rheumatic heart disease, rheumatoid arthritis,multiple sclerosis, Graves disease, Sjogrensyndrome, autoimmune thyroiditis, psoriasis,Kawasaki disease, Crohn disease, and insulin-dependent diabetes mellitus (6,18-24).

Although it is accepted that superantigenshave a role in autoimmune disorders, theacceptance is based on extensive animal modelstudies (6,18-24) but limited human clinicalstudies. In human diseases where superantigenshave been clearly demonstrated as the cause, forexample, toxic shock syndrome, initial T-cell


Special Issue

proliferation and T-cell receptor–mRNA up-regulation have been observed, but the long-termsequelae in terms of T-cell function are unknown(22). Recent studies suggest that superantigensmay also cause an acute flare of a disease withinpatients in remission from a preexistingautoimmune disorder.

Renal DiseaseAfter colitis caused by E. coli O157:H7 and

other enterohemorrhagic strains of E. coli,hemolytic uremic syndrome develops in somepatients (1,2,25). The syndrome is characterizedby a triad of symptoms: acute renal failure,thrombocytopenia, and microangiopathichemolytic anemia. Acute renal failure is theleading cause of death in children, andthrombocytopenia is the leading cause of death inadults. Hemolytic uremic syndrome is a world-wide problem that mirrors the distribution of E.coli O157:H7 and other Shiga and Shiga-liketoxin–producing microorganisms. Outbreaks ofhemorrhagic colitis and subsequent cases ofhemolytic uremic syndrome have developed as aresult of various food vehicles. Besides O157:H7,other Shiga-like toxin–producing E. coli,Citrobacter, Campylobacter, Shigella, Salmo-nella, and Yersinia have been linked to thedisorder (1,2,25,26).

The toxin-mediated damage to the kidneysmay not be limited to the glomerular endothelialcells as once thought but may include the tubularepithelial cells (26-28). Binding studies showedthe toxins to be specific for the glycosphingolipidglobotriaosylceramide (Gb3), which is present onrenal but not umbilical endothelial cells. This mayaccount for the differential sensitivity of renalcells to toxin-induced damage, since Gb3 waspresent in the glomeruli of infants under 2 yearsof age but not in the glomeruli of adults. Thus, thepresence of Gb3 in the pediatric renal glomerulusmay be a risk factor for development of hemolyticuremic syndrome (28). Characterization of theShiga toxin receptor has led to a potentialpreventive treatment (4).

Neural and Neuromuscular DisordersGuillain-Barré syndrome is a subacute,

acquired, inflammatory demyelinatingpolyradiculoneuropathy that frequently occursafter acute gastrointestinal infection. The diseaseis characterized by alexia, motor paralysis withmild sensory disturbances, and an acellular

increase in the total protein content in thecerebrospinal fluid. The disease occurs worldwideand is the most common cause of neuromuscularparalysis. Cases have three dominant character-istics: the predilection to nerve roots, mono-nuclear infiltration of peripheral nerves, andeventual demyelination (primary axonal degen-eration) (29). Severe cases tend to occur in thesummer and have been linked to previousinfection with C. jejuni, although other entericpathogens may trigger the syndrome.

Some controversy exists regarding whetherGuillain-Barré syndrome is an autoimmunedisease. Although adequate data exist to classifythe syndrome as an autoimmune disease (fourmajor Rose-Witebsky criteria are almost com-pletely met), the immunologic mechanisms atwork in Guillain-Barré syndrome triggered by C.jejuni are likely to be complex (29-31). Studies ofthe relationship between Guillain-Barré syn-drome and C. jejuni support the hypothesis ofmolecular mimicry, since peripheral nerves mayshare epitopes with surface antigens of C. jejuni(32). This has been supported by studies in whichanti-GM1 IgG antibodies recognized surfaceepitopes on intact C. jejuni, and the reaction wasstrain-specific for certain Penner serotypes.There are inconclusive data with regard toGuillain-Barré syndrome and HLA, althoughsome studies have shown a predilection for theHLA-B35 haplotype (29-31). Cytokines may beresponsible for inducing the inflammatoryprocess and probably play a role in the responseleading to nerve demyelination. Complement hasa role in the process leading to nerve damage,possibly through the production of activationproducts, which lead to an increase in thepermeability of the blood nerve barrier, whichperpetuates the inflammation. Although Guillain-Barré syndrome might be considered anautoimmune response, it also serves as anexample of a disease with an infectious origin, adisease that entails the integrated actions of bothhumoral and cellular immunity.

Ciguatera poisoning is the most commonfoodborne disease related to the consumption offin fish; this distinctive clinical syndrome ischaracterized by a plethora of gastrointestinal,neurologic, and sometimes cardiovascular fea-tures (33,34). Two toxins are involved in toxicosis.Ciguatoxin-1 (cig-1), the principal toxin, is a heatstable, lipid-soluble polyether that is notinactivated by heat, cold, or gastric juices, nor


Special Issue

eliminated by drying, salting, smoking, ormarinating. Cig-1 induces membrane depolariza-tion in nerve and muscle tissue by openingvoltage-dependent sodium channels. A secondtoxin, maitotoxin, is water-soluble and openscalcium channels. The role of this second toxin inthe pathophysiology of the disease is less wellunderstood. The acute symptoms of the toxicosisare varied and include paresthesia of theextremities, circumoral paresthesia, reversal ofhot and cold sensations, dental pain, myalgias,arthralgias, generalized pruritus, cranial nervedysfunction, and dysuria. Severe acute symptomsrequire urgent care with parenteral atropine forbradyarrhythmias. Mannitol is often adminis-tered to counter the effect of the toxin on thesodium channels; however, the mechanism ofaction is unknown, and the therapy is uselessafter 24 hours. Many of these symptoms mayremain chronic and are often misdiagnosed aschronic fatigue syndrome, brain tumors, ormultiple sclerosis. The management of thechronic symptoms is frustrating for the patientand clinician. Interventions include amitrip-tyline, tocaidine, or mexilitine to modulatesodium channels in conjunction with calciumchannel blockers such as nifedipine. Antidepres-sants such as Prozac also appear to be useful.Patients with chronic symptoms frequently reportwaxing and waning of symptoms. Activities such assexual intercourse and drinking alcohol signifi-cantly exacerbate expression. Some women withchronic symptoms report worsening duringmenses. Mood levels, weather conditions, anddietary constituents often exacerbate symptoms.Some clinicians advocate a strict diet that avoids allseafood, fish byproducts, nuts, and alcohol, and insome cases, patients are asked to abstain from sex.One distinctive feature in this toxicosis is that oneepisode of ciguatera poisoning does not conferimmunity. In fact, it is likely to sensitize the patientto otherwise subthreshold doses of toxin (33,34).

Amnesic shellfish poisoning is caused bydomic acid, a conjugate of kainic and glutamicacid (35). In small quantities domic acid has anexcitatory effect, but in large amounts it isneurotoxic. The toxicosis is first characterized bygastrointestinal symptoms followed by neurologicdysfunction. Severe cases may be prolonged andchronic; sequelae include confusion with disorien-tation, paucity of speech, lack of response to deeppain due to blocking of receptors in the spinalcord, autonomic nervous system dysfunction,

seizures, abnormal ocular movements, grimacingposture, myoclonus, loss of reflexes, and coma.Other prominent chronic sequelae include loss ofvisual-spacial recall and mononeuropathies with-out dementia, mimicking Alzheimer’s disease.The toxicosis is particularly serious in the elderly,and any deaths usually occur within thispopulation. Valium, calcium channel blockers,phenobarbital, diazapam, thiobarbiturates, andhypothermia are treatments for patients withsevere and chronic cases.

General Immunity, Organ Impairment, andNeurologic Disorders

Toxoplasmosis due to Toxoplasma gondii is achronic latent parasitic infection (36,37). Inhumans the parasite exists in two forms: thetachyzoite, the rapidly multiplying stage thatactively invades host cells and represents theprincipal form of the acute phase of the disease;and the bradyzoite, the form that multiplies veryslowly in host cells, resulting in the formation ofcysts that persist in tissues. Toxoplasma infectionin humans is usually asymptomatic because ofeffective immunity involving antibodies, T cells,and cytokines. Activated macrophages, CD4 andCD8 lymphocytes, and the cytokines IFN-γ,TNF-α, and IL-6 play a major role in control ofboth the acute infection and maintenance orprevention of the chronic stage (37,38). Indeed,treatment with IFN-γ is used to control passageinto the chronic stage, and treatment with anti-IFN-γ reactivates chronic infection (39). Theproduction of nitric oxide may have opposingeffects. Nitric oxide production protects against T.gondii and at the same time limits the immuneresponse, probably contributing to the establish-ment of the chronic state (40).

The incidence of congenital toxoplasmosis isuncertain but may be as high as 9,500 cases a year(1). The percentage attributable to food isuncertain; however, consumption or contact withcontaminated meat is more important as a causethan is contact with cats (1). Congenitalimpairments associated with maternal toxoplas-mosis infection passed to the fetus includehearing loss, visual impairment (retinal lesions,strabismus), and slight to severe mentalretardation. These impairments are still presentin 80% of persons who reach the age of 20 years(1). Chronic toxoplasmic encephalitis may occurwhen a person’s immune system is impaired.Indeed, toxoplasmic encephalitis marked by


Special Issue

dementia and seizures has become the mostcommonly recognized cause of central nervoussystem opportunistic infections in AIDS patients.Additionally, it appears that certain cancertreatments weaken the immune system, and oldinfections in the muscles can become reactivated,causing severe complications or death (1,41,42).

Helminth parasites can cause serious diseasein infected persons (42). The impact of helminthinfections is due less to the severity of the diseasesthey cause, than to the vast number of personsinfected. For example, more than one billionpeople are infected with the largest intestinalnematode, Ascaris lumbricoides. Although thereis usually no overt clinical sign of infection,disease can arise from an overwhelming infectionor an inappropriate immune response. Addition-ally, infected persons frequently harbor morethan one parasite for years. Most intestinalhelminth parasites have direct life cycles, with nointermediate host or vector, and are transferredby contaminated food. Some species, such asTrichuris (whipworm) and Enterobius (pinworm),are restricted to the gut, but others, such asAscaris, have tissue-migrating phases. All,however, induce a dramatic expansion of the Th2lymphocyte subset. It remains unclear whetherthese Th2-derived responses (induction ofinterleukin-4 [IL-4] and down-regulation ofIFN-γ), resulting in stimulation of IgG1 and IgEisotypes, eosinophilia, and mastocytosis areresponsible for the immune-mediated pathologicresponse. Immunologic lesions may occur whereearly infection is associated with a strong T-cellproliferative response that becomes down-regulated in established chronic disease (evidenceof a Th1 defect in the chronic disease). Inascariasis, an allergic response generated by thelung migratory phase (chronic immune sensitiza-tion) can cause pneumonia and, in animal models,spontaneous development of idiopathic bronchialasthma. A formative influence on the response ofthe immune system is the antigenic environmentduring pregnancy. Children born to infectedmothers may have significantly higher suscepti-bility to the same infection later in life (42).

Viral agents induce autoimmune disorders,and one potential mechanism of induction ismolecular mimicry (43,44). Hepatitis A virusinfection is a well-recognized cause of acutehepatitis with jaundice in adults. In most affectedpersons, the course is usually relatively short-lived and benign, and symptoms are usually

resolved within weeks. Occasionally, relapsesoccur after initial recovery, or recovery is markedby severe or prolonged cholestasis. However,even in these cases recovery is usual. Chronicsequelae of hepatitis A virus infections are rareand poorly defined; however, several recentstudies suggest that hepatitis A virus infectiontriggers the onset of (idiopathic) autoimmunechronic active hepatitis within a geneticallypredisposed subgroup. Apparently, the chronicdisorder may develop despite normal serologicresponse to hepatitis A virus infection. Thetriggering factors and mechanism of actionremain ill-defined; however, in the most recentstudy, the authors concluded that hepatitis Avirus infection may be the precipitating event inthe pathogenesis of this disorder (45).

Metabolic activation and detoxification play acrucial role in determining the toxic response ofhumans to mycotoxin exposure. These highlytoxic secondary metabolites are produced by awide variety of molds including Aspergillus,Fusarium, and Penicillium. Mycotoxins exhibitproperties of acute, subacute, and chronictoxicities with some molecules being carcinogenic,mutagenic, and teratogenic. Because mycotoxinsare resistant to food processing and do notdegrade at high temperatures, they enter thehuman food supply (46-49).

In many cases, the relationship betweenmycotoxins as the causative agent of disease inhumans is difficult to determine. Acute effects ofgastroenteritis may be easily identified; however,chronic effects often result from ingestion of low tomoderate levels and can be difficult to recognize(46-50). The most threatening effects of ochra-toxin A are its nephrotoxicity and carcinogenicity.Ochratoxin A is increasingly involved in anendemic nephropathy, a human chronic intersti-tial neuropathy that is usually associated withurinary tract tumors. Aflatoxins have beenimplicated in both acute and chronic liver diseasein humans; however, other organs (kidney,spleen, pancreas) may also be affected (51). Thebest studied chronic effects are those induced bythe fumonisins, zearalenone, and trichothecenemycotoxins produced by Fusarium sp. (49).Fumonisin levels in corn-based foods have beenstatistically associated with an increased risk ofhuman esophageal cancer. Zearalenone is anestrogenic mycotoxin. Ingestion by animals,especially swine, causes hyperestrogenism withsymptoms of enlargement and prolapse of the


Special Issue

uterus, atrophy of ovaries and testicles, enlarge-ment of mammary glands, and infertility. Thismycotoxin might add to the estrogen load ofhumans. Human consumption of trichothecene-contaminated foods causes acute symptoms ofheadaches, chills, severe nausea, vomiting, andvisual disturbances, which may last 7 to 10 days.Since trichothecenes modulate immune function,over time mycotoxicosis could reduce immuneresistance to infectious diseases, facilitate tumorgrowth through reduced immune function, andcause autoimmune disease (48).

Heart and Vascular DiseasesSeveral foodborne pathogens have been

either directly or indirectly associated withendocarditis and myocarditis, and any heartdamage appears to be permanent (2,52). Personswith ankylosing spondylitis linked to entericpathogens as the trigger show a high incidence ofcardiac conduction abnormalities, which may besequelae to other seronegative arthropathies. Apossible connection between foodborne gram-negative bacteria and atherosclerosis has beenproposed, suggesting that the bacteria gainaccess to the lymphoid and general circulationwith relative frequency. Endotoxin from degrad-ing bacteria in macrophages may act in concertwith the inflammatory factors (cytokines)induced by endotoxin from endothelium andsmooth muscle cells. Although the process ofatherogenesis is complex and involves manyfactors, the hypothesis is attractive and providesa model system for further study. Oxidativestress responses by E. coli and S. Typhimuriumand the induction of the peroxide stimulon andthe superoxide stimulon have also been recentlyimplicated in atherosclerosis, rheumatoid arthri-tis, and inflammatory bowel disease (53).

Nutritional and GastrointestinalDisturbances

Enteric pathogen-induced diarrhea may leadto a variety of conditions including loss of fluids,anorexia, and malabsorption of nutrients, allforms of malnutrition. The enteric pathogens thatcause malabsorption and nutrient loss vary andinclude Enterobacteriaceae, Rotavirus, Amoeba,Cryptosporidium, and Giardia (2,54-58). Unlesstreated with antimicrobial drugs, many diarrhealepisodes become chronic; however, the stress ofeven short periods of diarrhea may result insubtle changes in immunologic status. The extent

of the diseases depends mostly on the immunestatus of the person and may last for severalyears or for life. Death due to diarrheal illness inthe immunosuppressed and in persons withAIDS is nearly 80%. No effective treatment isavailable, although treatment with severalantibiotics in combination shows promisingresults. However, AIDS patients may alsodevelop further sequelae. Cryptosporidium arehost-adapted, which may lead to pulmonary ortracheal cryptosporidiosis accompanied by cough-ing and frequent low-grade fever. In these casesthere is no effective treatment. Similarly, incyclosporidiosis, AIDS patients’ enteric infectionis chronic. Long-term prophylaxis withtrimoxazole is required, and discontinuation ofthe treatment causes severe relapse (2,54-58).

Severe cases of diarrhea lasting months oryears and characterized by dysentery, with foul-smelling, mucous bloody stools accompanied byflatulence and abdominal distention, may resultfrom C. jejuni, Citrobacter, Enterobacter, orKlebsiella enteric infections. These infectionsalways require extensive antibiotic therapy andusually result in failure to thrive. Entericinfections may alter bowel permeability whichallows absorption of otherwise excluded foodcomponents. Proteins that can modulate theimmune system can be absorbed possibly withdeleterious consequences such as the induction ofautoimmunity and atopy. Several studies of bothhuman and porcine models indicate thatsignificant quantities of unwanted proteins can beabsorbed by damaged gut tissue and thatmaximum expression of diarrhea correspondswith peak protein uptake (2).

Helicobacter pylori is the undisputed cause ofchronic gastritis. Environmental sources indicatethat H. pylori can survive in water, chilled foods,milk, and fresh vegetables for several daysbecause of fecal contamination. The species hasnever been isolated from these sources; however,infectious, viable but nonculturable (nonspiralcoccoid) bodies may survive in fresh water formore than a year. H. pylori can be found in humanfeces and can be transmitted directly from personto person by the fecal-oral or oral-oral route. H.pylori can be found in several animal reservoirs;however, the possibility of animal-to-animal orzoonotic transmission is unknown (59,60).Ingestion of H. pylori leads to acute gastritis, andcolonization of the stomach is virtually alwaysaccompanied by chronic inflammation that


Special Issue

disappears within 6 to 12 months aftereradication of the infection. Infections aregenerally acquired during childhood or adoles-cence and result in chronic gastritis lasting fordecades or life. On the basis of histologic andserologic follow-up studies, this chronic gastritishas been suggested as an important risk factor(odds ratio 9.0; p = 0.001) or first stage in amultistep process leading to gastric mucosalatrophy, intestinal metaplasia, and eventuallygastric cancer (61,62).

Chronic Sequelae and PersonalityChanges

One area that has received scant interest isthe effect of a chronic infection on humanpersonality factors. Personality changes might bepredicted: continual pain from arthritis, anirritable bowel, or chronic diarrhea would beenough to make anyone temperamental, moody,or depressed. Studies using Cattell’s 16 Personal-ity Factor questionnaire showed highly signifi-cant correlations (p < 0.01 - 0.002) betweenchronic toxoplasmosis and several personalityfactors (63,64). Men and women showed distinctdifferences in behavioral states. For men, lowsuperego strength, protension, guilt-proneness,and group dependency were positively associated,whereas in women the related factors wereaffectothymia, alexia, untroubled adequacy, andself-sufficiency. A correlation of the intensity ofthe personality factor-shifts with the duration ofthe infection suggested that the infection per seinduced the shift in personality, not vice versa.

An exploratory study using the 16 Person-ality Factor questionnaire and the Holmes andRahe Social Readjustment Rating Scale, whichmeasures stressful life events, was made ofpatients with rheumatoid arthritis (65). As agroup, patients with rheumatoid arthritisexhibited higher stress at disease onset (p <0.01); a large high-stress-at-onset subgroup ofrheumatoid arthritis patients had a worseprognosis. Although there were importantpersonality changes in the high-stress-at-onset-rheumatoid arthritis patients, the studyconcluded that the interaction between thevariables that determined personality changeswere very complex and could not simply bereferred to as the “rheumatoid arthritispersonality” complex.

ConclusionFoodborne diseases are for the most part

preventable; however, there is an inherent riskassociated with the consumption of certain typesof uncooked foods. Recognition by the publichealth community and the public that manyfoodborne illnesses may have serious chronicsequelae would help eliminate many illnesses andreduce health-care cost. Public health authoritiescould make a substantial impact by reducing pooror unhygienic food production or food-handlingpractices and by educating the public about howharmful microorganisms enter the food chain andhow they can be avoided.

References 1. Foegeding PM. Foodborne pathogens: risks and

consequences. Council for Agricultural Science andTechnology. Task Force Report 1221; 1994.

2. Archer DL, Young FE. Contemporary issues: diseaseswith a food vector. Clin Microbiol Rev 1988;1:377-98.

3. Bunning VK. Immunopathogenic aspects of foodbornemicrobial disease. Food Microbiology 1994;11:89-95.

4. Bunning VK, Lindsay JA, Archer DL. Chronic healtheffects of foodborne microbial disease. World HealthStat Quart 1997;50:51-6.

5. Miller FW. Genetics of autoimmune diseases. Exp ClinImmunogenet 1995;12:182-90.

6. Behar SM, Porcelli SA. Mechanisms of autoimmunedisease induction: the role of the immune response tomicrobial pathogens. Arthritis Rheum 1995;38:458-76.

7. Yu DT, Thompson GT. Clinical, epidemiological andpathogenic aspects of reactive arthritis. FoodMicrobiology 1994;11:97-108.

8. Sieper J, Braun J. Pathogenesis of spondylarthropathies.Arthritis Rheum 1995;38:1547-54.

9. Gaston JH. How does HLA-B27 confer susceptibility toinflammatory arthritis? Clin Exp Immunol 1990;82:1-2.

10. McGuigan LE, Prendergast JK, Geczy AF, EdmondsJP, Bashir HV. Significance on non-pathogenic cross-reactive bowel flora in patients with anklosingspondylitis. Ann Rheum Dis 1986;45:566-71.

11. Kapasi K, Inman RD. HLA-B27 expression modulatesGram-negative bacterial invasion into transfected L-cells. J Immunol 1992;148:3554-9.

12. Wolf MW, Misaki T, Bech K, Tvede M, Silva JE, IngbarSH. Immunoglobulins of patients recovering fromYersinia enterocolitica infections exhibit Graves’disease-like activity in human thyroid membranes.Thyroid 1991;1:315-20.

13. Luo G, Seetharamaiah GS, Niesel DW, Zhang H, PetersonJW, Prabhakar BS, Klimpel GR. Purification andcharacterization of Yersinia enterocolitica envelopeproteins which induce antibodies that react with humanthyrotropin receptor. J Immunol 1994;152:2555-61.

14. Seppel T, Rose F, Schlaghecke R. Chronic intestinalgiardiasis with isolated levothyroxine malabsorptionas reason for severe hypothyroidism: implications forlocalization of thyroid hormone absorption in the gut.Exp Clin Endocrinol Diabetes 1996;104:180-2.


Special Issue

15. Chiodini RJ. Crohn’s disease and the mycobacteriosis:a review and comparison of two disease entities. ClinMicrobiol Rev 1989;2:90-117.

16. Millar D, Ford J, Sanderson J, Withey S, Tizard M,Doran T, Hermon-Taylor J. IS900 PCR to detectMycobacterium paratuberculosis in retail supplies ofwhole pasteurized cow’s milk in England and Wales.Appl Environ Microbiol 1996;62:3446-52.

17. Liu Y, van Kruiningen HJ, West AB, Cartun RW, CortotA, Colombel JF. Immunocytochemical evidence ofListeria, Escherichia coli, and Streptococcus antigens inCrohn’s disease. Gastroenterology 1995;108:1396-401.

18. Goodacre JA, Brownlee CED, Ross DA. Bacterialsuperantigens in autoimmune arthritis. Br JRheumatol 1994;33:413-9.

19. Kotb M. Bacterial pyrogenic exotoxins as superantigens.Clin Microbiol Rev 1995;8:411-26.

20. Kotb M. Infection and autoimmunity: a story of thehost, the pathogen and the copathogen. Clin ImmunolImmunopathol 1995;74:10-22.

21. Johnson HM, Torres BA, Soos JM. Superantigens:structure and relevance to human disease. Proc SocExp Biol Med 1996;212:99-109.

22. Leung DY, Meissner HC, Fulton DR, Murray DL,Kotzin BL, Schlievert PM. Toxic shock syndrome toxinsecreting Staphylococcus aureus in Kawasaki syndrome.Lancet 1993;342:1385-8.

23. Kay RA. The potential role of superantigens ininflammatory bowel disease. Clin Exp Immunol1995;100:4-6.

24. Conrad B, Weidmann E, Trucco G, Rudert WA, BehbooR, Ricordi C, et al. Evidence for superantigeninvolvment in insulin dependent diabetes mellitisaetiology. Nature 1994;371:351-5.

25. Tarr PI. Escherichia coli O157:H7: clinical, diagnostic,and epidemiological aspects of human infection. ClinInfect Dis 1995;20:1-10.

26. Tesh VL, O’Brien AD. The pathogenic mechanisms ofShiga toxin and the Shiga-like toxins. Mol Microbiol1991;5:1817-22.

27. Pudymaitis A, Armstrong G, Lingwood CA. Verotoxin-resistant cell clones are deficient in the glycolipidglobotriosylceramide: differential basis of phenotype.Arch Biochem Biophys 1991;286:448-52.

28. Lingwood CA. Verotoxin-binding in human renalsections. Nephron 1994;66:21-8.

29. Shoenfeld Y, George J, Peter JB. Guillain-Barré as anautoimmune disease. Int Arch Allergy Immunol1996;109:318-26.

30. Peterson MC. Clinical aspects of Campylobacter jejuniinfections in adults. West J Med 1994;161:148-52.

31. Bolton CF. The changing concepts of Guillain-BarréSyndrome. New Engl J Med 1995;333:1415-7.

32. Oomes PG, Jacobs BC, Hazenberg MP, Banffer JR,van der Meche FG. Anti-GM1 IgG antibodies andCampylobacter bacteria in Guillain-Barré Syndrome:evidence of molecular mimicry. Ann Neurol1995;38:170-5.

33. Swift AE, Swift TR. Ciguatera. J Toxicol Clin Toxicol1993;31:1-29.

34. Lange WR. Ciguatera fish poisoning. Am FamPhysican 1994;50:579-84.

35. Todd ECD. Domic acid and amnesic shellfishpoisoning: a review. Journal of Food Protection1993;56:69-83.

36. Smith JL. Foodborne toxoplasmosis. Journal of FoodSafety 1991;12:17-57.

37. Derouin F. Pathology and immunological control oftoxoplasmosis. Braz J Med Biol Res 1992;25:1163-9.

38. Khalil SS, Rashwan EA. Tumor necrosis factor-alpha(TNF-alpha) in human toxoplasmosis. J Egypt SocParisitol 1996;26:53-61.

39. Olle P, Bessieres MH, Malecaze F, Seguela JP. Theevolution of ocular toxoplasmosis in anti-interferongamma treated mice. Curr Eye Res 1996;15:701-7.

40. Hayashi S, Chan CC, Gazzinelli R, Roberge FG.Contribution of nitric oxide to the host parasiteequilibrium in toxoplasmosis. J Immunol1996;156:1476-81.

41. Shrittematter C, Lang W, Wiestler OD, Kleihues P.The changing patter of HIV associated cerebraltoxoplasmosis. Acta Neuropathol (Berl) 1992;8:475-81.

42. Allen JE, Maizels JM. Immunology of human helminthinfection. Int Arch Allergy Immunol 1996;109:3-10.

43. Moyer L, Warwick M. Mahoney FJ. Prevention ofhepatitis A virus infection. Am Fam Physician1996;54:107-14.

44. von Herrath MG, Oldstone MBA. Virus-inducedautoimmune disease. Curr Opin Immunol 1996;8:878-85.

45. Rahyaman SM, Chira P, Koff RS. Idiopathicautoimmune chronic hepatitis triggered by HepatitisA. Am J Gastroenterol 1994;89:106-8.

46. Robens JF, Richard JL. Aflatoxins in animal andhuman health. Rev Environ Contam Toxicol1992;127:69-94.

47. Coulombe RA. Biological actions of mycotoxins. J DairySci 1993;76:880-91.

48. Pestka JJ, Bondy GS. Immunologic effects ofmycotoxins. In: Miller JD, Trenholm HL, editors.Mycotoxins in grain. St Paul (MN): Eagan Press; 1994.pp. 339-58.

49. Jackson LS, De Vries JW, Bullerman LB, editors.Fumonisins in food. New York: Plenum Press; 1996.

50. Neal GE. Genetic implication in the metabolism andtoxicity of mycotoxins. Toxicol Lett 1995;82-83:861-7.

51. Creppy EE, Baudrimont I, Betbeder AM. Prevention ofnephrotoxicity of ochratoxin A, a food contaminant.Toxicol Lett 1995;83-83:869-77.

52. Archer DL. Foodborne gram-negative bacteria andatherosclerosis: is there a connection? Journal of FoodProtection 1987;50:783-7.

53. Farr SB, Kogoma Y. Oxidative stress response inEscherichia coli and Salmonella typhimurium.Microbiol Rev 1991;55:561-85.

54. Bruckner DA. Amebiasis. Clin Rev Microbiol1992;5:356-69.

55. Wolfe MS. Giardiasis. Clin Microbiol Rev 1992;5:93-100.

56. Chappell CL, Matson CC. Giardia antigen detection inpatients with chronic gastrointestinal disturbances. JFam Pract 1992;35:49-53.

57. Smith JL. Cryptosporidium and Giardia as agents offoodborne disease. Journal of Food Protection1994;56:451-61.


Special Issue

58. Keusch GT, Hamer D, Joe A, Kelley M, Griffiths J,Ward H. Cryptosporidia—who is at risk. Schweiz MedWochenschr 1995;125:899-908.

59. Owen RJ. Bacteriology of Helicobacter pylori.Bailliere’s Clin Microbiol 1995;9:415-46.

60. Fox JG. Non-human reservoirs of Helicobacter pylori.Aliment Pharmacol Ther 1995;9(Suppl 2):93-103.

61. Kuipers EJ, Uyterlinde AM, Pena AS, Roosendaal R,Pals G, Nelis GF, et al. Long term sequelae ofHelicobacter pylori gastritis. Lancet 1995;345:1525-8.

62. Sipponen P, Kekki M, Seppala K, Siurala M. Therelationship between chronic gastritis and gastricsecretion. Aliment Pharmacol Ther 1996;10(Suppl1):103-18.

63. Flegr J, Hrdy I. Influence of chronic toxoplasmosis onsome human personality factors. Folia Parasitol(Praha) 1994;41:122-6.

64. Flegr J, Zitkova S, Kodym P, Frynta D. Induction ofchanges in human behaviour by the parasiticprotozoan Toxoplasma gondii. Parasitology1996;113:49-54.

65. Latman NS, Walls R. Personality and stress: anexploratory comparison of rheumatoid arthritis andosteoarthritis. Arch Phys Med Rehabil 1996;77:796-800.


Special Issue

U.S. Fisheries SystemCoastal estuaries serve as a breeding ground

and provide habitats for more than 75% ofcommercial landings and 80% to 90% of therecreational catch of fish and shellfish. Fromthese habitats, hundreds of species of seafood areproduced. Aquacultured species now contributeup to 15% of the U.S. supply (1,2). Wild speciesare harvested by 17,000,000 recreational anglersand nearly 300,000 commercial harvesters.Commercial harvesters deploy 93,000 vessels,while recreational fishermen have millions ofrecreational fishing boats. Nearly 5,000 domesticplants are located in every state throughout theUnited States, not just in the coastal areas (3).Current per capita consumption of commerciallyharvested species averages 15 pounds; it isestimated that per capita consumption ofrecreationally harvested seafood approaches anadditional 3 to 4 pounds per person (4).

The seafood business community—in theUnited States and in other industrializedcountries—cannot rely solely on domesticallyproduced stock. For a number of years, more thanhalf of U.S. seafood consumption has relied onimported stock. Currently, the United Statesimports more than 50% of the consumed seafood,which originates in 172 countries around theworld (3). This trend toward economic reliance onimported stock has steadily increased over thepast 10 years so that now the United States is theworld’s second largest importer of seafood. Theprincipal seafood imports are tuna, shrimp,salmon, lobster, and groundfish (3).

It is difficult to determine where importedfish was harvested. For example, the UnitedStates imports salmon from Switzerland andPanama, although neither Switzerland norPanama is noted for vast salmon resources. U.S.participation in the international seafood trade isvery complex, since in addition to being theworld’s second largest importer, the UnitedStates is also the world’s second largest exporterof seafood (3). This dichotomy requires that U.S.marketing and import/export food control

Public, Animal, and Environmental HealthImplications of Aquaculture

E. Spencer Garrett,* Carlos Lima dos Santos,†and Michael L. Jahncke‡

*National Seafood Inspection Laboratory, Pascagoula, Mississippi, USA;†Food and Agriculture Organization of the United Nations, Rome, Italy;

‡Virginia Tech Seafood Research and Extension Center,Hampton, Virginia, USA

Address for correspondence: Michael Jahncke, Virginia TechSeafood Research and Extension Center, 102 King Street, P.O.Box 369, Hampton, VA 23669 USA; fax: 757-727-4871; e-mail:[email protected].

Aquaculture is important to the United States and the world’s fishery system. Bothimport and export markets for aquaculture products will expand and increase asresearch begins to remove physiologic and other animal husbandry barriers.Overfishing of wild stock will necessitate supplementation and replenishment throughaquaculture. The aquaculture industry must have a better understanding of the impactof the “shrouded” public and animal health issues: technology ignorance, abuse, andneglect. Cross-pollination and cross-training of public health and aquaculture personnelin the effect of public health, animal health, and environmental health on aquaculture arealso needed. Future aquaculture development programs require an integrated Gestaltpublic health approach to ensure that aquaculture does not cause unacceptable risks topublic or environmental health and negate the potential economic and nutritionalbenefits of aquaculture.


Special Issue

inspection strategies be carefully planned. Forexample, the United States exports seafood to 162countries, which has come about with the fulldevelopment of northwest and Alaska fisheries andimproved efficiency in processing techniques.Major U.S. exports are salmon, crab, surimi, fishblocks, groundfish, flatfish, shrimp, and lobster (3).

Current Aquaculture StatusIn 1996, U.S. aquaculture production of

nearly 227,000 metric tons consisted of baitfish,catfish, salmon, trout, clams, crawfish, mussels,oysters, fresh and saltwater shrimp, andmiscellaneous species such as ornamental fish,alligators, algae, aquatic plants, tilapia, andhybrid striped bass. The United States exportedprincipally rainbow trout, Atlantic salmon,tilapia, catfish, freshwater crawfish, and livemussels to 19 countries in Europe, North andSouth America, and Asia. Freshwater crawfishled the export seafood market at slightly over $8million, with the other species accounting for lessthan $1 million each (3). The United States alsoimports large volumes of aquacultured products,approximately $2.5 billion in cultured products,primarily shrimp and salmon. Imported culturedseafood accounts for most of the current U.S.trade deficit for edible fishery products, whichwas approximately $3.5 billion in 1995.

The Safety of SeafoodMost seafood is safe; however, like all foods, it

carries some risk. The food safety issues forseafood are highly focused, well-defined, andlimited to a very few species. For seafood-borneillnesses (in which the cause was known)reported to the Centers for Disease Control andPrevention, more than 90% of the outbreaks and75% of the individual cases were associated withciguatoxin (from a few reef fish species) andscombrotoxin (from tuna, mackerel, bluefish, anda few other species) and the consumption ofmollusks (mostly raw) (5-12).

Hazards associated with the consumption ofall food (including seafood) can be categorizedinto three areas: product safety; food hygiene(clean vs. dirty plants, wholesome vs. unwhole-some products); and mislabeling or economicfraud. Traditionally, the food safety risks ofseafood products (aquacultured and wild-caught)have been subcategorized by environment,process, distribution, and consumer-inducedrisk; the environmental risk category is further

subdivided into natural hazards (e.g., biotoxins)and anthropogenic contaminants (e.g., polychlo-rinated biphenyls) (13).

“Shrouded” Aquaculture HazardsThe future of aquaculture is bright;

aquaculture products are as safe and wholesomeas wild-caught species. However, in addition tothe consumer hazards listed above, there aresome less obvious “shrouded” public healthhazards associated with ignorance, abuse, andneglect of aquaculture technology.

Technology IgnoranceA common practice in many developing

countries is the creation of numerous small fishpond impoundments. However, this approachmay have a greater adverse effect on humanhealth than the creation of a single largeimpoundment (14). Small impoundments greatlyincrease the overall aggregate shoreline of ponds,causing higher densities of mosquito larvae andcercaria, which can increase the incidence andprevalence of diseases such as lymphaticfilariasis and schistosomiasis, respectively.Centralized planning approaches for newfreshwater and marine aquaculture sites shouldinclude discussions of the potential effect of largeor small impoundments on such issues as diseasetransmission, water supply, irrigation, andpower generation (14).

Ignorance of the microbial profile ofaquaculture products can also affect humanhealth as evidenced by the recent transmission ofstreptococcal infections from tilapia to humans,which resulted in several meningitis cases inCanadian fish processors (15). A change inmarketing strategies to sell live fish in smallcontainers, instead of ice-packs, resulted inhuman Vibrio infections from live tilapia inIsrael in 1996. Such bacteria can be present inother aquacultured and wild-caught species inaddition to tilapia.

Ignorance of the hazards associated with theuse of untreated animal or human waste inaquaculture ponds to increase production alsohas tremendous human health implications (16).For centuries, food growers have cultured speciesin waste-water–fed ponds and grown secondaryvegetable crops in waste water and sedimentmaterial in integrated aquaculture operations.However, the potential for transmission ofhuman pathogens to cultured species and


Special Issue

secondary vegetable crops is rarely considered byfishery aquaculturists. For example, of morethan 250 presentations at the 1997 WorldAquaculture Society meeting held in Seattle,Washington, few referred to the potential humanhealth implications of aquaculture (17).

The potential transmission of animal patho-gens from exotic aquacultured species to wild-stock species also affects animal health. Recentoutbreaks of taura, yellow spot, and white headviruses have occurred in aquaculture shrimp inSouth Carolina and Texas. Recent studiesindicate that native wild white shrimp may alsobe susceptible to these exotic viruses (18).

Technology AbuseTechnology abuse includes the willful misuse

of therapeutic drugs, chemicals, fertilizers, andnatural fishery habitat areas. The widespreaduse and misuse of antibiotics to control diseasesin aquaculture species is worldwide and willprobably increase as aquaculturists movetowards more intensive animal husbandry–rearing techniques and stocking densities. Forexample, the illegal use of chloramphenicol inshrimp culture to control diseases may result inviolative levels in the harvested product.Similarly, the improper or illegal use ofchemicals (e.g., tributyl tin) to control pond pestssuch as snails can also result in human healthhazards. The abuse and misuse of raw chickenmanure as pond fertilizer may result in thetransmission of Salmonella from manure to thecultured product (16).

The destruction of mangrove areas to buildaquaculture ponds can have a drastic impact onthe survival of wild aquatic species through thedegradation of essential fish habitats andnurseries. In Brazil, destruction of mangroveareas for shrimp ponds affected climatic changesto such an extent that the aquaculture operationshave been terminated because consequentreduced rainfall resulted in excessive pondsalinity (19).

Technology NeglectThe final “shrouded” hazard associated with

aquaculture involves technology neglect, whichincludes such events as the abandonment ofsmall aquaculture ponds in tropical countries,leading to increased mosquito habitats andconcomitant increases in malaria (14). Facilitymanagement can be responsible for technology

neglect if employees are not trained in the properuse and application of therapeutics andchemicals, for example. Finally, from an animalhealth perspective, ignorance or willful neglect ofthe International Council for Explorations of theSea/European Inland Fisheries Advisory Com-mission Code of Practice for the Introduction andTransfer of Marine Organisms can result in theescape of exotic species and animal pathogensinto the environment with a potential tragicimpact on native aquatic species (20).

Health Control Considerations

Human HealthProcedures to help protect humans from

aquaculture-associated risks include bettereducation and training of aquaculture personnelon the proper use and storage of therapeutics andchemical compounds. Additional research on newmore effective and, we hope, safer, antibiotic andvaccine treatment of aquaculture species isunder way. Likewise, certain extralabel useapplications for selected antibiotics are underconsideration. Streamlined enforcement effortsare being developed to ensure compliance withnew food safety regulations and new regulatorycontrol procedures such as Hazard Analysis andCritical Control Points (HACCP) and theapplication of HACCP principles to animal andenvironmental control procedures (21,22).

The Food and Agriculture Organization andWorld Health Organization recommend that theHACCP concept be applied to fresh wateraquaculture programs to control foodbornedigenetic trematode infections in humans.Experiments are being carried out in Asia by amultidisciplinary team of experts in publichealth, parasitology, aquaculture, fisheriesextension, and fish inspection (22). In one studyin Vietnam, experimental activities were con-ducted in two side-by-side fish ponds. In theexperimental ponds, fish were cultured inconjunction with HACCP principles, and controlpond fish were cultured according to conven-tional local aquaculture practices (22). Watersupply, fish fry, fish feed, and pond conditions inthe experimental pond were identified as criticalcontrol points. The HACCP principles of hazardanalysis, preventative measures, critical limits,monitoring, recordkeeping, and verificationprocedures relating to the critical control pointswere applied; study results showed Clonorchis


Special Issue

sinensis eggs and fish infected with the parasitemetacercaria and the first intermediate host(Melanoides tuberculata) in the experimentalponds (22). Forty-five percent of control pond fishwere infected with C. sinensis metacercaria,while white fish from the experimental pondmonitored according to HACCP principles werecompletely free of trematode infection (22).Preliminary results indicate that application ofHACCP-based principles to silver carp culture inNorth Vietnam is an effective way to prevent andcontrol C. sinensis. Similarly, the application ofthese principles to fresh water aquacultureponds in Thailand and Laos to controlOpisthorchis viverrini infections has also beensuccessful. Additional studies are recommendedto confirm these preliminary results (22).

Animal HealthProcedures to safeguard animal health are

set out in the International Council forExplorations of the Sea and the European InlandFisheries Advisory Commission Codes of Prac-tice, which describe how to prevent the adverseeffects of introducing new and exotic species andemerging animal pathogens. Education and on-site training programs for aquaculture employ-ees will help them understand the detrimentalimpact of introduction of exotic species andanimal pathogens, misuse and abuse oftherapeutics and chemicals, and willful habitatdestruction. High priority issues also includeimplementation of biosecurity procedures inaquaculture operations to prevent the escape andspread of exotic species and pathogens into thefacility and surrounding natural environmentand the use of the HACCP principles to helpcontrol the spread of exotic pathogens to wildaquatic populations (17,23).

The application of HACCP principles tocontrol transmission of exotic shrimp virusesfrom cultured to wild shrimp was proposed at ashrimp pathogen workshop held in June 1996 inNew Orleans, Louisiana (21). Natural resourceregulatory agencies are concerned about thepossible transmission of exotic shrimp patho-genic viruses, recently found in shrimp aquacul-ture ponds in Texas and South Carolina, to wildnative shrimp populations. The principles ofHACCP, in conjunction with International Councilfor Explorations of the Sea and the EuropeanInland Fisheries Advisory Commission Codes ofPractice, were proposed to control the spread of

exotic animal viruses into the environment.Shrimp aquaculture has the following proposedcritical control points: pond site selection; watersupply quality; pond management techniques;and transportation, especially as it relates to thelive transport of aquaculture shrimp species (21).

Approximately 600 million pounds of shrimpare also imported for further processing into theUnited States on a yearly basis, half of which areaquacultured species (3). Natural resourcemanagers, particularly at the state level, areconcerned about the possible transmission ofexotic shrimp pathogens into the environmentfrom shrimp processing plant wastewaterdischarge and solid waste material landfill leakage.Proposed HACCP shrimp processing plant criticalcontrol points include unload/receive; de-ice/wash;thaw; dehead/peel/devein; wash; re-ice; de-ice/wash; re-ice; and dip/glaze (21).

Application of HACCP principles at aquacul-ture site and processing plant locations has thepotential to control transmission of exotic humanand animal pathogens. However, to our knowledge,except for the application of HACCP principles tocontrol of human pathogens in Asia (17), noresearch has been conducted on this issue.

References 1. Ratafia M. Aquaculture today: a worldwide status

report. Aquaculture News 1994;3:12-13,18-19. 2. Rhodes RJ. Status of world aquaculture. Aquaculture

Magazine 17th Annual Buyers Guide 1987;4-18. 3. National Marine Fisheries Service. Fisheries of the United

States 1995. Current Fisheries Statistics No. 9500.Washington (DC): U.S. Department of Commerce; 1996.

4. Krebs-Smith SM. Report on per capita consumption andintake of fishery products [letter]. In: National Academy ofSciences Committee on the Evaluation of the Safety ofFishery Products 1991 Seafood Safety Report. Washington(DC): National Academy Press; 1991.

5. Centers for Disease Control. Annual summary offoodborne diseases (1978). Atlanta (GA):CDC;1979.



8. Centers for Disease Control. Annual summary offoodborne diseases (1980 and 1981). Atlanta(GA):CDC;1983.


10. Centers for Disease Control. Annual summary offoodborne diseases, unpublished data. Atlanta(GA):CDC;1989.

11. Centers for Disease Control. Foodborne diseaseoutbreaks, 1983-87. MMWR CDC Surveill Summ1990;39(SS-1):1-45.


Special Issue

12. National Marine Fisheries Service. The draft report ofthe model seafood surveillance project. A report toCongress. Pascagoula (MS): Office of Trade andIndustry Services, National Seafood InspectionLaboratory;1995.

13. Garrett ES. Role of diseases in marine fisheriesmanagement. Transactions of the 50th North AmericanWildlife and Natural Resources. NOAA Conference 1986.Tech. Memo. NMFS, F/NWR-16. Washington (DC):National Marine Fisheries Service;1986.

14. Mott KE. A proposed national plan of action forschistosomiasis control in the United Republic ofCameroon. Geneva: World Health Organization(unpublished document WHO/SCHISTO/86,88); 1996.

15. Centers for Disease Control and Prevention. Invasiveinfection with Streptococcus iniae-Ontario, 1995-1996.MMWR Morb Mort Wkly Rep 1996; 45(30):650-3.

16. Buras N. Microbial safety of produce from wastewaterfed aquaculture. In: Pullin RVC, Rosenthal H, MacleanJL, editors. Environment and aquaculture indeveloping countries. Proceedings of the 31stInternational Center for Living Aquatic ResourcesManagement Conference; 1993; Manila, Philippines.Manila: The Center; 1993. p. 285-95.

17. World Aquaculture Society. Abstracts of the AnnualInternational Conference and Exposition of the WorldAquaculture Society and The National AquacultureConference and Exposition of National AquacultureAssociation; 1997 Feb 19-23; Seattle, WA.

18. Virus re-emerges at shrimp farm. The AquacultureNews 1996;4(11):1-15.

19. Pullin RSV. Discussion and recommendations onaquaculture and the environment in developingcountries. In: Pullin RSV, Rosenthal H, Maclean JL,editors. Environment and aquaculture in developingcountries. Proceedings of the 31st ICLARM Conference1993. p. 312-338.

20. Turner GE, editor. Codes of practice and manual ofprocedures for consideration of introductions andtransfers of marine and freshwater organisms. EIFACOccasional Paper 23. Rome: FAO; 1988.

21. Jahncke JL. The application of the HACCP concept tocontrol exotic shrimp viruses. Proceedings of theNMFS Workshop on Exotic Shrimp Viruses. 1996June; New Orleans, LA.

22. Son TQ, Hoi VS, Dan TV, Nga C, Toan TQ, Chau LV, et al.Application of hazard analysis critical control point(HACCP) as a possible control measure against Clonorchissinensis in cultured silver carp Hypophthalmichthysmolitrix. Paper presented at 2nd Seminar on Food-BorneZoonoses: Current Problems, Epidemiology and FoodSafety; 1995 Dec 6-9; Khon Kaen, Thailand.

23. Austin B. Environmental issues in the control ofbacterial diseases of farmed fish. In: Pullin RVC,Rosenthal H, Maclean JL, editors. Environmental andaquaculture in developing countries. Proceedings ofthe 31st International Center for Living AquaticResources Management Conference; 1993; Manila,Philippines. Manila: The Center; 1993. p. 237-51.


Special Issue

Advances in agronomic, processing, preser-vation, packaging, shipping, and marketingtechnologies on a global scale have enabled thefresh fruit and vegetable industry to supplyconsumers with a wide range of high-qualityproduce year round. Some of the same technologiesand practices have also introduced an increasedrisk for human illness associated with pathogenicbacteria, mycotoxigenic molds, viruses, andparasites. The use of manure rather thanchemical fertilizer, as well as the use of untreatedsewage or irrigation water containing pathogens,viruses, or parasites, undoubtedly contributes tothis increased risk. Changes in the produceindustry, social demographics, food consumptionpatterns, and awareness of fresh fruits andvegetables as potential vehicles of infection mayalso be contributing to an increase in documentedproduce-associated outbreaks of human illness.

Changing factors that contribute to theepidemiology of diseases that may be associatedwith fresh fruits and vegetables were discussedby Hedberg et al. (1). Increases in foodborneillness during the summer are not fullyunderstood, although fresh produce is likely toplay a role since it is consumed in higherquantities during the summer. The per capitaconsumption of fresh produce has increased inthe United States in recent years (Figure 1), notonly in the summer but also in other seasons,partly because of increased importation.Knowledge of the presence and numbers ofspecific pathogens on produce imported to theUnited States from countries that may havelower sanitation standards is minimal. However,produce from a single grower, packinghouse, orshipper, whether located outside or within theUnited States, may be routinely distributedthroughout the country, thus facilitatingwidespread dissemination of potential pathogens.The epidemiology of foodborne diseases is greatlyinfluenced by these global changes. Control or

Produce Handling andProcessing Practices

Larry R. Beuchat and Jee-Hoon RyuUniversity of Georgia, Griffin, Georgia, USA

Address for correspondence: Larry R. Beuchat, Center forFood Safety and Quality Enhancement, University ofGeorgia, Griffin, Georgia 30223 USA; fax: 770-229-3216; e-mail: [email protected].

In the past decade, outbreaks of human illness associated with the consumptionof raw vegetables and fruits (or unpasteurized products produced from them) haveincreased in the United States. Changes in agronomic, harvesting, distribution,processing, and consumption patterns and practices have undoubtedly contributed tothis increase. Pathogens such as Listeria monocytogenes, Clostridium botulinum, andBacillus cereus are naturally present in some soil, and their presence on fresh produceis not rare. Salmonella, Escherichia coli O157:H7, Campylobacter jejuni, Vibriocholerae, parasites, and viruses are more likely to contaminate fresh produce throughvehicles such as raw or improperly composted manure, irrigation water containinguntreated sewage, or contaminated wash water. Contact with mammals, reptiles, fowl,insects, and unpasteurized products of animal origin offers another avenue throughwhich pathogens can access produce. Surfaces, including human hands, which comein contact with whole or cut produce represent potential points of contaminationthroughout the total system of growing, harvesting, packing, processing, shipping, andpreparing produce for consumption. Treatment of produce with chlorinated waterreduces populations of pathogenic and other microorganisms on fresh produce butcannot eliminate them. Reduction of risk for human illness associated with rawproduce can be better achieved through controlling points of potentialcontamination in the field; during harvesting; during processing or distribution; orin retail markets, food-service facilities, or the home.


Special Issue

elimination of pathogenic microorganisms fromfresh fruits and vegetables can be achieved onlyby addressing the entire system, from the field,orchard, or vineyard to the point of consumption.

We reviewed some of the practices (particularlypreharvest practices) used by the fresh fruit andvegetable industry that may promote contamina-tion of produce with pathogenic microorganisms.

Sources of ContaminationThe presence of pathogenic bacteria,

viruses, and parasites on fresh fruits andvegetables has been extensively documented(3). Contamination of produce can occur in thefield or orchard; during harvesting, postharvesthandling, processing, shipping, or marketing;or in the home (Table).

Preharvest SourcesSpores of Clostridium species, including C.

botulinum and C. perfringens, as well as spores ofenterotoxigenic Bacillus cereus, are commonlyfound in soil, so their occasional presence onfruits and vegetables should not be unexpected.

Numbers of clostridial spores on some types ofvegetables appear to increase during the summer(4). Perhaps the most prevalent disease-causingmicroorganism in soil is Listeria monocytogenes(5,6). Twenty-seven strains were isolated fromsoil and vegetation taken from 19 sites in theNetherlands (7). Plant materials from which theorganism was isolated included dead and decayedcorn and soybean plants and wild grasses,indicating its preference to exist in nature as asaprophyte. A study of soil and domestic animalfeces has shown that Listeria is more oftenpresent during July to September than othermonths (8). L. monocytogenes and L. innocuawere predominant in feces, whereas L. ivanoviand L. seeligeri were most common in soil.

Vegetation in a rural area in Virginiawhere clinical listeriosis is rare was analyzed

Table. Sources of pathogenic microorganisms on freshfruits and vegetables*

PreharvestFecesSoilIrrigation waterWater used to apply fungicides, insecticidesGreen or inadequately composted manureAir (dust)Wild and domestic animals (including fowl and reptiles)InsectsHuman handling

PostharvestFecesHuman handling (workers, consumers)Harvesting equipmentTransport containers (field to packing shed)Wild and domestic animals (including fowl and reptiles)InsectsAir (dust)Wash and rinse waterSorting, packing, cutting, and further processing equipmentIceTransport vehiclesImproper storage (temperature, physical environment)Improper packaging (includes new packaging technologies)Cross-contamination (other foods in storage, preparation, and display areas)Improper display temperatureImproper handling after wholesale or retail purchase

*Adapted from Beuchat (3)

Figure 1. Per capita consumption of fresh fruits andvegetables in the United States. From USDA,Economic Research Service (2).


Special Issue

for L. monocytogenes (9). Dead soybean plantmaterial and stalks, leaves, and tassels of cornwere collected in April following the previousplanting year. Eight of twelve sampling sitesyielded plant materials positive for L. monocyto-genes. Only 25% of the strains were pathogenicfor mice, a low frequency compared with thepercentage of pathogenic strains isolated fromListeria-positive humans and animals in Virginiaand the United States as a whole. Theseobservations suggest that the predominance ofcertain serotypes of L. monocytogenes may beinfluenced by the environment and that somestrains indigenous to decaying plant vegetationare incapable of causing human illness.

Weiss and Seeliger (10) isolated 154 strains ofL. monocytogenes in Germany from soil andplants, 16 from feces of deer and stag, nine frommoldy fodder and wildlife feeding grounds, andeight from birds. Corn, wheat, oats, barley, andpotato plants and soils from the fields in whichthey were growing were among the materialsanalyzed. Nearly 10% of the corn plants and 13%of the grain plants were infected with L.monocytogenes. Plants from cultivated fields hada lower incidence (12.5%) than plants fromuncultivated fields (44%). Twenty-three percentof samples collected from wildlife feedinggrounds were positive for L. monocytogenes. Itwas suggested that L. monocytogenes is asaprophyte that lives in a plant-soil environmentand could therefore be contracted by humans andanimals through many possible routes frommany sources. Birds and animals are unlikely tobe the only sources responsible for thedistribution of L. monocytogenes in nature and itspresence on fruits and vegetables.

The presence of other pathogenic bacteria,viruses, and parasites in soil likely results largelyfrom application of feces or untreated sewage,either by chance or design. Whatever the case,soil on the surface of fruits and vegetables mayharbor pathogenic microorganisms that remainviable through subsequent handling to the pointof consumption unless effective sanitizingprocedures are administered.

Irrigation and surface run-off waters can besources of pathogenic microorganisms thatcontaminate fruits and vegetables in the field.Irrigation water containing raw sewage orimproperly treated effluents from sewage treat-ment plants may contain hepatitis A, Norwalkviruses, or enteroviruses (poliomyelitis, echovi-

ruses, and Coxsackie viruses) (11). Rotavirusesare known to retain viability on the surface ofvegetables held at 4°C for up to 30 days (12).

Listeria and other potentially pathogenicbacteria have been reported in sewage. Watkinsand Sleath (13) analyzed 52 sewage, river water,and industrial effluents for pathogens. Effluentswere from abattoirs, cattle markets, and poultrypacking plants. L. monocytogenes was isolatedfrom all samples. In many instances, populationsof L. monocytogenes were higher than those forsalmonellae and, in some instances, L. monocyto-genes was isolated when no salmonellae weredetected. Application of sludge containing L.monocytogenes and salmonellae to soil showedthat L. monocytogenes could survive longer.Populations of L. monocytogenes in soil remainedessentially unchanged during 7 weeks afterapplication.

Treatment of sewage does not always yield asewage sludge cake or a final discharge free ofListeria (14). The use of sewage as a fertilizercould contaminate vegetation destined forhuman consumption. MacGowan et al. (8)examined sewage at 2-month intervals in 1991 to1992 and found 84% to 100% contained L.monocytogenes or L. innocua.

Application of sewage sludge or irrigationwater to soil is one avenue through whichparasites can contaminate fruits and vegetables.Ascaris ova sprayed onto tomatoes and lettuceremain viable for up to 1 month, while Endamoebahistolytica could not be recovered 1 week afterspraying (15,16). If sewage irrigation or night soilapplication is stopped 1 month before harvest, theproduce would not likely be vectors for transmissionof diseases caused by these parasites.

Wang and Dunlop (17,18) recovered Salmo-nella, Ascaris ova, and Endamoeba coli cystsfrom more than half of irrigation water samplescontaminated with either raw sewage orprimary-treated, chlorinated effluents. Only oneof 97 samples of vegetables irrigated with thiswater yielded Salmonella, but Ascaris ova wererecovered from two of 34 vegetable samples.Barbier et al. (19) concluded that application ofsewage sludge containing Taenia saginataeggs offers a serious risk for cattle even aftera 3-week no-grazing period.

Feces have been suspected as sources ofpathogens on contaminated fruits, vegetables,or minimally processed produce that havesubsequently been associated or confirmed as


Special Issue

causes of human disease outbreaks (3). Amongthe more recent outbreaks are those linkingunpasteurized apple juice to Escherichia coliO157:H7 infections. This pathogen can remainviable in bovine feces for up to 70 days, dependingon inoculum level and temperature (17).Cryptosporidium infection linked to consumptionof unpasteurized apple juice was hypothesized tohave been caused by contamination of apples bycalf feces (20). Contact of fruits and vegetables bypickers and handlers at the time of harvest alsooffers a mechanism by which pathogens in fecescan contaminate raw produce.

Wild birds are known to disseminate Campy-lobacter (21,22), Salmonella (22,23), Vibriocholerae (24), and Listeria species (25). Morerecently, E. coli O157:H7 has been isolated fromwild bird feces. In a survey of wild birds (mainlygulls), 0.9% of the bacterial isolates from fecalsamples at an urban landfill and 2.9% of bacterialisolates from fecal samples on intertidalsediments were Vero cytotoxin-producing E. coliO157:H7 (26). Pathogenic bacteria are appar-ently picked up as a result of birds feeding ongarbage, sewage, fish, or lands that are grazedwith cattle or have had applications of freshmanure. Control of preharvest contamination offruits and vegetables with pathogenic bacteria bywild birds would be exceptionally difficult.

Postharvest SourcesSome of the possible preharvest sources of

pathogenic microorganisms may also bepostharvest sources (Table). The fecal-oral routeof transmission of pathogens broadens to includeworkers handling fruits and vegetables from thepoint of removal from the plant through all stagesof handling, including preparation at the retailand food service levels and in the home. Changesin eating habits, particularly the increasedconsumption of meals away from home, must beconsidered when attempting to provide reasonsfor increased frequency of outbreaks associatedwith fresh produce. Proper training of food-service workers in hygienic practices is essential.One cannot assume that newly hired personnelhave even rudimentary knowledge of foodmicrobiology. This is particularly critical amongteenagers who, partly because they and theirparents are eating more meals away from home,have had minimal or no exposure to proper food-handling practices. Instruction in elementaryprinciples of food hygiene at the high school or

middle school levels has diminished greatly inthe past two decades.

Traditionally recognized postharvest controlpoints for access of pathogens to whole or cutproduce include transport containers andvehicles and sorting, packing, cutting, andfurther processing equipment. The developmentof new processing equipment and technologiesshould include a team of experts in foodmicrobiology as well as engineering. Too often,aspects of sanitizing equipment are not consid-ered or are an afterthought and can increase therisk for contaminated end products. Temperaturecontrol is absolutely critical at every stage ofpostharvest handling if any success is to beachieved in minimizing the growth of pathogens.

Removal of PathogensSanitizers that can be used to wash or to

assist in lye peeling of fruits and vegetables areregulated by the U.S. Food and Drug Administra-tion in accordance with the Federal Food, Drugand Cosmetic Act as outlined in the Code ofFederal Regulations, Title 21, Ch. 1, Section173.315. As noted by Barmore (27), no chlorinesubstitute effective for washing fruits andvegetables is available. Numerous alternativesfor sanitizing equipment (28) can be used in atotal sanitation program, but none has as broad aspectrum of activity as chlorine.

Chlorine is routinely used as a sanitizer inwash, spray, and flume waters used in the freshfruit and vegetable industry. Antimicrobialactivity depends on the amount of free availablechlorine (as hypochlorous acid) in water thatcomes in contact with microbial cells. The efficacyof chlorine in killing pathogenic microorganismshas been extensively studied. Possible uses inpackinghouses and during washing, cooling, andtransport to control postharvest diseases of wholeproduce have been reviewed by Eckert andOgawa (29). The effect of chlorine concentrationon aerobic microorganisms and fecal coliforms onleafy salad greens was studied by Mazollier (30).Total counts were markedly reduced withincreased concentrations of chlorine up to 50ppm, but a further increase in concentration up to200 ppm did not have an additional substantialeffect. A standard procedure for washing lettuceleaves in tap water was reported to remove 92.4%of the microflora (31). Including 100 ppmavailable free chlorine in wash water reduced thecount by 97.8%. Adjusting the pH from 9 to 4.5 to


Special Issue

5.0 with inorganic and organic acids resulted in a1.5- to 4.0-fold increase in microbicidal effect.Increasing the washing time in hypochloritesolution from 5 to 30 minutes did not decreasemicrobial levels further, whereas extendedwashing in tap water produced a reductioncomparable to hypochlorite. The addition of 100ppm of a surfactant (Tween 80) to a hypochloritewashing solution enhanced lethality but ad-versely affected sensory qualities of lettuce.

Dipping Brussels sprouts into chlorinesolution (200 ppm) for 10 seconds decreased thenumber of viable L. monocytogenes cells byabout 2 log10 CFU/g (32). The maximum log10reduction of L. monocytogenes on shreddedlettuce and cabbage treated with 200 ppmchlorine for 10 minutes was 1.3 to 1.7 log10 CFU/gand 0.9 to 1.2 log10 CFU/g, respectively (12).Numbers decreased only marginally withincreased exposure time from 1 to 10 minutes,which agrees with observations made byBrackett (32) that the action of chlorine againstL. monocytogenes occurs primarily during thefirst 30 seconds of exposure. Nguyen-the andCarlin (33) concluded that the elimination ofL. monocytogenes from the surface of vegetablesby chlorine is unpredictable and limited.

Populations of Salmonella Montevideo on thesurface and in the stem core tissue of tomatoescan be substantially reduced by dipping fruits 2minutes in a solution containing 60 or 110 ppmchlorine, respectively; however, treatment in asolution containing 320 ppm chlorine does notresult in complete inactivation (34). Theineffectiveness of 100 ppm chlorine against S.Montevideo injected into cracks in the skin ofmature green tomatoes was demonstrated by Weiet al. (35). Treatment of alfalfa seeds injectedwith Salmonella Stanley (102 to 103 CFU/g) in 100ppm chlorine solution for 10 minutes has beenreported to cause a substantial reduction inpopulation, and treatment in 290 ppm chlorinesolution resulted in a substantial reductioncompared with treatment with 100 ppm chlorine(36). Initial free chlorine concentrations up to1,000 ppm, however, did not result in furtherreductions. Treatment of seeds containing 101 to102 CFU/g of S. Stanley for 5 minutes in a solutioncontaining 2,040 ppm chlorine reduced thepopulation to less than 1 CFU/g.

We have studied the efficacy of chlorine,hydrogen peroxide, and ethanol in removingSalmonella from injected alfalfa sprouts. Sprouts

were dipped in solutions containing 200, 500, or2,000 ppm chlorine for 2 minutes. The pathogenwas reduced by about 2 log10 CFU/g aftertreatment with 500 ppm chlorine, compared withthe control, and to an undetectablelevel (<1 CFU/g) after treatment with 2,000 ppmchlorine (Figure 2). Chlorine treatment (2,000ppm) of cantaloupe cubes injected with thesame five-serotype cocktail of Salmonellaresulted in less than 1 log10 reduction in viablecells (Figure 2). The very high level of organicmatter in the juice released from cutcantaloupe tissue apparently neutralizes thechlorine before its lethality can be manifested.

As noted by Lund (37), the inaccessibility ofchlorine to microbial cells in crevices, creases,pockets, and natural openings in the skin alsoundoubtedly contributes to the overall lack ofeffectiveness of chlorine in killing pathogens. Thehydrophobic nature of the waxy cuticle on tissuesurfaces protects surface contaminants fromexposure to chlorine and other produce sanitizersthat do not penetrate or dissolve these waxes.Surface-active agents lessen the hydrophobicityof fruit and vegetable skins as well as the surfaces

Figure 2. Efficacy of chlorine, hydrogen peroxide, andethanol in killing Salmonella on alfalfa sprouts andcantaloupe cubes. Bars not noted by the same letterare significantly different (p ≤ 0.05).


Special Issue

of edible leaves, stems, and flowers, but they mayalso cause deterioration of sensory qualities(31,38). Sanitizers that contain a solvent thatwould remove the waxy cuticle layer, and with itenmeshed contaminants, without adverselyaffecting sensory characteristics would holdgreater potential than chlorinated water inreducing microbial populations on whole rawproduce. Such sanitizers may be limited to use onproduce that will be further processed into juiceor cut products, or on whole fruits, vegetables, orplant parts destined for immediate consumption,since their application could adversely affectvisual appearance. Clearly, chlorine, at concen-trations currently permitted for use by theindustry to wash fresh fruits and vegetables,cannot be relied upon to eliminate pathogens.

References 1. Hedberg CW, MacDonald KL, Osterholm MT.

Changing epidemiology of food-borne disease: aMinnesota perspective. Clin Infect Dis 1994;18:671-82.

2. United States Department of Agriculture. Vegetablesand specialties/VGS-269/July. Fruits and tree nuts/FTS-278/October. Washington (DC): USDA EconomicResearch Service; 1996. p. 22,81.

3. Beuchat LR. Pathogenic microorganisms associatedwith fresh produce. Journal of Food Protection1996;59:204-6.

4. Ercolani GL. Occurrence and persistence of culturableclostridial spores on the leaves of horticultural plants.Journal of Applied Microbiology 1997;82:137-40.

5. Beuchat LR. Listeria monocytogenes: incidence invegetables. Food Control 1996;7:223-8.

6. Welshimer HJ. Survival of Listeria monocytogenes insoil. J Bacteriol 1960;80:316-20.

7. Welshimer HJ, Donker-Voet J. Listeria monocytogenesin nature. Appl Environ Microbiol 1971;21:516-9.

8. MacGowan AP, Bowker K, McLauchlin J, Bennett PM,Reeves DS. The occurrence and seasonal changes in theisolation of Listeria spp. in shop bought food stuffs,human feces, sewage and soil from urban sources. Int JFood Microbiol 1994;21:325-34.

9. Welshimer HJ. Isolation of Listeria monocytogenesfrom vegetation. J Bacteriol 1968;95:300-3.

10. Weiss J, Seeliger HPR. Incidence of Listeriamonocytogenes in nature. Appl Microbiol 1975;29:29-32.

11. Bagdasargan GA. Survival of viruses of the enterovirusgroup (poliomyelitis, ECHO, Coxsackie) in soil and onvegetables. Journal of Hygiene, Epidemiology,Microbiology and Immunology 1964;8:497-505.

12. Badaway AS, Gerba CP, Kelly LM. Survival ofrotavirus SA-11 on vegetables. Food Microbiology1985;2:199-205.

13. Watkins J, Sleath KP. Isolation and enumeration ofListeria monocytogenes from sewage sludge and riverwater. J Appl Bacteriol 1981;50:1-9.

14. Al-Ghazali MR, Al-Azawi SK. Detection andenumeration of Listeria monocytogenes in a sewagetreatment plant in Iraq. J Appl Bacteriol1986;60:251-4.

15. Rudolfs W, Falk LL, Ragotzkie RA. Contamination ofvegetables grown in polluted soil. III. Field studieson Ascaris eggs. Sewage and Industrial Waste1951;23:656-60.

16. Rudolfs W, Falk LL, Ragotzkie RA. Contamination ofvegetables grown in polluted soil. II. Field and waterstudies on Endamoeba cysts. Sewage and IndustrialWaste 1951;23:478-85.

17. Wang G, Zhao R, Doyle MP. Fate of enterohemorrhagicEscherichia coli O157:H7 in bovine feces. Appl EnvironMicrobiol 1996;62:2567-70.

18. Dunlop SG, Wang W-LL. Studies on the use of sewageeffluent for irrigation of truck crops. Journal of FoodProtection 1961;24:44-7.

19. Barbier D, Perrine D, Duhamel C, Doublet R, GeorgesP. Parasitic hazard with sewage sludge applied to land.Appl Environ Microbiol 1990;56:1420-2.

20. Millard PG, Gensheimer KF, Addiss DG, Sosin DM,Beckett GA, Houck-Jankoski A, Hudson A. Anoutbreak of cryptosporidiosis from fresh-pressed applecider. JAMA 1994;272:1592-6.

21. Luechtefeld N, Blaser M, Reller L, Wang W. Isolationof Campylobacter fetus subsp. jejuni from migratorywildfowl. J Clin Microbiol 1980;12:406-8.

22. Quessy S, Messier S. Campylobacter spp. and Listeriaspp. in ring-billed gulls (Larus delawarensis). J WildlDis 1992;28:526-1.

23. Jones F, Smith P, Watson DC. Pollution of a watersupply catchment by breeding gulls and the potentialenvironmental health implications. Journal of theInstitute of Water and Engineering Science1978;32:469-82.

24. Lee JV, Basford D, Donovan T, Furniss A, West D. Theincidence of Vibrio cholerae in water, animals and birdsin Kent, England. J Appl Bacteriol 1982;52:281-91.

25. Fenlow DR. Wild birds and silage as reservoirs ofListeria in the agricultural environment. J ApplBacteriol 1985;59:537-44.

26. Wallace JS, Cheasty T, Jones K. Isolation of Verocytotoxin-producing Escherichia coli O157:H7 fromwild birds. J Appl Microbiol 1997;82:399-404.

27. Barmore CR. Chlorine—are there alternatives?Cutting Edge 1995; Spr 1995; 4-5.

28. Cords BR, Dychdala GR. Sanitizers: halogens, surface-active agents and peroxides. In: Davidson PM, BranenAL, editors. Antimicrobials in Foods. 2nd ed. NewYork: Marcel Dekker, Inc.; 1993. p. 469-537.

29. Eckert JW, Ogawa JM. The chemical control ofpostharvest diseases: deciduous fruits, berries,vegetables and root/tubers crops. Annu Rev Phytopathol1988;26:433-63.

30. Mazollier J. Ivè gamme. Lavage-desinfection dessalades. Infros-Crifl 1988;41:19.

31. Adams MR, Hartley AD, Cox LJ. Factors affecting theefficiency of washing procedures used in the productionof prepared salads. Food Microbiology 1989;6:69-77.

32. Brackett RE. Antimicrobial effect of chlorine onListeria monocytogenes. Journal of Food Protection1987;50:999-1003.


Special Issue

33. Nguyen-the C, Carlin F. The microbiology of minimallyprocessed fresh fruits and vegetables. Crit Rev FoodSci Nutr 1994;34:371-401.

34. Zhuang R-Y, Beuchat LR, Angulo FJ. Fate ofSalmonella montevideo on and in raw tomatoes asaffected by temperature and treatment with chlorine.Appl Environ Microbiol 1995;61:2127-31.

35. Wei CI, Huang TS, Kim JM, Lin WF, Tamplin ML,Bartz JA. Growth and survival of Salmonellamontevideo on tomatoes and disinfection withchlorinated water. Journal of Food Protection1995;58:829-36.

36. Jaquette CB, Beuchat LR, Mahon BE. Efficacy ofchlorine and heat treatment in killing Salmonellastanley inoculated onto alfalfa seeds and growth andsurvival of the pathogen during sprouting and storage.Appl Environ Microbiol 1996;62:2212-5.

37. Lund BM. Bacterial spoilage. In: Dennis C, editor.Post-harvest pathology of fruits and vegetables.London: Academic Press; 1983. p. 219.

38. Zhang S, Farber JM. The effects of variousdisinfectants against Listeria monocytogenes on fresh-cut vegetables. Food Microbiology 1996;13:311-21.


Special Issue

Fresh, Preservative-Free Foods ThatPromote Health

Food industry marketers perceive thatconsumers want foods that are convenient; fresh(less-processed and less-packaged); all natural—with no preservatives (a so-called “clean label”);without a perceived negative (i.e., foods withouthigh fat, high salt, and high sugar); and healthy.The industry perception is that consumers wantfoods that not only cause no harm but also remedyailments from heart disease, osteoporosis, andfatigue to memory loss. Categories of foods thatpromote health are fortified foods, performance-enhancing food additives, probiotics, andprebiotics.

Food fortification is an old process. Milk (withvitamins A and D), bread (with iron and niacin),or salt (with iodine) have long been fortified toreplace nutrients thought to be lost duringprocessing. Newer foods fortified with nutrientsneeded by the body to stave off the progression ofdiseases associated with aging or enhancephysical performance attract the consumer’sattention and sell well in today’s marketplace.For example, marketers are promoting all sorts offoods fortified with calcium to women concernedabout osteoporosis. Performance-enhancing foodsare popular. Such foods range from beverages toreplace electrolytes and prolong physical endur-ance to amino acids and fatty acids to improvealertness and memory. Probiotics and prebioticsare two paths to the same result. Research

studies suggest that a desirable intestinalmicroflora causes the host to be less susceptible tointestinal pathogens. Probiotics create thisdesirable state by incorporating the microorgan-ism directly into the food, either as a stableculture or as part of food fermentation. Thisprocess is costly, and the microorganisms often donot survive well in the food. Thus, manufacturersmust add 10 to 100 times the needed number ofmicroorganisms to account for a loss of viabilityduring the product’s normal shelf life. Prebioticsovercomes the limitations of probiotics by addingspecific nutrients, usually a particular carbohy-drate, to the food. When ingested as part of thediet, these specific nutrients “select” for abeneficial microflora in the intestinal tract.

Food Processing and Food ProductDevelopment

The consumer’s quest for health is having agreat impact on the food processor. Comparedwith the marketplace of 25 years ago, today’smarketplace has more perishable products,including fruits and vegetables, and moreinnovative packaging. In addition, consumeraversion to traditional chemical preservativeshas left food processors with less flexibility inchoosing preservation methods. To find atechnologic edge in the marketplace, foodprocessors are exploring new processing andpreservation technologies. Some of these tech-nologies include ohmic heating, high-pressure,pulsed electric field, bright light, and asepticprocessing. Ohmic heating involves passing anelectric current through the food to create heatdue to electrical resistance within the food. With

The Impact of Consumer Demands andTrends on Food Processing

Don L. ZinkNestlé USA, Inc., Glendale, California, USA

Address for correspondence: Don L. Zink, Nestlé USA, Inc.800 N. Brand Blvd., Glendale, CA 91203, USA; fax: 818-549-6908; e-mail: [email protected].

In the United States, consumer demand for new foods and changes in eating habitsand food safety risks are affecting the food processing industry. The population isbecoming older on average; moreover, consumers want fresh and minimally processedfoods without synthetic chemical preservatives. To address the need for safer food andcompete for consumer acceptance, manufacturers are exploring new food processingand preservation methods.


Special Issue

ohmic heating, food particles heat at the samerate as the carrier medium or sauce. Ohmicheating can enhance food quality by limiting heatdamage to the sauce and food particles. High-pressure processing uses very high pressure,often thousands of atmospheres, to pasteurizefoods without heat. This technology is ideal forheat-sensitive foods, but some enzymes aredifficult to inactivate with high-pressure process-ing. Pulsed electric field processing uses a verystrong pulsed electric current to disruptmicrobial cells and pasteurize foods with little orno heating. Bright light processing uses anintense white light to kill bacteria on the surfaceof foods; this light does not penetrate deeply intofoods and can only be used for surfacepasteurization.

Aseptic processing dates back to at least themid-1940s but has yet to realize its full potential.The most widely used of these new technologies,aseptic processing involves sterilizing a foodproduct in a continuous process through a heatexchanger and then filling that food in an asepticfiller. The aseptic filler is a highly specializedpiece of equipment designed to sterilize thepackaging material, fill the sterile product intoits container in a sterile environment, and thenseal the package.

Food processors have also explored novel foodpreservation systems. An ideal food preservativewould come from a natural source and preservefood without being labeled a synthetic chemicalpreservative. Such preservatives include bacte-riocins, dimethyl dicarbonate (Velcorin), com-petitive microbial inhibition, controlled andmodified atmospheres, and irradiation. Bacterio-cins are not new; however, like nisin, they arenow being used to extend shelf life and enhancethe safety of a variety of food products. The use ofbacteriocins is likely to be expanded in the future.Dimethyl dicarbonate, a relatively new preserva-tive used in beverages such as wine, tea, andjuices, is particularly effective in preventingspoilage caused by yeasts. Competitive microbialinhibition relies on the fact that many harmlessbacteria, notably lactic acid bacteria, can inhibitthe growth of both spoilage bacteria andpathogens. Inhibitory strains of lactic acidbacteria can be selected for use in dairy culturesor be added to refrigerated foods to extend shelflife and enhance safety. Modified and controlledatmosphere packaging are already widely usedby the food industry. They have the potential for

even wider use, particularly with fresh fruits andvegetables sold at retail. These methods rely oninhibiting microbial growth by excluding oxygenor by inhibitory concentrations of carbon dioxide.Carefully selected gas mixtures can also delay theripening of certain fruits and vegetables andextend the shelf life of fresh meats. Finally,irradiation, also not a new technology, is poisedfor widespread use to enhance the safety andshelf life of many foods. With proper controls,irradiation could be a valuable means of reducingSalmonella contamination of poultry and Es-cherichia coli O157:H7 contamination of groundbeef.

A Scientist’s View of Consumer TrendsOne of the most obvious consumer trends is a

dramatic increase in the consumption of freshfoods, particularly fruits and vegetables. Thisincrease is the result of the well-publicized valueof a high-fiber diet and betacarotenes as an aid inpreventing colon cancer. The number of mealseaten away from home has increased dramati-cally. The trend toward dining outside the homeis likely rooted in lifestyle changes such ashouseholds with two working parents. Thenumber of home-delivered meals, the ultimateconvenience food, has also increased, eventhough the most popular foods consumed today(pizza and hamburgers) are generally the sameas those of 20 years ago. This indicates that thetypes of foods consumed do not change rapidly,but the way these foods are consumed haschanged. Finally, the population is getting olderon average. Aging may not be a consumer trend,but it has a profound effect on food safetyconsiderations. An older population means amore susceptible population.

New food processing and preservationtechnologies and wider applications of oldertechnologies have, for the most part, had littleimpact on most processed foods. Adoption of newtechnologies will likely continue at a slow pace.Consumers consistently buy foods on the basis ofvalue and taste, not processing technology.Technologies that add value will be the first togain consumer acceptance. The demand forconvenience foods will probably increase. De-mands on our time are increasing, and we haveless time to spend on food preparation, and moremeals will be eaten away from home, in partbecause of convenience but also because of atrend for new tastes and variety in the diet.


Special Issue

Finally, the trend toward foods that claim toenhance performance, rooted in an agingpopulation’s need for better health during longerlife-spans, will continue. With increased demand,

the pressure on the food industry for betterprocessing and preservation methods will alsoincrease and may result in safer food.


Special Issue

Reporting of foodborne and waterbornediseases in the United States began more than 50years ago (1). At that time, state and territorialhealth offices were concerned about the levels ofmorbidity and mortality caused by typhoid feverand infantile diarrhea; cases were to beinvestigated and reported. The underlyingpurpose of reporting was to obtain informationregarding the role of food, milk, and water inoutbreaks of intestinal illness to provide a basisfor public health action.

In 1923, the Public Health Service beganpublishing summaries of outbreaks of gas-trointestinal illness attributed to milk; in 1938, itadded summaries of outbreaks due to any foods.In 1966, the present system of surveillance offoodborne and waterborne diseases began toincorporate into an annual summary all reportsof enteric disease outbreaks attributed tomicrobial or chemical contamination of food orwater. Comprehensive surveillance should resultin greater awareness of the most important food-protection methods.

Between 1983 and 1987, the etiologic agent infoodborne disease outbreaks was not determinedin 62% of the outbreaks (2); between 1988 and1992, the foodborne disease was of unknownetiology in 59% of the outbreaks (1). Bacterialpathogens caused the largest percentage ofoutbreaks (79%) when etiology was known—Salmonella caused 69% of bacterial outbreaks.For each year from 1983 through 1992, the mostcommonly reported food preparation practicethat contributed to foodborne disease concernedimproper holding or storage temperatures. Thesecond most common practice was poor personalhygiene of the food handler. Food from unsafesources was the least commonly reported factorin each of the 10 years of reporting. It is nowtime to examine food handling and determinehow to reverse the trend.

Foodborne disease surveillance has tradi-tionally served three purposes. The first isdisease prevention and control. Prevention andcontrol measures include early identification andremoval of contaminated products from thecommercial market and correction of faulty food-preparation practices in both food-serviceestablishments and the home. Surveillance alsoprovides knowledge of disease causation. Theresponsible pathogen is not identified in more

Impact of Changing Consumer Lifestyleson the Emergence/Reemergence of

Foodborne Pathogens

Janet E. CollinsAmerican Meat Institute, Arlington, Virginia, USA

Address for correspondence: Janet E. Collins, American MeatInstitute 1700 N. Moore Street, Suite 1600, Arlington, VA22209 USA; fax: 703-527-0938; e-mail: [email protected].

Foodborne illness of microbial origin is the most serious food safety problem in theUnited States. The Centers for Disease Control and Prevention reports that 79% ofoutbreaks between 1987 and 1992 were bacterial; improper holding temperature andpoor personal hygiene of food handlers contributed most to disease incidence. Somemicrobes have demonstrated resistance to standard methods of preparation andstorage of foods. Nonetheless, food safety and public health officials attribute a rise inincidence of foodborne illness to changes in demographics and consumer lifestyles thataffect the way food is prepared and stored. Food editors report that fewer than 50% ofconsumers are concerned about food safety. An American Meat Institute (1996) studydetails lifestyle changes affecting food behavior, including an increasing number ofwomen in the workforce, limited commitment to food preparation, and a greater numberof single heads of households. Consumers appear to be more interested in convenienceand saving time than in proper food handling and preparation.


Special Issue

than half of the foodborne disease outbreaks forvarious reasons, including late initiation oflaboratory investigation, inability of availabletechnology to identify the pathogen, and lack ofidentification of the pathogen with a particularfood. Finally, surveillance assists in administra-tive guidance. Information enables assessment oftrends in prevalence of outbreaks caused byspecific etiologic agents and in vehicles oftransmission. This information assists in identi-fying common errors in food handling. In July1995, the Centers for Disease Control andPrevention (CDC), Food and Drug Administra-tion (FDA), and Food Safety and InspectionService (FSIS) began a comprehensive effort totrack major bacterial pathogens that causefoodborne illnesses (3). CDC provides the overallmanagement and coordination with state healthdepartments in the five survey sites of the FSIS/CDC/FDA Sentinel Site Study. The programactively seeks out specific cases of foodborneillness to identify whether a food was of concernand to better establish frequency and source offoodborne disease outbreaks and cases. CDCwill use the data to identify emergingfoodborne pathogens and monitor incidence offoodborne illness; FSIS will use the data toevaluate the effectiveness of new food-safetyprograms and regulations to reduce foodbornepathogens in meat and poultry; FDA will usethe data to evaluate its efforts to reducefoodborne pathogens in seafood, dairy prod-ucts, fruits, and vegetables.

According to a recent report to Congressionalcommittees (4), experts believe that the risk forfoodborne illness is increasing. The food supply ischanging in ways that can promote foodborneillness, and there are no comprehensive data toexplain at what point pathogens are introducedinto food. Further, because of demographicchanges, more people are at a greater risk ofcontracting a foodborne illness.

According to Ollinger-Snyder and Matthews(5), changes in agricultural practices, a growingpopulation susceptible to infectious diseases,lifestyle changes, the emergence of newfoodborne pathogens, and the high turnover ratereported for workers in the food-service industryindicate that new approaches are needed to allayconsumers’ fears and to prevent the spread offoodborne disease in the United States. Theyrecommend implementation of Hazard Analysis

and Critical Control Points (HACCP) systemsand certification of food-service managers.

The food processing industries are develop-ing and implementing HACCP systems; the meatand poultry industries are mandated to do sobeginning in January 1998 (6). Hazard analysishas been defined as the identification of sensitiveingredients, critical processing points, andhuman factors that affect product safety. Criticalcontrol points have been described as processingdeterminants whose loss of control would resultin an unacceptable food-safety risk.

Most contend that the HACCP systemapproach must be implemented at each stage ofthe farm-to-family continuum. Where are thecritical control points and the HACCP systemdevelopment in the home, food-service or retailestablishments, or the car when food is carriedfrom one location to another? The consumer is acomplex and critical control point in the process.

Take the case of the barbecued chickenserved to 260 guests at an outdoor barbecue in1983. Guests were served chicken that wasparboiled in the morning by one set of cooks andthen placed in a large container and refrigerated.The evening cooks assumed the chicken had beenadequately cooked, so they basted it in barbecuesauce and warmed it over the fire. Some 71% ofthe guests got sick from the chicken that wasinsufficiently cooked and improperly held (5).What of the infected bakery worker who stirred avat full of buttercream frosting with a bare handand arm? Some 5,000 cases of viral gastroenteri-tis were caused by the infected worker whoclaimed he had washed his hands. Other morerecent outbreaks (7) appear in Table 1.

Recent data (1) indicate that 80% of reportedfoodborne illness outbreaks occur outside thehome. Even though illnesses would be expected tobe reported more often when they occur as aresult of eating in restaurants, the numbers arelarge. National standards for restaurant safetyare contained in the Food Code (8). FDA has thelegal authority to impose the standards on stateand local jurisdictions. The Food Code, which isupdated every 2 years, includes temperatures forcooking, cooling, refrigeration, reheating, andholding food in food-service establishments.County or city employees are generally chargedwith responsibility for inspecting restaurants;each state or locality has its own laws governingrestaurant safety.


Special Issue

Food service outside the home is big business,with sales of more than $300 billion (9) and nearly10 million employees. The restaurant industry’sshare of the food dollar is 43%, and the typicalconsumer more than 8 years of age had more thanfour meals per week away from home in 1996.Given those statistics, it is clear that food-serviceestablishments play a critical role in food safety.The Center for Science in the Public Interest (7)conducted a survey of 45 agencies across thecountry to determine if state and local agencieswere enforcing 12 key food-safety standards inthe FDA Food Code. The standards chosen for thestudy affect consumer health and safety andinclude such areas as food cooking andrefrigeration temperatures, frequency of inspec-tions, and consumer warnings for raw foods. Notone of the 45 agencies surveyed was following allof the Food Code recommendations.

In the survey, only 13% of agencies enforcedthe Food Code and recommended cookingtemperatures for pork, eggs, fish, and poultry;only 64% of agencies required hamburgers to becooked to 155°F. Recommendations for coolingcooked food were followed by only 20% of theagencies, and only 11% required refrigeration offood at FDA-recommended temperatures.

Every restaurant can take steps to ensure thesafety of the food it prepares and serves to itscustomers. Continuous employee training andinstitution of HACCP-type systems should assistrestaurants and other food-service institutions inimproving their food-safety records. Programsavailable through the national restaurant tradeorganization could assist even the smallestestablishments in achieving food-safety goals.

For more than 25 years, the Food MarketingInstitute (10) has surveyed consumers abouttheir changing needs and priorities in foodattitudes and behavior. The 1996 trends reporthas an expanded focus on the primary grocery storeor supermarket, including questions to helpretailers learn more about take-out foods (Table 2).In 1996, nearly 40% of the 2,000 shopperssurveyed purchased fresh deli items from theirprimary supermarket at least once per week, andmore than 10% reported purchasing ready-to-eattake-out foods as frequently. Three-fourths ofthese shoppers purchased food from the deli atleast once per month, and half bought take-outfood from the supermarket as often.

According to the survey, fast-food restau-rants dominate (48%) all food outlets as theprimary source of take-out food; only 12%

Table 1. Foodborne illness reports from restaurants, 1996Date Description Cause6/96 Salmonella-contaminated food, 38 cases Employees did not wash hands before handling food9/95 Escherichia coli O157:H7 “beef,” 11 cases Raw food cross-contaminated other8/95 Salmonella Newport “chicken,” >850 cases Raw meat on cutting board with vegetables1/95 Hepatitis A, contaminated food, 95 cases Human fecal matter from handling—handwashing8/94 Salmonella, hollandaise sauce, 56 cases Holding temperature too low for 9 hours1/93 Clostridium botulinum, canned Unrefrigerated storage of opened container

cheese sauce, 7 cases, 1 deathSource: Center for Science in the Public Interest, 1996

Table 2. Sources of take-out food (%)Source 86 87 88 89 90 91 92 93 94 95 96Fast-food restaurant 43 44 41 41 46 51 55 46 46 41 48Restaurant 38 33 38 33 27 23 24 27 25 22 25Supermarket 10 9 11 12 14 14 12 15 15 17 12Deli/pizza parlor/bagel * * * * * * * * * 8 4 shop/coffee shop/donut shopGourmet or specialty shop * * * * * * * * * * 3Convenience store * * * * 2 2 2 2 2 2 1Some other place 2 7 3 6 2 1 * 1 1 1 2It varies 1 1 1 3 4 5 3 5 3 2 0Don’t eat out * * * 7 6 4 4 4 4 3 2Not sure 3 3 4 1 * 1 1 1 2 3 2*Data not collected for this year.Source: Food Marketing Institute, 1996


Special Issue

purchase take-out foods from the supermarket. Arecent article in Food Processing Magazine (11)states, “Somewhere on their way to thesupermarket, consumers have been getting lost.”Home-meal replacement, ready-made mealsapproximating what Mom used to make, havebegun to rapidly compete for the food dollars oftime-pressed consumers. According toHollingsworth (12), consumers are eating moremeals at home, but they are not cooking more.Consumers want to get food in a take-out locationand go home to eat it (Figure).

These take-out or eat-at-home foods havebuilt-in food-safety hazards. Consumers aretime-pressed, and they are buying these foods.Are they treating them as perishable? The U.S.Department of Agriculture (13) has expressedconcerns about these foods; they say that take-out foods need to be handled with care. Hot foodsneed to be picked up or received hot and eatenwithin 2 hours. If eaten later, hot foods should bedivided into shallow containers, covered loosely,and refrigerated immediately.

Are consumers ready for all of this foodhandling? Most consumers are confident that thefood they purchase is safe to eat (10). Spoilage offoods is considered the greatest threat to foodsafety by the largest group (49%) of respondents.They count on freshness and expiration dates(22%) and increasingly see bacteria andcontamination as threats (17%). It is interestingto note, however, that between 1992 and 1996these shoppers were less likely (15% vs. 7%) to seespoilage as a threat; similarly, processing andpreparation of foods was less an issue in 1996than in 1992 (8% vs. 10%).

Consumers are concerned about handling offoods by other shoppers and by supermarketemployees. Consumers rely increasingly on foodstores (16%), manufacturers (21%), government(21%), and themselves (25%) for food-safetyprotection. Consumers apparently are willing toshare responsibility for food safety with others,but they want to know that steps are takenduring the processing and distribution of foods toreduce the likelihood of pathogen or otherbacterial contamination.

According to Technomics (14), these super-market issues noted in the Food Market Institutetrends data (10) will be shared with food-serviceoperators as the share of consumer foodexpenditures changes from 51% vs. 49%, 48% vs.52%, and 45% vs. 55% (projected) for retailexpenditures versus food-service expenditures in1991, 1996, and 2001, respectively.

The number of households earning more than$75,000 annually continues to grow, and thesehouseholds exhibit the highest levels ofspending on food service. Consumer demandsare changing the way that food-serviceoperators and suppliers of food services mustreact. The area of convenience, highly prized byconsumers today, has profound implications forfood. Consumers want fast service with easy-to-eat foods and no stress, which means a fargreater emphasis on portable foods.

Technology has conditioned us to demandand receive near-immediate satisfaction. Therewill be even greater emphasis on faster service,meaning more emphasis on convenient foodformats to expedite preparation. Packaging andstorage will greatly affect product quality andsafety. According to Technomics (14) packagingwill need to be temperature-tolerant andbreathable. Preparation and processing tech-nologies will need to have greater ability torapidly cool and chill. And then there is the food-safety concern associated with dispensingequipment. Food will be required to have anextended shelf life. The safety factors associatedwith these new formats will also change.

Consumers want easy access to portablefoods. Accessibility to variety in food optionstranslates to a proliferation in nontraditionallocations. These smaller sites may include back-of-house preparation facilities. This easy accessto smaller operations also suggests a need formore of such operations and more variety in

Figure. Annual meals (including snacks) purchasedat commercial restaurants per person and consumedat home.


Special Issue

menu options. While to the consumer this maytranslate to upscale menus with indulgence foodssuch as new and different bakery items,microbrewery beverages, and gourmet coffees, tothe food-service operator it may mean greatercross-contamination with cream fillings, unpas-teurized fermented drinks, and spoiled milk.New menu options create new challenges forservice and for safety.

According to Steve Harrison, brewmaster ofthe Sierra Nevada Brewing Company (Chico,CA), “The concept that a beer will automaticallygo bad in ‘X’ number of days is a very untrue one.”Consumers do not know that. What is “skunkybeer”? Starting in late 1996, Anheuser-Buschbegan a freshness strategy in their advertising.Other large brewers are catching on, so freshnessis associated with quality and safety. Imaginefreshness dating, “born-on dating,” as a qualityparameter in brewing.

Consumers’ increased emphasis on food-safety issues directly affects food service. Theperceived healthfulness and quality of foodsaffects food sales; the increasing considerationsof cleanliness as healthfulness and quality assafety become even greater shared responsibili-ties as food-service operators take over the roleshistorically associated with home kitchens. “On-the-spot exhibition” cooking is of increasinginterest to today’s consumers.

In June 1996, the Food Marketing Institute(15) published a review of foodborne illness.They note that the organisms that causefoodborne illnesses are found throughoutnature and that mishandling and poorrefrigeration are responsible for most contami-nation. The most common causes are cross-contamination of cooked foods with raw foods,contaminated utensils or serving plates, poorhygiene of food handlers, and time ortemperature abuse.

Agreement is widespread that the mostserious food-safety problem is foodborne illness ofmicrobial origin (Table 3). Foodborne pathogensinclude a wide array of microorganisms, whichhave various physiologic effects on people,ranging from mild to severe, and are associatedwith a wide array of foods. Cross-contaminationand association of foods within mixed dishescomplicate environmental control. Further, someof the microbes have evolved and become moreresistant to food preparation and storagetechniques. Several industry and government

publications (1,2,8,15,16) summarize biologichazards associated with foodborne illness.

Mishandling can occur at any point in thefood chain—in processing, at supermarkets orrestaurants, or in homes. Many food manufactur-ers and retailers have HACCP plans in place, andover the next few years that number willincrease. Consumers, however, must assumeresponsibility for the safety of food in the home.Proper preparation and sanitation methods arekey to preventing foodborne illness in the homeas in other areas of food handling. The messagesfor each of the segments of the food chain are thesame—keep it clean (e.g., wash your hands) andcontrol the temperatures (keep hot things hotand cold things cold) (Table 4).

For the food-service industry, a number ofprograms have been developed to educate food

Table 3. Sources of reported outbreaks with confirmedcauses (%)

Restaurant Other Known Place1983- 1988- 1983- 1988- 1996

87 92 87 92Salmonella 50 60 46 58 30Escherichia <1 <1 2 1 5 coliHepatitis A 6 7 3 4 virusStaphylococcus 2 3 10 5 aureusCampylobacter 1 2 6 3 45Shigella 8 2 6 2 17Sources: Centers for Disease Control and Prevention and U.S.Department of Agriculture for 1996, the 1996 Sentinel SiteStudy

Table 4. Pathogen control in foods to reduce foodborneillnessPathogen Control mechanismCampylobacter Heat foods ≥ 140oF

Proper handlingSalmonella Rapid chilling <40oF

Hot storage >140oFCooking >165oF

Escherichia coli O157:H7 Heat foods >155oFAvoid cross- contamination

Staphylococcus aureus Rapid cooling <40oFPersonal hygiene

Clostridium botulinum Boil food 10-15 minutesClostridium perfringens Refrigerate <40oF

Proper handlingListeria Pasteurization of milk

Adequate cookingSource: Food Marketing Institute, 1996


Special Issue

handlers about food-related and personal behav-iors that affect the safety of foods. For example,the Food Marketing Institute (17) has a FoodProtection Certification Program for supermar-ket personnel to learn about the FDA Food Coderequirements regarding food handling andhygiene. Similarly, the National RestaurantAssociation has developed a food-safety programcalled Serve Safe, intended to educate food-service workers about safe food handling.

Who or what teaches the average consumerabout food safety? Common sense? Family?Health and fitness magazines? In May 1996, theFood Marketing Institute (17) conducted a seriesof consumer focus groups to establish theimportance of food safety to consumers and toidentify barriers to consumers’ safe foodpurchase, handling, and preparation. Theyreport that how consumers manage food safetyreflects years of conditioning, observation, andreinforcement from mothers and grandmothers.In some cases, the more often consumers shop,the less concerned they seem to be about foodsafety when it comes to shopping, storage, andhandling. Consumers link safety to fresh food,and they assume that when they shop more often,they purchase food in smaller quantities and foodsafety is less an issue. Respondents in the studyalso tended to think that cooked food wasgenerally “safer” than raw food. For example,they believed that recontamination ofunrefrigerated food was less a problem withcooked than with raw food.

Some safe food practices are observed forconvenience, esthetics, or taste rather than forfood safety. Thawing meat is messy; covering foodprevents it from drying out; separating foods inthe refrigerator is tidier. These kinds of behaviorimprove safety, but consumers may not under-stand the food-safety implications.

Overall, the consumers in the Food Market-ing Institute study (17) find food-safety messagesgenerally are “common sense,” “basic,” “practi-cal,” and “believable.” Messages about suchsubjects as the order in which to choose foods inthe supermarket, sell-by dates, storage andfreezing of products, ways of keeping hot foodshot and cold foods cold are not considered tooelementary. They also believe that storage timesfor food safety do not apply equally across foodgroups; they do not understand hazards fromvegetables or fruits. Barriers to safe food-handling behavior in this study included

historical (and cultural) practices, feeling ofinvulnerability, taste preferences, timing andplanning, and space and convenience.

A 1992 survey conducted at Cornell Univer-sity and designed to assess consumer food-safetyawareness documented a substantial lack ofknowledge about safe home food preparationpractices. Seventy-five percent of those surveyedknew that Salmonella is associated with meat,poultry, and eggs, but only 65% would refrigeratea roasted chicken breast immediately; 29% wouldleave it on the kitchen counter until it reachedroom temperature. Further, 18% said they wouldnot be concerned or were not sure about the safetyof cooked meat left unrefrigerated for more than4 hours; 14% said the same for cooked poultry.

In April 1996, the American Meat Institute(16) commissioned a study of 1,000 adults in theUnited States. Compared with 98% of respon-dents in the study who know that harmfulbacteria can be present on meat and poultryproducts, only 74% made the link to dairyproducts and eggs; two in five respondents (43%)recognized that fruits and vegetables maycontain harmful bacteria. These conclusionscould be drawn for consumers who responded tothe American Meat Institute (1996) questions.While the U.S. population is growing (up 10%since 1980), households are becoming smaller. Inthe 1980s, the number of households grew 17%,while the average household size decreased from2.8 to 2.6 persons. This shift in family size and theincrease in single heads of households hasresulted in increased stress in the family withless time for shopping and food preparation. Inaddition, more women are in the workforce.Today, 70% of women ages 25 to 44 years are inthe workforce; 75% work full time. Therefore,no adult is likely to be in the home for 70% ofAmerican households, and many children arepreparing food for themselves. Finally, con-sumers spend less time on food preparation.More than 85% of employed women shop andcook, but most spend less than 30 minutespreparing every meal and 20% spend less than15 minutes. Consumers are using conveniencefoods and quick methods of food preparation,including partially cooked foods that mayrequire special handling.

The study results provided further documen-tation that the risk for foodborne illness isincreasing, largely because of societal changesthat affect the way consumers purchase and


Special Issue

prepare food. Contributing to this are changes inthe family structure, more women in the workforce,and less available time for food preparation.Consumers in this study were not able to correctlyseparate home preparation issues from foodservice, nor did they know correct cookingtemperatures to use in their own homes.

The ways in which consumers spreadmicroorganisms to one another and to themselvesinclude more than just coughing and sneezing.Not washing hands before, during, and afterhandling foods clearly contributes to the spreadof foodborne infections and intoxication. Handscan spread disease-causing microbes to foodsfrom other foods and from infected persons.

In a comprehensive review of 91 scientificarticles published after 1986, Bryan et al.(18)attempted to link hand washing and infections.They report that hand washing has become anintegral component of the tradition and ritual ofprevention practice for the spread of infection,but several factors confound the ability toestablish the effectiveness of hand washing forreducing infectious disease. Hand-washing prac-tices were shown to significantly reduceinfections transmitted by the fecal-oral route andin situations of poor personal hygiene. Handwashing is clearly a critical step in reducingpersonal contamination of food and cross-contamination between foods. Hand washing isbut one practice that could dramatically affectrisk, if not incidence, of foodborne disease.

According to data provided by the AmericanSociety for Microbiology (19), people do not washtheir hands as often as they think they do (Table5). In telephone surveys, 94% of respondents

claim they always wash up after using the restroom; however, researchers contend that almostone-third of people do not wash their hands afterusing the bathroom. Of the more than 7,000people nationwide who participated in the study,81% said they wash their hands before handlingor eating food. However, most say they do notwash up after petting an animal (48%), coughingor sneezing (33%), or handling money (22%).

In early 1997 (8), the U.S. Departments ofAgriculture and Health and Human Services andthe U.S. Environmental Protection Agencydeveloped a program intended to coordinate afood-safety initiative among federal agencies,immediately after an announcement by U.S.President Clinton (January 1997) to promote aninitiative designed to improve the safety of thenation’s food supply. The president charged thefederal agencies to work with consumers,producers, industry, states, tribes, universities,and the public to identify ways to improve foodsafety through government and private sectoraction, including public-private partnerships.The interagency response is a multifacetedprogram designed to include surveillance,coordination of activities within the variousprograms and agencies, risk assessment, re-search, inspections, and education. The underly-ing premise upon which this program wasdeveloped is that foodborne infections remain amajor public health problem. Further, sources offood contamination are said to be almost asnumerous and varied as the contaminants;bacteria and other infectious organisms arepervasive in the environment.

The current systems for protecting food in theUnited States include a broad range ofgovernment agencies and industries, many ofwhich have been discussed in this paper.Responsibilities are shared among the U.S.Department of Agriculture (Food Safety andInspection Service and Animal and Plant HealthInspection Service), the U.S. Department ofHealth and Human Services (FDA and CDC), andthe U.S. Environmental Protection Agency. Theseresponsibilities include oversight on the farm, inthe processing facilities, during transportation anddistribution (including food from foreign countries),and in food marketing channels includingrestaurants, supermarkets, and institutional foodservices (such as schools and hospitals).

Surveillance of foodborne illness outbreaksand their causes is a responsibility of FDA and

Table 5. American Society for Microbiology/Bayer hand-washing survey, 1996Behavior/Location What they What they

saya (%) dob (%)Wash hands:After using public restroom 94 68 Women 74 Men 61New York (Penn Station) 60Chicago (Navy Pier) 78New Orleans (casino) 71San Francisco 69 (Golden Gate Park)Atlanta (Braves game) 64

Women 89Men 46

a1,004 adults; b6,330 adultsSource: American Society for Microbiology


Special Issue

CDC. Education is shared among the agenciesand is not the primary concern or responsibility ofany one of the agencies. Pivotal to this newinitiative is the element of education. Specifi-cally, the program is intended to reinvigorateeducation of all those involved in food prepara-tion, focusing on the use of safe practices.According to USDA et al. (8), educating peopleabout steps they must take to prevent and controlfoodborne illness is a vital link in the foodpreparation chain. In spite of the educationefforts of the government, both state and federal,consumer groups, and industry, which haveoccurred historically, foodborne illness occursfrom a lack of knowledge of the risks involved atall stages of food preparation. Choicesconsumers make about how they handle food athome and about eating food that increases therisk for illness can have an important effect onfoodborne disease incidence.

USDA et al. (8) will develop a program toimprove consumer education; retail, food service,and institutional education; veterinary andproducer education; and industry education inthe transportation area. They propose developingan alliance among industry, consumer groups,and governmental agencies to mount a compre-hensive food-safety awareness campaign forconsumers. Highly focused messages and tacticsfor the general public and consumers at high riskwill be developed. This thrust is in perfectharmony with the strategies and tactics proposedby the American Meat Institute (16) as anoutcome to a series of studies and roundtablediscussions held with medical doctors, dieti-tians, educators, and others. The ability ofindustry and consumer groups to work with thegovernment in a program with common themesand elements is critical to the positive outcomeof the effort. As one of the focus group membersin the Food Marketing Institute (17) said, themore often the message is repeated, the morelikely is the listener to hear it.

The broad-based approach to education,which includes data from surveillance andinspections, should provide the foundation forchanges in consumer behavior. It is critical thatconsumers not only take responsibility for theiractions regarding food safety, but that they alsotake seriously the learning that must occur forconsumers of all ages to prevent contamination,cross-contamination, and mishandling of foods at

home and in restaurants. Convenience, taste, andvariety are welcome qualities in foods that weenjoy; safety in foods is critical to the publichealth and safety of consumers and to thegovernment and businesses that support thoseconsumers.

References 1. Centers for Disease Control and Prevention.

Surveillance for foodborne disease outbreaks, UnitedStates, 1988—1992. MMWR CDC Surveill Summ1996;45:1-66.

2. Centers for Disease Control. Foodborne diseaseoutbreaks, 5-year summary, 1983—1987. MMWRCDC Surveill Summ 1990;39:15-57.

3. Food Safety and Inspection Service. FSIS/CDC/FDAsentinel site study: the establishment andimplementation of an active surveillance system forbacterial foodborne diseases in the United States.Washington (DC): USDA Report to Congress; 1997.

4. United States General Accounting Office. Food safety:information on foodborne illnesses. Report toCongressional Committees. Washington (DC): GAO/RCED-96-96; 1996.

5. Ollinger-Snyder P, Matthews ME. Food safety issues:press reports heighten consumer awareness ofmicrobiological safety. Dairy, Food and EnvironmentalSanitation 1994;14:580.

6. FSIS Pathogen Reduction/HACCP. Washington (DC):Federal Register 1996 Jul 6.

7. Center for Science in the Public Interest. Dine at yourown risk: the failure of local agencies to adopt andenforce national food safety standards for restaurants.Washington (DC): The Center; 1996.

8. United States Department of Agriculture, UnitedStates Department of Health and Human Services,United States Environmental Protection Agency. Foodsafety from farm to table: a new strategy for the 21ST

century. Discussion Draft. Washington (DC): UnitedStates Department of Agriculture; 1997.

9. National Restaurant Association. In: 1997 RestaurantIndustry Pocket Factbook. Washington (DC): TheAssociation; 1997.

10. Food Marketing Institute. Trends in the United States:consumer attitudes and the supermarket. Washington(DC): The Institute; 1996.

11. Neff J. Will home meal replacement replace packagedfoods? Food Processing 1996;57:35.

12. Hollingsworth P. The changing role of fast food. FoodTechnology 1997;51:24.

13. United States Department of Agriculture, Food Safetyand Inspection Service. Take out foods—handle withcare. The Food Safety Educator 1996;1:2.

14. Management summary update: critical strategicissues 1996-2001. Chicago: Technomics, Inc.; 1997.

15. Food Marketing Institute. Backgrounder: foodborneillness. Washington (DC): The Institute; 1996.

16. American Meat Institute. Putting the food-handlingissue on the table: the pressing need for food safetyeducation. Washington (DC): American Meat Instituteand Food Marketing Institute; 1996.


Special Issue

17. Food Marketing Institute. Food safety: a qualitativeanalysis. Washington (DC): The Institute; 1996.

18. Bryan JCL, Chorine J, Larson EL. Handwashing: aritual revisited. Infection and Control in Critical Care1995;7:617-24.

19. American Society for Microbiology. Americans getcaught dirty handed. Fact Sheets. Washington (DC):The Society; 1996.


Special Issue

When I was growing up on my parents’ farmin East Texas, we never thought about foodpoisoning or unsafe food. The only foods webought were sugar, salt, flour, and oatmeal;everything else we produced and preserved onthe farm. My mother spent all summer canningfruits and vegetables for winter. We had norefrigeration; we cured our own meat and drankraw milk. But I never heard of botulism, staphpoisoning, or salmonellosis or perfringenspoisoning until I studied bacteriology in college.Only then did I wonder how we survived with norefrigeration in a hot climate. Finally, the answercame to me. We just did not give the bacteria timeenough to develop so they could hurt us.Leftovers from breakfast—hot biscuits, eggs,ham, bacon or sausage, oatmeal, coffee or milk—went right out to the chickens. Lunch leftovers—biscuits, cornbread, vegetables, or fried chicken—were saved for a cold supper 4 or 5 hours later.Any food left went to the pigs. The bacteria hadonly a maximum of 3 or 4 hours to grow, andthat usually is not enough. I survived and wenton to study food microbiology, which includedwhat was known then about food poisoning.The guru of food poisoning in those days wasprofessor Gail M. Dack at the University ofChicago. Dr. Dack was a protégé of Professor

E.O. Jordan, who in 1917 published a 107-pagebook entitled Food Poisoning. Dr. Dack tookover the book and published his first version ofFood Poisoning in 1943. In 1949 and 1956,subsequent editions appeared in which certaintruisms became apparent.

Botulism was considered a problem ofcanners, both home and commercial. Thus,adequate heat processing would seem to solve theproblem. Perhaps it did for the canner, but nowwe know that heating will not eliminate allbotulism. Many foods, including salmon eggs,smoked fish, garlic in oil, vacuum packaged lotusroots, and baked potatoes, can support growthand botulinum toxin formation if the storagetemperature is suitable. Similarly, we thoughtstaphylococcal poisoning was limited to cream-filled pastries and cured ham. In recent years,outbreaks of staphylococcal poisoning have beentraced to cheese, whipped butter, ham salad,fermented sausages, and canned corned beef. Wenow know how to prevent staphylococcal poisoning,but not all food handlers understand and fullycomply with the appropriate control measures.

Salmonellosis was once thought a problemwith meat from infected animals. Now we knowthat a variety of food products can serve asvehicles of this disease. As early as World War II,we found that dried eggs from the United Statescould transmit this disease to our British allies.Thousands of cases of human salmonellosis in theUnited States and other industrialized countries

Historical Overview of Key Issuesin Food Safety

E. M. FosterMadison, Wisconsin, USA

Address for correspondence: E. M. Foster, Professor Emeritus,Food Research Institute, University of Wisconsin, Madison,WI 53706 USA; fax: 608-263-1114.

Foodborne transmission of pathogenic and toxigenic microorganisms has been arecognized hazard for decades. Even half a century ago we knew about the dangers ofbotulism from underprocessed canned foods; staphylococcal poisoning fromunrefrigerated cream-filled pastries, sliced ham, meat, and poultry salads; andsalmonellosis from infected animal products. Despite new protective measures,changes in preservation techniques and failure to follow recognized procedures havecreated new dangers. Moreover, we now recognize new organisms that can causefoodborne illness—Listeria monocytogenes, Escherichia coli O157:H7, Campylobacterjejuni, Vibrio parahaemolyticus, Yersinia enterocolitica, and others. Controlling theseorganisms will require widespread education and possibly new regulatory initiatives.


Special Issue

have been transmitted by ice cream, chocolate,potato salad, cheddar cheese, raw milk, blackpepper, pâté, aspic, ham, pasteurized milk, anddrinking water.

Clostridium perfringens, known since the1940s, causes a problem only when there is grosstemperature abuse of cooked food. Clostridiumbotulinum, Staphylococcus aureus, C. perfringens,and the salmonellae were well known in Dr.Dack’s day, although the food vehicles mighthave changed. Not so well known were many ofthe organisms that preoccupy us today. Forexample, we used to think of Escherichia coli asmerely an indicator organism that suggestedinsanitary handling. Now we know forms of E.coli can kill. Thirty years ago, Listeriamonocytogenes, Campylobacter jejuni, Aeromonashydrophila, Plesiomonas shigelloides, Vibrioparahaemolyticus, and Yersinia enterocoliticawere not known; now these are well-establishedfoodborne pathogens that we must control.

Although not part of a historical overview,other key issues deserve attention during thismeeting. For example, we once thought thatfresh, uncracked eggs were essentially sterileand safe to eat. We did not recognize the ability ofSalmonella Enteritidis to invade the laying henand thereby the yolk of an egg. An outbreak of S.Enteritidis at a Chicago hotel taught us not to

rely on the safety of eggs merely because the shellwas intact. S. Enteritidis in shell eggs is still aserious health problem and a growing concern toegg and poultry producers.

Of equal, if not greater, concern isSalmonella Typhimurium strain DT 104. Widelydistributed in cattle herds of England, Scotland,and Wales, this organism is resistant to severalantibiotics, including ampicillin, chlorampheni-col, streptomycin, sulfonamides, and tetracy-cline. Between 1990 and 1995, the number of S.Typhimurium DT 104 isolated from humans inBritain increased from 259 to 3,837 per year—a15-fold increase. Moreover, the percentage ofdrug-resistant isolates increased from 39% in1990 to 97% in 1995. S. Typhimurium DT 104 hasbeen isolated in the United States from sheep,pigs, horses, goats, emus, cats, dogs, elk, mice,coyotes, ground squirrels, raccoons, chipmunks,and birds. American egg and poultry producersare concerned about its entry into U.S. poultryflocks. S. Typhimurium DT 104 infection inhumans has been associated with the consump-tion of chicken, sausage, and meat paste as wellas with the handling of sick animals. More thanone-third of the patients have required hospital-ization, and 3% have died; these figures are veryunusual for ordinary Salmonella infections andindicate serious problems ahead.


Special Issue

The changing epidemiology of foodbornediseases is a result of complex interactions andchanges in pathogens, foods, food distribution,food consumption, and population immunity (1-3). Predicting the impact of a trend in one part ofthe food continuum presupposes understandingof the whole system. Aspects of the foodprocessing and distribution system can amplifyor attenuate the trend as it grows into a potentialhealth hazard. While a full understanding ofpathogen contamination, infection, and survivalis difficult, a systematic approach to assessing theimpact of the pathogen on health may improve thequality of public health decisions (4,5).

Quantitative risk assessment is a possibleapproach for designing programs to addressemerging foodborne diseases. The use of riskassessment in environmental toxicology illus-trates the potential advantages of applyingquantitative risk assessment in a new field.

Risk Assessment DefinedThe essence of microbial risk assessment is

describing a system in which a microbial hazardreaches its host and causes harm. Risk assessmentconsists of four steps: hazard identification,exposure assessment, dose-response assessment,and risk characterization (6). The knowledge in

each step is combined to represent a cause-and-effect chain from the prevalence and concentra-tion of the pathogen to the probability andmagnitude of health effects. In risk assessment,risk consists of both the probability and impact ofdisease. In this way, risk reduction can beachieved in either dimension—by reducing theprobability of disease or by reducing its severity.

Hazard IdentificationIn hazard identification, an association

between disease and the presence of a pathogenin a food is documented. The information maydescribe conditions under which the pathogensurvives, grows, causes infection, and dies.Epidemiologic and surveillance data, challengetesting, and scientific studies of pathogenicityalso contribute information. Data collectedduring hazard identification are later used inexposure assessment, where the impact ofprocessing, distribution, preparation, and con-sumption of the food are incorporated.

Exposure AssessmentExposure assessment describes the pathways

through which a pathogen population isintroduced, distributed, and challenged in theproduction, distribution, and consumption offood. This step differs from hazard identificationin that it describes a particular food-processingpathway. Depending on the scope of the riskassessment, exposure assessment can begin with

Quantitative Risk Assessment:An Emerging Tool for Emerging

Foodborne Pathogens

Anna M. Lammerding* and Greg M. Paoli†*Health Canada, Guelph, Ontario, Canada; and

†Decisionalysis Risk Consultants, Ottawa, Ontario, Canada

New challenges to the safety of the food supply require new strategies for evaluatingand managing food safety risks. Changes in pathogens, food preparation, distribution,and consumption, and population immunity have the potential to adversely affect humanhealth. Risk assessment offers a framework for predicting the impact of changes andtrends on the provision of safe food. Risk assessment models facilitate the evaluation ofactive or passive changes in how foods are produced, processed, distributed, andconsumed.

Address for correspondence: A. M. Lammerding, HealthCanada, Health of Animals Laboratory, 110 Stone Road West,Guelph, Ontario, Canada N1G 3W4; fax: 519-822-2280; e-mail: [email protected].


Special Issue

pathogen prevalence in raw materials (e.g., a“farm-to-fork” risk assessment), or it can beginwith the description of the pathogen populationat subsequent steps (e.g., as input to a food-processing step). In any case, the intent of riskassessment is to track the pathogen populationand estimate the likelihood of its being ingestedby the consumer. By completing the pathway tothe consumer, we incorporate the importantissues of dose-response assessment.

Dose-Response AssessmentDose-response assessment is used to trans-

late the final exposure to a pathogen populationinto a health response in the population ofconsumers. This step is very difficult because ofthe shortage of data on pathogen-specificresponses and because those responses dependon the immune status of the host (consumer).However, even limited knowledge of the shapeand boundaries of a dose-response function canbe informative in comparing the efficacy ofalternate controls. The differences in responseamong various susceptible populations areimportant features in this step (7).

Risk CharacterizationRisk characterization involves integrating

the information gathered in the previous steps toestimate the risk to a population, or in somecases, to a particular type of consumer. In thisstep, by modifying the assumptions in theparameters of previous steps, we can study theeffects of these alternate assumptions onultimate health risk. Assumptions can bechanged to study the impact of lack of knowledgeand the potential gains through further researchor to suggest the impact of a suspected trend. Forthis type of analysis, risk assessments aretypically done in a computer environment to easethe computational burden and provide rapidresponses to “what-if” questions using alternateassumptions and situations. Current spread-sheet applications and available “add-ins” allowgeneration of complicated probabilistic modelsthat had previously only been available throughexpensive custom software.

Risk Assessment in EnvironmentalToxicology

In environmental toxicology, quantitativerisk assessment has emerged as the predominantparadigm for describing the public health

consequences of human exposure to environmen-tal contaminants (8). Within this paradigm,existing situations are measured and comparedaccording to a measure of population health risk.Similarly, proposed interventions are comparedaccording to the reduction in population healthrisk that each intervention confers.

The adoption of risk assessment wasprimarily a result of legal and administrativechallenges to regulatory authority during the1970s (6). Regulatory agencies were required toprovide a clear connection between an imposedregulation and an expected health benefit. If theexpected health benefit could be quantified, theregulatory agencies were required to demon-strate that it was substantive. Quantitative riskassessment has since become widespread fordifferent reasons. It is now used proactively tosupport decisions such as selection of wastetreatment technologies, contaminated site cleanupoperations, and state and municipal prioritysetting for public health initiatives.

The shift of environmental health issues intoa framework of risk reduction opened the field toa broader set of analytic tools and prompted abroader spectrum of professionals to examine thecomplex problems in the field. Scientific societieshave emerged with the sole mission of focusing onthe general techniques of risk assessment andtheir role in public health decisions. In addition,environmental health risks can be compared withconcurrent public health risks from other sourcesthrough the use of common measures. While thistype of comparison is not always performed,scrutiny of the cost-effectiveness of variousregulatory programs is increasingly required onthe basis of risk reduction. Microbial food safety,as a relative latecomer to the field of riskassessment, can take advantage of its successesand failures and the wealth of constructivecriticisms of frameworks, decisions, and methodsfor addressing pervasive uncertainties (4,8).

Opportunity for Technology Transfer toMicrobial Health Risks

Quantitative risk assessment is an emergingtool in the field of microbial food and water safety(9-12). Recognizing the deficiencies of currentapproaches to evaluating the risk for humanillness from pathogens in food, the Council forAgricultural Science and Technology recom-mended that risk assessment provide the basisfor establishing food safety priorities and policies


Special Issue

(5). Because of recent initiatives advocating thewidespread implementation of Hazard Analysisand Critical Control Points (HACCP) systems,quantitative risk assessment has been proposedas a means of providing health-outcome–basedspecification of microbial criteria for HACCPplans (12-14). Concurrently, international tradeagreements have advocated that demonstrationof increased domestic health risk (in a riskassessment) is the only acceptable basis forbarriers to international trade in food (15-18).However, one of the most important benefits inthe adoption of quantative risk assessment isimproved understanding of the many factors thatdetermine the safety of the food supply.

Some resistance to the adoption of riskassessment is likely. Good manufacturingpractices and standard operating procedurescarry a long history of reasonably safe productionwhen properly applied. The return on investmentin producing a quantitative risk assessment maynot be high for an individual food company with avery conservative production process. However,good manufacturing practices and outbreak dataare not particularly useful in predicting theimpact of new products, newly recognizedpathogens, and changes in food processing or incomparing international food systems. Whetherchanges in the food supply are planned (as inrefocused inspection systems and minimallyprocessed foods) or are occurring passively (as inchanged pathogens, demographics, and con-sumer behavior), tools are required to assemblethe information that describes the impact.Quantitative risk assessment may provide theonly systematic means to interpret the impact ofchanges or trends before they become a source ofepidemiologic data.

In a quantitative risk assessment of broadscope, there is a place for all the data from diverseinformation gathering activities relevant tomicrobial food safety. Recent analyses ofpasteurized liquid egg (19) and ground beefcontamination (20) incorporated evidence fromfarm-based studies of pathogen prevalence,technology assessments comparing decontami-nation methods, process-specific parameters oflot size and raw material mixing, growth anddeath models from predictive microbiology,monitoring studies of transportation and retailtemperature control, and studies of consumptionamounts and cooking preference.

By designing the quantitative risk assess-ment process as an intelligent information bank,we can develop a model to accommodate thebreadth of available information. The modelprovides a focus for discussions among workersfrom diverse disciplines: farmers, veterinarians,food-processing experts, microbiologists, andconsumer behavior experts. The model alsoallows for consideration and comparison of controlstrategies for which experimentation would bevery difficult in a “live” environment. The impact,for example, of an aging population or a shift incooking practices can be simulated by a variety ofassumptions that reflect the extent of the change.By placing all of the information together, we candelineate gaps in knowledge and provideestimates of the benefits of proposed research.

The most obvious users for quantitative riskassessment as applied to microbial food safety areagencies responsible for food inspection, diseasesurveillance, and food standards. These agencieshave the most to gain from models thatincorporate existing and new data, captureknowledge of the relevant features of the foodprocessing and distribution continuum, andcapture knowledge of the variability in consumerbehavior and immune system responses. Ifmodels are constantly updated and improved,decisions made to research, monitor, and controlfoodborne pathogens can be made with informa-tion that lends itself to multidisciplinarydiscussion and best describes what is currentlyknown and unknown. Without such a model,there is little common ground for the type ofcollaboration often advocated for addressing theinherent complexity of foodborne disease.

Risk Assessment Case ExamplesTwo case examples illustrate the prospects of

using risk assessment to support decisionsregarding emerging foodborne diseases.

Escherichia coli O157:H7 in Ground BeefA model of E. coli O157:H7 in ground beef has

been developed to support comparative assess-ment of control strategies (20). The modeldescribes the pathogen population from theproduction of ground beef (including carcassprocessing) to consumer cooking and consump-tion. The variability and uncertainty in the modelare accommodated through the use of probabilis-tic representations for many of the parameters.


Special Issue

To generate a representative distribution of risk,the model is simulated many times with differentvalues selected from the probability distribu-tions. This is a technique known as Monte Carlosimulation (20-22).

While the direct output of the model is adistribution of health risk from eating groundbeef hamburger patties, a more important use ofthe model is to describe the changes in health riskassociated with changes in various parameters.By changing parameters describing, for example,pathogen prevalence and concentration in rawmaterial, temperature abuse in transportationand retail, consumer cooking preference, infec-tious dose, and size of susceptible populations, wecan study the impact of trends in disease riskfactors. Because this model includes the farm-to-fork continuum, it is possible to assess theefficacy of interventions that would otherwise notbe compared in the same analysis. In addition,the importance of improved data at differentpoints in the process can be estimated.

ToxoplasmosisA probabilistic model describing the inci-

dence of toxoplasmosis was generated (23). Whilethis model did not begin at the raw material level,valuable insights were gained in studying theimpact of trends in exposure to Toxoplasmagondii. In congenital toxoplasmosis, the impact ofmaternal exposure to T. gondii depends onwhether the mother has previously been infected(24). If this is the first exposure, the impactfurther depends on the trimester of pregnancy. Ifdetected at an early stage and treated withcertain drug therapies, the infection may have asmaller impact.

With such a model, the impact of varying riskfactors can be studied. Since the most seriousconsequences of toxoplasmosis occur duringpregnancy, a key variable is seroprevalence as afunction of age. The protection offered by priorinfection complicates disease therapy; a reduc-tion in exposure to T. gondii could increaseincidence of congenital toxoplasmosis by reduc-ing the prevalence of immune women ofchildbearing age. This may be further compli-cated by changes in the age profile of pregnancysince younger women are less likely to have beenexposed. In addition to the complexities of thepopulation immunity profile, various trends inrisk factors can be simulated, such as trends incat ownership, consumption of implicated

products, and the age distribution of pregnancy.The emergence of toxoplasmosis as one of theleading causes of death in the human immunode-ficiency virus–positive population can be studiedconcurrently. The effectiveness of mitigationstrategies (e.g., education and screening pro-grams designed for pregnant women) can becompared to food-processing strategies intendedto reduce overall exposure.

The model of T. gondii infection providesinsight into the importance of detailed hazardidentification to understand the complex mecha-nisms of disease, exposure modeling to under-stand the time-dependent nature of exposure,and intervention modeling to understand thepotential negative consequences of a reduction inoverall exposure. Moreover, the results underlinethe importance of performing all of the above tasksin the same overall exercise if the implications oftrends and interventions are to be fully understood.It is unlikely that a sound decision could be madewithout a full microbial risk assessment involvingmodeling of the complex nature of populationimmunity and exposure.

ConclusionsOne of the key benefits of quantitative risk

assessment is the development of modelsdescribing the complex nature of pathogenpopulations in the food supply. Improvedunderstanding of the efficacy of pathogenreduction is the most important side effect of thisapproach. Studies assessing the health impact ofa foodborne pathogen often include extensivedocumentation of pathogen levels at unconnectedpoints in the food and consumer pathway. Incontrast, a microbial risk assessment based on amodel provides a repository of knowledgedescribing health risk outcomes and controlstrategies. The model improves with each newrelated study and each critical review as more andmore relevant data are uncovered. Furthermore,when a decision is required, a description of thesystem is already available in which assumptionsand proposed interventions can be tested.

Initially, models can be expected to be crude.However, as a base for discussion, a model can bevery effective at soliciting input from experts inthe food industry and the public healthcommunity. Input from epidemiologists, microbi-ologists, and industry safety managers can bemerged into the model until it represents the bestavailable understanding of the interacting


Special Issue

features of the food supply and their effect on thedistribution of health risk. Once the model hasbeen developed, the impact of various controlstrategies and trends can be simulated. Our currentinability to compare control strategies at differentpoints of the food supply chain is evidence of theneed for a system-level understanding that willimprove decision-making capacity.

Decisions to address foodborne pathogenscannot wait for scientific certainty. Large degreesof uncertainty require that decisions be madewith great caution; however, there is no excusefor not making the best decision on the basis ofavailable information. Model-based quantitativerisk assessment can provide the decision-maker additional insights not typically evidentin “piece-meal” considerations of data. Theability to represent the essentially probabilisticnature of emerging foodborne disease isanother risk assessment attribute not typicallyachieved by traditional approaches.

Many gains in decision support can beachieved through model-based risk assess-ment. Given that many current concerns arefocused on emerging pathogens, it may betimely to adopt risk assessment as a tool that iswell equipped for studying changes andinterventions in the race against pathogens.

References 1. Altekruse SF, Cohen ML, Swerdlow DL. Emerging

foodborne diseases. Emerg Infect Dis 1997;3:285-93. 2. Hedberg CW, MacDonald KL, Osterholm MT.

Changing epidemiology of food-borne disease: aMinnesota perspective. Clin Infect Dis 1994;18:671-82.

3. Smith JL, Fratamico PM. Factors involved in theemergence and persistence of food-borne diseases.Journal of Food Protection 1995;58:696-708.

4. Rodricks JV. Risk assessment, the environment, andpublic health. Environ Health Perspect 1994;102:258-64.

5. Foegeding PM, Roberts T, Bennett JM, Bryan FL,Cliver DO, Doyle MP, et al. Foodborne pathogens: risksand consequences. Ames (IA): Council for AgriculturalScience and Technology (CAST) 1994; Task ForceReport No.:122.

6. National Research Council. Risk assessment in thefederal government: managing the process. Washington(DC): National Academy Press; 1983.

7. Gerba CP, Rose JB, Haas CN. Sensitive populations:who is at the greatest risk? Int J Food Microbiol1996;30:113-23.

8. Commission on Risk Assessment and Risk Management.Framework for environmental health risk management.Washington (DC): The Commission;1997.

9. Jaykus L-A. The application of quantitative riskassessment to microbial food safety risks. Crit RevMicrobiol 1996;22:279-93.

10. Rose JB, Sobsey MD. Quantitative risk assessment forviral contamination of shellfish and coastal waters.Journal of Food Protection 1993;56:1043-50.

11. ILSI Risk Science Institute Pathogen Risk AssessmentWorking Group. A conceptual framework to assess therisks of human disease following exposure topathogens. Risk Anal 1996;16;841-8.

12. Rose JB, Haas CN, Gerba CP. Linking microbialcriteria for foods with quantitative risk assessment. In:Sheridan JJ, Buchanan RL, Montville TJ, editors.HACCP: an integrated approach to assuring themicrobiological safety of meat and poultry. Trumbull(CT): Food Nutrition Press;1996.p.159-70.

13. Notermans S, Gallhoff G, Zwietering MH, Mead GC.The HACCP concept: specification of criteria usingquantitative risk assessment. Food Microbiol1995;12:81-90.

14. Buchanan RL. The role of microbiological criteria and riskassessment in HACCP. Food Microbiol 1995;12:421-4.

15. World Health Organization. Application of riskanalysis to food standards issues. Report of the JointFAO/WHO Expert Consultation 1995.WHO/FNU/FOS/Report No.95.3.

16. Hathaway S. Harmonization requirements underHACCP-based control systems. Food Control1995;6:267-76.

17. Dawson RJ. The role of the Codex AlimentariusCommission in setting food standards and the SPSagreement implementation. Food Control 1995;6:261-5.

18. Hathaway SC, Cook RL. A regulatory perspective on thepotential uses of microbial risk assessment ininternational trade. Int J Food Microbiol 1997;36:127-33.

19. Whiting RC, Buchanan RL. Development of aquantitative risk assessment model for Salmonellaenteritidis in pasteurized eggs. Int J Food Microbiol1997;36:111-26.

20. Cassin MH, Lammerding AM, Todd ECD, Ross W,McColl S. Quantitative risk assessment of Escherichiacoli O157:H7 in ground beef hamburgers. Int J FoodMicrobiol. In press 1997.

21. Thompson KM, Burmaster DE, Crouch EAC. MonteCarlo techniques for quantitative uncertainty analysisin public health risk assessments. Risk Anal1992;12:53-63.

22. Vose D. Quantitative risk analysis: a guide to MonteCarlo simulation modelling. New York: John Wiley &Sons, Inc.;1996.

23. Cassin MH, Lammerding AM, Paoli GM, McColl RS.Population response and immunity to Toxoplasmagondii [abstract]. In: Proceedings of the AnnualConference of the Society for Risk Analysis andInternational Society for Exposure Analysis; 1996 Dec8-12; New Orleans, LA. MacLean (VA):Society for RiskAnalysis. p. 129.

24. Dixon BR. Prevalence and control of toxoplasmosis - aCanadian perspective. Food Control 1992;3:68-75.


Special Issue

The guardians of the world’s food supply facea communication challenge of extraordinarycomplexity. They need to be ready at short noticeto deal with various crises, often involvingbaffling combinations of foods, pathogens, handlingand distribution practices, dietary norms, andinteractions with medical conditions and medica-tions. Their response to this challenge may haveimportant health, economic, and even politicalimplications. Conflicting pressures may comefrom groups that bear the cost when the publichealth response is too swift or too slow. Quick,confident explanations are expected after out-breaks that may never be fully understood. Whenconsumers (and producers) need information,they cannot wait for more research. Consumerscan read between the lines, especially when theyperceive their lives or livelihoods at risk. If theymisread messages, the communicators may stillbe held responsible. Moreover, consumers knowthat silence is also a form of communication.

At the same time, the guardians of the foodsupply must wage a continuing struggle toimprove the handling of food. In the UnitedStates, campaigns are under way for cooking beefmore thoroughly, separating raw meat fromsalad ingredients, and improving the sanitationof food handlers (e.g., Operation Clean Hands).To some extent, these campaigns are theincarnations of old messages that have not been

communicated effectively. At the same time, thecampaigns are responses to changes in the foodsupply that have increased the risks associatedwith conventional practices. For example, as theincidence or severity of foodborne diseasepathogens increases, the effectiveness of custom-ary food-handling practices decreases.

This article briefly reviews risk perceptionand communication research as a possibleresource for better understanding (and perhapsmeeting) the public’s needs (1-3). Communicationresearch provides a set of general tools andtheories, as well as a body of results, showing acomplex picture of strengths and weaknesses inlay understanding of risk. We explore here theimplications for anticipating public response toemerging foodborne pathogens and offer aproposal for how an effective communicationcampaign might be organized.

Although risk communication research doesnot directly address emerging foodborne patho-gens, it is compatible with the model of riskassessment that the food industry seems to beadopting (4). Drawn from the National ResearchCouncil’s (5) volume, Improving Risk Communi-cation, the model involves overlapping processesof assessing the magnitude of risks (throughanalytical procedures), managing their level(through practical measures), and communicat-ing with the public about them.

Like many other risks, emerging foodbornepathogens are of primary concern to somespecialists but one more thing to worry about forordinary citizens. The thought processes that

Communicating Foodborne Disease Risk

Baruch Fischhoff and Julie S. DownsCarnegie Mellon University, Pittsburgh, Pennsylvania, USA

Address for correspondence: Baruch Fischhoff, Department ofSocial and Decision Sciences, Department of Engineering andPublic Policy, Carnegie Mellon University, Pittsburgh, PA15213, USA; fax: 412-268-6938; e-mail: [email protected].

The food industry, like many others, has a risk communication problem. Thatproblem is manifested in the public’s desire to know the truth about outbreaks offoodborne diseases; ongoing concern about the safety of foods, additives, and food-processing procedures; and continued apathy regarding aspects of routine foodhygiene. If these concerns are addressed in a coherent and trustworthy way, the publicwill have better and cheaper food. However, sloppy risk communication can itself causepublic health damage. Because citizens are ill-equipped to discriminate amonginformation sources, the food industry as a whole bears responsibility for the successesand failures of its individual members. We review risk communication research andpractice for their application to the food industry.


Special Issue

people rely on for making decisions are the focusof much research (Table; 6).

How much does the public know andunderstand? The answer to this question

depends on the risks consumers face and theopportunities they have to learn about them.The next section discusses strategies forimproving those opportunities.

Communicating RiskAn overarching theme of risk communication

is that people understand risks that draw theirattention and are presented comprehensibly.Whether the public’s attention is arousedspontaneously or as a result of a message, theopportunity must be seized. The right informa-tion must be selected and communicatedappropriately (1,16,17).

The hallmarks of effective communicationshould be used. Match the audience’s level oftechnical sophistication. Do not talk down.Clarify terms (e.g., virus) that are used ineveryday speech but not very precisely (e.g., risk).Organize information. Provide the audience witha quick logical overview. Make the desired level ofdetail easy to read. Use numbers to communicatequantities. Avoid ambiguous verbal quantifiers,such as “rare” or “likely”. Ensure sourcecredibility. Realize that messengers are a part ofthe message and essential to its interpretation.Use knowledgeable sources that will notmisrepresent the message. Avoid risk compari-sons with rhetorical implications. Comparing oneuncontrollable accident risk with another, morefamiliar one (e.g., half as likely as being injuredby lightning) can be useful; however, peopledislike comparisons that imply they shouldaccept one risk because they accept another, e.g.,comparing the risks of nuclear power with thoseof eating peanut butter (from aflatoxin).

However useful communication researchmay be, there is no substitute for empiricaltesting of messages. With heterogeneous audi-ences, any fixed message will work better forsome people than for others. In such cases,universal understanding may require provid-ing the opportunity for the public to askquestions through public information sessions,agricultural extension services, science teach-ers, or toll-free numbers.

What To SayThe effort to communicate is wasted if the

information is not worth communicating, eitherbecause people already know it or because itmakes no difference to them. Indeed, communica-tion can backfire if consumers think that their

Table. Thought processes involved in decision-making

People simplify. Many decisions require people to dealwith more details than they can readily handle at anyone time. To cope with the overload, people simplify.People want to know if foods are “safe,” rather thantreating safety as a continuous variable; they demandproof from scientists who can provide only tentativefindings; and they divide the participants in riskdisputes into good guys and bad guys. Suchsimplifications help people cope, yet also lead topredictable biases (7).

Once people’s minds are made up, it’s hard to changethem. People are adept at maintaining faith in theirbeliefs unless confronted with overwhelming evidenceto the contrary. One psychologic process that helpspeople to maintain their current beliefs is underesti-mating the need to seek contrary evidence. Anotherprocess is exploiting the uncertainty surroundingnegative information to interpret it as consistent withexisting beliefs (8).

People remember what they see. People are good atkeeping track of events that come to their attention(9,10). As a result, if the appropriate facts reach peoplein a credible way before their minds are made up, theirfirst impression is likely to be the correct one.Unfortunately, it is hard for people to gain firsthandknowledge of many risks, leaving them to decipher theincomplete reports they get.

People cannot readily detect omissions in the evidencethey receive. It is unusual both to realize that one’sobservations may be biased and to undo the effects ofsuch biases. Thus people’s risk perceptions can bemanipulated in the short run by selective presenta-tions. People will not know and may not sense howmuch has been left out (11). What happens in the longrun depends on whether the missing information isrevealed by other experiences or sources.

People may disagree more about what “risk” is thanabout how large it is. One obstacle to determining whatpeople know about specific risks is disagreement aboutthe definition of “risk” (12-15). For some risk experts,the natural unit of risk is an increase in probability ofdeath; for others, it is reduced life expectancy; for stillothers, it is the probability of death per unit ofexposure. If lay people and risk managers use the term“risk” differently, they may agree on the facts of ahazard, but disagree about its riskiness.Abridged from Fischhoff & Svenson (15).


Special Issue

time is being wasted with useless messages whilethey are being denied pertinent information. Ananalytical effort to determine what is worthknowing and a coordinated empirical effort todetermine what people know already arerequired. These efforts take different forms insituations where consumers face well-formulateddecisions and need only a few quantitativeestimates before making choices, and in situationswhere consumers are trying to understand theprocesses creating and controlling a risk, in order tofollow public discussion, devise decision options, orunderstand quantitative estimates.

Identifying Relevant EstimatesThe tools of decision analysis provide ways to

determine how sensitive well-structured choicesare to uncertainty in different decision param-eters (18,19). The more sensitive parametersshould receive more attention, unless consumersknow them already (and need no reminder). Ifconditions do not permit sensitivity analyses forindividual decision makers, one can model theinformation needs of a population similar to theintended audience. Merz et al. (20) demonstratedthis approach for communicating to carotidendarterectomy candidates. Scraping out themain artery to the brain reduces the probabilityof stroke for patients with arteriosclerosis.However, the procedure can cause manyproblems, including strokes. Decision analysiscomputed the attractiveness of surgery for ahypothetical population of patients, with adistribution of physical states (e.g., stroke risks)and values (e.g., time horizons). The analysisfound that three of the potential complications(stroke, facial paralysis, and persistent head-aches) posed sufficient risk that learning aboutthem should dissuade about 30% of candidatesfrom surgery. Learning about the other sideeffects should affect few additional patients.Therefore, physicians trying to secure informedconsent should (while not hiding other informa-tion) make sure that patients understand therisks of these three complications.

Identifying Relevant ProcessesRisk analysis provides one way to identify the

critical processes in creating and controllingrisks. Figure 1 shows a simple model for the risksof foodborne pathogens. It uses the formalism ofthe influence diagram (21,22). Such a model canbe used both to assess risks and to characterize

the comprehensiveness of lay understanding. Inthis model, people incur food-related risks as aresult of decisions, which possibly lead to actionsor exposures. These decisions concern suchactions as eating a bite of suspicious food,choosing a particular diet, or opting for school (orhome) lunch. Those decisions depend, in part, onthe perceived risks of those actions as well asother nonrisk factors (i.e., other costs andbenefits). Exposure may follow, if a pathogen isactually present; it can lead, in turn, totransmission of the pathogen and to changedhealth states, depending on the resistance todisease that the person’s health provides.

Figure 2 elaborates on this model. It showsthat food pathogenicity depends on both theprevalence of pathogens in the environment andthe quality of food handling. A person’s ownhealth influences risk perceptions through theintermediate variable of perceptions of health,which in turn is influenced by the person’s historyof food consumption (or avoidance). Actualpathogenicity influences risk perceptions throughawareness, a variable that communicators mightaffect. Nonrisk factors include visceral factors(e.g., hunger), external social factors (e.g., socialpressure to eat any food offered by a host), norms(e.g., not eating dog), and the expected benefits ofconsumption (e.g., taste, texture, and othergustatorial pleasures). Computing risks withthis model would require specifying eachvariable and estimating the contingencies byusing statistical sources or expert judgment.For risk communication purposes, even aqualitative model can define the universe ofdiscourse and allow approximate estimates ofthe most important relationships (23).

Identifying Current KnowledgeDetermining what people already know

about quantitative estimates is relativelystraightforward, although there are variouspitfalls (24,25). Eliciting knowledge of processesis more difficult. Respondents should be given thefocus of the problem and maximum freedom toexpress their ideas and reveal which of theprocesses in Figure 2 are on their minds. Studiesusing open-ended techniques often find thatpeople speak the language of risk withoutunderstanding its terms. For example, in a studyabout radon, we found that respondents oftenknew that it was a colorless, odorless, radioactivegas that caused lung cancer. However, when


Special Issue

Figure 1. Diagram of the general model.

pressed, respondents often revealed inappropri-ate notions of radioactivity, believing thatanything radioactive would permanently con-taminate their homes. Some told us that theywould not test for radon because there wasnothing that they could do if they found a problem(26). In studies with adolescents, we found otherforms of false fluency; for example, teens usedterms such as “safe sex” and “clean needles”without understanding them (23). Without open-ended probing, we miss misconceptions that atechnical expert never would have imagined, orwe use language that does not communicateeffectively with our audience (27).

Food Industry Communication StrategiesTechnical experts (in any industry) generally

want to get the facts in order before sayinganything. Although that is an appropriate normwithin the scientific community, refusing toaddress a concerned public can evoke mistrustand anger, as can failing to arouse an apatheticpublic. To steer an appropriate course, communi-cators need an explicit policy that balances the

risks of saying too much with the risks of sayingtoo little. The policy must consider both what tosay and when to say it. From a decision theoryperspective, citizens need information critical toidentifying actions that will help them achievepersonal goals. As a result, any recommenda-tions should reflect both scientific knowledgeand citizens’ values. That is, consumers need toknow what is the best gamble, given the trade-offs between, for example, the risks of throwingout good food and the risks of eating food thatmight make them sick.

At times, there may be a temptation not to tellit like it is. For example, one might argue thatrisks should be exaggerated when a frank reportwould leave people unduly apathetic, as judgedby their own standards. That is, people shouldsay “Thanks for getting my attention” once thegrounds for the overstatement were made clear.Such gratitude requires a public that not onlyrecognizes the limits of its own understanding,but also accepts paternalistic and manipulativeauthorities. That acceptance seems morelikely for misrepresentations intended to get a


Special Issue

Figure 2. Diagram of the foodborne model.

complacent public moving than for ones intendedto allay a hysterical public’s fears. In either case,once the secret is out, all future communicationsmay be subjected to second guessing (“Howseriously should we take them this time?”).

One situation in which paternalistic author-ity is needed arises when a single message mustbe sent to a heterogeneous audience—forexample, when officials must decide whether todeclare a particular food “safe.” Safety is acontinuous variable, and any cut-off represents avalue judgment. For any given food, differentgroups may face different risks, derive differentbenefits, and want to make different trade-offs. Forexample, a few people are strongly allergic to sulfurdioxide as a dried food preservative. Marketingsuch foods signals their safety to all. Labels thatdeclare preservatives in foods allow consumers tocustomize their risk levels, but only if they knowtheir own risks (i.e., whether they are stronglyallergic, which they may learn only through a badreaction whose source they identify).

The Food and Drug Administration faces asimilar challenge in its effort to standardize risklabels for over-the-counter drugs. For example,other things being equal, producing bilinguallabels will require either reducing print size oromitting information about some side effects.These modifications would, in turn, increase therisks for consumers with limited vision (e.g.,some of the elderly) or those particularlysensitive to the omitted effects. Whateverlabeling, warning, or communication strategy ischosen will leave some residual risk, with anuneven distribution depending on the heteroge-neous sensitivities of the audience. Thus, thestrategy reflects the authorities’ notion of the“acceptable level of misunderstanding” (28).

What that acceptable level should be is apolitical and ethical question, which could beresolved by properly constituted public or privategroups, and a scientific question, partiallyresolvable by research of the sort described here.Rigorous empirical testing is needed to deter-


Special Issue

mine whether communications fulfill the hopesplaced in them (27). Emerging foodbornepathogens provide a particular challenge tosafety communications—and a particular needfor evaluation. Their novelty and ability to produceoutbreaks in diverse places in the world and thefood chain encourage treating them as unique. If acommunication strategy is improvised only when acrisis hits, or as it evolves, the chances for amisstep increase. Those chances are especiallylarge if the outbreak is the first major riskproblem for the health authorities involved (16).

As a result, communications about theseunique situations should be routine. A standardformat for reporting risk information should beadopted. Funtowicz and Ravetz (29) propose anotation that includes a best-guess risk estimate(expressed in standard units), a measure ofvariability, and a “pedigree” (indicating the qualityof the research). Although new, such notationmight become familiar, much as degreesFahrenheit, miles per gallon, probability ofprecipitation, and recommended daily allow-ance have become familiar.

Another part of communication planning is toadopt standard scripts for reporting complexprocedural information regarding what citizensshould do and what food specialists are doing.The adoption process should include empiricallytesting the comprehensibility of concrete mes-sages with an audience like the intendedaudience. Influence diagrams offer one templatefor organizing procedural information. Riskanalyses provide one way to identify the crisesmost likely to occur and may allow not onlytesting the most likely messages, but alsoidentifying the persons most likely to do thecommunicating and preparing them accordingly.The chemical industry’s Community Awarenessand Emergency Response program might providesome useful lessons in how to organize forunlikely events, although the challenges ofdealing with the relatively identifiable commu-nity surrounding a chemical plant are differentfrom those presented by dealing with the diffusenational (or even international) audience con-cerned about a food. The chemical industry’sexperience may also provide guidance on howto achieve voluntary industry compliance witha set of communication principles. Publicgoodwill is eroded every time an industryspokesperson violates the public trust by

misrepresenting, or just explaining inadequately,the state of affairs. Reducing misrepresentationrequires institutional discipline; reducing inad-equate communication requires a scientificapproach to communication.

AcknowledgmentsOur research was supported in part by the National

Institute for Allergy and Infectious Diseases, the NationalInstitute of Alcohol Abuse and Alcoholism, and the NationalScience Foundation.

References 1. Fischhoff B, Bostrom A, Quadrel MJ. Risk perception

and communication. In: Detels R, McEwen J, OmennG, editors. Oxford textbook of public health. London:Oxford University Press; 1997. p. 987-1002.

2. Krimsky S, Golding D, editors. Social theories of risk.Westport (CT): Praeger; 1992.

3. Slovic P. Perception of risk. Science 1987;236:280-5. 4. Lammerding AM. Quantitative microbial risk

assessment. Emerg Infect Dis. In press 1997. 5. National Research Council. Improving risk

communication. Washington (DC): National AcademyPress; 1989.

6. Carter MW, editor. Radionuclides in the food chain.Berlin: Springer-Verlag; 1988.

7. Kahneman D, Slovic P, Tversky A, editors. Judgmentunder uncertainty: heuristics and biases. New York:Cambridge University Press; 1982.

8. Gilovich T, editor. How we know what isn’t so. NewYork: Free Press; 1993.

9. Hasher L, Zacks RT. Automatic and effortful processesin memory. J Exp Psychol Gen 1984;108:356-88.

10. Peterson CR, Beach LR. Man as an intuitivestatistician. Psychol Bull 1967;69:29-46.

11. Fischhoff B, Slovic P, Lichtenstein S. Fault trees:sensitivity of assessed failure probabilities to problemrepresentation. J Exp Psychol Hum Percept Perform1978;4:330-44.

12. Crouch EAC, Wilson R, editors. Risk/benefit analysis.Cambridge (MA): Ballinger; 1981.

13. Fischhoff B, Watson S, Hope C. Defining risk. PolicySciences 1984;17:123-39.

14. Slovic P, Fischhoff B, Lichtenstein S. Rating the risks.Environment 1979;21:14-20,36-9.

15. Fischhoff B, Svenson O. Perceived risk of radionuclides:understanding public understanding. In: Harley JH,Schmidt GD, Silini G, editors. Radionuclides in the foodchain. Berlin: Springer-Verlag; 1988. p. 453-71.

16. Fischhoff B. Risk perception and communicationunplugged: twenty years of process. Risk Analysis1995;15:137-45.

17. Schriver KA. Evaluating text quality: the continuumfrom text-focused to reader-focused methods. IEEETransactions on Professional Communication1989;32:238-55.

18. Bursztajn HJ, Feinbloom RI, Hamm RA, Brodsky A,editors. Medical choices, medical chances: howpatients, families, and physicians can cope withuncertainty. New York: Routledge; 1990.


Special Issue

19. Raiffa H, editor. Decision analysis: introductorylectures on choices under uncertainty. Reading (MA):Addison-Wesley; 1968.

20. Merz J, Fischhoff B, Mazur DJ, Fischbeck PS.Decision-analytic approach to developing standards ofdisclosure for medical informed consent. Journal ofToxics and Liability 1993;15:191-215.

21. Burns WJ, Clemen RT. Covariance structure modelsand influence diagrams. Management Science1993;39:816-34.

22. Howard RA. Knowledge maps. Management Science1989;35:903-22.

23. Fischhoff B, Downs J. Accentuate the relevant.Psychological Science 1997;8:154-8.

24. Morgan MG, Henrion M, editors. Uncertainty. NewYork: Cambridge University Press; 1990.

25. Poulton EC, editor. Bias in quantifying judgment.Hillsdale (NJ): Lawrence Erlbaum; 1989.

26. Bostrom A, Fischhoff B, Morgan MG. Characterizingmental models of hazardous processes: a methodologyand an application to radon. Journal of Social Issues1992;48:85-100.

27. Bruhn CM. Consumer education. Emerg Infect Dis. Inpress 1997.

28. Fischhoff B, Riley D, Kovacs D, Small M. Whatinformation belongs in a warning label? Psychologyand Marketing. In press 1998.

29. Funtowicz S, Ravetz J, editors. Uncertainty andquality in science for policy. London: Kluwer; 1990.


Special Issue

In recent years, public concern regardingfood safety has increased as a consequence of theoutbreak of bovine spongiform encephalopathy(BSE) in cattle, the prevalence of Salmonellaserotype Enteritidis illnesses (from poultry,meat, eggs), and the more localized outbreaks ofillnesses associated with Listeria monocytogenes(from dairy products, pâté, salads) and Escherichiacoli O157:H7 (from ground or minced beef,unpasteurized apple juice, vegetables). Emergingpathogens and the appearance of problems suchas BSE have resulted in enactment of specificcontrols in many countries, while the generalheightening of interest internationally hasprompted health professionals and the foodindustry in many countries to scrutinize thecontrol of emerging infectious agents.

Animal Feeding and Food SafetyThe Food and Agriculture Organization

(FAO) of the United Nations has had a long-standing interest in the area of food safety andfood quality. Because of problems such as BSEand emerging pathogens, FAO convened anExpert Consultation on Animal Feeding andFood Safety in Rome in March 1997 to address

these issues and provide the scientific basis forimproving practices in the feeding of animalsfor the production of food.

The ultimate objective of food industry andsafety regulators is to ensure that food reachingthe consumer is safe and wholesome. Thisobjective does not imply that food can ever becompletely free of risk but rather that the level ofrisk to the consumer can be acceptable. Foodsgenerally expected to be safe may become unsafeas a result of hazards introduced duringproduction, processing, storage, transport, orfinal preparation by the consumer. For foodderived from animal sources, the hazards mayoriginate from a number of sources, including theconsumption of contaminated feed.

Hazards in food that may relate to animalfeed include salmonellosis, mycotoxicosis, andingestion of unacceptable levels of veterinarydrugs and agricultural and industrial chemicals.In addition, if the postulated link between BSEand new variant–Creutzfeldt-Jakob disease isestablished, this disease would also be an exampleof contamination originating from animal feed.

The FAO consultation limited itsconsiderations to food safety matters thatpertained strictly to animal feeds; it did notconsider plant toxins, radionuclides, or parasitesspread by human sewage. The risk to humanhealth from other infectious agents that may

Animal Diseases of PublicHealth Importance

Gregory D. OrrissFood and Agriculture Organization of the

United Nations, Rome, Italy

The Food and Agriculture Organization’s (FAO) interest in emerging diseasescaused by foodborne pathogens derives from its role as the leading United Nationsagency with a mandate for food quality and safety matters. The Food Quality andStandards Service of FAO’s Food and Nutrition Division is active in all areas related tofood safety and implements the FAO/World Health Organization Food StandardsProgram. Its activities include providing assistance to FAO’s member nations inaddressing problems,strengthening infrastructure, promoting standardization as ameans of facilitating trade, and safeguarding the interests of consumers. This paperconsiders the importance of emerging foodborne diseases from the perspectives of theconsumer, international trade in food, producers and processors, and developingcountries and addresses prevention and control measures.

Address for correspondence: Gregory D. Orriss, Chief, FoodQuality and Standards Service; Food and Nutrition Division;Food and Agriculture Organization of the United Nations; fax:39-6-5705-4593; e-mail: [email protected].


Special Issue

contaminate either feed or forage appears to benegligible or nonexistent and was, therefore, notconsidered by the consultation. Only the standarddomestic animals from which food is derived inlarge quantities, such as meat and meat products,milk and milk products, and eggs and eggproducts, as well as fish products derived fromaquaculture that involves the feeding of fish,were considered. All aspects of animal feed, otherthan natural unrestricted grazing, were considered.The consultation concluded that emerging pathogensare generally not identified through traditionalanimal surveillance and epidemiology.

Hazards Associated with Animal FeedMycotoxins are secondary metabolites

produced by fungi of various genera when fungigrow on agricultural products before or afterharvest or during transportation or storage.Some fungi such as Aspergillus spp. andPenicillium spp. can invade grain after harvestand produce mycotoxins, while others, such asFusarium spp., typically infest grains andproduce mycotoxins before harvest. In somecircumstances, Aspergilli can grow and producemycotoxins before the crop is harvested.

Both intrinsic and extrinsic factors influencefungal growth and mycotoxin production on asubstrate. Intrinsic factors include wateractivity, pH, and redox potential; extrinsicfactors are relative humidity, temperature, andavailability of oxygen.

Many mycotoxins with different chemicalstructures and widely differing biologic activitieshave been identified. Mycotoxins may becarcinogenic (e.g., aflatoxins B1, ochratoxin A,fumonisin B1), estrogenic (zearalenone and I andJ zearalenols), nephrotoxic (ochratoxins, citrinin,oosporeine), dermonecrotic (trichothecenes), orimmunosuppressive (aflatoxin B1, ochratoxin A,and T-2 toxin). Much of the publishedinformation on toxicity comes from studies inexperimental animals, and these may not reflectthe effects of mycotoxins on humans and otheranimals. In addition, their significance in humanfoods of animal origin is incompletely understood.Mycotoxins are regularly found in animal feedingredients such as maize, sorghum grain, ricemeal, cottonseed meal, groundnuts, legumes,wheat, and barley. Most are relatively stablecompounds, are not destroyed by feed processing,and may even be concentrated in screenings.

Various animal species metabolize mycotoxinsin different ways. In pigs, ochratoxin A canundergo enterohepatic circulation and iseliminated very slowly, whereas in poultryspecies it is rapidly excreted. The polarmycotoxins such as fumonisins tend to beexcreted rapidly. Mycotoxins, or their metabolites,can be detected in meat, visceral organs, milk,and eggs. However, their concentration in thesefood products is usually considerably lower thanin the feed consumed by the animals; at theselevels, mycotoxins are unlikely to cause acuteintoxication in humans consuming these products.Residues in animal products of carcinogenicmycotoxins, such as aflatoxin B1, M1, andochratoxin A, pose a threat to human health, andtheir levels should be monitored and controlled.

In most instances, the principal source ofmycotoxins for humans is contaminated grainsand cereals, rather than animal products. Thismeans that the hazard is much greater indeveloping countries in which maize and othergrains form the staple diet and the intake ofanimal products, including meat, is low.

Only limited information is availableregarding mycotoxin residues in animal productsintended for human consumption. The metabolismof mycotoxins by animals and the residues ofmycotoxins and their metabolites in animaltissues should be studied further.

Infectious Agents

Agent Causing Transmissible SpongiformEncephalopathies in Ruminants

Transmissible spongiform encephalopathiesare nonfebrile neurologic diseases with a longincubation period and are fatal. These diseasesare associated with incompletely defined agentstermed prions, which are resistant to normalheat treatments of feed and food. Sheep scrapiehas been recognized for over 250 years. BSE wasfirst recognized in the United Kingdom during1986. For BSE, the infectious agent enters thefeed primarily through rendered infected tissues(notably the central nervous system and thereticuloendothelial system) under insufficientheat to reduce the concentration of the infectiousagent to an ineffective dose. In the case of sheepscrapie, infection is naturally maintained bytransmission between sheep. Humans havelikely been exposed to the scrapie agent by eating


Special Issue

brain and other tissues, although there is noevidence that Creutzfeldt-Jakob disease inhumans has been associated with scrapie.

Humans can potentially be exposed to BSEthrough consumption of infected tissues. Theoccurrence of a new variant of the humantransmissible spongiform encephalopathy,Creutzfeldt-Jakob disease, has raised thepossibility of an association with the BSE agent.With the limited number of cases now, there is noproven link between this new variant and thepossible transmission of the agent from infectedbovine tissue to humans. The FAO consultationrecommended risk reduction measures toaddress the elimination of BSE from cattle.

Salmonella entericaThe more than 2,000 Salmonella serotypes

can be divided into three groups: species-specific,such as gallinarum (in poultry); invasive, whichmay cause systemic infections in their host, suchas Enteritidis (in laying hens); and noninvasive,which tend to remain within the intestinal tract.Members of the first group are infrequentlyfeedborne pathogens. Among the second group,the principal manifestation of human infection isgastroenteritis, with septicemia occurring insome patients. The third group may be associatedwith subclinical infections in farm livestock; itsometimes causes disease in livestock and isassociated with food poisoning in humans.

Salmonellae are widely distributed, andanimal feed is only one of many sources ofinfection for farm animals. Animal feedingredients of both animal and plant origin arefrequently contaminated with salmonellae,although the most common serotypes associatedwith human disease, Enteritidis and Typhimurium,are rarely isolated from animal feed. Feed can becontaminated from raw ingredients.

Toxoplasma gondiiThe protozoon T. gondii is found in cats and,

according to serologic surveys, also in birds andother domesticated species including sheep, pigs,goats, and horses. The primary source ofinfection for animals is feed contaminated withfeces of cats and possibly with rodent tissues.

Cats are an important source of infection forhumans; however, some human infections maybe due to the handling or consumption of rawmeat. Pregnant women may miscarry or give

birth prematurely, and infants often get centralnervous system disorders and ocular disease.

Trichinella spiralisT. spiralis is a nematode that parasitizes the

intestinal tract of mammals, particularly pigs.The larvae encyst in the tissues, particularly themuscles, which act as a source of infection forhumans who consume raw or partially cookedmeat. The clinical manifestations include fever,muscle pain, encephalitis, meningitis,myocarditis, and (rarely) death.

The cysts in infected carcasses can be killedby freezing (-18°C for 20 days) or traditionalrendering temperatures. Adequate cooking ofraw meat and table scraps before feeding to farmanimals would eliminate this hazard.

The FAO consultation also addressedpotential hazards associated with veterinarydrugs and agricultural and other chemicals andrecommended risk reduction measures toprevent, eliminate, or reduce the hazards toacceptable levels. The consultation participantsprepared a draft Code of Practice for GoodAnimal Feeding to be considered by the CodexAlimentarius Commission (CAC).

Codex Alimentarius CommissionSince 1962, CAC has been responsible for

implementing the Joint FAO/World HealthOrganization (WHO) Food Standards Program.“Codex Alimentarius,” whose name is taken fromLatin and translates literally as “food code” or “foodlaw,” was founded in response to the worldwiderecognition of the importance of international tradeand the need to facilitate trade while ensuring thequality and safety of food for the world consumer.

It follows, therefore, that the commission’sprimary objectives are the protection of thehealth of consumers, the assurance of fairpractices in the food trade, and the coordinationof all food standards. Food standards, guidelines,and recommendations are the work of CAC. Withthe adoption of the World Trade Organization’sAgreement on the Application of Sanitary andPhytosanitary Measures and the Agreement onTechnical Barriers to Trade, a new emphasis anddimension have been placed on Codex standards.

Codex Committee on Food HygieneThe Codex Committee on Food Hygiene

(CCFH) has overall responsibility for all


Special Issue

provisions of food hygiene prepared by Codexcommodity committees and contained incommodity standards, codes of practice, andguidelines. CCFH also develops general principles,codes of practice, guidelines for food hygiene, andmicrobiologic criteria for food to be appliedhorizontally across Codex committees. Foodhygiene is defined as “all conditions andmeasures necessary to ensure the safety andsuitability of food at all stages of the food chain.”

According to the deliberations at the 29thsession of CCFH, the microbiologic safety of foodsis principally ensured by control at the source,product design, process control, and good hygienicpractices during production, processing, handling,distribution, storage, sale, preparation, and use,preferably in conjunction with the application ofthe Hazard Analysis and Critical Control Points(HACCP) system. This preventive system offersmore control than end-product testing because ofthe limited effectiveness of microbiologicexamination to assess the safety of food.

When they have been established by Codex ornational risk managers, objectives for food safetycan be taken up by industry; by applying HACCP(or an equivalent food safety managementsystem), industry can ensure that theseobjectives are met. This is the use of HACCP as acorrective risk management option: a risk isidentified, and a management option is selectedand implemented. HACCP is also used as apreventive risk management tool. In this case,hazard analysis identifies potential hazards inraw materials, production line, and line-environments to the consumer. Hazard analysisis defined as “The process of collecting andevaluating information on hazards and conditionsleading to their presence to decide which aresignificant for food safety and therefore should beaddressed in the HACCP plan.” Input concerningthe potential hazards and their control couldcome from risk analysis, but often suchinformation is not available and industries needto apply their best judgment.

The Revised Principles for the Establishmentand Application of Microbiological Criteria ForFoods states, “Microbiological criteria should beestablished according to these principles, and bebased on scientific analysis and advice, andwhere sufficient data are available, on a riskanalysis appropriate to the foodstuff and itsuses.” These criteria may be relevant to theexamination of foods, including raw materials

and ingredients of unknown or uncertain origin,and may be used when no other means ofverifying the efficacy of HACCP-based systemsand good hygienic practices are available.Microbiologic criteria may also be used todetermine that processes are consistent with theGeneral Principles of Food Hygiene. Microbiologiccriteria are not normally suitable for monitoringcritical limits as defined in the HACCP system.

Establishing microbiologic criteria and foodsafety objectives in general is difficult because ofthe considerable knowledge gap relating tobiologic hazards and their relationship to humanillness. This has led to many evaluations byCCFH, which are based on subjective orqualitative assessments and serve as the basisfor recommendations. Although aware of theselimitations, CCFH is now developing a frameworkof principles and guidelines for the application ofmicrobiologic risk assessment. CCFH’s actionwas in response to the recommendation of the1995 Joint FAO/WHO Consultation on theApplication of Risk Analysis to Food Standardsrelating to the application of risk assessmentwithin the Joint FAO/WHO Food StandardsProgram. International Commission forMicrobiological Specifications for Foods andCCFH delegations are also in the process ofdeveloping background papers on a number offoodborne pathogens to better conductquantitative risk assessments and set subsequentfood safety objectives. Notwithstanding thedevelopment of risk analysis approaches by thesegroups, the work of CCFH and all Codexcommittees would benefit from advice from anexpert body on foodborne biologic hazards forpurposes of risk management. The committeecould be modeled on the FAO/WHO Joint ExpertCommittee on Food Additives and Joint Meetingon Pesticide Residues, allowing for the uniqueconsideration of epidemiologic and clinical datarelated to pathogens causing human illness, andof the dynamics of microbial populations in foodthroughout the food chain.

Control of Listeria monocytogenes in foods isan example of the need to consider a structuredrisk management approach. Listeria arefrequently consumed in small amounts by thegeneral population without apparent ill effects.Only higher levels of Listeria are thought tocause serious disease problems. It is believedthat Listeria will always be present in theenvironment. Therefore, the critical issue may


Special Issue

not be how to prevent Listeria in foods, but how tocontrol its survival and growth to minimize thepotential risk. In many foods, complete absenceof Listeria is unrealistic and unattainable; tryingto achieve this goal can limit trade withouthaving any appreciable benefit to public health.A relevant risk management option, therefore, isto focus on foods that have historically beenassociated with human disease and support thegrowth of Listeria to high levels, rather thanfocusing on foods that do not support growth.Thus, establishing tolerably low levels of Listeriain specific foods may be one food safety objectiveachieved by risk managers after a rigorous andtransparent risk analysis. Such an approach isnow being considered by CCFH after an initialrisk assessment by the InternationalCommission for Microbiological Specificationsfor Foods and CCFH delegations.

Although Listeria presents unique challengesin terms of its widespread occurrence and theparticular susceptibility of vulnerable groups,pathogens such as E. coli O157:H7, Salmonella,and Campylobacter are also being addressed.These microbial pathogens produce acutefoodborne illnesses and can cause severe chronicsequelae, creating an important public healthproblem and food safety concerns.

Codex Codes of Hygienic Practice are basedon good manufacturing practices, HACCP principles,and risk analysis. CCFH is responsible forcoordinating and overseeing the work of specificCommodity Committees in this area. In the specificarea of food hygiene, Codex has revised its maindocument, Recommended International Code ofPractice: General Principles of Food Hygiene, toincorporate risk assessment principles and includespecific references to the HACCP system.

FAO Programs on Food Quality and SafetyThe Food Quality and Standards Service is a

service within the Food and Nutrition Division ofthe FAO, located in Rome. The Secretariat ofCAC is also located there. The Regular Programof the Food Quality and Standards Serviceprovides the technical and scientific basis forFAO for all food quality matters, including foodsafety. This includes providing the Secretariatfor the Joint Expert Committee on FoodAdditives and participation in both the JointMeeting of the FAO Panel of Experts on PesticideResidues in Food and the Environment and theWHO Expert Group on Pesticide Residues and in

the Joint Expert Committee on Food Irradiation.The Food Quality and Standards Service

develops and publishes guidelines and manuals(including the FAO Food and Nutrition PaperSeries and Manuals of Food Quality Control),arranges expert consultations and conferences(e.g., the Joint FAO/WHO Expert Consultation onBiotechnology and Food Safety, September 30 toOctober 4, 1996; the Joint FAO/WHO ExpertConsultation on the Application of RiskManagement to Food Safety Matters, January 27-31, 1997; the Joint FAO/WHO Consultation onFood Consumption and Exposure Assessment toChemicals, February 10-14, 1997; and the FAOConsultation on Animal Feeding and FoodSafety, March 10-14, 1997), and has a major andcontinuing program of providing technicalassistance regarding food standards and foodcontrol to member countries, particularlydeveloping countries and countries in transitionfrom a centrally planned to a market economy.

The Joint Expert Committee on FoodAdditives, the Joint Meeting on PesticideResidues, and the Joint Expert Committee onFood Irradiation are expert committees thatprovide independent scientific advice that formsthe basis for the development of food safetyrecommendations used in international trade.These committees are forums in whichindependent, invited experts assess the state ofscientific knowledge of food additives, pesticideand veterinary drug residues in food, mycotoxins,other chemical contaminants in food, and foodirradiation treatments and make recommendationsto member governments and to Codex.

FAO’s Food Quality and Standards Servicealso develops and publishes Manuals of FoodQuality Control. These manuals providerecommendations for the development andoperation of food quality and safety systems.While aimed primarily at providing advice todeveloping countries, the manuals documentmodern approaches, including the developmentof quality control programs throughout the foodchain that apply to all countries. Such anapproach is instrumental in facilitatinginternational trade in food. Key titles in theseries include Food Inspection, Food for Export,Management of Food Control Programs, ImportedFood Inspection, and Quality Assurance in theFood Control Laboratory.

The program of technical assistance projectsundertaken by the Food Quality and Standards


Special Issue

Service handles assistance in food qualitycontrol, including safety; such projects haveestablished or strengthened the food controlsystems in a number of developing countries.Typically, they assist in establishing theinfrastructure for an enhanced food controlprogram, assessing laboratory servicerequirements, providing guidance to developlegislation and procedural manuals, setting upreputable inspection and certification systems,and providing training and staff development. Inthese assistance projects, the standardsestablished by the CAC are basic guides tointernational requirements.

ConclusionFood will always represent some biologic

risk; it is the task of the food industry tomaintain the level of risk at the minimum thatis practical and technologically feasible. It

should be the role of regulatory bodies to userisk assessment to determine realistic andachievable risk levels for foodborne hazardsand to base their risk management and foodsafety policies on the practical application ofthe results of these analyses.

Foodborne illnesses are preventable.Adherence to good manufacturing practices andgood hygienic practices and application of theHACCP system can result in food safety andensure food quality. Food safety is the sharedresponsibly of governments, academia, the foodindustry, and the consumer.

Codex standards, guidelines, andrecommendations have the objective of protectingthe consumer and facilitating international foodtrade. Adherence to Codex provides the basis forfood safety and quality and meets therequirements of international trade.


Special Issue

In the past two to three decades, public healthauthorities in industrialized countries have beenfaced with an increasing number of food safetyproblems. In 1983, a Joint Food and AgricultureOrganization/World Health Organization ExpertCommittee on Food Safety concluded that illnessdue to contaminated food was perhaps the mostwidespread health problem in the contemporaryworld and an important cause of reduced economicproductivity (2). More recent data from industri-alized countries indicate that annually up to 10%or more of the population may have a foodbornedisease. The situation is equally serious indeveloping countries, where infant diarrhea causesmany illnesses and deaths. In addition to knownfoodborne diseases, public health communitiesare being challenged by the emergence of new ornewly recognized types of foodborne illnesses,often with serious and chronic health conse-quences. Certain populations (e.g., pregnant

women, the elderly, infants and children,immunocompromised persons, and the under-nourished) are particularly vulnerable. In economicterms, foodborne illnesses are very costly forindustry, health services, and society as a whole.

Many factors have contributed to the increasein foodborne disease. Industrialization, leading toincreased wealth and urbanization, has revolu-tionized the food supply system, resulting in massproduction and an explosive increase in thenumber of food service establishments and foodoutlets. Mass production, environmental factors,and inadequate knowledge on the part of foodhandlers have contributed to increased contami-nation of primary foodstuffs.

The increase in international trade hasincreased the risk for cross-border transmissionof infectious diseases. The globalization of food(and feed) trade, facilitated by the liberalizationof world trade, while offering many benefits andopportunities, also presents new risks (3). Food, amajor trade commodity, is also an importantvehicle for transmission of infectious diseases.Because food production, manufacturing, and

Foodborne Disease Control:A Transnational Challenge

F. K. Käferstein, Y. Motarjemi, and D. W. BettcherWorld Health Organization, Geneva, Switzerland

“Disease knows no boundaries and borders are porous to disease.”(1)

Address for correspondence: F.K. Käferstein, Food Safety Unit, WorldHealth Organization, 20 Avenue Appia, CH-1211 Geneva 27, Swit-zerland; fax: 41-22-791-4807; e-mail: Kä[email protected].

In the globalized political economy of the late 20th century, increasing social,political, and economic interdependence is occurring as a result of the rapid movementof people, images, values, and financial transactions across national borders. Anotherconsequence of the increase in transnational trade, travel, and migration is the greaterrisk of cross-border transmission of infectious diseases. As the world becomes moreinterconnected, diseases spread more rapidly and effectively. With more than onemillion people crossing international borders every day, and with the globalization offood production, manufacturing, and marketing, the risk of infectious diseasetransmission is greater. Economic globalization has also increased the need forgovernmental budget austerity, and consequent national preparedness has beeneroded. The emergence of new infectious diseases, as well as the reemergence of oldones, thus represents a crucial transnational policy issue. These problems cannot beresolved by national governments alone; they require international cooperation. Thisarticle analyzes the role of foodborne disease surveillance programs, nationally andinternationally, in the control of foodborne diseases.


Special Issue

marketing are now global, infectious agents canbe disseminated from the original point ofprocessing and packaging to locations thousandsof miles away. This multinational approach tofood production and distribution and theprogressive opening up of world markets haveallowed the international food trade to flourish.The value of food trade, U.S. $266 billion in 1994,was more than 300% greater than it was 20 yearsago and continues to grow rapidly (4).

The globalization of foodborne diseases alsoresults from increased travel. Internationaltravel is more accessible today. The WorldTourism Organization estimates world touristarrivals at 567 million in 1995, and this figure isexpected to rise to 660 million by the year 2000.Over the past 200 years, the average distancetraveled and the speed of travel have increased1,000 times while incubation periods for diseaseshave not changed. As a result, a person can beexposed to a foodborne illness in one country andexpose others to the infection in a locationthousands of miles from the original source of theinfection (5). Depending on their destination,travelers are estimated to run a 20% to 50% riskof contracting a foodborne illness.

As international trade and travel increase,foodborne disease outbreaks of the same originare more likely to occur in different parts of theglobe. Food safety in the late 20th centuryrepresents a transnational challenge requiringenhanced levels of international cooperation insetting standards and regulations and instrengthening surveillance systems. Effectivefood safety programs, built on a clear under-standing of the epidemiology of foodbornedisease, must be developed and implemented.The globalization of the world’s economy has beenaccompanied by intense economic competitionand increased pressure on governments todownsize. Public sector austerity has reduceddisease surveillance in many countries (6). Forexample, in Great Britain, the failure to maintainpublic health infrastructures has, in the words ofthe British Medical Association, resulted in“Britain returning to the 19th century in terms ofpublic health, with problems such as dirty water,contaminated food, and old infectious diseasesreemerging” (7). Failing a reversal of this trend,public health authorities and health services maybe overwhelmed in the near future by outbreaksor epidemics of foodborne diseases. The 1991epidemic of cholera in Peru and the 1996

outbreak of Escherichia coli O157 in Japandemonstrate how one single foodborne diseaseepidemic or outbreak may disrupt the function-ing of a health-care system.

Epidemiologic surveillance of foodborne illnessis fundamental to the planning of food safetyprograms and the development of a strategy forprevention and control. There are differentmethods of surveillance: death registrations andhospital discharges; disease notification; labora-tory-confirmed cases; sentinel surveillance;surveillance of investigated outbreaks; popula-tion-based surveillance; and case-control studiesof sporadic cases (8). This article examines therole of foodborne disease surveillance programs,nationally and internationally, in the control andprevention of foodborne disease.

Foodborne Disease Awareness of PublicHealth Authorities

Data on the incidence of foodborne illnessescollected through notifications, laboratory con-firmations, and sentinel or population-basedstudies can provide a measure of the magnitude ofthe foodborne disease problems, their economicconsequences, and over the years, an indicationof the trend. Although several weaknesses areassociated with the collection of such data—particularly those collected through notifica-tion and laboratory confirmations (since theyrepresent only the tip of the iceberg)—they cannevertheless be useful in raising the awarenessof public health authorities about the impor-tance of food safety.

Surveillance data collected in some industri-alized countries confirm that foodborne diseasesconstitute one of the most widespread healthproblems and that they have increased over thelast two or three decades (Figures 1-4). Part of theincrease may be attributable to recent improve-ments in information reporting and collectionsystems, improved diagnoses, or greater public-ity and concern about food safety in general.However, a real increase of foodborne diseaseincidence is not disputed. First, the increase hasbeen steady and cannot be explained by a one-time improvement in the surveillance system.Second, increases have been observed in differentcountries, including those with no improvementin reporting and surveillance programs. Thegeneral increase, as demonstrated by the resultsof surveillance data, has led many public healthauthorities to take stringent regulatory and


Special Issue

educational measures to improve food safety,with some successful results (9). For instance, inthe United States, active surveillance offoodborne listeriosis has led to concerted effortsby industry and government to prevent thedisease. As a consequence, the number of casesand deaths has decreased by 44% and 48%,respectively (10).

Public health authorities must be aware ofthe magnitude and trend of foodborne illness sothat necessary resources can be mobilized toimprove food safety programs. Lack of reliableepidemiologic data in many parts of the world hasimpeded the recognition of the public healthimportance of food safety and consequently theemphasis on food safety programs.

Early Detection Of Foodborne DiseaseOutbreaks

Surveillance of foodborne diseases plays animportant role in the early detection offoodborne disease outbreaks and their control.Early identification of the source of theoutbreak is becoming increasingly important ascountries move towards industrialization.Increased mass production means outbreakscan change from being small and confined to afamily to large, affecting hundreds or eventhousands of people (Table).

Rapid investigation of foodborne diseaseoutbreaks is crucial to prevent them from takingon massive proportions. In the 1993 Frenchoutbreak of listeriosis due to potted minced pork(affecting 39 persons and causing eight miscar-riages and one death), public health authoritiestraced its source within 1 week and thusprevented the outbreak from spreading byremoving the implicated food product from themarket and informing the group at risk about itsunsafe nature (11). In an outbreak of botulism in

Figure 1. Laboratory reports of gastrointestinalinfections in England and Wales.

Figure 2. Incidence of foodborne diseases inVenezuela.

Figure 3. Incidence of salmonellosis in the UnitedStates, Japan, and Australia.

Figure 4. Incidence of infectious enteritis andtyphoid and paratyphoid fevers in Germany.

Table. Examples of large foodborne disease outbreaksCountry Year Disease No. casesUnited Kingdom 1985 Salmonellosis 1,000United States 1985 Salmonellosis >168,000United States 1993 Salmonellosis 224,000China 1988 Hepatitis A >310,000Germany 1993 Salmonellosis 1,000Australia 1991 Norwalk-like >3,050

agentUnited States 1992-93 E. coli O157 >500

infectionJapan 1996 E. coli O157 >6,000

infection


Special Issue

the United Kingdom traced to hazelnut yogurt,the source was identified within 3 days, and theproduct was withdrawn from the market (12).

Because of global food distribution andworldwide travel, an international exchange ofinformation on foodborne disease incidences andoutbreaks and the foods involved is extremelyimportant to identify international clustersoriginating from a common source. For instance,Salm-Net, a network for the internationalsurveillance of human salmonellosis, has demon-strated the value of such an interactiveinternational collaboration. Individual countrieswith apparently isolated outbreaks can feed theirinformation into the network and ascertainwhether the outbreak is confined to their countryor is of wider international importance. Theidentification and investigation of severalinternational outbreaks have been simplified bythe Salm-Net network.

Food, the Transmission of Diseases, andthe Identification of Associated RiskFactors

Information collected through investigationof foodborne disease outbreaks or case-controlstudies of sporadic cases provides a betterunderstanding of the role of food in thetransmission of communicable diseases and inthe identification of risk factors leading todisease. Epidemiologic data from foodbornedisease surveillance can provide public healthauthorities with important information about thetypes of food implicated in outbreaks; populationsat risk; practices that lead to contamination,growth, and survival of foodborne pathogens; andplaces where foods are often mishandled. Suchdata are essential for designing effectiveintervention programs. Such programs inindustrialized countries, for example, havedemonstrated the relatively greater prevalenceand incidence of foodborne diseases of microbialorigin over those of chemical origin and the role offood handlers in the transmission of diseases;they have identified campylobacteriosis andsalmonellosis (particularly infections caused bySalmonella Enteritidis) as the leading foodbornediseases. The emergence of other diseases, suchas infections due to E. coli O157 and Listeriamonocytogenes—often with serious sequelae—has been pinpointed as a major public healthproblem. These surveillance programs have alsoalerted public health authorities to the foods

most often implicated and the major riskfactors in food preparation.

Because of the lack of epidemiologic data, therole of food in the transmission of diseases hasbeen poorly acknowledged, particularly in develop-ing countries. Diarrheal diseases in infants andchildren and diseases such as shigellosis andcholera have been perceived as being water-bornefor many years. For instance, after the choleraepidemic in Peru (where epidemiologic investiga-tions implicated, among other foods, seafood, andan embargo was placed on trade in foodstuffs),the role that food plays in the transmission of thedisease began to be fully recognized.

Increased trade in food, international traveland migration, and economic and technologicdevelopment have changed dietary habits. Newfoods, food preparations, and dietary habits areintroduced into different regions, and as aconsequence, foodborne diseases are emerging orreemerging. Dietary habits are also changing as aresult of nutritional recommendations andcampaigns or may be influenced by food policy,production systems, or environmental changesthat lead to increased access to certain foods.These changes in dietary habits influence theepidemiology of foodborne illnesses and contrib-ute to the emergence of foodborne diseases. In theUnited States, public information campaignspromote an increased consumption of fruits andvegetables. To meet the increased demand, theseproducts have to be imported on a seasonal basis.At certain times of the year, more than 75% of thefresh fruits and vegetables available in grocerystores and restaurants are imported (13).Epidemiologic data have shown that, partly as aconsequence of the increased consumption offruits and vegetables, the proportion of foodbornedisease outbreaks has doubled (14).

Data collected through foodborne diseasesurveillance programs permit the monitoring ofchanges in the epidemiology of foodbornediseases and the identification of new pathogensand new dietary or food preparation habits thatmay present a health risk. The data can alsodetermine if existing programs need to bereadjusted to ensure that the food safety programis adequate and relevant.

A method used in recent years to complementepidemiologic data in identifying risky practicesand behavior is the Hazard Analysis and CriticalControl Points system (HACCP). Application ofHACCP to food preparation permits the


Special Issue

identification of practices that may be potentiallyhazardous and need to be modified or those thatare critical for ensuring the safety of foods andrequire specific monitoring. However, the firstprinciple of HACCP—to conduct a hazardanalysis—calls for epidemiologic data on foodbornediseases, as the process involves an appraisal ofthe possibility of hazards and the severity oftheir effects; the qualitative and quantitativeevaluation of the presence of hazards; thesurvival and multiplication of microorganismsof concern; the production or persistence oftoxins, chemicals, or physical agents in foods;and, conditions leading to the above.

As demonstrated in the decision tree forhazard analysis (Figure 5) (15), access toinformation would be difficult without epidemio-logic surveillance of foodborne diseases. Simi-larly, epidemiologic data are also needed todevelop sampling plans of food, as demonstratedin the decision tree for Listeria monocytogenessampling plans of foods (Figure 6) (16).

Planning and Evaluating the Effectivenessof Food Safety Programs

The collection of epidemiologic data isimportant in planning interventions and settingpriorities. Countries with scarce resources,facing an abundant number of foodborne diseasesand food safety problems, need to prioritize foodsafety issues. Epidemiologic data provide a basisfor identifying foodborne diseases, groups at risk,or even priority points in the food chain.

Evaluating the effectiveness and impact of anintervention is an important element of any plan.Data collected through disease notification orsentinel studies permit an evaluation of theeffectiveness of interventions and their impact onhealth, and if necessary, the adjustment of aprogram to improve its efficacy and impact. Data

Figure 5. Hazard identification: identification ofpotentially hazardous microorganisms (15).

Figure 6. Listeria monocytogenes sampling plansof foods that did not receive an in-pack listericidaltreatment (16).


Special Issue

on the rising incidence of foodborne illnesses inmany countries demonstrate that presentprevention strategies, mainly based on regula-tory measures, are inadequate and emphasizethe need for additional measures (e.g.,additional regulatory initiatives and healtheducation about food safety).

Risk Assessment and International FoodStandards

The movement of ever-increasing quantitiesof food across borders has resulted in atransnationalization of disease risk (17). There-fore, the globalization of food trade and the openaccess to foreign markets need to be accompaniedby effective means of health protection forpopulations. In the food sector, internationalregulatory instruments need to be integratedwith strengthened surveillance and monitoring.

As a result of the Uruguay Round ofMultilateral Trade Negotiations and the in-creased liberalization of trade facilitated by thisagreement, concern about the safety of importedfood has grown. However, provisions in theAgreement on the Application of Sanitary andPhytosanitary Measures, which entered intoforce with the establishment of the World TradeOrganization on January 1, 1995, are designed toaddress these concerns: according to the work ofthe Codex Alimentarius Commission, its stan-dards, guidelines, and recommendations arerecognized as the reference for national foodsafety requirements. Countries that are mem-bers of the World Trade Organization may nolonger be able to reject foods that meet Codexstandards, guidelines, and recommendationswithout providing justification.

Moreover, the increased volume of the globalfood trade underscores the need for soundepidemiologic information and internationalrisk assessment. In this regard, Article 5 of theSanitary and Phytosanitary Measures agree-ment explicitly requires World Trade Organi-zation members to conduct scientific andconsistent risk assessments. Furthermore, theWorld Health Organization has recommendedthat the application of the HACCP system atevery stage of the food chain represents aneffective approach for governments to meet theterms outlined in the agreement (18).

Another issue receiving more attention fromregulatory agencies and underlined during theFood and Agriculture Organization/World Health

Organization Conference on Food Standards,Chemicals in Food, and Food Trade (1991), is thescientific basis of the Codex standards. TheConference recommended that the Codex, in itsnorm-setting work on health and safety, placegreater emphasis on risk assessment (19).Epidemiologic data on foodborne diseases havean important role in risk assessment. Oneexample is assessing the risk of contractinglisteriosis associated with different levels ofListeria monocytogenes in smoked fish andmeat products (16). However, the need for riskassessment as the basis for setting standardshas shown a great gap in knowledge aboutfoodborne pathogens and their relation tohuman illness (20-22). To address the national/transnational risks caused by foodbornediseases, this gap must be narrowed.

Risk Assessment ApproachRisk assessment is defined as a scientifically

based process that has the following steps: 1)Hazard identification—The identification ofbiologic, chemical, and physical agents present ina particular food or group of foods that can causeillness. 2) Hazard characterization—The qualita-tive or quantitative evaluation of the nature ofthe illness associated with biologic, chemical, andphysical agents that may be present in food. Forchemical agents, a dose-response assessmentshould be performed. For biologic or physicalagents, a dose-response assessment should beperformed if the data are obtainable. 3) Exposureassessment—The qualitative or quantitativeevaluation of the likely intake of biologic,chemical, and physical agents in food as well asexposures from other sources. 4) Risk character-ization—The qualitative or quantitative estima-tion, including uncertainties, of the probability ofand severity of known or potential illness in agiven population on the basis of hazardidentification, hazard characterization, andexposure assessment.

In many cases, data are not available tosupport a quantitative risk assessment of biologichazards. We discuss next the types of challengesthat make quantitative risk assessment difficultfor pathogenic organisms associated with foodand the role of epidemiologic surveillance.

Hazard IdentificationBecause only some foodborne disease out-

breaks are adequately investigated and have the


Special Issue

etiologic agents identified, many foodbornepathogens remain unidentified. Most of theavailable epidemiologic data are furnished byindustrialized countries, while the situation indeveloping countries is largely unknown. Theepidemiologic database must be extended toinclude information from developing countries.However, investigation and surveillance systemsin developing countries need to be strengthenedbefore the database can expand.

Hazard CharacterizationFor many foodborne pathogens, dose-re-

sponse data are limited or nonexistent. Informationon which dose-response estimates can be based isdifficult to obtain and may be inaccurate for variousreasons: host susceptibility to pathogens is highlyvariable, attack rates from a specific pathogen mayvary widely, virulence of a pathogenic species ishighly variable, pathogenicity is subject to geneticvariation resulting from frequent mutation,antagonism from other bacteria in foods or thedigestive system may influence pathogenicity, andfoods may modulate the ability of bacteria to infector otherwise affect the host.

Exposure AssessmentAn exposure assessment will give an estimate

of either the number of pathogenic organisms orthe level of toxins consumed in food. Although thelevels of chemical agents in food may change onlyslightly due to processing, the population ofbacterial agents is dynamic and may increase ordecrease dramatically. Changes in populations ofbacteria are affected by complex interactions ofthese factors: ecology of the bacterial pathogen;processing, packaging, and storing of food;preparation steps, such as cooking, which mayinactivate bacterial agents; and cultural factorsrelating to consumers.

In addition, for some of the emergingfoodborne pathogens, the sources of exposure arestill not fully understood. Information onfoodborne disease outbreaks provides an oppor-tunity to learn about the types of foods that mayharbor the pathogen.

Risk CharacterizationCharacterizing the risk associated with

biologic pathogens depends on informationgained in the previous steps. Risk characteriza-tion will result in a qualitative or quantitative

estimate of the potential for adverse effects froma particular pathogen on a specific population.Whether a quantitative risk assessment approachis possible and appropriate for characterization ofrisks associated with foodborne pathogens is notknown. Thus, the qualitative approach tocharacterizing risk may be the only alternative.

International TravelInternational travel and migration are

contributing factors in the spread of foodbornediseases in some countries. For instance, 80% to90% of the incidence of salmonellosis inScandinavian countries is attributed to interna-tional travel. Surveillance of travel-relatedfoodborne diseases provides a mechanism forappreciating the relative prevalence of foodbornediseases in various countries. It also provides abasis for informing physicians and healthservices about unfamiliar diseases contracted bytravelers returning from distant places. In thisway, advice on precautionary measures can alsobe given to travelers. The only foodborne diseasenow covered by the International HealthRegulations is cholera, which is reported to theWorld Health Organization. Since the purpose ofthese regulations is to help provide maximumsecurity against the international spread ofdiseases with a minimum of interference withworld traffic (i.e., trade and travel) (23), it istimely to consider whether the regulationsshould cover additional foodborne diseases.

ConclusionThe globalization of the risks associated with

foodborne illness, specifically increased interna-tional travel and trade in food, has resulted ingreater interdependence in terms of food safety.Therefore, internationally agreed-upon foodsafety standards and other types of agreementsare becoming increasingly important in address-ing the complex transnational challenge offoodborne disease control. Epidemiologic dataprovide a common ground for reaching interna-tional consensus on food safety issues.

As Morris Potter has said, “If onerecognizes that ensuring food safety isinherently uncertain, foodborne illnesses be-come opportunities to learn rather than failures topredict. Foodborne disease will occur, and wemust be prepared to react quickly to reduce therisk of new foodborne hazards” (24).


Special Issue

References 1. Kemel, W. Health dilemma at the borders: a call for

global action. In: Proceedings of the 34th Session of theWHO Advisory Committee on Health Research; 1996Oct; Geneva, Switzerland. Geneva: World HealthOrganization; 1996.

2. FAO/WHO. The role of food safety in health anddevelopment. Report of Joint FAO/WHO ExpertCommittee on Food Safety. World Health Organ TechRep Ser; 1984;705.

3. Fidler D. Globalization, international law andemerging infectious diseases. Emerg Infect Dis1996;2:77-84.

4. FAO. The world food summit. FAO technicalbackground document no. 12;1996.

5. WHO. International response to epidemics andapplications of the International Health Regulations:report of a WHO consultation. Geneva: World HealthOrganization unpublished document;1996;WHO/EMC/IHR/96.1.

6. Berkelman RL, Bryan RT, Osterholm MT, LeDuc JW,Hughes JM. Infectious disease surveillance: acrumbling foundation. Science 1994:264:368-70.

7. Ferriman A. Doctors warn of a return of past plagues.The Independent 1997 Mar 7; News section:5.

8. Borgdorff MW, Motarjemi Y. Foodborne diseasesurveillance: What are the options? World Health StatQuart 1997;50(1/2):12-23.

9. Motarjemi Y, Käferstein FK, Moy G, Miyagawa S,Miyagishima K. Importance of HACCP for publichealth and development. The role of the World HealthOrganization. Food Control 1996;7:77-85.

10. Tappero JW, Schuchat A, Deaver KA, Mascola L, WengerJD. Reduction in the incidence of human listeriosis in theUnited States. JAMA 1995;273:1118-22.

11. Goulet V. Investigation en cas d’épidemie de listériose,Méd Mal Infect 25 Spécial 1995;184-90.

12. O’Mahony M, Mitchell E, Gilbert RJ, Hutchinson DN,Begg NT, Rodhouse JC, Morris JE. An outbreak offoodborne botulism associated with contaminatedhazelnut yoghurt. Epidemiol Infect 1990;104:389-95.

13. Hedberg CW, MacDonald KL, Osterholm MT.Changing epidemiology of food-borne diseases: aMinnesota perspective. Clinical Infect Dis 1994;18:671-80.

14. Wachsmuth K, Kruse H, Tauxe R, Hedberg C, PotterM. Microbial hazards and emerging issues associatedwith produce. Journal of Food Protection. In press1997.

15. Notermans S, Mead GC. Incorporation of elements ofquantitative risk analysis in the HACCP system. Int JFood Microbiol 1996;30:157-73.

16. Van Schothorst M. Sampling plan for Listeriamonocytogenes. Food Control 1996;7:203-8.

17. Nakajima H. Global disease threats and foreign policy.Brown Journal of World Affairs. In press 1997.

18. WTO. Selected World Health Organization activitiesrelevant to the application of sanitary and phytosanitarymeasures. Geneva: WTO;1995;G/SPS/W/37.

19. FAO/WHO. Report of the Joint FAO/WHO Conferenceon Food Standards, Chemicals in Food and Food Trade.Rome: FAO;1991.

20. FAO/WHO. Report of the Joint FAO/WHO ExpertConsultation on Application of Risk Analysis to FoodStandards Issues, Geneva, 1995 13-17 Mar. Geneva:World Health Organization, unpublisheddocument;1995;WHO/FNU/FOS/95.3.

21. FAO/WHO. Report of the Joint FAO/WHO ExpertConsultation on Risk Management and Food SafetyIssues, Rome, 1997 27-31 Jan. FAO Food Nutr Pap1997;62.

22. Bernard DT, Scott VN. Risk assessment and foodbornemicro-organisms: the difficulties of biological diversity.Food Control 1995;6:329-33.

23. WHO. International Sanitary Regulations: WorldHealth Regulation No. 2. Geneva: World HealthOrganization;1951.

24. Potter M, Gonzalez-Ayala S, Silarug N. Theepidemiology of foodborne diseases. In: Doyle MP,Beuchat L, Montville T, editors. Fundamentals of FoodMicrobiology. Washington (DC): American Society forMicrobiology; 1997.


Special Issue

Consumer Knowledge and ConcernConsumers are receptive to information

about microbiologic hazards. Nationwide surveysby the Food Marketing Institute indicate thatmore people volunteer concerns about microbio-logic hazards than about any other potential foodsafety issue. From 1992 to 1996, volunteeredconcern about microbiologic safety increased from36% to 49% (1). Specifically, concern aboutcontamination by bacteria or other microorgan-isms was 77%, more than concern about pesticideresidues (66%), product tampering (66%), antibi-otic residues (42%), or any other food safety risk.

Food-Handling PracticesAlthough many consumers recognize the

potential seriousness of foodborne bacteria, theylack information on safe handling and storage offood products (2). Williamson et al. (3) found thatconsumers under 35 years of age knew less aboutfood safety terms and concepts than those over35. Specific safe food handling was not practiced

by 15% to 30% of survey respondents. Forexample, consumers did not cool cooked foodrapidly, with 29% indicating they would letroasted chicken cool completely before refrigerat-ing. Only 32% indicated they would use small,shallow containers to refrigerate leftovers.Consumers did not know that failure torefrigerate may jeopardize safety, with 18% notconcerned or uncertain about the safety of cookedmeat and 14% not concerned about poultry leftunrefrigerated for more than 4 hours. The needfor sanitation was not recognized, with only54% indicating they would wash a cuttingboard with soap and water between cutting rawmeat and chopping vegetables.

Food safety experts have identified the mostcommon food-handling mistakes made byconsumers at home. These mistakes includeserving contaminated raw food, cooking orheating food inadequately, obtaining food fromunsafe sources, cooling food inadequately,allowing 12 hours or more between preparationand eating, and having a colonized person handleimplicated food or practice poor hygiene (4). Thesame factors were identified in mishandlingassociated with specific pathogens (5).

Consumer Concerns:Motivating to Action

Christine M. BruhnUniversity of California, Davis, California, USA

Microbiologic safety is consumers’ most frequently volunteered food safetyconcern. An increase in the level of concern in recent years suggests that consumers aremore receptive to educational information. However, changing lifestyles have lessenedthe awareness of foodborne illness, especially among younger consumers. Failure tofully recognize the symptoms or sources of foodborne disease prevents consumers fromtaking corrective action. Consumer education messages should include the ubiquity ofmicroorganisms, a comprehensive description of foodborne illnesses, and preventionstrategies. Product labels should contain food-handling information and warnings forspecial populations, and foods processed by newer safety-enhancing technologiesshould be more widely available. Knowledge of the consequences of unsafe practicescan enhance motivation and adherence to safety guidelines. When consumersmishandle food during preparation, the health community, food industry, regulators, andthe media are ultimately responsible. Whether inappropriate temperature control, poorhygiene, or another factor, the error occurs because consumers have not been informedabout how to handle food and protect themselves. The food safety message has notbeen delivered effectively.

Address for correspondence: Christine M. Bruhn, ConsumerBehavior, University of California, Center for ConsumerResearch, University of California, Davis, CA 95616-8598,USA; fax: 916-752-3975; e-mail: [email protected].


Special Issue

Changing LifestylesMany factors have contributed to consumers’

lack of familiarity with safe food handling andincreased foodborne illnesses. Increased partici-pation in the paid labor force has lessened theexposure of young people to food-handlingpractices in the home; few schools offer or requirefood preparation classes; and partially preparedfoods may have different, less familiar handlingrequirements (2,6).

People eat out more frequently today,thereby increasing their exposure to the foodservice industry, noted for high turnover ratesand minimal job training in personal hygiene(7). Furthermore, the population is shifting,with an increased percentage of persons athigher risk for foodborne illness because of ageor health status (8,9). Additionally, some foodsafety recommendations related to tempera-ture and acidity do not eliminate risks fromsome pathogens (2).

Nature and Source of Foodborne IllnessConsumer perceptions and behavior re-

lated to foodborne illness changed littlebetween 1988 and 1993 (10). Consumersmisperceived the nature of foodborne illnessand the most likely pathogen source. Consumerbelief about the type of food responsible forfoodborne illness—meat, poultry, seafood,eggs—was consistent with expert opinions;however, consumers believed that foodborneillness was generally mild, without fever, andoccurred within a day of eating contaminatedfood. Infections caused by Salmonella andCampylobacter, the most common foodborneillnesses in the United States (11), are notconsistent with the symptoms consumersdescribed, because these organisms havelonger latency and cause fever.

Most consumers believed that their foodborneillness was caused by food prepared somewhereother than the home. Williamson (3) found thatabout one-third of consumers thought food safetyproblems most likely occurred at food manufac-turing facilities, and one-third blamed unsaferestaurant practices. Only 16% thought mishan-dling was most likely to occur in the home. Fein etal. (10) found that 65% of consumers attributedfoodborne illness to food prepared at arestaurant, 17% to mishandling at the supermar-ket, and 17% to mishandling at home. In contrast,food safety experts believe sporadic cases and

small outbreaks in the home are far morecommon than recognized outbreaks (2).

Failure to recognize the home as a likelysource of foodborne illness is not unexpectedbecause illness traced to a food establishmentaffects many people and may receive widespreadpublicity (12). Illness that occurs at home israrely reported unless severe (2).

If consumers misperceive the nature andorigin of foodborne illness, they underestimatethe frequency of serious consequences and areless motivated to change. Schafer et al. (13) foundthat motivation for proper food handling requiresviewing the mishandling of food as a direct threatto one’s health. The failure to associatemishandling of food in the home with foodborneillness interferes with foodborne diseaseeducation efforts (10).

Ubiquity of OrganismsConsumers do not seem to be aware of the

ubiquity of microorganisms in the environment.During foodborne disease outbreaks, pressaccounts focus on fecal contamination of food.Government standards classify natural microor-ganisms as contaminants, which suggests thatmicroorganisms are only present as a result ofmishandling. In contrast, Hazard Analysis andCritical Control Points (HACCP) programsrecognize and attempt to control potentialdangers related to pathogenic microorganisms.

When consumers in a national sampling wereasked on whom they rely for product safety, thepercentage responding “myself as an individual”decreased from 48% in 1989 to 25% in 1996 (1). Asself-reliance decreased, consumer reliance on foodmanufacturers and supermarkets increased.This may be a response to the message that ifraw food contains microorganisms, it iscontaminated. It suggests some consumers areshifting the responsibility for safe food tomanufacturers and retailers.

Consumers may not realize they canintroduce pathogens during food handling. In1990, the Food Marketing Institute askedconsumers what steps they took at home toensure the safety of food (14). Respondentsvolunteered refrigeration (58%), proper stor-age (35%), checking expiration dates (26%),washing and cleaning the food (25%), cookingproperly (22%), and wrapping food properly(20%). No one volunteered washing hands orpreventing cross-contamination.


Special Issue

LabelingProducts must contain safety labels in-

structing consumers how to handle food. In1989, the National Advisory Committee onMicrobiological Criteria for Foods recom-mended a mandatory uniform logo for perish-able refrigerated foods, uniform labeling forfrozen food, “Use by” dates, and time/temperature indicators wherever possible (15).

Although products are currently labeledwhen they require refrigeration, the label isineffective because the warning is difficult to findor read. As the proportion of older peopleincreases, print must be larger. Labels shouldalso display symbols to further enhance theeffectiveness of the message.

Safe-handling labels on meat productsappear to have made a difference. The FoodMarketing Institute (1) found that 60% ofsurvey respondents had seen the labels. Ofthose aware of the labels, 65% said the labelsincreased their awareness of safety, and 43%said they changed their behavior as a result ofthe information. The most common volunteeredchange was washing the counter and utensilsafter contact with meat (approximately 40%),followed by washing hands more frequently(approximately 20%) and cooking to the propertemperature (approximately 20%).

Labels do not consistently contain neededinformation. When foodborne illness was relatedto consumption of unpasteurized apple juice,consumers were not able to determine from thelabel which products were pasteurized. Manymajor manufacturers do not indicate whethertheir fresh juice product is pasteurized.

Consumers are not advised about potentialrisks for special populations. Raw milk sold inCalifornia must contain a warning statement,but other states may not have this requirement.Because of inconsistencies in labeling, unpas-teurized juice products may be given to infants.Also, products that contain honey do not include awarning about potential risk for infant botulism.

Processing TechnologyConsumers do not realize that pathogens can

survive minimal processing, as evidenced by arecent Escherichia coli outbreak associated withfresh apple juice, which demonstrated thatprocessors also may not recognize potential risks.A fresh apple juice manufacturer in northernCalifornia claimed its product was safe because

the juice was squeezed in small batches andfrozen immediately.

Freezing is not effective against E. coliO157:H7, but other methods are protective.Several methods have been developed to reducepathogens and increase the safety of foods. Oncethese methods are verified as effective and safe,the food industry should be free to use them, andconsumers should have the opportunity to selectsafer foods. In some cases, the regulatoryapproval process appears to hinder rather thanfacilitate the safer handling of food.

Food irradiation, exposing food to high levelsof electromagnetic energy for specific purposes,has been approved for selected uses. A petitionbefore the Food and Drug Administration topermit irradiation of meat and other muscle foodsappears to have satisfied safety concerns, butapproval has not yet been granted. Therequirement to seek approval for each applica-tion of irradiation prevents rapid response incases of foodborne outbreaks. Although thisregulatory procedure may have been reasonablewhen irradiation was first introduced, itwarrants a fresh look in view of the wealth of datanow available on the safety and wholesomenessof food irradiation (16).

Attitude surveys and marketing experienceconsistently demonstrate that consumers willpurchase irradiated food (17). National surveysindicate consumer concern about irradiation waslower than other food-related concerns. Whenspecifically asked what they considered a serioushealth hazard, 29% identified irradiation, 77%identified bacteria, and 66% identified pesticides(1). The percentage of consumers concernedabout irradiation has decreased significantlyover time. In the late 1980s, 42% to 43% classifiedirradiation as a serious concern, decreasing to29% in 1996. A relative ranking of food processingmethods surveyed by the Gallup Organizationfound that irradiation, food preservatives, andchlorination generated similar concern (18).

In a nationwide Food Marketing Institutesurvey, 69% of consumers indicated they werevery or somewhat likely to purchase productsirradiated to kill bacteria or other microorgan-isms (1). Surveys completed in several areas ofthe country indicate 60% to 70% of consumerswould prefer irradiated food (17). In one study,information about irradiation increased interestin purchasing to 90%, and education plus foodsamples increased purchase intent to 99%.


Special Issue

Consumers have purchased irradiated food inselect locations across the United States since1992 (17). Fruits from the Mainland and Hawaiihave sold well in the Midwest and California (L.Wong, pers. comm.). Irradiated chicken gainedover 60% of market share when priced 10% lowerthan nonirradiated chicken and 47% when pricedthe same (J.A. Fox, pers. comm.).

Irradiated food is not widely available.Special interest groups threaten companies thatexchange information about irradiation process-ing. Consumers, however, appear to preferirradiated foods when the benefits of these foodsare endorsed by health professionals. Foodmanufacturers and retailers should offer con-sumers the choice of safety-enhanced, energy-pasteurized irradiated food.

Communicating with the PublicTo respond to consumers’ need for informa-

tion, a multifaceted program is needed. TheHACCP strategy, which teaches consumers tocritically think through the food safety process todetermine how foodborne illness could occur, hasbeen effective (19,20). The HACCP approach tohome food preservation is logical and highlightskey control points (21).

Consumers should be informed that micro-organisms are ubiquitous in the environment,found on raw products of animal or plant origin.Pathogens may survive minimal processing andpreservation treatments. People may introducepathogens during any stage of food processing orhandling, including just prior to consumption.Foodborne illness can range from mild to severeand life-threatening with chronic complications.People have control and can reduce risks.

Communicating food safety information to thepublic effectively is another challenge. Consumersobtain most of their information on food, nutrition,and science from the media; television is cited mostfrequently, and newspapers and magazines follow(22,23). Brochures enforce messages and serve asuseful references, although they are not as widelyseen as media stories.

Developing messages with the press shouldbe a primary activity of a food safety educationprogram. Consumers judge a message by thecredibility of the person conveying it, its appeal totheir common sense, and the frequency of themessage (24). Media presentations can motivatepeople to listen and change behavior. Consult-ants from the USDA hotline say, “We’ve seen an

explosion in media coverage of food safety, andcallers want more detailed explanations of thingsthey read and hear” (25).

Information on safe food handling must bemotivating and memorable. Stories that capturethe public’s attention are personal. They relatelife experiences of people with whom the publiccan identify. Stories of the consequences ofmistakes are memorable. They can be touching,humorous, or grotesque. It is easy to visualizeand remember the infected bakery worker whomade 5,000 people ill when he mixed a vat ofbuttercream frosting with his bare hands andarms despite bouts with diarrhea (26).

Stories can be heartrending, as in theexperiences of a family who lost a child to E. coliO157:H7 infection. It is difficult to document theeffectiveness of vivid accounts of doing thingsright or wrong. However, when Washington statecarried extensive coverage linking the outbreakto undercooked hamburgers, 13% of the men saidthey ate undercooked hamburger, compared with38% in Colorado (25).

ConclusionsConsumer concerns about foodborne illness

can motivate change in regulatory and industryuse of technology, product labeling, andconsumer education.

New TechnologiesThe food industry has both a right and a

responsibility to use safe and effective technology toenhance the safety of the food supply. Regulatoryauthorities should expediently evaluate andfacilitate new technologies, such as food irradiation,laser light treatment, and high-pressure process-ing, which enhance food safety. Health profession-als, the food industry, and regulators shouldchallenge special interest groups that distortinformation and strive to limit consumer choice.

Improved LabelingProcessed and packaged food should bear

labels that clearly indicate how food should behandled. Labels should include warnings aboutspecial risks to select populations. Benefits fromspecial processing that can reduce microbesshould also be encouraged.

Consumer EducationConsumers need to appreciate the serious-

ness of foodborne diseases. They must learn to


Special Issue

recognize unsafe food-handling practices, thelatency period for some microbes, and thesymptoms of foodborne disease. They need tounderstand how to protect themselves throughkitchen and personal hygiene, including thor-oughness and frequency of hand washing,temperature control, and safe food choices suchas foods processed by heat or energy pasteuriza-tion. Young people should be reached throughage-specific school curricula, such as personalhygiene and special “living skills” units thataddress food safety and diet. Food industry andhealth educators should work with the media todevelop interesting and timely messages toincrease consumer knowledge about safe foodhandling. Messages must be consistent, science-based, frequent, and personalized.

References 1. Abt Associates Inc. Trends in the United States,

consumer attitude and the supermarket 1996.Washington (DC): Food Marketing Institute; 1996.

2. Institute of Food Technologists’ Expert Panel on FoodSafety and Nutrition. Scientific status summary,foodborne illness: role of home food-handling practices.Food Technology 1995;49:119-31.

3. Williamson DM, Gravani RB, Lawless HT. Correlatingfood safety knowledge with home food-preparationpractices. Food Technology 1992;46:94-100.

4. Bryan F. Risks associated with vehicles of foodbornepathogens and processes that lead to outbreaks offoodborne diseases. Journal of Food Protection1988;51:663-73.

5. Bean NH, Griffin PM. Foodborne disease outbreaks inthe United States, 1973-1987; pathogens, vehicles, andtrends. Journal of Food Protection 1990;53:804-17.

6. Dumagan JC, Hackett JW. Almost half of the foodbudget is spent eating out. Food Review 1995;18:37.

7. Martin R. Foodborne disease threatens industry.National Restaurant News 1990;24:27,30.

8. Doyle MP. Food safety in the 21st century. Dairy FoodEnviron Sanit 1993;13:383.

9. Food Institute. Population trends 1996-2050. FairLawn (NJ): The Institute; 1996. p. 31.

10. Fein SB, Jordan CT, Levy AS. Foodborne illness:perceptions, experience, and preventive behaviors inthe United States. Journal of Food Protection1995;58:1405-11.

11. Tauxe RV. Epidemiology of Campylobacter jejuniinfections in the United States and other industrializednations. In: Nachamkin J, Blaser MJ, Tompkins LS,editors. Campylobacter jejuni: current status andfuture trends. Washington (DC): American Society forMicrobiology; 1992. p. 9-19.

12. Ollinger-Snyder P, Matthews ME. Food safety issues:press reports heighten consumer awareness ofmicrobiological safety. Dairy, Food and EnvironmentalSanitation 1994;14:580-9.

13. Schafer RB, Schafer E, Bultena GL, Hoiberg EO. Foodsafety: an application of the health belief model.Journal of Nutritional Education 1993;25:17-24.

14. Food Marketing Institute. Opinion research, trends 90,consumer attitudes & and the supermarket 1990.Washington (DC): The Institute; 1990.

15. Rhodes ME. Educating professionals and consumersabout extended-shelf-life refrigerated foods. FoodTechnology 1991;45:162-4.

16. Diehl JF. Safety of irradiated foods. New York: MarcelDekker, Inc.; 1995.

17. Bruhn CM. Consumer attitudes and market responseto irradiated food. Journal of Food Protection1995;58:175-81.

18. The Gallup Organization. Irradiation: consumers’attitudes. Princeton (NJ): Prepared for the AmericanMeat Institute; 1993.

19. Hoban T. Consumer awareness and acceptance ofbovine somatotropin. Survey conducted for the GroceryManufacturers of America; 1994.

20. Hoban T, Kendall PA. Consumer attitudes about theuse of biotechnology in agriculture and foodproduction. Raleigh (NC): North Carolina StateUniversity; 1992.

21. Bruhn CM, Peterson S, Phillips P, Sakovich N.Consumer response to information on integrated pestmanagement. Journal of Food Safety 1992;12:315-26.

22. United States Department of Agriculture. The FoodSafety Educator 1996;1:3.

23. Beware the Norwalk virus. Food Talk (Winter) 1989;4.24. Reicks J, Bosch A, Herman M, Krinke UB.

Effectiveness of a food safety teaching strategypromoting critical thinking. Journal of NutritionalEducation 1994;26:97-100.

25. Medeiros LC, George RT, Gruns K, Chandler C, CruseyS, Fittro J, et al. The safe food handling for occasionalquantity cooks curriculum. Journal of NutritionalEducation 1996;28:39-42.

26. United States Department of Agriculture, Food Safetyand Inspection Service. A margin of safety: the HACCPapproach to food safety education. Washington (DC):United States Department of Agriculture Informationand Legislative Affairs; 1990.


Special Issue

The microbiologic safety of food has beenadvanced substantially by the introduction andimplementation of Hazard Analysis and CriticalControl Points (HACCP). HACCP provides asystematic conceptual framework for identifyinghazards and focusing efforts on the properfunctioning of key food production, processing,and marketing steps. When applied appropri-ately, HACCP is a cost-effective means ofcontrolling known hazards in foods. Its success-ful implementation depends on knowledge ofsuch issues as the pathogenic microorganisms’virulence, cultural characteristics, ways in whichthey contaminate the food, effects of foodprocessing and preparation on their survival, andfood consumption patterns. Because it requiressubstantial knowledge, HACCP cannot beexpected to control unknown hazards, such asemerging foodborne pathogens. Therefore, con-trolling a new foodborne microbial threatrequires moving the hazard as quickly as possiblefrom being unknown to being known. The key tothis transition is the timely acquisition of neededresearch data. This article identifies classes ofresearch information needed and discusses aconceptual approach for addressing unknownmicrobial threats.

Anticipating the Next Emerging PathogenTwo types of emergence are encountered with

pathogenic foodborne microorganisms. A trueemergence, where a microorganism that had notbeen identified as a public health threat begins tocause disease, is relatively rare. More common isreemergence, where a known microorganismcauses disease in a new way, for example, bycausing new types of infections, being associatedwith new foods, or appearing in new geographiclocations. For both types, the operationalrequirement is to control an unanticipated publichealth threat. The timeliness of response is criticalsince the public health and economic costs of anemerging pathogen are directly related to the timebetween its emergence and its control.

The events that lead to emergence are oftencomplex, with the cause often being obscure andonly indirectly related to the new agent. Pastemergence of foodborne pathogens has beenassociated with changes in microbial genotypes,demographics, food production and processingmethods, marketing and preparation practices,medical diagnostics, globalization of the foodindustry, changes in consumer education, andgeneral socioeconomic trends (1-3). Planning fora microbial threat is a challenge because one doesnot know what the agent will be, what food it willbe associated with, or where or when emergencewill occur. While there are several potential waysof anticipating and responding to microbial

Identifying and Controlling EmergingFoodborne Pathogens: Research Needs

Robert L. BuchananUSDA ARS Eastern Regional Research Center, Wyndmoor, PA, USA

Systems for managing the risks associated with foodborne pathogens are based ondetailed knowledge of the microorganisms and the foods with which they areassociated—known hazards. An emerging pathogen, however, is an unknown hazard;therefore, to control it, key data must be acquired to convert the pathogen from anunknown to a known hazard. The types of information required are similar despite theidentity of the new agent. The key to rapid control is rapid mobilization of researchcapabilities targeted at addressing critical knowledge gaps. In addition, longer-termresearch is needed to improve our ability to respond quickly to new microbial threats andhelp us become more proactive at anticipating and preventing emergence. The type ofcontingency planning used by the military in anticipating new threats serves as a usefulframework for planning for new emergence.

Address for correspondence: Robert L. Buchanan, USDA ARSEastern Regional Research Center, 600 East Mermaid Lane,Wyndmoor, PA, 19038, USA; fax: 215-233-6445; e-mail:[email protected].


Special Issue

threats, the contingency planning used by themilitary to anticipate threats seems well suitedfor emerging pathogens. Military contingencyplanning can be viewed as having four majorcomponents: intelligence, personnel and facili-ties, rapid response, and strategic planning.

Intelligence is the gathering of medical,scientific, and other information that allowsemergence to be identified. In the United States,this role is filled to a great extent by the Centersfor Disease Control and Prevention (CDC). Inaddition to providing information on knownfoodborne pathogens, CDC works with localpublic health agencies and the medical andscientific communities to investigate new diseasesyndromes and identify unrecognized foodbornepathogens. This type of intelligence gatheringplayed a pivotal role in the recent recognition ofCyclospora as a cause of foodborne gastroenteri-tis. CDC’s new sentinel site program, FoodNet, isexpected to greatly enhance the identification ofnew foodborne diseases. However, these surveil-lance activities are largely limited to the UnitedStates, whereas an effective intelligence systemfor foodborne disease must be worldwide. Forexample, Cyclospora was identified as a likelyfoodborne or waterborne pathogen in Asia andSouth America before an outbreak was reportedin the United States. Intelligence related tofoodborne disease can be acquired from severalsources: the World Health Organization’ssurveillance program, the U.S. military’s inter-national network of laboratory and medicalinvestigators, medical and scientific reports, andthe Internet. The Internet is increasingly animportant source of intelligence related toemerging pathogens; through news groups andbulletin boards such as ProMed, scientists andpublic health practitioners share their experi-ences on almost a real-time basis. Such advancesin intelligence gathering are critical to reducingthe time between emergence and control.However, limiting intelligence to medicalconsiderations is not enough: intelligencegathering must include awareness of changesand advances in food production methods,agricultural practices and conditions, veterinarymedicine, environmental and water microbiol-ogy, food technologies, consumer trends, andgeneral socioeconomic conditions.

The second component of contingencyplanning is ensuring sufficient personnel andfacilities to characterize a new biologic agent and

develop control strategies. The inability topredict the agent or the associated food, coupledwith the degree of specialization required ofinvestigators, requires a broad range ofcapabilities and resources. However, no oneorganization is likely to maintain the capabilitiesneeded to deal with all contingencies. If we wereto follow the military pattern, we would havereserve groups that could be mobilized as needed.However, even this approach requires planningand support to ensure the needed expertise andfacilities. For example, the number of research-ers and laboratories studying Clostridiumbotulinum has dropped to a point where it wouldbe difficult to rapidly mobilize a research team,despite this pathogen’s history of reemerging in asurprisingly wide range of foods.

Rapid mobilization of resources is the thirdcomponent. This component is particularlyimportant for free-living infectious agentsbecause one goal is to limit their dissemination toprevent them from establishing secondaryreservoirs. It is much easier to fight a small,contained war than a global one. The mobiliza-tion of resources to respond to an emergencemust be appropriate to the severity of the threat.Overreacting hurts the credibility of the entiresystem, while underreacting increases both thepublic health and economic impact. Rapidresponse efforts have focused at identifyingnew agents and removing suspect food from themarketplace, two key initial steps. However,research to prevent another occurrence of theemerging pathogen has been much lessorganized and timely.

The fourth component of contingency plan-ning, strategic planning, is actually the firstchronologically. This is the phase wheremembers of war colleges pose “what we would doif” scenarios and plan appropriate responses.This type of contingency planning has generallyreceived attention in relation to emergingpathogens only in connection with the use ofbiologic warfare agents. This process relies onfuturist thinking to consider how changes insociety, economics, technology, agriculture,medicine, and international trade may affect themicrobiologic safety of the food supply. Such abroad view is needed because more generalevents or trends in society cause most diseaseemergence. This type of strategic planning isundertaken with the realization that theprobability of any specific “what if” scenario is


Special Issue

low, but the probability that one scenario willmaterialize is extremely high.

Research NeedsResearch, an integral part of responding to a

new foodborne microbial threat, is the key formoving a new or reemerging biologic agent frombeing an unknown pathogen to being one forwhich control measures are available. Two areasof research can be classified on the basis of timeconstraints. Acute research needs are deficien-cies in knowledge that must be addressed toestablish control of an emerging pathogen. Thisresearch is highly targeted and specific for themicroorganism and food of concern; it must beaccomplished as quickly as possible. Acuteneeds generally require applied research,although basic research may have to beconducted if the deficiencies in knowledge aregreat. The second class encompasses longer-term basic and applied research needs notmandatory to immediate control.

Acute NeedsWhile the data needed for any single

emerging biologic agent are highly specific, acuteresearch needs fall into general categories thatare virtually the same for all new pathogens.Common research questions include the follow-ing: Are methods available for detecting andcategorizing the agent? What food is the vehiclefor the pathogen? How do the implicated foodsbecome contaminated? What is the pathogen’sreservoir in nature? Is the pathogen’s presence incontaminated food the result of an error orbreakdown in normal controls? Does thepathogen grow in foods? Does the pathogensurvive normal food processing, distribution, andpreparation? How infectious/toxigenic is thepathogen? Are there subpopulations of consum-ers at increased risk for this pathogen? Is thepathogen’s ability to cause disease restricted tospecific strains with identifiable virulencecharacteristics? Answering these questionsrequires specific data that do not differsubstantially from pathogen to pathogen (Table).

The criteria for classifying needs as acute arereasonably straightforward: Is the researchneeded to prevent a recurrence of the disease orto modify current HACCP plans? However, thesequestions have different priorities, which dependon when the information is needed. To deal withemerging pathogens, we should learn from

modern business practices, especially the conceptof “just-in-time” research. Little considerationhas been given to how to assess and set researchpriorities for emerging foodborne pathogens. Oneattempt was provided as an appendix of the U.S.Pathogen Reduction Task Force. A relativelysimple decision tree used a series of questions to

Table. Research data needed for most emergingfoodborne pathogens

Research area Knowledge gapsDetection Sampling and enrichmentmethods techniques

CultivatingBiochemical/taxonomic char.Antibodies for capture and

differentiationSubtypingVirulence-associated

char.Detecting injured or viable-

but-nonculturable cellsMicrobial Contaminated foodsecology Reservoirs and routes of

transmissionLife cyclesGeogr. range and seasonalityRoute of contamination and

location of pathogen in foodPathogenicity Dis. char. and diagnosis

SequelaeHost rangeInfectious doseSubpopulations at riskAnimal models

Growth Free-living vs. obligatecharacteristics parasite

Growth requirementsTemperaturepHWater activityOxygen

Survival Heat resistancecharacteristics D-values

Z-valuesSusceptibility to anti-microbial food additives

Acid resistanceSensitivity to disinfectants

or dessicationSensitivity to radiation

UVIonizing

Control Effectiveness of foodpreservation

Inspection systems to segregatecontaminated materials


Special Issue

identify what research was the limiting step inresponding to the foodborne pathogen (4).

The timeliness of addressing research needsmust be an integral part of the planning process,but has been generally overlooked. Past researchmobilization efforts for new foodborne microbialthreats can be best described as haphazard,likely because they reflect the way research isfunded. The traditional means of ensuring strongresearch programs, competitive funding ofprojects after proposals have undergone exten-sive peer review, is time consuming and often notappropriate for the acute phase of responding toan emerging foodborne pathogen. Further, thepeer-review process tends not to select the oftenmundane research needed during the acutephase of an emergence. Two alternativeapproaches may be more effective. The first is tohave a series of designated laboratories that haveas part of their mission and funding the task ofbeing able to modify their research programs toaddress acute research needs. Such laboratorieswould need to have a critical mass of facilitiesand expertise in various aspects of food safetymicrobiology. The second approach is to have agroup of reserve scientists with unique expertiseor access to facilities not available at thedesignated laboratories or needed to supplementthose capabilities. Funds could be earmarked tononcompetitively fund such reserve scientists onan as-needed basis, with the understanding thatresearch needs designated as acute would takeprecedence over other research needs.

Longer-Term NeedsThe three areas of longer-term research

associated with emerging pathogens are ame-nable to more traditional funding. The first area,specific to the new pathogen, consists of researchfor improvements or alternatives to the detectionand control methods initially devised. Withinitial disease control established, basic andapplied research can seek to understand themicroorganism and develop more optimalapproaches for its prevention, control, orelimination. The second area concerns activitiesto help reduce the time between the emergence ofa pathogen and its initial control (e.g., improvedsurveillance through the development of newdiagnostic methods and further identificationand characterization of virulence determinantsand modes of pathogenicity to acceleratedetection of new agents). Just as important as

acquiring research data is rapid data dissemina-tion. The continuing development of computer-based information networks is a component ofthis second research area.

The third area focuses on identifyingresearch factors that will allow new microbialthreats to be anticipated. Of necessity, thecurrent response to emerging pathogens isalmost entirely reactive. The public healthcommunity detects a new syndrome, and onlythen is research mobilized, often during a crisis.While reactive response will always be part ofdealing with emerging microbial threats, a moreproactive approach is needed if prevention is tobe even partially realized. In military terms, waris the last resort and represents the failure ofdiplomats to predict and prevent a crisis.Microbial threats, like wars, do not spontane-ously emerge but are the result of a series ofevents or conditions. There is a need toreexamine how food is produced, processed,marketed, and prepared to identify conditionsthat contribute to emergence. For example,organic acids are used extensively throughoutthe food industry to control spoilage andpathogenic microorganisms. Archer (5) hypoth-esized that over time, exposure to pH conditionsthat stress but do not kill may lead to theemergence of hardier and possibly more virulentfoodborne pathogens. It is already well estab-lished that the induction of acid tolerance canenhance both the survival and virulence offoodborne pathogens (6). Further, one of the basictenets of microbial genetics is that conditionsthat kill most, but not all, of a bacterialpopulation foster the development of resistance.This is supported by recent studies that suggestthat bacterial stress responses may select forhypermutability (7,8). While these findings donot mean that organic acids should not be used asa tool for controlling foodborne pathogens, theysuggest that proactive research should beconducted to find ways of using these agents thatminimize the potential for resistance. Proactiveresearch, including research that might appearunrelated to the emergence of foodbornepathogens, can draw on the already substantialbody of basic research related to the conditionsand requirements for gene transfer amongbiologic agents. For example, Baur et al. (9)reported on the conditions that led to thecompetence of Escherichia coli for genetictransformation in freshwater environments.


Special Issue

Maximal competence occurred when the bacte-rium was exposed to 2 mM Ca2+ as temperaturesincreased from 10°C to 20°C. With suchinformation, researchers could examine foodprocessing operations to determine the presenceand importance of such conditions. For example,fruits and vegetables are often treated withcalcium under fluctuating temperatures toenhance the texture during later processing.

A key to being more proactive in addressing thethreat of microbial foodborne pathogens—consider-ation of root causes—will likely require foodmicrobiologists to become involved in nontradi-tional research areas. If new biologic agents arise asthe result of changes in technology, society, orglobal economics, predicting and preventingemergence will ultimately require better under-standing of how such factors influence pathogenintroduction and dissemination.

ConclusionsOne of the critical lessons of the past 10 years

is that we cannot become complacent aboutinfectious diseases (1). Only a few diseases (e.g.,smallpox) have actually been eliminated. Therest, including virtually all foodborne diseases,we hold in check, winning battles but not the war.Eventually, our weapons (e.g., antibiotics)become obsolete; pathogens (e.g., E. coli) becomemore dangerous; or we become complacent.Contingency planning must be developed andundertaken with a long-term commitment.Without that commitment and without under-standing that planning is successful whenproblems are avoided or minimized, programs of

this type lapse quickly. In the long term, the costsof planning, both in terms of economics andhuman suffering, are a fraction of thoseincurred as the result of the emergence of amajor microbial threat.

References 1. Lederberg J, Shope RE, Oaks S. Emerging infections,

microbial threats to health in the United States.Washington (DC): U.S. National Academy of SciencesPress; 1992.

2. Nathanson N. The emergence of infectious diseases:societal causes and consequences. American Society ofMicrobiology News 1996;62:83-8.

3. Potter ME. Factors for the emergence of foodbornedisease. In: Amgar A, editor. Food Safety ‘96:Proceedings of the 4th ASEPT International Conference.Laval, France: ASEPT; 1996:185-95.

4. United States Department of Agriculture.Subcommittee report for the pathogen reductiontaskforce: identifying research and education needs.Washington (DC): USDA Food Safety and InspectionService; 1994.

5. Archer DL. Preservation microbiology and safety:evidence that stress enhances virulence and triggersadaptive mutations. Trends in Food Science Technology1996;7:91-5.

6. Foster JW. Low pH adaptation and the acid toleranceresponse of Salmonella typhimurium. Crit Rev Microbiol1995;21:215-37.

7. Moxon ER, Rainey PB, Nowak MA, Lenski RE.Adaptive evolution of highly mutable loci in pathogenicbacteria. Curr Biol 1994;4:24-33.

8. LeClerc JE, Li B, Payne WL, Cebula TA. Highmutation frequencies among Escherichia coli andSalmonella pathogens. Science 1996;274:1208-11.

9. Baur B, Hanselmann K, Schlimme W, Jenni B. Genetictransformation in freshwater: Escherichia coli is ableto develop natural competence. Appl Environ Microbiol1996;62:3673-8.


Special Issue

Most countries aim to regulate food and thefood industry to ensure a safe and wholesomesupply. Although the technology of food processingis based on tradition (drying, freezing, fermenta-tion), research has resulted in a wide range ofprocesses to meet the ever-changing requirementsof consumers. Canning of low-acid foods set anenviable example of microbiologic safety. Today,consumers demand less heavily processed foodsthat contain a minimum of chemical preservatives,have a long shelf life, and are “fresh” yetmicrobiologically safe. Competitive food retailingrequires such products to be no more expensivethan existing products. In the United Kingdom,there has been a swing toward convenience foodsand foods with ethnic flavors, sometimes with apremium price. Commercial success is evident, butrelatively few of those foods have been evaluatedexperimentally for microbiologic safety. Emphasisis on risk assessment, on Hazard Analysis andCritical Control Points (HACCP) and processcontrol, all of which rely heavily on existingknowledge. That knowledge, and the raw datafrom which it is derived, could be betterorganized and easier to use.

The International Commission onMicrobiological Specifications forFoods (ICMSF)

The history of microbiologic aspects of foodcontrol is reflected in the publications of ICMSF.ICMSF was established in 1962 as foodbornedisease greatly increased microbiologic testing offoods. Increased testing, in turn, created wide-spread practical and regulatory problems in theinternational food trade. At that time, analyticalmethods were not standardized, sampling planswere of doubtful statistical validity, and theinterpretation of results applied differentconcepts of biologic significance and acceptancecriteria that confused and frustrated industryand regulatory agencies alike.

ICMSF was founded to assemble, correlate,and evaluate evidence about the microbiologicsafety and quality of foods; to consider whethermicrobiologic criteria would improve andensure the microbiologic safety of particularfoods; to propose, where appropriate, suchcriteria; and to recommend methods ofsampling and examination.

The primary purpose of the commissionremains to give guidance on appraising andcontrolling the microbiologic safety of foods,with due attention to microbiologic quality.

Maximizing the Usefulness ofFood Microbiology Research

Terry A. RobertsFood Safety Consultant; Chairman, International Commission on

Microbiological Specifications for Foods

Address for correspondence: Terry A. Roberts, Ph.D., 59Edenham Crescent, Reading RG1 6HU, United Kingdom;fax: 44-118-956-6998.

Funding for food microbiology research often follows disease outbreaks: botulismfrom vacuum-packed white-fish chubs, listeriosis from soft cheeses, or illness due toSalmonella Enteritidis or Escherichia coli. As a consequence of research, detection,identification, and subtyping methods improve, and more is learned about pathogenicityand virulence. Research also explores the organisms’ capacity to multiply or survive infood and to be killed by established or novel processes. However, rarely is there a criticaloverview of progress or trustworthy statements of generally agreed-on facts. Thatinformation is not maintained in a form that can readily be used by regulatorydepartments and the food industry to ensure a safe food supply. A centralized system isurgently needed that is accessible electronically and carries information in astandardized format on the essential properties of the organisms, includingpathogenicity, methods of detection, enumeration and identification, alternativeprevention and control methods, and growth and survival characteristics.


Special Issue

Meeting the commission’s objectives wouldfacilitate international trade; the work ofnational control agencies, the food industry,and international agencies concerned with thehumanitarian aspects of food distribution; andthe health of consumers.

An advisory body, the ICMSF provides basicinformation through extensive study and experi-ence and makes recommendations, withoutprejudice, based on that information. Results ofthe studies are published as books or papers.

At its meetings, the ICMSF functions as aworking group. Its membership consists of 18food microbiologists from 11 countries whoseinterests include research, public health, officialfood control, education, product and processdevelopment, and quality control. Members arefrom government laboratories in public health,agriculture, and food technology; universities; andthe food industry. ICMSF also seeks assistancefrom consultants specializing in particular areas ofmicrobiology. Four subcommissions (Balkan andDanubian, Latin American, Middle East andNorth African, and Southeast Asian) promoteactivities similar to those of ICMSF among foodmicrobiologists on a regional scale and facilitateworldwide communication.

ICMSF raises its own funds and pays for itsactivities and meetings. The commission ob-tained support from government agencies, theWorld Health Organization, International Unionof Microbiological Sciences and InternationalUnion of of Biological Sciences, and the foodindustry (more than 80 food companies andagencies in 13 countries).

During its first 25 years, ICMSF devoted themajor portion of its effort to methodology research,improved standardization of methods, and 17refereed publications (1). One important findingwas that in analyzing salmonellae, samples couldbe composited, making it practical to collect andanalyze the large number of samples recom-mended in some stringent sampling plans. Withthe evolution of alternative methods, rapid testkits, and the everexpanding list of biologic agentsinvolved in foodborne illness, the commissionreluctantly discontinued its program of compari-son and evaluation of methods.

The long-term objective of enhancing themicrobiologic safety of foods in internationalcommerce was addressed initially in booksrecommending uniform analytical methods (2),

sound sampling plans and criteria (3), and themicrobial ecology of foods (4,5), which familiar-ized the analyst and food technologist withprocesses used in the food industry and foodssubmitted to the laboratory. Knowledge of themicrobiology of major foods and the factorsaffecting their microbial content helps inter-pret analytical results.

At an early stage, the commission concludedthat no food sampling plan could ensure theabsence of a pathogen. Testing foods at ports ofentry, or end-product testing elsewhere in thefood chain, cannot guarantee food safety. This ledthe commission to investigate the potential valueof HACCP for enhancing food safety. A jointICMSF/World Health Organization meeting in1980 led to a report on the use of HACCP forcontrolling microbiologic hazards in food, par-ticularly in developing countries (6-8).

The commission published the principles ofHACCP and procedures for developing HACCPplans (9); the publication discussed the relativeimportance of controlling the production andharvesting, preparing, and handling of foods.Recommendations are given for the application ofHACCP at each step in the food chain—fromproduction and harvest to consumption.

The commission recognized that a majorweakness in the development of HACCP plans ishazard analysis; knowing about the manybiologic agents responsible for foodborne illnesshas become difficult. The commission’s latestbook (1) proposes to facilitate the developmentand assessment of HACCP plans, improve thesafety of foods, and facilitate risk assessment. Athorough concise review of reports on growth anddeath responses of foodborne pathogens, it isintended as a relatively quick reference manualto assist in making decisions. The commission isrevising Microbial Ecology of Foods: (Vol. 2) FoodCommodities with publication planned asMicroorganisms in Foods 6 in late 1997.

ICMSF submitted its recommendations forsampling foods and acceptance criteria forListeria monocytogenes to the Secretariat ofCodex Alimentarius in 1993. To stimulatediscussion, the recommendations were initiallyprovided to the Codex; they were subsequentlypublished (10) and revised after furtherdiscussion (11). At the request of the Secretariatof Codex, the commission developed recommen-dations for the revision of Principles for the


Special Issue

Establishment and Application of MicrobiologicalCriteria for Foods, which appears in theProcedural Manual of Codex.

Members of ICMSF believe that the originalobjectives of the commission still apply. Theformation of the European Union, the many otherworldwide political changes, the growth ofdeveloping countries seeking export markets,and the increased worldwide interest ininternational trade, as evidenced by the passageof the General Agreement on Tariffs and Tradeand the North American Free Trade Agreement,all point to the continuing need for unprejudicedguidance and advice. Import and export policiesmust be as uniform as possible and based onscientific facts. The commission will strive tomeet this goal through a combination ofeducational materials, promotion of HACCP, andrecommendations, where appropriate, of sam-pling plans and microbiologic criteria. In addition,commission members will continue to transferinformation on food safety through symposia,meetings, committee activities, and in their dailywork. The future success of ICMSF will continueto depend on extensive support from consultantsand those who provide the financial supportessential to the commission’s activities.

Research and Its Current Uses

Research Procedures and FundingIn most countries, funding for research has

become more competitive, and has evendecreased in recent years. The followingobservations are based on experience in theUnited Kingdom and the European Union.

The researcher first identifies the sources offunds and writes a proposal. If funds are granted,the research proceeds. The output may includeconfidential reports to the sponsor, refereedscientific publications, and, increasingly, a patentor a licensed process or technique. Because fundinghas so many different sources, writing proposalstakes a substantial proportion of the researchers’time. In many cases, the best researchers also writethe most persuasive proposals, but is this anefficient use of the researchers’ time?

In the United Kingdom, research generallyfalls into two categories: research funded by theparent body and research supported by acontract. The Institute of Food Research fallsunder the Biotechnology & Biological SciencesResearch Council (BBSRC), formerly the

Agriculture and Food Research Council. Aproportion of the institute’s funds comesdirectly from the BBSRC, although these fundsare declining and facing increased competitionamong BBSRC institutes.

In addition, funds may be obtained from thecentral government: the Ministry of Agriculture,Fisheries, and Food (MAFF) and the Departmentof Health, both of which have research programsfocusing on food microbiology. In addition,Scotland and Northern Ireland have separateresearch budgets, which traditionally have beenspent in those countries. Funds from thesesources sometimes support projects for 3 years,but more recently for shorter periods.

The European Union supports researchacross very wide areas, of which only a small partis food technology and food microbiology, e.g.,Food-Linked Agro-Industrial Research, nowended, and European Cooperation in Scientificand Technical Research. European Unionprograms run for 4 or 5 years, but funding isusually less than 100% (often 50% and dependenton financial support from the institute or onparallel funding from one of the above sources).

In the United Kingdom, research funds fromthe food industry are usually restricted either tovery short trouble-shooting projects or toconfidential investigations. MAFF encouragesindustrial support of research in LINK programs,with the government contributing up to 50% ofthe total funding, provided that the researchcontains elements of novelty and a consortium ofcompanies, sometimes only three or four.

Not only is the researcher faced withattracting funds from a variety of sources, but therequirements of those funding sources differgreatly. Research on genetic or physiologicreasons for phenomena, such as the unusual acidtolerance of Vero cytotoxin-producing Escheri-chia coli strains, might meet BBSRC’s stringentrequirements of scientific quality. MAFF pub-lishes a Requirements Document each year,identifying areas and topics that fit into its policyto maintain a safe and wholesome food supply.Some of the research is basic science, e.g.,developing novel methods of detecting oridentifying microbes; some is more mundane,e.g., improving the effectiveness of sanitizingsurfaces in food processes. MAFF also supports alimited program of surveillance for particularfoodborne pathogens. These funds are availablefor research institutes, universities, and other


Special Issue

laboratories, and proposals are judged by bothMAFF staff and independent assessors fromoutside. Research is organized into programs onhygienic food processing, separation and detec-tion of pathogens and their toxins, physicochemi-cal principles underlying microbial growth,growth conditions for pathogens, ending in 1997-98, and new programs on assessing microbiologichazards and risks and managing microbiologichazards and risks of salmonellae andcampylobacters in poultry. The Department ofHealth also supports food safety research—partly as surveillance and partly as programstargeted at a specific organism, e.g., E. coliO157:H7. The Department of Health alsoresponds to topics identified by the AdvisoryCommittee on Microbiological Safety of Foods.

University departments in England aresubjected to a different form of scrutiny todetermine the amount of research funding theyreceive from the central government, judged by theHigher Education Funding Council for England.The quality of publications is judged against thenumber of researchers, and the funding is directedtoward research excellence. Newly establisheddepartments find it difficult to compete.

Research OutputOutput from research has been regarded as

the report to the contractor, sometimes confiden-tial and sometimes public. More recently, fundingagencies have encouraged exploitation of re-search through patents or licensing agreements.To the researcher, expertise must be recognizedthrough refereed publications. Although inaddition to detailed papers, overviews andsometimes summaries for the industry areprovided, there is generally less consideration forthe research user. In recent years, many reviewshave become mere compilations of completedresearch. Moreover, most computerized data-bases terminate in about 1970, thus ignoring awealth of earlier information. HACCP and riskassessment would be greatly facilitated if data onmicrobial responses (growth, survival, death)were collated in a more standardized form.Details should include the food commodity (e.g.,meat, fish, vegetables, or fruit), the subgroup(e.g., for meat, beef, pork, or lamb), and theprocess applied (e.g., cooked, cured, smoked, orfermented). Details of the packaging shouldinclude the pack atmosphere (%CO2, %O2, %ballast

gas). In the longer term, it might becomeimportant to know the exact identity of theorganism including serotype, phage type, strain,and, in time, virulence factors and evensequences for 16S rRNA and for toxin genes.Other items that should be tabulated includeinoculum concentration; prior history of theinoculum, especially if that history might alter theresponse; incubation temperature(s); measure ofwater (% brine, aw); pH; other factors (e.g., nitrite,sorbate); microbiologic method(s); response (growthrate, D value, time to toxin); and estimate of thereliability of the measures of the responses.

Many readers assume that these details areincluded in most, if not all, publications.However, our experience in compiling publisheddata for comparisons with models of growth,thermal death, or survival (nonthermal death)illustrates that much desirable information islacking (12-15). Attempts to define the boundarybetween growth and no growth are not new (16-19). Despite different methods and strains ofmicrobes, compilations of reported data oftenshow a trend (Figure).

Research UsersDespite an overall neglect of the research

user’s needs, some useful listings are available

sp.g

r.(1/

h)

sp.gr.(1/h)

Figure. Specific growth rate of Listeria monocytogenespredicted by different models and independent observers(provided by J. Baranyi, Institute of Food Research,Reading Laboratory). Estimates of specific growth ratefrom published papers are shown vertically. Theresponse surfaces were generated from predictivemodels and are not a surface fitted to the values.


Special Issue

(20) as are helpful overviews and discussions (21-26). There remains, however, an opportunity forthe researcher to make a product more attractiveto the research user, e.g., the food industry (bothnational and international), regulatory bodies,and informed consumers.

Additional Research NeedsRequired research includes developing a

database of reliable information on microbialresponses to food processing conditions. Thedatabase needs to be easy to access and use. Inaddition, because understanding of growthresponses and thermal death is greater than ourunderstanding of survival, i.e., little or no loss ofviability over many months, we need to have abetter understanding of the factors determiningthat survival and of the underlying microbialphysiology. In recent years, we have been facedwith such new microbiologic challenges asSalmonella Enteritidis Phage Type 4, multidrug-resistant S. Typhimurium Definitive Type 104,verocytotoxigenic E. coli, Shiga-like toxin-produc-ing E. coli, and increasingly vancomycin-resistantEnterococcus faecium, and we lack the capacity toanticipate emergence of new pathogens. Finally,new technologies make it possible to study geneticstability and variation in populations, perhapsidentifying why particular strains are, or become,more virulent and more pathogenic.

Control of Food SafetyFood safety has traditionally been controlled

by inspection and compliance with ordinances orcodes of practice. Much faith has been placed insampling plans and associated microbiologiccriteria. However, this approach is retrospective,depending on the microbiologic examination ofproduct that has often been dispatched orconsumed. The number of samples that can beexamined statistically is too small to detect lowlevels of “defectives.” HACCP was a substantialstep forward, with the concept that safety couldbe designed into a food process. However, HACCPdemands substantial knowledge about thecharacteristics of the food, the process, and themicrobes of concern. The knowledge required toimplement HACCP effectively is not organizedinto systems that are easy to use.

References 1. International Commission on Microbiological

Specifications for Food. Microorganisms in foods. 5.Characteristics of microbial pathogens. London:Blackie Academic & Professional;1996. p. 511-3.

2. International Commission on MicrobiologicalSpecifications for Food. Microorganisms in foods. 1.Their significance and methods of enumeration. 2nded. Toronto: Univ. of Toronto Press;1978.

3. International Commission on MicrobiologicalSpecifications for Food. Microorganisms in foods. 2.Sampling for microbiological analysis: principles andspecific applications. 2nd ed. Toronto: Univ. of TorontoPress;1986.

4. International Commission on MicrobiologicalSpecifications for Food. Microbial ecology of foods. Vol.1. Factors affecting life and death of microorganisms.New York: Academic Press;1980.

5. International Commission on MicrobiologicalSpecifications for Food. Microbial ecology of foods. Vol.2. Food commodities. New York: Academic Press;1980.

6. International Commission on MicrobiologicalSpecifications for Food. Report of the WHO/ICMSFmeeting in hazard analysis: critical control pointsystem in food hygiene. Geneva: World HealthOrganization;1982. VHP/82.37.

7. International Commission on MicrobiologicalSpecifications for Food. Prevention and control offoodborne salmonellosis through application of thehazard analysis critical control point system. Report ofan International Commission on MicrobiologicalSpecifications for Foods (ICMSF) ad hoc committee.Geneva: World Health Organization;1986. WHO/CDS/VPH/86.65.

8. Simonsen B, Bryan FL, Christian JHB, Roberts TA,Tompkin RB, Silliker JH. Report from the InternationalCommission on Microbiological Specifications of Foods(ICMSF). Prevention and control of food-bornesalmonellosis through application of hazard analysiscritical control point (HACCP). Int J Food Microbiol1987;4:227-47.

9. International Commission on MicrobiologicalSpecifications for Food. Microorganisms in foods. 4.Application of the Hazard Analysis Critical ControlPoint (HACCP) System to ensure microbiologicalsafety and quality. London: Blackwell ScientificPublications, Ltd.;1988.

10. International Commission on MicrobiologicalSpecifications for Food. Choice of sampling plan andcriteria for Listeria monocytogenes. Int J FoodMicrobiol 1994;22:89-96.

11. van Schothorst M. Sampling plans for Listeriamonocytogenes. Food Control 1996;7:203-8.

12. Graham AF, Mason DR, Peck MW. Predictive model ofthe effects of temperature, pH and sodium chloride ongrowth from spores of non-proteolytic Clostridiumbotulinum. Int J Food Microbiol 1996;31:69-85.

13. Sutherland JP, Bayliss AJ. Predictive modelling ofgrowth of Yersinia enterocolitica—the effects oftemperature, pH and sodium chloride. Int J FoodMicrobiol 1994;21:197-215.


Special Issue

14. Sutherland JP, Bayliss AJ, Roberts TA. Predictivemodelling of growth of Staphylococcus aureus: the theeffects of temperature, pH and sodium chloride. Int JFood Microbiol 1994;21:217-36.

15. Sutherland JP, Bayliss AJ, Braxton DS. Predictivemodelling of growth of Escherichia coli O157:H7: theeffects of temperature, pH, and sodium chloride. Int JFood Microbiol 1995:25:29-49.

16. Riemann H, Lee WH, Genigeorgis C. Control ofClostridium botulinum and Staphylococcus aureus insemi-preserved meat products. Journal of Milk andFood Technology 1972;35:514-23.

17. Bean PG. Developments in heat treatment processesfor shelf-stable products. Food microbiology: advancesand prospects. In: Roberts TA, Skinner FA, editors.London: Academic Press; 1983. p. 97-112.

18. Brown KL. Use of a computer data file for storage ofheat resistance data on bacterial spores. J ApplBacteriol 1988;65:49-54.

19. Brown KL. (1992) Heat resistance of bacterial spores[dissertation]. Nottingham,England: Univ ofNottingham;1992.

20. Bryan FL. Diseases transmitted by foods: aclassification and summary. 2nd ed. Atlanta (GA):Centers for Disease Control and Prevention;1982.

21. Council for Agricultural Science and Technology.Foodborne pathogens: risks and consequences. Ames(IA): the Council; 1994; Task Force Rep. No. 122.

22. Advisory Committee on the Microbiological Safety ofFood. Report on vacuum packaging and associatedprocesses. London: HMSO; 1992.

23. Advisory Committee on the Microbiological Safety ofFood. Interim report on Campylobacter. London:HMSO;1993.

24. Advisory Committee on the Microbiological Safety ofFood. Report on Salmonella in eggs. London:HMSO;1993.

25. Advisory Committee on the Microbiological Safety ofFood. Report on verotoxin-producing Escherichia coli.London: HMSO;1995.

26. Advisory Committee on the Microbiological Safety ofFood. Report on poultry meat. London: HMSO;1996.


Special Issue

Emerging infectious diseases can be definedas infections that have newly appeared in apopulation or have existed but are rapidlyincreasing in incidence or geographic range (1).Agents for which a particular route oftransmission is newly recognized and agents(previously unidentifiable) that are now knownbecause of advances in detection methods shouldalso be included in this definition. Advances inepidemiologic and detection methods during thelast 10 to 20 years have placed food andwaterborne human enteric viruses and protozoalparasites within this category.

Human enteric viruses and protozoa areparasitic agents that replicate in the intestines ofinfected hosts and are excreted in the feces. Ingeneral, the viruses are limited to human hosts,while the parasitic agents (in the form of cysts oroocytes) have a variety of human and nonhumananimal hosts. Both are transmitted primarily bythe fecal-oral route, and as a result, the majorsource of contamination for foods and water isthrough contact with human and animal fecalpollution. This contamination may occur directly,through contaminated meat carcasses or poorpersonal hygiene practices of infected food

handlers, or indirectly, through contact withfecally contaminated water or other cross-contamination routes. Viruses and parasitesdiffer from foodborne bacterial pathogens inimportant ways. Because they are environmen-tally inert, they do not replicate in food, water, orenvironmental samples. Additionally, unlikebacterial pathogens, human enteric viruses andprotozoal parasites are environmentally stable(2), are resistant to many of the traditionalmethods used to control bacterial pathogens (2),and have notably low infectious doses (3,4). Thisallows virtually any food to serve as a vehicle fortransmission and enables these agents towithstand a wide variety of commonly practicedfood storage and processing conditions (2,5).

Epidemiology

Human Enteric VirusesHuman enteric viruses are increasingly

recognized as important causes of foodborneillness. A recent report issued by the Council forAgricultural Science and Technology rankedhuman enteric viruses as fifth and sixth amongidentified causes of foodborne disease in theUnited States (6). A review of U.S. nationalsurveillance data for 1979 showed that 14 (44%)of 32 foodborne disease outbreaks in institutionalsettings were epidemiologically typical of viral

Epidemiology and Detection as Optionsfor Control of Viral and Parasitic

Foodborne Disease

Lee-Ann JaykusNorth Carolina State University, Raleigh, North Carolina, USA

Address for correspondence: Lee-Ann Jaykus, Department ofFood Science, North Carolina State University, Box 7624,Raleigh, NC 27695, USA; fax: 919-515-7124; e-mail:[email protected].

Human enteric viruses and protozoal parasites are important causes of emergingfood and waterborne disease. Epidemiologic investigation and detection of the agentsin clinical, food, and water specimens, which are traditionally used to establish the causeof disease outbreaks, are either cumbersome, expensive, and frequently unavailable orunattempted for the important food and waterborne enteric viruses and protozoa.However, the recent introduction of regulatory testing mandates, alternative testingstrategies, and increased epidemiologic surveillance for food and waterborne diseaseshould significantly improve the ability to detect and control these agents. We discussnew methods of investigating foodborne viral and parasitic disease and the future ofthese methods in recognizing, identifying, and controlling disease agents.


Special Issue

gastroenteritis (7). Viral gastroenteritis wasreported as the most common foodborne illness inMinnesota from 1984 to 1991, predominantlyassociated with poor personal hygiene of infectedfood handlers (8). Furthermore, recent dataindicate that 10% of the 4,617 outbreaks offoodborne disease of unconfirmed etiologyreported from 1973 to 1987 met at least two of theclinical criteria for outbreaks of acute viralgastroenteritis (8,9). The apparent failure toconfirm a viral etiology in such outbreaks hasbeen due largely to the lack of available tests andthe reluctance of public health officials to useepidemiologic criteria in the classification offoodborne viral disease (9-11). The unavailabilityof food specimens and the failure to reportoutbreaks of mild gastrointestinal disease havealso contributed to reporting difficulties. All ofthese factors have resulted in a drasticunderestimate of the true scope and importanceof foodborne viral infection (8).

The most common types of foodborne viraldisease are infectious hepatitis due to hepatitis Avirus and acute viral gastroenteritis associatedwith the Norwalk agent and other related small,round-structured gastrointestinal viruses (12).The human enteroviruses may also be transmit-ted by foodborne routes (13) and are the mostcommonly isolated agents in surveys of naturallyoccurring viral contamination in foods (14).Foodborne outbreaks due to small round viruses,parvoviruses, and astroviruses are occasionally

reported (15). Rotaviruses, some adenoviruses,and hepatitis E virus are important causes ofwaterborne disease outbreaks, particularly indeveloping countries (12). Foodborne outbreaksassociated with human enteric viruses arealmost always due to the consumption of fecallycontaminated raw or undercooked shellfish andready-to-eat products contaminated by infectedfood handlers (12). Postrecovery and secondarytransmission are a concern (16,17). Recentoutbreaks are summarized in Table 1 (18-24).

Parasitic ProtozoaSince 1981, enteric protozoa have become the

leading cause of waterborne disease outbreaksfor which an etiologic agent could be determined(5). A recent study reported that 21% of drinkingwater–associated outbreaks between 1991 and1992 were attributable to parasitic agents (25).Furthermore, these agents are frequent contami-nants of potable water supplies (26,27). Thepotential for transmission of these agents byfoodborne routes is increasingly recognized(28). For instance, from 1988 to 1992, sevenfood-associated outbreaks of giardiasis, com-prising 184 cases, were reported in the UnitedStates (29). However, since foodborne trans-mission is only recently documented and morethan half of all reported foodborne diseaseoutbreaks have undetermined etiology, thetrue importance of foodborne transmission ofparasitic protozoa is unknown.

Table 1. Recent outbreaks of foodborne viral disease

Agent Location Date No. Cases Food Confirmationa Ref.HAV Shanghai, Jan. 1988 300,000 Raw clams Yes 18

China (4% total (IEM, Hybridization, population) Cell culture,

Experimental infection)HAV U.S. (AL, GA, July-Aug. 61 Raw oysters Yes 19

FL, TN, HI) 1988 (Antibody capture- RT-PCR)

SRSV U.S. (LA) Nov. 1993 40 Raw oysters No 21SRSV U.S. (LA, MD, Nov. 1993 180 Raw/steamed No 22

MS, FL, NC) oystersSRSV U.S. (GA) Dec. 1994 34 clusters Steamed/ No 22

roasted oystersSRSV U.S. (FL, TX) Jan. 1995 3 Oysters No 20Norwalk U.S. (DE) Sept. 1987 191 Commercial ice No 23Norwalk U.S. (CO) July, 1988 1440 Celery/ No 24

chicken saladaConfimation of the virus, viral antigen, or viral nucleic acid in food specimensHAV = hepatitis A virus; IEM = immune electron microscopy; RT-PCR = reverse transcriptase polymerase chain reaction;SRSV = small round-structured gastrointestinal virus.


Special Issue

The most common human enteric parasiticinfections in the United States are caused byCryptosporidium parvum and Giardia lamblia.Cyclospora is also an emerging enteric protozoonthat has recently been associated with theconsumption of contaminated fruit (30). Largecommunitywide waterborne outbreaks of para-sitic protozoa are usually associated with surfacewater supplies that are either unfiltered orsubjected to inadequate flocculation and filtra-tion (5). Two large waterborne outbreaks haveoccurred in the United States within the last 10years (31,32); one of these was the largest recordedwaterborne disease outbreak in U.S. history (32).

LimitationsMost of the information about viral and

parasitic food and waterborne disease comesfrom outbreak investigations by state and localhealth departments and surveillance programsdirected by the Centers for Disease Control andPrevention (CDC). However, since many of thesediseases are not reportable and surveillance isbased on voluntary reporting by state healthdepartments, the magnitude of this diseaseproblem is underestimated. This is exacerbatedby a reluctance to use epidemiologic criteria inthe classification of foodborne viral disease andthe failure to report mild outbreaks ofgastrointestinal disease. Furthermore, sinceinvestigation generally follows an outbreak,important information and samples may havebeen destroyed, consumed, or lost to inaccuraterecall. Since many of these outbreaks are smalland confined, epidemiologic investigation may belimited by the resources available to state andlocal health departments. Difficulties in investi-gating and reporting are further complicated bythe fact that the enteric protozoa cause commonopportunistic infections in the immunocompro-mised, and the role of foods in these diseases hasnot been studied. Likewise, the role of thefoodborne transmission route in sporadic diseaseand the importance of carrier states andsecondary illness after a primary foodbornedisease outbreak are poorly characterized.

Detection

Clinical SamplesIllness caused by human enteric viruses can

be suspected epidemiologically by considering

incubation period and illness duration analysis,classic viral gastroenteritis symptoms, and theabsence of bacterial or parasitic pathogens instool samples (10,11). Laboratory confirmation ofhuman enteric viral infection has been based on arise in specific antibody to the virus, oralternatively, the demonstration of virus par-ticles, antigen, or nucleic acid in stools. Thedetection methods most often applied to clinicalsamples have included immune electron micros-copy, radioimmunoassay, and enzyme immu-noassay (33). The usefulness of these assays hasbeen reduced by low detection limits (>104-105

particles/ml) and the inability to cultivate Norwalk-like viruses in vitro, which has limited the supplyof viral antigen available for developing reagents(8). In addition, the Norwalk agent is only one ofseveral small round-structured gastrointestinalviruses that cause outbreaks with similar clinicaland epidemiologic features (8).

Before 1981, parasitic disease in humans wasdiagnosed histologically by identifying the life cyclestages of parasitic agents in the intestinal mucosa(34). More recently, clinical diagnosis has involvedmethods to concentrate parasitic agents from stoolspecimens followed by a variety of fluorescent orimmunofluorescent staining techniques and subse-quent microscopic examination (34). Serodiagnos-tic methods have been developed (35,36), butadditional evaluations are needed to confirm thediagnostic utility of these methods.

Environmental, Food, and Water SamplesFailure to confirm a viral and parasitic

etiology in foodborne outbreaks has also been dueto the lack of adequate methods to detect thecausative agents in environmental samples. Likebacterial pathogens in food and water, virusesand parasites are frequently present in smallnumbers. However, unlike traditional foodmicrobiologic techniques, which have relied oncultural enrichment and selective plating toincrease cell numbers and differentiate patho-gens in background microflora, techniques todetect human enteric viruses and parasiticprotozoa require live mammalian cells forgrowth. For this reason, standard methods todetect enteric bacteria in foods cannot be used;instead, detection requires an initial concentra-tion step, often from large volumes of food orwater, followed by mammalian cell cultureinfectivity assay or immunofluorescent staining.


Special Issue

Concentration methods are usually cumbersome,and yields are less than optimal. Both cell cultureinfectivity and immunofluorescent staining areexpensive and slow and require highly trainedpersonnel. Furthermore, mammalian cell culturelines are largely unavailable for the epidemio-logically important foodborne viruses. Alterna-tive detection methods based on immunologicand molecular methods have been reported;recent methodologic developments have focusedon overcoming barriers to detection, such asimproving recovery efficiencies and detectionlimits and preventing inhibitions due to food-related compounds.

Human Enteric Viruses in Foods

ConcentrationTwo general schemes for the concentration of

human enteric viruses from foods have beenreported: extraction-concentration methods andadsorption-elution-concentration methods (Fig-ure). The general purpose of concentration is toprovide a high recovery of infectious virus in alow-volume aqueous solution free of cytotoxicmaterials. Both schemes employ conditionsfavoring the separation of viruses from shellfishtissues (most of the developmental work has usedshellfish as model food commodities), primarilythrough the use of filtration, precipitation,polyelectrolyte flocculation, and solvent extrac-tion. While either method can be used,adsorption-elution-precipitation methods havebeen favored in recent years (2,37). Virus yieldsafter concentrations are 10% to 90% (38).

DetectionTraditional methods to directly detect

viruses in foods after concentration have beenbased on the ability of enteric viruses to infectlive mammalian cells in culture. Quantal andenumerative methods using a variety ofmammalian cell culture lines, generally fromprimate kidneys, have been reported. Suchapproaches have been limited because levels ofcontaminating virus generally are low (1-200infectious units per 100 grams of shellfish) (39),residual food components interfere with assays(38), and the epidemiologically important virusesdo not replicate (small round-structured gas-trointestinal viruses) or replicate poorly (hepati-tis A virus) in mammalian cell culture (2).Alternative methods such as enzyme-linked

immunosorbent assay (ELISA) and DNA/RNAprobes have been reported but are limited by highdetection limits (>103 infectious units), unavail-ability of reagents, and poor sample quality (2).These difficulties are illustrated by the confirma-tion of viral contamination in a food in only tworeported instances (18,19).

The in vitro enzymatic amplification methodof polymerase chain reaction (PCR) offers anopportunity to enrich a single specific nucleicacid sequence up to a millionfold and henceprovides a sensitive and specific method with atheoretical detection limit of one virus unit. Thismethod is readily adaptable to the detection ofRNA viruses by preceding the PCR with a briefreverse transcription (RT) step, hence thedesignation RT-PCR. The recent cloning of theNorwalk agent and related small round-structured gastrointestinal viruses has providedan opportunity to develop effective molecular

Figure. General steps in the isolation of human entericviruses and parasitic protozoa from foods.


Special Issue

detection methods for these previouslynondetectable agents (40,41).

The application of PCR methods to thedetection of human enteric viruses in foods is anarea of active research. However, the develop-ment of such methods is complicated by low levelsof contamination, high sample volumes, and thepresence of food components, which mayinterfere with enzymatic amplification reactions.To address these issues, three alternativeapproaches have been used to simultaneouslyreduce sample volumes and the level ofinterfering compounds. The most frequentlyapplied approach involves isolating and purify-ing nucleic acids (RNA) from the food samplebefore RT-PCR (42-47). A second approachcombines capture of the virus with specificantibody followed by nucleic acid amplificationby using RT-PCR (19,48). In the third approach,the intact virus particle is concentrated andpurified from the complex food matrix resultingin sample volume reduction and removal ofinhibitors, followed by subsequent heat release ofviral nucleic acid from the virion capsid and RT-PCR (49-51). All three methods have beenapplied to various shellfish species, and in somecases, to other food commodities and naturallycontaminated field shellfish specimens. Themethods are summarized in Table 2, along withoptimized virus detection levels (19,42-51). Acombined approach was reported by Chung et al.(51), who successfully detected human enterovi-ruses and hepatitis A virus in naturallycontaminated oyster samples after viral amplifi-cation in mammalian cell cultures. Despiteenormous strides in the ability to detect humanenteric viruses with PCR, the technique is stilllimited by the absence of effective concentrationmethods, the presence of enzymatic inhibitors,and the inability to distinguish betweeninfectious and noninfectious virions.

Enteric Parasitic Agents

ConcentrationLike viruses, parasitic protozoa are usually

present in low concentrations in contaminatedwater and hence must be concentrated from largevolumes of water before detection (Figure).Giardia cysts and Cryptosporidium oocysts areconcentrated from 100 to 1,000 liters of water byfiltration through yarn-wound filters. Retainedparticulates are eluted from the filters and

reconcentrated by centrifugation. The pelletedcysts and oocysts are then separated fromparticulate debris by flotation on Percoll-sucrosegradients, followed by subsequent detection.With this method, a concentrate of less than 5 mlcan be provided for final detection (52). Whilerecovery efficiencies as high as 100% have beenreported (53), recovery is generally poor andgreatly affected by water quality and particulatematter (53,54). Alternative methods usingcalcium carbonate precipitation can concentrateCryptosporidium oocysts from water withconcentration efficiencies as high as 63% (55).

Table 2. Emerging detection methods for human entericviruses in foods

Detection Field

Pathogen Sample limit app. Ref.Nucleic Acid Extraction/RT-PCRHAV Clams 2000 particles/g No 44

(10 PFU/g)Poliovirus Oysters 38 PFU/20 g No 42

(2 PFU/g)HAV Clams/ 100 PFU/1.5 g No 43

Oysters (67 PFU/g)Norwalk Clams/ 5-10 PCRU/1.5 g No 43

Oysters (3-7 PCRU/g)Poliovirus Oysters/ 10 PFU/5 g No 45

Mussels (2 PFU/g)SRSV Oysters/ Not Specified Yes 46

MusselsPoliovirus, Clams 100 PFU/50 g No 50 HAV (2 PFU/g)Norwalk Clams 1000 PCRU/50 g No 50

(20 PCRU/g)Norwalk Various 20-200 PCRU/10 g No 47

(2-20 PCRU/g)

Antibody-Capture/RT-PCRHAV Clams/ Not Specified No 48

OystersHAV Oysters Not Specified Yes 19

Virion ConcentrationPoliovirus, Oysters 10 PFU/50 g Yes 49, HAV (0.02 PFU/g) 51Norwalk Oysters 4500 PCRU/50 g No 49

(90 PCRU/g)Poliovirus, Clams 1000 PFU/50 g No 50 HAV (20 PFU/g)Norwalk Clams 100 PCRU/50 g No 50

(2 PCRU/g)HAV = hepatitis A virus; RT-PCR = reverse transcriptasepolymerase chain reaction; PCRU = PCR-amplifiable units;SRSV = small round-structured gastrointestinal virus.


Special Issue

DetectionTraditional methods to detect parasitic

agents from water sample concentrates havebeen based on immunofluorescent staining offiltered sample concentrates (52). The sampleconcentrate is filtered through cellulose acetatefilters and commercial kits that use fluoresceinisothiocyante–labeled monoclonal antibodies ap-plied for immunofluorescent staining. Thestained filters are examined under an ultravioletmicroscope, and cysts and oocysts are classifiedaccording to immunofluorescence, size, shape,and internal morphologic characteristics. Theresults are reported as presumptive andconfirmed cysts and oocysts per 100 liters ofwater (52). Confirmation is based on the ability tovisualize organelles under light microscopy. Thismethod is extremely limited because it is time-consuming and expensive, requires highly skilledpersonnel, and does not indicate viability of thecysts or oocysts; in addition, cross-reactions ofmonoclonal antibodies with algal cells and debrisinterfere with the interpretation of results (56).

Because of the limitations of immunofluores-cence assays, alternative strategies for thedetection of protozoal parasites in environmentaland water samples are being sought. Most of theapproaches use the traditional methods ofconcentration, in conjunction with alternativedetection methods. In many cases, the inclusion/exclusion of fluorogenic dyes is used to enhancemorphologic examination; in particular, 4,6-diamidino-2-phenyl indole and propidium iodidehave been used to facilitate detection and alsoassess viability (57). Two recent methodologicdevelopments include cell sorting/particle count-ing approaches and molecular approaches. Inmany cases, a combination of approaches is used.Particle-counting approaches include the Fluo-rescence-Activated Cell Sorting system, a laser-based particle counter that is able to simulta-neously sort particles, sense fluorescence, anddetermine size (58). This system is being used todetect Giardia and Cryptosporidium in watersamples in England and Australia (55).Differentially stained cysts and oocysts may bevisualized microscopically with cooled chargecouple devices (58,59). Molecular approachesthat apply DNA hybridization to the detection ofGiardia species have been reported (60). Amethod that combines fluorescence and in situhybridization with confocal microscopy has beenapplied to both detect and speciate Giardia cysts

(61). PCR methods have been developed to detectGiardia and Cryptosporidium (62-64). WhilePCR methods have the potential to detect onesingle infectious unit and may be applied todiscriminate pathogenic from nonpathogenicspecies, they remain limited because of enzy-matic inhibition, the inability to discriminatebetween viable and nonviable organisms, and thecurrent absence of quantitative assay.

Several methods in development are combi-nations of multiple detection approaches. Acombined method, designated the electrorotationassay, couples filtration and subsequent elutionwith affinity immunocapture by using paramag-netic beads. By inducing an electric field on amicroscope with a special stage attachment, theorganism-bead complexes rotate in a characteris-tic pattern that enables detection of parasiticprotozoa (52). Several investigators are alsoworking on methods that couple cell cultureinfectivity with immunostaining, thereby provid-ing detection with simultaneous indication ofviability (M. Sobsey, pers. comm.). ClinicalELISA kits have been evaluated for use inenvironmental water samples with reporteddetection limits of fewer than 10 cysts or oocyts(65); however, cross-reaction with algae contin-ues to be a problem. Emerging detectionapproaches are summarized in Table 3 (58-64).Methods to concentrate and detect parasiticprotozoa specifically from foods are underdevelopment.

ConsiderationsAlthough prototype alternative and rapid

methods to detect human enteric viruses andparasitic protozoa in foods and water have beenreported, multiple barriers must be overcomebefore these methods are applicable to routinemonitoring. To obtain sample representation anddetection sensitivity adequate for the low levelsof contamination found with these entericpathogens in naturally contaminated environ-mental specimens, large sample volumes of foodand water need to be processed. While some of thedetection methods begin with large samples,many do not consider sample size, which limitsthe sensitivity of the assay procedure from thebeginning. The approaches then applied toconcentrate and purify the pathogens from thesamples do not only need to be reasonablyefficient but also need to produce a concentratelow in volume and free of inhibitory compounds.


Special Issue

To complicate matters further, different foodcommodities have differing types and levels ofinhibitors, many of which have remainedrecalcitrant to almost all removal processes.

The relationship between detection bymolecular and immunologic approaches and thesubsequent viability or infectivity of these entericpathogens remains a concern. While investiga-tors have addressed these issues using in vitroexcystation, inclusion and exclusion of vital dyes,animal infectivity (parasitic protozoa) (57,66,67),and assays combining mammalian cell cultureinfectivity with alternate detection strategies(viruses) (51; M. Sobsey, pers. comm.), thesemethods are not useful for the nonculturablehuman enteric viruses; they are expensive, time-consuming, and not readily amenable to routinediagnostic work. Establishing quantitative de-tection methods also needs further research.

The development of detection methods is alsolimited by the current state of knowledge. Forinstance, while recent sequencing evidenceindicates that the Norwalk-like small round-structured gastrointestinal virus group consistsof multiple members having the same physicaland genomic characteristics as other viruses inthe family Caliciviridae (40,41), considerablesequence and antigenic diversity remain among

the members of this group (68-70). While PCRprimers for these genetically diverse agents havebeen reported (46,71), development of universaldetection methods is clearly limited until morecomplete information about this group of humanenteric viruses is available.

Research is needed to develop and refine theprototype protocols into collaboratively testedmethods that could be routinely and expeditiouslyused to evaluate the microbiologic safety of foodproducts. In general, future research needs for theroutine application of alternative methods to detectenteric viral and parasitic protozoal contaminationin foods requires development of the following: 1)simple, rapid, and cost-effective extraction andconcentration procedures; 2) simple and reliablemethods for the removal of inhibitors; 3) methodsthat are not restricted by food product; and 4)quantitative approaches for assessing the relativelevels of contamination.

ConclusionsImproved epidemiologic surveillance, pre-

dominantly through the creation of population-based centers that will focus on the epidemiologyand prevention of food and waterborne infectiousdisease (72), should improve knowledge regard-ing viral etiology in foodborne disease outbreaks

Table 3. Emerging detection methods for parasitic protozoa in water

Pathogen Detection limit Viability Differentiation Ref.

Flow Cytometry with Fluorescent Imaging and CCD

Cryptosporidium NR Yes, using differential NR 58,fluorogenic vital dyes 59

DNA Hybridization

Giardia 1 cyst NR No 60 (16s-like rDNA)Giardia NR NR Yes-in situ hybridization 61 (16s-like rDNA) with differential fluorescence

G. lamblia; G. muris

Polymerase Chain Reaction

Giardia 1 cyst Partial, using mRNA NR 62 (Giardin gene) as target; Depends on

inactivation methodGiardia 1 cyst NR Yes-primer sequence 63 (Giardin gene) G. duodenalisCryptosporidium 20 oocysts No Yes-by hybridization 64 (Target not specified) C. parvum

CCD = cooled charge couple device; NR = not reported


Special Issue

and the clinical and economic importance of viraland parasitic disease agents. Data obtained fromthese studies may elucidate the role of foods insporadic disease as well as in secondary spread.

Current research programs under way atCDC, U.S. Department of Agriculture, and Foodand Drug Administration, as well as extramuralfunding through the National Institutes ofHealth and U.S. Department of Agricultureprograms, should promote the development,testing, and dissemination of rapid and accuratedetection methods for viral and parasiticfoodborne disease agents. Such research willcontinue to be important as testing regulations,such as the landmark Environmental ProtectionAgency Information Collections RequirementRule, and the new U.S. Department ofAgriculture Pathogen Reduction: Hazard Analy-sis and Critical Control Points (HACCP) SystemsRule, emerge in the future. From a clinicalstandpoint, the development of inexpensive andwidely available reagents has been improvedthrough recent developments in molecularbiology (33). This will increase the number offacilities able to perform diagnostic testing,which will ultimately facilitate epidemiologicinvestigation. From a food safety standpoint,improved detection methods should eventuallyprovide a regulatory option for monitoring foodsafety, particularly in the economically impor-tant shellfish species and in detecting viralcontamination due to human handling, orparasitic protozoal contamination from animalwastes. The development of rapid detectionmethods will also aid in the evaluation of controlstrategies for viral and protozoal contaminationof foods. For instance, depuration, thermalprocessing, and irradiation for the control offoodborne viral contamination in shellfish needto be further evaluated. In the case ofcontamination by infected food handlers andanimal wastes, the availability of rapid methodswill increase our understanding of this transmis-sion route and help determine critical controlpoints for HACCP approaches to control thetransmission of these foodborne disease agents.This will allow the integration of a true farm-to-table food safety approach for the control of viraland parasitic food and waterborne disease.

In the United States, there is currently aclear regulatory mandate to evaluate food safetyrisks within a systematic conceptual frameworkin the form of quantitative microbial risk

assessment. The application of quantitative riskassessment to food safety issues has beenhampered by the lack of available data regardingprevalence, transmission, infectious dose, andbehavior of microorganisms in foods; this hasbeen particularly true for emerging pathogens.Improved epidemiologic and detection methodsshould dramatically affect the ability to evaluatefood safety risks through risk assessmentstrategies. For instance, by using emergingepidemiologic data in conjunction with epidemio-logic modeling, statistical overview analysis(meta-analysis), and geographic informationsystems, scientists should be able to improvehazard analysis and exposure assessment andprovide a clearer picture of disease transmissionpatterns. More rapid, accurate, and readilyavailable detection methods should allowdetermination of prevalence of contaminationand disease, and when quantitative, shouldprovide a means of assessing dose-responserelationships. Together these will improvesubsequent risk characterization, allowing regu-lators to quantitate the magnitude of theproblem, evaluate risk reduction strategies, andprioritize competing risks.

The combination of increased surveillance,improved detection methods, and testing require-ments should result in a marked improvement inthe ability to detect, investigate, and control foodand waterborne enteric viral and parasiticprotozoal agents. Taken together, these ap-proaches promise to provide increased informa-tion necessary to assess risks, control disease,and ultimately improve public health in the nextcentury.

References 1. Morse SS. Factors in the emergence of infectious

disease. Emerg Infect Dis 1995;1:5-10. 2. Jaykus LA, Hemard MT, Sobsey MD. Human enteric

pathogenic viruses. In: Hackney CR, Peirson MD,editors. Environmental indicators and shellfish safety.New York: Chapman and Hall; 1994.

3. DuPont HL, Chappell CL, Sterling CR, Okhuysen PC,Rose JB, Jakubowski W. The infectivity ofCryptosporidium parvum in healthy volunteers. NEngl J Med 1995;332:855-9.

4. Haas CN. Estimation of risk due to low doses ofmicroorganisms: a comparison of alternative methodologies.Am J Epidemiol 1983;118:573-82.

5. Moe CL. Waterborne transmission of infectious agents.In: Hurst CJ, Knudsen GR, McInerney MJ,Stenzenbach LD, Walter MV, editors. Manual ofenvironmental microbiology. Washington (DC):American Society for Microbiology; 1996.


Special Issue

6. Council for Agricultural Science and Technology.Foodborne pathogens: risks and consequences.Washington (DC): Library of Congress No. 122; 1994.

7. Jenkins S, Horman JT, Isreal E, Cukor G, BlacklowNR. An outbreak of Norwalk-related gastroenteritis ina boys’ camp. American Journal of Diseases of Children1985;139:787-9.

8. Hedberg CW, Osterholm MT. Outbreaks of foodborneand waterborne viral gastroenteritis. Clin MicrobiolRev 1993;6:199-210.

9. Bean NH, Griffin PM. Foodborne disease outbreaks inthe United States, 1973-1987: pathogens, vehicles, andtrends. Journal of Food Protection 1990;53:804-17.

10. Kaplan JE, Feldman R, Campbell DS, Lookabaugh C,Gary W. The frequency of a Norwalk-like pattern ofillness in outbreaks of acute gastroenteritis. Am JPublic Health 1982;72:1329-32.

11. Kaplan JE, Gary GW, Baron RC, Singh N, SchonbergerLB, Feldman R, Greenberg HB. Epidemiology ofNorwalk gastroenteritis and the role of Norwalk virusin outbreaks of acute nonbacterial gastroenteritis. AnnIntern Med 1982;96:756-76.

12. Cliver DO, Ellender RD, Fout GS, Shields PA, SobseyMD. Foodborne viruses. In: Vanderzant C, SplittstoesserDF, editors. Compendium of methods for themicrobiological examination of foods. Washington(DC): American Public Health Association; 1992.

13. Centers for Disease Control. Foodborne and waterbornedisease outbreaks annual summary 1976. Washington(DC): U.S. Department of Health, Education, andWelfare Publication No. CDC 78-8185; 1978.

14. Gerba CP, Goyal SM. Detection and occurrence ofenteric viruses in shellfish: a review. Journal of FoodProtection 1978;41:743-54.

15. Appleton H. Small round viruses: classification androle in food-borne infections. In: Ciba FoundationSymposium. Novel Diarrhoea Viruses. Chinchester,England: Wiley Publishers; 1987.

16. Heun EM, Vogt RL, Hudson PJ, Parren S, Gary GW.Risk factors for secondary transmission in householdsafter a common-source outbreak of Norwalkgastroenteritis. Am J Epidemiol 1987;126:1181-6.

17. White KE, Osterholm MT, Mariotti JA, Korlath JA,Lawrence DH, Ristinen TL, Greenberg HB. Afoodborne outbreak of Norwalk virus gastroenteritis:evidence for post-recovery transmission. Am JEpidemiol 1986;124:120-6.

18. Xu ZY, Li ZH, Wang JX, Xiao ZP, Dong DH. Ecology andprevention of a shellfish-associated hepatitis A epidemic inShanghai, China. Vaccine 1992;(Suppl 1):S67-8.

19. Desenclos JCA, Klontz KC, Wilder MH, Nainan OV,Margolis HS, Gunn RA. A multistate outbreak ofhepatitis A caused by the consumption of raw oysters.Am J Pub Health 1991;81:1268-72.

20. Centers for Disease Control and Prevention. Multistateoutbreak of viral gastroenteritis associated withconsumption of oysters-Apalachicola Bay, Florida,December 1995-January 1995. MMWR Morb MortWkly Rpt 1995;44:37-9.

21. Centers for Disease Control and Prevention. Viralgastroenteritis associated with consumption of rawoysters-Florida, 1993. MMWR Morb Mort Wkly Rpt1994;43:446-9.

22. Centers for Disease Control and Prevention. Stateoutbreak of viral gastroenteritis related to consumptionof oysters-Louisiana, Maryland, Mississippi, andNorth Carolina, 1993. MMWR Morb Mort Wkly Rpt1993;42:945-8.

23. Cannon RO, Poliner RJ, Hirshhorn RB, RodeheaverDC, Silverman PR, Brown EA, et al. A multistateoutbreak of Norwalk virus gastroenteritis associatedwith consumption of commercial ice. J Infect Dis1991;164:860-4.

24. Warner RD, Carr RW, McClesky FK, Johnson PC, GoldyElmer LM, Davison VE. A large nontypical outbreak ofNorwalk virus: gastroenteritis associated with exposingcelery to nonpotable water and with Citrobacter freundii.Arch Intern Med 1991;151:2419-24.

25. Centers for Disease Control and Prevention.Surveillance for waterborne disease outbreaks-UnitedStates, 1991-1992. MMWR Morb Mort Wkly Rpt1993;42:1-22.

26. LeChevallier MW, Norton WD, Lee RG. Giardia andCryptosporidium spp. in filtered drinking watersupples. Appl Environ Microbiol 1991;57:2617-21.

27. Rose JB, Gerba CP, Jakubowski W. Survey of potablewater supplies for Cryptosporidium and Giardia.Environ Sci Technol 1991;25:1393-400.

28. Bean NH, Griffin PM, Goulding JS, Ivey CB. Foodbornedisease outbreaks, 5-year summary, 1983-1987. MMWRMorb Mort Wkly Rpt 1990;39(SS-1):15-57.

29. Bean NH, Goulding JS, Lao C, Angulo FJ. Surveillancefor foodborne disease outbreaks-United States, 1988-1992. MMWR Morb Mort Wkly Rpt 1996;45(SS-5):1-66.

30. Ortega YR, Sterling CR, Gilman RH, Cama VA, Diaz F.Cyclospora species—A new protozoan pathogen ofhumans. N Engl J Med 1993;328:1308-12.

31. Hayes EB, Matte TD, O’Brien TR, McKinley TW,Logsdon GS, Rose JB, et al. Large community outbreakof cryptosporidiosis due to contamination of a filteredpublic water supply. N Engl J Med 1989;320:1372-6.

32. MacKenzie WR, Hoxie NJ, Proctor ME, Gradus MS,Blair KA, Peterson DE, et al. A massive outbreak inMilwaukee of Cryptosporidium infection transmittedthrough the public water supply. N Engl J Med1994;331:161-7.

33. Kapikian AJ, Estes MK, Chanock RM. Norwalk groupof viruses. In: Fields BN, Knipe DM, Howley PM,editors. Fields Virology. Philadelphia: Lippencott-Raven Publishers, Inc.; 1996.

34. Current WL, Garcia LS. Cryptosporidiosis. ClinMicrobiol Rev 1991;4:324-58.

35. Campbell PN, Current WL. Demonstration of serumantibodies to Cryptosporidium sp. in normal andimmunodeficient humans with confirmed infections. JClin Microbiol 1983;18:165-9.

36. Ungar BLP, Soave R, Fayer R, Nash TE. Enzymeimmunoassay detection of immunoglobulin M and Gantibodies to Cryptosporidium in immunocompetentand immunocompromised persons. J Infect Dis1986;153:570-8.

37. DeLeon R, Jaykus L. Detection of bacteria and virusesin shellfish. In: Hurst CJ, Knudsen GR, McInerney MJ,Stenzenbach LD, Walter MV, editors. Manual ofenvironmental microbiology. Washington (DC):American Society for Microbiology; 1996.


Special Issue

38. Sobsey MD, Carrick RJ, Jensen HR. Improved methodsfor detecting enteric viruses in oysters. Appl EnvironMicrobiol 1978;36:121-8.

39. Cole MT, Kilgen MB, Reily LR, Hackney CR. Detectionof enteroviruses, bacterial indicators, and pathogens inLouisiana oysters and their overlying waters. Journalof Food Protection 1986;49:596-601.

40. Jiang X, Graham DY, Wang K, Estes MK. Norwalkvirus genome cloning and characterization. Science1990;250:1580-3.

41. Lambden PR, Caul EO, Ashley CR, Clarke IN.Sequence and genome organization of a human smallround-structured (Norwalk-like) virus. Science1993;259:516-9.

42. Atmar RL, Metcalf TG, Neill HF, Estes MK. Detectionof enteric viruses in oysters by using the polymerasechain reaction. Appl Environ Microbiol 1993;59:631-5.

43. Atmar RL, Neill HF, Romalde JL, LeGuyader F,Woodley CM, Metcalf TG, Estes MK. Detection ofNorwalk virus and hepatitis A virus in shellfish tissueswith the PCR. Appl Environ Microbiol 1995;61:3014-8.

44. Goswami BB, Koch WH, Cebula TA. Detection ofhepatitis A in Mercenaria mercenaria by coupledreverse transcription and polymerase chain reaction.Appl Environ Microbiol 1993;59:2765-70.

45. Lees DN, Henshilwood K, Dore WJ. Development of amethod for detection of enteroviruses in shellfish byPCR with poliovirus as a model. Appl EnvironMicrobiol 1994;60:2999-3005.

46. Lees DN, Henshilwood K, Green J, Gallimore CI,Brown DWG. Detection of small round structuredviruses in shellfish by reverse transcription-PCR. ApplEnviron Microbiol 1995;61:4418-24.

47. Gouvea V, Santos N, Carmo Timenetsky M, Estes MK.Identification of Norwalk virus in artificially seededshellfish and selected foods. J Virol Methods1994;48:177-87.

48. Deng MY, Day SP, Cliver DO. Detection of hepatitis Avirus in environmental samples by antigen-capturePCR. Appl Environ Microbiol 1994;60:1927-33.

49. Jaykus L, DeLeon R, Sobsey MD. A virionconcentration method for detection of human entericviruses in oysters by PCR and oligoprobe hybridization.Appl Environ Microbiol 1996;62:2074-80.

50. Dix A. Development of methods to extract humanenteric viruses from hard-shelled clams for detection byreverse transcriptase-polymerase chain reaction (RT-PCR) and oligoprobe hybridization (OP) [thesis].Raleigh (NC): North Carolina State University; 1997.

51. Chung H, Jaykus L, Sobsey MD. Improved detection ofhuman enteric viruses in field oyster specimens by invivo and in vitro amplification of nucleic acids. ApplEnviron Microbiol 1996;62:3772-8.

52. Schaefer III FW. Detection of protozoan parasites insource and finished drinking waters. In: Hurst CJ,Knudsen GR, McInerney MJ, Stenzenbach LD, WalterMV, editors. Manual of Environmental Microbiology.Washington (DC): American Society for Microbiology;1996.

53. LeChevallier MW, Norton WD, Siegel JE, AbbaszadeganM. Evaluation of the immunofluorescence procedurefor detection of Giardia cysts and Cryptosporidiumoocyts in water. Appl Environ Microbiol 1995;61:690-7.

54. Nieminski EC, Schaefer III FW, Ongerth JE.Comparison of two methods for the detection of Giardiacysts and Cryptosporidium oocytes in water. ApplEnviron Microbiol 1995;61:1714-9.

55. Versey G, Slade JS, Byrne M, Shepard K, Fricker CR. Anew method for the concentration of Cryptosporidiumoocysts from water. J Appl Bacteriol 1993;75:82-6.

56. Rodgers MR, Flanigan DJ, Jakubowski W. Identificationof algae which interfere with the detection of Giardiacysts and Cryptosporidium oocysts and a method foralleviating this interference. Appl Environ Microbiol1995;61:3759-63.

57. Campbell AT, Robertson LJ, Smith HV. Viability ofCryptosporidium parvum oocytes: correlation of invitro excystation with inclusion/exclusion of fluorogenicvital dyes. Appl Environ Microbiol 1992;58:3488-93.

58. Campbell AT, Robertson LJ, Smith HV. Novelmethodology in the detection of Cryptosporidium parvum:a comparison of cooled charge coupled device (CCD) andflow cytometry. Water Science Technology 1993;27:89-92.

59. Campbell AT, Haggart R, Robertson LJ, Smith HV.Fluorescent imaging of Cryptosporidium using a cooledcharge couple device (CCD). J Microbiol Methods1992;16:169-74.

60. Abbaszadegan M, Gerba CP, Rose JB. Detection ofGiardia cysts with a cDNA probe and applications towater samples. App Environ Microbiol 1991;57:927-31.

61. Erlandsen SL, VanKeulen H, Gurien A, JakubowskiW, Schafer III FW, Wallis P, et al. Molecular approachto speciation and detection of Giardia: fluorochromerDNA probes for identification of Giardia lamblia,Giardia muris, and Giardia ardeae in laboratory andenvironmental samples by in situ hybridization. In:Thompson RCA, Reynoldson JA, Lymbery AJ, editors.Giardia: from molecules to disease. Oxford: CABInternational; 1994.

62. Mahbubani MH, Bej AK, Perlin M, Schaefer III FW,Jakubowski W, Atlas RM. Detection of Giardia cysts byusing polymerase chain reaction and distinguishing livefrom dead cysts. Appl Environ Microbiol 1991;57:3456-61.

63. Mahbubani MH, Bej AK, Perlin M, Schaefer III FW,Jakubowski W, Atlas RM. Differentiation of Giardiaduodenalis and other Giardia spp. by using polymerasechain reaction and gene probes. J Clin Microbiol1992;30:74-8.

64. Webster KA, Pow JDE, Giles M, Catchpole J,Woodward MJ. Detection of Cryptosporidium parvumusing a specific polymerase chain reaction. VetParasitol 1993;50:35-44.

65. de la Cruz AA, Sivaganesan M. Detection of Giardiaand Cryptosporidium spp. in source water samples bycommercial enzyme-immunosorbant assay. In:Proceedings 1994 Water Technology Conference; 1994Nov 6-10; San Francisco, California. Denver (CO):American Water Works Association, 1995.

66. Sauch JF, Flanigan D, Galvin ML, Berman D,Jakubowski W. Propidium iodide as an indicator ofGiardia cyst viability. Appl Environ Microbiol1991;57:3243-7.

67. Schupp DG, Erlandsen SL. A new method to determineGiardia cyst viability: correlation of floresceindiacetate and propidium iodide staining with animalinfectivity. Appl Environ Microbiol 1987;53:704-7.


Special Issue

68. Moe CL, Gentsch J, Ando T, Grohmann G, Monroe SS,Jiang X, et al. Application of PCR to detect Norwalkvirus in fecal specimens from outbreaks ofgastroenteritis. J Clin Microbiol 1994;32:642-8.

69. Green J, Norcott JP, Lewis D, Arnold C, Brown DWG.Norwalk-like viruses: demonstration of genomicdiversity by polymerase chain reaction. J ClinMicrobiol 1993;31:3007-12.

70. Wang J, Jiang X, Madore HP, Gray J, Desselberge U,Ando T, et al. Sequence diversity of small, round-structured viruses in the Norwalk virus group. J Virol1994;68:5982-90.

71. Ando T, Monroe SS, Gentsch JR, Jin Q, Lewis DC,Glass RL. Detection and differentiation of antigenicallydistinct small round-structured viruses (Norwalk-likeviruses) by reverse transcription-PCR and southernhybridization. J Clin Microbiol 1995;33:64-71.

72. Centers for Disease Control and Prevention. Addressingemerging infectious disease threats: a preventionstrategy for the United States. Atlanta (GA): U.S.Department of Health and Human Services, U.S.Public Health Service; 1996.


Special Issue

Ensuring the microbial safety and shelf life offoods depends on minimizing the initial level ofmicrobial contamination, preventing or limitingthe rate of microbial growth, or destroyingmicrobial populations. With many foods, thesestrategies have been practiced successfully forthousands of years. However, in the last decade,the incidence of foodborne disease has increasedin the industrialized world (1), despite theintroduction of the Hazard Analysis and CriticalControl Points (HACCP) concept and thepromulgation of regulations in food safety. Theincreased incidence of foodborne disease iscaused by changes in agricultural and foodprocessing practices, increasing internationaltrade in foods, and social changes (which includechanged eating habits and increased populationmobility) (2).

This article develops the propositions thatavailable quantitative information, properlyapplied, is a basis for improved food safety; theinformation available, largely an empiric descrip-tion of microbial behavior in foods, highlights alack of understanding of the physiology offoodborne pathogens; and knowledge of thephysiology may lead to more precise control offoodborne bacteria and novel protocols to ensurethe microbiologic safety of foods.

Characteristics of BacteriaBacteria have inhabited the earth for

approximately three and a half billion years andhave colonized almost every conceivable habitat(3). In fact, the development of microbialpopulations is probably precluded only whenliquid water is absent or conditions are soextreme that rapid denaturation of proteinsoutpaces their rate of replacement. The majorcharacteristics that underpin the success ofprokaryotes are small size and ease of dispersal,physiologic diversity, and tolerance of extreme

Quantitative Microbiology:A Basis for Food Safety

T. A. McMeekin, J. Brown, K. Krist, D. Miles, K. Neumeyer,D.S. Nichols, J. Olley, K. Presser, D. A. Ratkowsky, T. Ross,

M. Salter, and S. SoontranonUniversity of Tasmania, Hobart, Australia

Because microorganisms are easily dispersed, display physiologic diversity, andtolerate extreme conditions, they are ubiquitous and may contaminate and grow in manyfood products. The behavior of microbial populations in foods (growth, survival, or death)is determined by the properties of the food (e.g., water activity and pH) and the storageconditions (e.g., temperature, relative humidity, and atmosphere). The effect of theseproperties can be predicted by mathematical models derived from quantitative studieson microbial populations. Temperature abuse is a major factor contributing to foodbornedisease; monitoring temperature history during food processing, distribution, andstorage is a simple, effective means to reduce the incidence of food poisoning.Interpretation of temperature profiles by computer programs based on predictive modelsallows informed decisions on the shelf life and safety of foods. In- or on-packagetemperature indicators require further development to accurately predict microbialbehavior. We suggest a basis for a “universal” temperature indicator. This articleemphasizes the need to combine kinetic and probability approaches to modeling andsuggests a method to define the bacterial growth/no growth interface. Advances incontrolling foodborne pathogens depend on understanding the pathogens’ physiologicresponses to growth constraints, including constraints conferring increased survivalcapacity.

Address for correspondence: Professor T.A. McMeekin, Depart-ment of Agricultural Science, University of Tasmania, GPO Box252-54, Hobart, TAS 7001, Australia; fax: 61-3-62-26-2642; e-mail:[email protected].


Special Issue

conditions (4). The temperature range over whichmicrobial populations develop is -12°C (thetemperature at which intracellular waterfreezes) to +112°C (the temperature at whichliquid water is maintained only under elevatedpressure). The pH range is pH 1 to pH 12, and thesalinity range is zero to saturated (4).

Langeveld et al. (5), who studied microbialdevelopment in biofilms in a tubular heatexchanger used to pasteurize milk, report theexploitation of different ecologic niches bybacteria. Through the ascending temperaturerange of the tube (~20°C to ~90°C), the dominantmicrobiota changed from gram-negative bacteriasuch as Acinetobacter, to coliforms to Streptococ-cus thermophilus to thermophilic bacilli. At thehighest temperature, the wall of the exchangerwas colonized by Thermus thermophilus. Thus, itappears that contaminants deposited along thelength of the tube were selected by the in situtemperature, with the fastest-growing organismdominating.

Factors Affecting Microbial Behavior inFoods

Most studies in food microbiology areconcerned with the rapid growth of populations,but in many ecosystems, the survival characteris-tics of the population also need to be considered.The longevity of bacterial spores and theirresistance to harsh conditions are well docu-mented. However, the ability of vegetative cellsto resist stressful conditions is increasinglyrecognized as an important ecologic trait (6).Attention also needs to be given to relativelyslow-growing populations in various situations,e.g., when the shelf life of a product is extended bycontrol of rapidly growing spoilage organisms.The behavior of foodborne microorganisms, be itthe growth or death of microbial populations, isbased on the time of exposure to environmentalfactors affecting population development; forexample, equivalent kills of bacteria in milk areachieved by low temperature–long time pasteur-ization (60°C/30 min) and high temperature–short time pasteurization (72°C/15 sec). Whenpopulations are in the biokinetic range, the rateat which they develop is determined by factorssuch as temperature, water availability, and pHapplied in food preservation procedures. Theextent of microbial growth is a function of thetime the population is exposed to combinations ofintrinsic food properties (e.g., salt concentration

and acidity) and extrinsic storage conditions (e.g.,temperature, relative humidity, and gaseousatmosphere).

Different factors assume dominance indifferent foods and preservation strategies. Inmany foods, the full preservation potential of asingle property is restricted because of consider-ations related to the esthetic, organoleptic, andnutritional properties of the product. However,several properties or conditions may be combinedto provide a desired level of stability (7). Insituations where the preservation strategy isdesigned to slow the rate of population growth,the effect will always be increased by storagetemperature. Temperature control in processing,distribution, and storage (the cold chain) iscrucial to ensure the adequate shelf life andsafety of many common foods, including meat,fish, poultry, and milk. Newer technologies,including modified atmosphere packaging andsophisticated products such as sous-vide meals,do not obviate the need for strict temperaturecontrol. Indeed, the requirement for vigilanceincreases with increased shelf life and thepossibility of growth of psychrotrophic pathogensover an extended period.

Predictive MicrobiologyPredictive microbiology involves knowledge

of microbial growth responses to environmentalfactors summarized as equations or mathemati-cal models. The raw data and models may bestored in a database from which the informationcan be retrieved and used to interpret the effect ofprocessing and distribution practices on micro-bial proliferation. Coupled with information onenvironmental history during processing andstorage, predictive microbiology provides preci-sion in making decisions on the microbiologicsafety and quality of foods. The term “quantita-tive microbial ecology” has been suggested as analternative to “predictive microbiology” (8).

The development, validation, and applicationof predictive microbiology has been extensivelyreviewed in the last decade (9,10). Modelingstudies have concentrated on descriptions of theeffect of constraints on microbial growth (ratherthan survival or death), often using a kineticmodel approach (rather than probability model-ing) and most often describing the effect oftemperature as the sole or one of a number ofcontrolling factors. For example, the temperaturedependence model for growth of Clostridium


Special Issue

botulinum demonstrated a good fit to data, butthe authors noted that “care must be taken atextremes of growth, as no growth may beregistered in a situation where growth is indeedpossible but has a low probability” (11).

The emphasis in modeling efforts ontemperature (often in combination with otherfactors) may be justified, given its crucial role inthe safe distribution and storage of foods.Surveys carried out over several decades in theUnited Kingdom, United States, Canada, andAustralia point to the predominant role oftemperature abuse in outbreaks of foodbornedisease (12-14).

Problems with Predictive Microbiologyand Research Needs

Several commonly perceived problems withpredictive modeling (8) are reviewed below.While considerable progress has been made indefining philosophic approaches and experimen-tal protocols for growth model development andmany models have been developed and pub-lished, more validation studies are required,particularly involving independent and industry-based trials. More emphasis should be placed onmodeling the death kinetics of foodbornepathogens with low infective doses.

Measurement of environmental factors (e.g.,temperature) can be achieved with precision, butin some situations, (e.g., in chilling of meatcarcasses), it is more difficult. Location of thesensor can be an important consideration (15,16).In meat chilling, where control of microbialdevelopment is a function of the combined effectsof falling temperature and water activity,development of a technique to measure wateractivity in situ at the carcass surface wouldprovide valuable information. Furthermore,development of techniques to measure con-straints such as water activity, pH, or redoxpotential on a microscale might provide usefulinformation for a complex food such as salami.This would allow definition of the role of themicroenvironment in determining microbialbehavior. The concept of a microenvironment iswell developed in soil microbiology (17) but hasbeen neglected in food microbiology.

The inherent variability of response times(generation time and lag phase duration) as anissue in predictive microbiology was first raisedby Ratkowsky et al. (18), who related the varianceof responses on a transformed rate scale such as

V(√k) or V(lnk) to the variances of responses on atime (θ) scale. The variance was shown to beproportional to the square or cube of the responsetime. It was later confirmed (19-21) thatnonnormal gamma and inverse Gaussian distri-butions described the distribution of responsetimes in predictive microbiology, which indicatethat variance is proportional to the square orcube of the response time, respectively.

The practical implication of these findings forthe application of kinetic models is that inherentbiologic variability increases markedly withincreasing response times, and thus theconfidence limits associated with predictions alsoincrease markedly. However, if the probabilitydistribution of the response time is known, onecan determine the probability that an organismwill grow more quickly than a predicted responsetime (21). Thus, kinetic models are appropriate todescribe consistent microbial growth responses,but under extreme conditions a probabilityapproach may be required.

Models are normally developed under staticconditions (growth rates and lag times aremeasured at a series of set temperatures, wateractivity values, and pH levels), and the resultsare combined to describe the effects of each factoror a combination of factors on populationdevelopment. Subsequently, models must bevalidated in foods under conditions that mimicsituations encountered in normal practice, e.g.,decreasing temperature and water activityduring active chilling of meat carcasses orfluctuating temperatures during the distributionand storage of many food commodities.

Shaw (22) and later several other authors(23-26) reported on the effect of fluctuatingtemperatures. Depending on the magnitude ofthe temperature deviation, the organism maychange its rate of growth to a rate characteristicof the new temperature, or it may stop growing ifa lag phase is introduced. In both situations,Salmonella Typhimurium has responded nearlyas predicted by the model (24,25). Baranyi et al.(26) presented similar results for the spoilagebacterium Brocothrix thermosphacta. Whencycled between 25°C and 5°C, the modelpredicted behavior well in both the growth rateand lag phase duration. However, a temperatureshift from 25°C to 3°C caused deviations frommodel predictions due to decline in viable cellnumbers or extended lag phases. During the finalextended phase of growth at 2.8°C, the rate


Special Issue

resumed at that predicted. Baranyi et al. (26) alsoexamined the perceived problem of modeling lagphase duration. The difficulty in predicting lagphase duration in foods is not due to the lack of asuitable model: the difficulty comes from the lackof knowledge of the physiologic status of theorganisms contaminating the food. The organ-isms may include cells that are actively growing,exhibiting a physiologic lag phase, damaged andunder repair, exhibiting physiologic (endospores)or exogenous dormancy (VNC cells), damaged butunable to reproduce because of ineffective repairmechanisms, and dead. At least part of theconfusion surrounding the measurement of lagphase duration arises from experiments in whichinocula of different physiologic statuses wereused (27,28).

Methods to define the physiologic status offoodborne contaminants under various condi-tions need to be developed. This will requireobservations on individual cells or smallpopulations of cells either directly by microscopyor an indicator of single-cell metabolic activity(26). Luminescent Salmonella strains have beenused as real-time reporters of growth andrecovery from sublethal injury (29). Alterna-tively, a parameter to describe the suitability ofcells to grow in a new environment may beincorporated in the model (26).

Current Status of Predictive MicrobiologySome problems with predictive microbiology

have been discussed, and, for each, neededresearch has been suggested. Opinions vary onthe efficacy of models to predict outcomes underreal life conditions. At one end of the scale,accuracy such as that described for the growth ofPseudomonas in minced beef (Figure 1) can be

obtained in trials conducted independently of thelaboratory in which the model was developed. Atthe other end, models developed in laboratorybroth systems have been reported to beinappropriate to describe growth on food (30).Dalgaard (31) reported similar discrepancies andsuggested an iterative approach to modeldevelopment using food, rather than laboratorymedia, as the growth substrate for modeldevelopment. Such reports emphasize the needfor rigorous validation of models under practicalconditions. Deviations from predictions do notnecessarily imply that the model is defective butmore likely that knowledge of some foodecosystems is incomplete and factors other thanthose used in model development have an effecton microbial behavior.

The common theme of the problems inpredictive microbiology discussed above is that ofuncertainty—uncertainty in terms of the startingconditions (e.g., initial microbial numbers andtypes) and the microbial response in a given orchanging environment. Uncertainty translates tovariability if the distribution of response times isunderstood and the variance can be described. Aswe have indicated above, the variabilityassociated with very long response times limitsthe utility of kinetic models and requires aprobability approach. Thus, while in the lastdecade predictive modelers were justified in theirselection of temperature as a primary factor tomodel in kinetic approaches, the next decade maysee a return to probability modeling as pioneeredby Genigeorgis (32) and Roberts (33). This shiftwill derive impetus from the emergence ofdangerous pathogens with very low infectivedoses, and continued kinetic modeling willconcentrate on survival and death rather thangrowth of populations.

The first kinetic death model to findwidespread use in the food industry was forthermal destruction (34). One can consider amodel describing a 12-log cycle reduction in C.botulinum spores in a short time withconsiderable certainty. However, as we movetoward less severe processes with longerresponse times and the added complications of“shoulders” and “tails” to define the growth/nogrowth interface, biologic variability will againdictate a probability approach to describe thesurvival and slow decline of microbial popula-tions.

Figure 1. Validation of Pseudomonas predictor in mincedbeef . Printed with permission of G. Thomson, Defence ForceFood Science Centre, Scottsdale, Tasmania, Australia.


Special Issue

Defining the Growth/No Growth InterfaceBecause growth of pathogenic bacteria in

foods always increases the risk for foodbornedisease, defining the conditions at which nogrowth is possible is of considerable practicalsignificance for food manufacturers and regula-tors. Bacterial growth/no growth interfacemodels quantify the combined effect of varioushurdles on the probability of growth and definecombinations at which the growth rate is zero.Increasing the level of one or more hurdles at theinterface by only a small amount will signifi-cantly increase the probability of “fail safe”events and decrease the probability that a fewcells in the population will resolve the lag phaseand begin to grow (a “fail dangerous” event) (7).The growth/no growth interface also has greatphysiologic significance because at that pointbiosynthetic processes are insufficient to supportpopulation growth, and survival mechanisms arein place.

A procedure to derive the interface wasproposed by Ratkowsky and Ross (35); it employsa logistic regression model to define theprobability of growth as a function of one or morecontroling environmental factors. From thismodel, the boundary between growth and nogrowth, at some chosen level of probability, canbe determined. The form of the expressioncontaining the growth limiting factors issuggested by a kinetic model, while the responseat a given combination of factors is eitherpresence or absence (i.e., growth/no growth) orprobabilistic (i.e., the fraction of positiveresponses in n trials). This approach representsan integration of probability and kineticapproaches to predictive modeling.

Microbial Responses to Stress andMicrobial Physiology

Bacteria have physiologic mechanisms en-abling them to survive in environments thatpreclude their growth. While some tolerance toenvironmental insults is adaptive, a wide rangeof protective mechanisms is induced when cellsenter a stationary phase or become starved.These phenomena are under the control of therpoS gene, which codes for a stationary-phase–specific sigma factor, expression of which triggersthe development of a semidormant state in whichbacteria can better resist multiple physicalchallenges (36). This factor and the gene products

whose expression it controls are of vitalsignificance to food microbiology; they form thebasis for a global stress response in which onestress can confer protection to a wide range ofother stresses. Under the influence of this factor,bacterial cells respond very quickly to unfavor-able changes in their environment. Often theseresponses are phenotypic and remain in placeonly during stress (37).

Low pH ToleranceBrown et al. (37) demonstrated “acid

habituation” (38), a phenotypic response to anenvironmental insult, for five strains of Escheri-chia coli. These strains showed a wide range ofintrinsic acid tolerance, which for each strain wasenhanced by exposure to nonlethal acidity (pH 5)before exposure to a lethal acid challenge (pH 3).Neutralization of the growth medium partiallyreversed tolerance to acid stress, underliningthat acid habituation is a phenotypic response.Furthermore, acid tolerance was correlated withchanges in the fatty acid composition of the cellmembrane. During acid habituation,monounsaturated fatty acids (16:1w7c and18:1w7c) in the phospholipids of E. coli wereeither converted to their cyclopropane deriva-tives (cy17:0 and cy19:0) or replaced by saturatedfatty acids. The degree of acid tolerance of the fivestrains of E. coli was highly correlated with themembrane cyclopropane fatty acid content,which may enhance the survival of cells exposedto low pH.

Low Water Activity ToleranceBacterial cells, when confronted by lowered

water activity, regulate the internal environmentby rapidly accumulating compatible solutes suchas glycine betaine or carnitine (39). The solutes,which may be scavenged by the cell, exert theirinfluence at very low concentrations; the effect isdemonstrated both in limits and rate growth.These compounds appear also to providecryotolerance as well as osmotolerance (40;Figure 2).

Energy DiversionMicrobial responses to stressful conditions

may constitute a drain on the energy resources ofthe cell, e.g., in relation to the accumulation ofcompatible solutes (41). Knochel and Gould (42)argued “that restriction of the availability ofenergy will interfere with a cell’s reaction to


Special Issue

osmotic stress.” The energy diversion hypothesiswas supported by McMeekin et al. (9) on the basisof observations on the growth of Staphylococcusxylosus at nine different levels of water activity.Though Tmin, the theoretic minimum temperaturefor growth, remained constant, the actualminimum temperature at which growth wasobserved increased with decreasing wateractivity, suggesting energy expenditure to copewith aw stress.

Krist and Ross (unpub. data), however,challenged this explanation because of findingsfrom growth rate and yield experiments on E. coligrowing in a glucose-limited minimal mineralsmedium. With both decreasing temperature andwater activity, the growth rate declinedgradually, but the yield was not greatly affecteduntil close to the point where growth ceased. Asthe substrate was converted to the same amountof biomass, this suggests that the stressesimposed by suboptimal temperature or wateractivity are not a major drain on the cell’s energyreserves. Compatible solutes likely amelioratethe effect of both factors by maintaining enzymesin an active configuration (39). With pH, thegrowth rate of many organisms is unaffectedacross a wide range of pH values. To maintainintracellular pH, the cell uses considerableenergy to export protons (43), and thus it is

anticipated that yield will be affected.Increasing knowledge of the physiologic

response of bacterial cells to individual con-straints and combinations of constraints willprovide greater precision in defining growth-limiting conditions and possibly allow develop-ment of novel protocols to ensure the microbialsafety of foods. As an example, the remarkableeffect of compatible solutes on the growth rateand growth range conditions is an obviousadvantage for bacteria growing under stressfulconditions (Figure 2). Compatible solutes, such asbetaine and carnitine, are widely distributed infoods of plant and animal origin and are easilyavailable to bacteria and rapidly taken up byspecific transport mechanisms (39,40). It isunlikely that growth might be controlled by“creating a hostile environment devoid ofosmolytes,” as suggested by Smith (44). However,it might be possible to use the specific uptakemechanism to deliver a compatible soluteanalogue with lethal effects on the cell.Alternatively, if the cell is moved from anenvironment in which growth is possible to onewhere growth ceases, compatible solutes mayalso allow enzymic reactions to continue withinthe cell, depleting energy reserves and inducing agreatly extended lag phase or death. Thishypothesis is supported by the observations ofmany authors that survival is better at low ratherthan ambient temperatures. For example,Clavero and Beuchat (45) state, “Regardless ofthe pH and aw, survival of E. coli O157:H7 wasbetter at 5°C than at 20°C or 30°C.” Furthermore,preliminary experiments in this laboratorysuggest that E. coli declines more rapidly in thepresence of betaine than in its absence (Krist,unpub. data).

Application of Predictive ModelsThe incorporation of predictive models into

devices such as temperature loggers has beendescribed for E. coli (46) and Pseudomonas (47),as has the development of expert systems frompredictive modeling databases (48,49).

The current food poisoning crisis indicatesthat existing quantitative information on micro-bial growth, survival, and death, if properlyapplied, would have an immediate impact on theincidence of foodborne disease in the industrial-ized world. Even without the synthesis of datainto mathematical models, simply logging thetemperature history of food processing, distribu-

50403020100.02

0.04

0.06

0.08

0.10

0.12

0.14 Low salt only

Low salt + betaine

High salt only

High salt + betaine

Temperature (°C)

Squ

are

root

gro

wth

rat

e

Figure 2. Effect of betaine on the growth of Escherichiacoli in glucose-minimal medium. Without added NaClthe growth rate yield and minimum growthtemperature are the same with and without betaine.With 4% NaCl the growth rate and yield are lowerwithout betaine and the actual minimum temperaturefor growth is approximately 9°C lower than withbetaine (K. Krist, unpub. data).


Special Issue

tion, and storage operations would provide muchuseful information. Loggers provide a hard copyof a temperature profile in real time and thusevidence of temperature abuse and the source ofthe abuse.

For loggers with appropriate software(46,47), the temperature profile may be inter-preted in terms of microbial growth. However,the interpretation must be based on an informedanalysis of the temperature history by a trainedoperator. The operator may, for example, berequired to enter default values for initialbacterial numbers or provide an estimate of lagphase duration under specified conditions.Estimates of both imply general knowledge offood microbiology and specific knowledge of theprocess and products under consideration. Theequivalence of an estimate of microbial growthderived from temperature profile to that obtainedfrom conventional microbiologic criteria may alsobe necessary (15).

An alternative to the use of temperatureloggers is the development of in- or on-packagetemperature tags as recommended in the U.S.Food Safety Initiative draft document FoodSafety from Farm to Table (50). With tempera-ture tags, informed interpretation is not requiredbecause abuse is indicated directly by the tagresponse. Therefore, the tag must indicate thesignificance of the environmental history formicrobial behavior. The time/temperature tagsavailable are based on physical or chemicalchanges that follow Arrhenius kinetics (9). Whilethese may give a reasonable approximation ofmicrobial growth in the normal range, thedeviation of microbial responses becomes in-creasingly large as conditions move from normalto stressful. The intriguing possibility of auniversal time/temperature indicator was flagged(51) on the basis of observations made oftemperature effects on foodborne pathogens inthis laboratory and by Snyder (52). The universalindicator is based on a relationship that describesthe maximum specific growth rate of a continuumof organisms from psychrophiles to thermophilesin terms of Arrhenius kinetics with an apparentactivation energy of ~80 kJ/mole. This value canbe related to the activation energy/growth rate atany other temperature by a relative rate functionderived from Belehradek (square root) kinetics.

ConclusionsWe have argued that a thorough understand-

ing of microbial ecology and physiology offers thebest opportunity to control microbial populationsin food and reverse the upward trend in theincidence of foodborne disease. Many foodpreservation strategies have their origin inempirical observations practiced for thousands ofyears. The systematic collection and collation ofdata on microbial behavior in foods spawned thediscipline of food microbiology, within whichpredictive microbiology (quantitative microbialecology) has accelerated our understanding of themicrobial ecology of foodborne bacteria. Studiesin microbial physiology will further enhance ourknowledge and offer new possibilities for foodpreservation.

The most disturbing aspect of the currentcrisis is that simple application of existingknowledge would lead to a marked reduction inthe incidence of foodborne disease. Education offood handlers and consumers in basic hygieneand the consequences of temperature abuse isurgently needed as is a greater depth ofunderstanding for those in technical positions inthe food industry or those with regulatoryresponsibilities. Furthermore, an appreciation ofthe need for shared responsibility for food safetywithin all sectors of the continuum from farm totable, including the consumer, has to bedeveloped. The U.S. Food Safety Initiative draftdocument emphasizes this point, as does thestructure of the Australian Meat ResearchCorporation’s Microbial Food Safety Key Pro-gram (53).

AcknowledgmentsThe work described in this manuscript was funded

largely by the Australian Meat Research Corporation (MRC)through a Core Funding Project and the Microbial FoodSafety Key Program. The authors are indebted to Dr. StefanFabiansson for his support and encouragement of researchinto the ecology and physiology of foodborne pathogens.

References 1. Maurice J. The rise and rise of food poisoning. New

Scientist 1994;144:28-33. 2. McMeekin TA, Olley J. Predictive microbiology and the

rise and fall of food poisoning. ATS Focus 1995;88:14-20.

3. Pace NR. Perspective on the natural microbial world:molecular microbial ecology. ASM News 1996;63:463-70.

4. Schelgel HG, Jannasch HW. Prokaryotes and theirhabitats. In: Balows A, Truper HG, Dworken M,Harder W, Scheifer K-H, editors. The prokaryotes.New York: Springer Verlag; 1992. p. 75-125.


Special Issue

5. Langeveld PLM, van der Klijn W, CB van der Waals.Bacterie groei in indampers voor milk en wei.Overdruk van Voedingsmiddelentechnologie1990;23:13-7.

6. Archer DL. Preservation microbiology and safety:evidence that stress enhances virulence and triggersadaptive mutations. Trends in Food Science Technology1996;7;91-5.

7. Leistner L. Food preservation by combined methods.Food Research International 1992;25:151-8.

8. Ross T, McMeekin TA. Predictive microbiology andHACCP. In: Pearson AM, Dutson TR, editors. HACCPin meat, poultry and fish processing. London: BlackieAcademic and Professional; 1995. p. 330-53.

9. McMeekin TA, Olley J, Ross T, Ratkowsky DA.Predictive microbiology: theory and application.Taunton, UK: Research Studies Press; 1993.

10. Whiting RC, Buchanan RLB. Predictive modeling. In:Doyle MP, Beuchat LR, Montville TJ, editors. Foodmicrobiology fundamentals and frontiers. Washington(DC): American Society for Microbiology Press 1997. p.728-39.

11. Graham A, Lund BM. The effect of temperature on thegrowth of non proteolytic type B Clostridiumbotulinum. Letters in Applied Microbiology 1993;16:158-60.

12. Davey GR. Food poisoning in New South Wales: 1977-84. Food Technology in Australia 1985;37:453-7.

13. Enneking U. Hazard analysis of critical control points(HACCP) as part of the Lufthansa in-flight servicequality assurance. International Food Safety News1993;2:52-3.

14. Adams M, Moss MO. Food microbiology. Cambridge:Royal Society of Chemistry; 1995.

15. Gill CO, Harrison JCL, Phillips DM. Use of atemperature function integration technique to assessthe hygienic efficiency of a beef carcass cooling process.Food Microbiology 1991;8:83-94.

16. Gill CO, Jones SDM, Tong AKW. Application of atemperature function integration technique to assessthe hygienic efficiency for spray chilling beef carcasses.Journal of Food Protection 1991;54:731-6.

17. Madigan MT, Martinko JM, Parker J. Microbialecology. In: Brock, editor. Biology of microorganisms.8th ed. New Jersey: Prentice-Hall; 1996.

18. Ratkowsky D, Ross T, McMeekin TA, Olley J.Comparison of Arrhenius-type and Belehradek-typemodels for the prediction of bacterial growth in foods. JAppl Bacteriol 1996;71:452-9.

19. Ratkowsky DA. Predicting response times inpredictive microbiology. Occasional paper. Tasmania,Australia: Research and Development Unit, BiometricsSection, Dept of Primary Industry, Fisheries andEnergy; 1991.

20. Alber SA, Schaffner DW. Evaluation of datatransformations used with the square root andSchoolfield models for predicting bacterial growth rate.Appl Environ Microbiol 1992;58:3337-42.

21. Ratkowsky DA, Ross T, Macario N, Dommett TW,Kamperman L. Choosing probability distributions formodelling generation time variability. J Appl Bacteriol1996;80:131-7.

22. Shaw MK. Effect of abrupt temperature shift on thegrowth of mesophilic and psychrophilic yeasts. JBacteriol 1967;93:1332-6.

23. Labuza TP, Fu B. Growth kinetics for shelf-lifeprediction: theory and practice. J of IndustrialMicrobiology 1993;12:309-23.

24. Mitchell GA, Brocklehurst TF, Parker R, Smith AC.The effect of transient temperatures on the growth ofSalmonella typhimurium LT2. I: Cycling within thegrowth region. J Appl Bacteriol 1994;77:113-9.

25. Mitchell GA, Brocklehurst TF, Parker R, Smith AC.The effect of transient temperature on the growth ofSalmonella typhimurium LT2. II: Excursions outsidethe growth region. J Appl Bacteriol 1995;79:128-34.

26. Baranyi J, Robinson TP, Kaloti A, Mackey BM.Predicting growth of Brocothrix thermosphacta atchanging temperature. Int J Food Microbiol 1995;27:61-75.

27. Walker SJ, Archer P, Banks JG. Growth of Listeriamonocytogenes at refrigeration temperatures. J ApplBacteriol 1990;68:157-62.

28. Dufrenne J, Delfgou E, Ritmeester W, Notermans S.The effect of previous growth conditions on the lagphase of some foodborne pathogenic microorganisms.Int J Food Microbiol 1997;34:89-94.

29. Chen J, Griffiths MW. Luminescent Salmonellastrains as real time reporters of growth and recoveryfrom sublethal injury in food. Int J Food Microbiol1996;31:27-44.

30. Gill CO, Greer GG, Dilts BD. The aerobic growth ofAeromonas hydrophila and Listeria monocytogenes inbroths and on pork. Int J Food Microbiol 1997;35:67-74.

31. Dalgaard P. Predictive microbiological modelling andseafood quality. In: Seafood from producer toconsumer, integrated approach to quality. Proceedingsof the International Seafood Conference, 1995Noordwijkerhout. The Netherlands: Elsevier,Amsterdam. In press 1997.

32. Genigeorgis CA. Factors affecting the probability ofgrowth of pathogenic microorganisms in foods. J AmVet Med Assoc 1981;179:1410-7.

33. Roberts TA, Gibson AM, Robinson A. Prediction oftoxin production by Clostridium botulinum inpasteurised pork slurry. Journal of Food Technology1981;16:337-55.

34. Stumbo CR, Prokit KS, Ramarkrishnan TV, Evans DA,Franaas FJ. CRC handbook of lethality guides for low-acid canned foods. 1. Boca Raton (FL): CRC Press;1983.

35. Ratkowsky DA, Ross T. Modelling the bacterialgrowth/no growth interface. Letters in AppliedMicrobiology 1995;20:29-33.

36. Lee IS, Lin J, Hall HK, Bearson B, Foster JW. Thestationary-phase sigma factor (RpoS) is required for asustained acid tolerance response in virulent Salmonellatyphimurium. Mol Microbiol 1995:17:155-67.

37. Brown J, Ross T, Nichols PD, McMeekin TA. Acidhabituation of Escherichia coli and the potential role ofcyclopropane and fatty acids in low pH tolerance. Int JFood Microbiol. In press 1997.


Special Issue

38. Goodson M, Rowbury RJ. Resistance of acid-habituated Escherichia coli to organic acid and itsmedical and applied significance. Letters in AppliedMicrobiology 1989;8:211-4.

39. Galinski EA. Compatible solutes of halophilic eubacteria:molecular principles, water solute interaction, stressprotection. Experientia 1993;49:487-96.

40. Ko R, Smith LT, Smith GC. Glycine betaine confersenhanced osmotolerance and cryotolerance on Listeriamonocytogenes. J Bacteriol 1994;176:426-31.

41. Csonka LN. Physiological and genetic responses ofbacteria to osmotic stress. Microbiol Rev 1989;53:121-47.

42. Knochel S, Gould G. Preservation microbiology andsafety: quo vadis? Trends in Food Science andTechnology 1995;6:127-31.

43. Davidson PM. Chemical preservatives and naturalantimicrobial compounds. In: Doyle MP, Beuchat LR,Montville TJ, editors. Food microbiology fundamentalsand frontiers. Washington (DC): American Society forMicrobiology Press, 1997.

44. Smith LT. Role of osmolytes in adaptation of osmoticallystressed and chill stressed Listeria monocytogenes grownin liquid media and on processed meat surfaces. ApplEnviron Microbiol 1996;62:3088-93.

45. Clavero MRS, Beuchat LR. Survival of Escherichia coliO157:H7 in broth and processed salami as influencedby pH, water activity and temperature and suitabilityof media for its recovery. Appl Environ Microbiol1996;61:2735-40.

46. Gill CO, Phillips DM. Hygienically appropriate time/temperature parameters for raw meat processing.Proceedings of the 36th International Congress ofMeat Science and Technology; 27 Aug–2 Sep 1990;Havana, Cuba; 1990. p. 458-70.

47. McMeekin TA, Ross T. Modeling applications. Journalof Food Protection 1996 (Suppl): 1-88.

48. Adair C, Briggs PA. The concept and application ofexpert systems in the field of microbiological safety.Journal of Industrial Microbiology 1993;12:263-7.

49. Jones JE. A real-time database/models base/expertsystem in predictive microbiology. J Ind Microbiol1993;12:268-72.

50. Glickman D, Salala DE, Browner CM. Food safety fromfarm to table: a new strategy for the 21st Century.Washington (DC): U.S. Department of Agriculture;1997.

51. McMeekin TA, Ross T. Shelf life prediction: status andfuture possibilities. Int J Food Microbiol 1996;33:65-83.

52. Snyder OP. Use of time and temperature specificationsfor holding and storing food in retail operations. DairyFood Environ Sanitat 1996;116:374-88.

53. Fabiansson S, Wrigley J, Sumner J, McMeekin TA, OrrC. Meat Research Corporation Business Plan,Microbial Food Safety Key Program, 1996. Sydney,Australia: Meat Research Corporation; 1996.


Special Issue

Mobilizing ResourcesIn the traditional paradigm of a foodborne

disease outbreak, the cases were from a smalllocal group, and the attack rate was high. Localhealth officials generally detected and investi-gated the outbreak. In the emerging paradigm, adiffuse outbreak may be spread over a very widearea, perhaps several counties or states, evenwith a low infective dose and a low attack rate.The outbreak may be registered only as anincrease in sporadic cases and detected onlybecause of specific laboratory-based subtypingsurveillance. Whether Salmonella serotyping oranother molecular method identifies a cluster ofrelated cases, the investigations are more complex,and often no obvious food-handling error is found.Industry contamination may be involved, andimplications may be industrywide or nationwide.

Detecting a widespread outbreak requiresincreased reliance on laboratory subtyping bystate public health laboratories at a time whensome states are considering eliminating orprivatizing their laboratories. Surveillance datamust be rapidly compared over increasinglybroad regions, not just at the county level but atthe state, regional, and national levels. Increased

awareness is needed throughout the system thata local outbreak may herald a broader problem.Moreover, investigations must be conductedquickly to prevent future cases. Because of thelow levels of microbiologic contamination, theright specimens and samples of food must beused. Available epidemiologic data should guidethis selection so that the most likely vehicles aresampled. Traceback must extend beyond theimmediate preparation of the implicated food tothe whole chain of preparation, i.e., sources ofingredients, processing, storage, and transporta-tion; cooperation at all levels of industry isrequired. The goals are to control the ongoingoutbreak, remove the contaminated product fromthe market, and learn how to prevent similaroutbreaks. Intervention in outbreaks mustchange. Emergency intervention must be basedon solid epidemiologic data (appropriate studydesign and sample size and clear statisticalassociation with logical and biologic plausibility)and cannot always wait for laboratory confirma-tion of a contaminated product. Illnesses will notwait for laboratory examination to yield thepathogen; the pathogen may not be detectablewith current technologies, the food may not beavailable, and the delay can be critical.

Strategies for Rapid Response toEmerging Foodborne Microbial Hazards 1

Jesse MajkowskiU.S. Department of Agriculture, Washington, D.C., USA

Address for correspondence: Jesse Majkowski, U.S.Department of Agriculture, Room 3715 Franklin Court, 14thStreet & Independence Avenue, S.W. Washington, DC 20250USA; fax: 202-501-7515; e-mail: [email protected].

The foodborne outbreak paradigm has shifted. In the past, an outbreak affected asmall local population, had a high attack rate, and involved locally prepared foodproducts with limited distribution. Now outbreaks involve larger populations and may bemultistate and even international; in many the pathogenic organism has a low infectivedose and sometimes is never isolated from the food product. Delay in identifying thecausative agent can allow the outbreak to spread, increasing the number of cases.Emergency intervention should be aimed at controlling the outbreak, stopping exposure,and perhaps more importantly, preventing future outbreaks. Using epidemiologic dataand investigative techniques may be the answer. Even with clear statistical associationsto a contaminated food, one must ensure that the implicated organism could logicallyand biologically have been responsible for the outbreak.

1The author has summarized the transcripts of panel discus-sions by Dennis Lang, Craig Hedberg, Eric Johnson, SuzanneBinder, and Philip Tarr.


Special Issue

The Human Side of Foodborne DiseasePublic health officials in Washington state

screened children with bloody diarrhea andhemolytic uremic syndrome (HUS) as the resultof a large outbreak of Escherichia coli O157:H7infection in 1993. One isolate of the organismcaused perhaps 75% of the cases, not 100%. As aresult, the source of the infection was identified,and regulatory action was taken to halt futurecases. Hospitals reported fewer new cases afterthis action. However, 500 Washington stateresidents, two-thirds of them children under 15years of age, became infected before theincriminating food could be removed from themarket. The HUS attack rate was approximately12% for children under the age of 16 years. Theorganism was recovered from the food product(hamburger), and DNA fingerprinting wasinitiated by several techniques. The number ofcolony-forming units of E. coli O157:H7 in thehamburger was relatively low.

This organism can cause severe life-threatening infection, even with an inoculumrate too low to be detected by testing. In theUnited States, the incidence of HUS isapproximately 1.7 cases per 100,000 childrenunder the age of 15 years. This figure is based ondata from King County, Washington, in the early1980s and indicates that there are 1,000 cases ofHUS in the United States per year or an averageof 2.8 cases per day in a population of 250 million.The Centers for Disease Control and Preventionhas recently reported that only half of the country’smicrobiologists screen for this organism.

Clostridium botulinum , A ReemergingPathogen

The outstanding property of C. botulinum isits ability to synthesize a neurotoxin ofextraordinary potency (lethal dose is approxi-mately one microgram). C. botulinum is anunusual foodborne pathogen in that it causesneuroparalytic rather than diarrheal disease.During illness, the first nerves affected are thecranial nerves in the head and eye; the paralysiscan descend and affect every peripheral nerve inthe body. The toxin can enter into foods, but inrecent years, C. botulinum has also been found tocolonize the intestinal tract of infants under 1year of age and of adults that have undergoneintestinal surgery or have been treated withantibiotics. The number of cases of adults with

unusual clostridia that produce botulinumtoxin is increasing.

Botulism occurs worldwide. The highestincidence is found in Poland and in Asia and isrelated to food handling (in Poland, homecanned meats).

New food processes and packaging have beenassociated with the reemergence of botulism. Aclam chowder outbreak in California involved aboxed food that was not properly stored.Because boxed foods generally do not requirerefrigeration, the food was kept at ambienttemperature; however, it was not shelf stableand should have been refrigerated.

A large outbreak (30 cases) of botulismoccurred in a restaurant in El Paso when potatoeswere cooked in foil, left wrapped, and then used inpotato salad. Baking had eliminated vegetativeorganisms, but the spores of botulinum were notkilled. In this case, a low-acid food was held underanaerobic conditions.

Botulism outbreaks are probably the mostreported type of foodborne illness. Changes inprocessing and ingredients in foods and formula-tions can inadvertently lead to the growth of C.botulinum. Failure to thoroughly heat a foodproduct may allow the botulinum toxin to survive.

DNA FingerprintingResponding to the threat of emerging

foodborne diseases requires public healthsurveillance that is based on epidemiologicmethods and close collaboration between epide-miologists and public health laboratories.Surveillance for foodborne diseases should includepathogen-specific surveillance to identify clustersof cases caused by the same organism andepidemiologic investigations to identify the source.

The Minnesota State Department of Healthis developing a new approach to foodbornedisease surveillance. A Salmonella Enteritidis(SE) outbreak was first recognized by the publichealth laboratory when the number of SE isolatessuddenly increased. Because many of the isolatescame from clinical laboratories in southeasternMinnesota, the outbreak initially seemed re-gional. However, within 48 hours of initiating acase-control study, a nationally distributed foodproduct was identified as the vehicle, and thenationwide scope of the outbreak was docu-mented. This was an example of consequentialepidemiology. When the association between food


Special Issue

consumption and SE was announced, theevidence implicating a particular brand waslimited to a single case-control study of 15matched pairs. Laboratory isolation of SE fromofficial samples was not reported until 10 dayslater. This prompt action, based on epidemiologicdata, prevented at least 10 days of potentialexposure for thousands of consumers. Consequen-tial epidemiology produces results that translateinto disease prevention. We need to continue todevelop models for how to rapidly evaluate and acton epidemiologic data and how to better coordinateactivities regionally and nationally.

A critical element of the success of this type offoodborne disease surveillance is the specificitywith which we can match isolates that may beepidemiologically linked. Molecular subtypingschemes such as pulsed-field gel electrophore-sis can greatly improve identification ofclusters of the same organisms such asparticular Salmonella serotypes.

Another example involved analysis of atypical epidemiologic curve for a seemingly singleoutbreak. When isolates were analyzed, however,a cluster of small outbreaks was found. One wascaused by infected food handlers at a restaurant.This outbreak would have continued, and theinfected food handlers would have provided anongoing source of infection to patrons had thecluster not been identified through subtype-specific surveillance. This incident serves as amodel for how foodborne disease surveillancesystems must be developed and used. Molecularsubtyping has revolutionized our ability toconduct meaningful surveillance. We consider itan integral part of disease prevention and controland continue to explore its usefulness.

ParasitesCyclospora cayetanensis is a protozoan

coccidian parasite. A one-celled organism, it isrelated to other organisms such as Toxoplasmaand Cryptosporidium. It is a prototypicalemerging pathogen. C. cayetanensis is unusual inthat it is not immediately infectious whenexcreted. Under optimal conditions, it matures indays to weeks, so direct person-to-person spreadis very unlikely. An outbreak following a meal isprobably not caused by the food handler. Theorganism appears to be seasonal, and in mostplaces where it has been studied, it occurs in thespring or summer and causes little or no diseaseduring the fall or winter. Infection has been

reported throughout the world, and the keystudies have been conducted in Peru and Nepal.Disease caused by C. cayetanensis is character-ized by watery stools, nausea, weight loss, low-grade fever, fatigue, or any combination of thesesymptoms. The disease (which is easilytreatable) can be quite protracted, and withouttreatment, relapse can occur. The meanincubation period of 1 week complicates theepidemiology; cases may not be recognizeduntil 2 weeks after people have been exposed.

In 1996, more than 1,450 cases of Cyclosporawere reported in the United States (87%) andCanada (13%). Approximately half of them werein clusters; the other half were sporadic (notepidemiologically linked to other cases). Morethan 65% of the 1,450 cases were laboratoryconfirmed; 22 infected patients were hospital-ized. Fifty-five clusters were reported, 47 in theUnited States and 8 in Canada. An average of 28attendees per event and a very high attack ratewere reported. The attack rate was 56% forattendees, not for people who ate the implicatedfood. At least one type of fresh berry was served atevery event and, despite other types of exposures,no other food was implicated in any clusterinvestigation. The berry did not always achievestatistical significance, largely because of thesmall number of attendees at a specific event. Theberry most likely linked to the cases was laterdetermined. That type of berry was served at 50%to 91% of the 55 events; it was the only kindserved at 10 or 11 of the events.

We had overwhelming epidemiologic evidence,but we never identified Cyclospora on anyraspberries. Two factors, at least, contributed tothis. The test for C. cayetanensis did not exist whenthis outbreak occurred; therefore, no implicatedraspberries were tested. The question in thisinvestigation as in others is how much epidemio-logic evidence is needed to implicate a food as thevehicle of disease? A review team may be needed tolook at the epidemiologic data and determine if theyare adequate to warrant informing the public abouta hazardous food. We are seeing new pathogens,new species. As outbreaks cross into other states,the need for coordination between health officials inthe states and in the federal government becomesmore urgent.

ConclusionsEarly identification of the outbreak and the

organism can prevent future cases. Recent


Special Issue

investigations have found that the presence of anorganism even at low levels can cause seriousconsequences. Fingerprinting organisms foridentification during outbreaks is extremelyimportant. In some instances, fingerprinting hashelped identify several small outbreaks thatinitially appeared to be one large outbreak. Wecan no longer afford to wait for all the evidenceand laboratory results to be collected andreported; we must use epidemiologic data. Oncethe outbreak is identified, DNA fingerprinting is

needed to identify whether other outbreaks areoccurring simultaneously.

A rapid and coordinated response is neededamong state officials and federal agencies.Interventions should stop outbreaks and identifyproducts causing illness so they can be removedfrom the market. Then health officials need totake the next step—investigate what happenedand determine the cause so that similaroutbreaks can be prevented.


Special Issue

I do not see microbiologic infections as totallydifferent from other food hazards, or microorgan-isms as totally different from other life forms. Weare all hosts and prey, parasites and predators.Even though we differ in size, complexity, andweaponry, we employ many similar strategies—the result of a shared planet, and a substantiallyshared genetic base. That is also why we havemuch to learn from animal disease surveillance.

Writing on emerging diseases, many argueforcefully for a broad and well-integrated view(1,2); however, aspects absolutely essential tounderstanding the problem are often omitted.This short article will doubtless make the sameerror. So, as Shakespeare had Prologue plead atthe opening of Henry the Fifth, “Piece out ourimperfections in your minds.”

Factors Contributing to EmergenceOutbreaks occur whenever pathogenic agents

in sufficient number or quantity encounter asusceptible population without effectiveinterceptive measures. Then, if we did not expectit, we say “it emerged.”

Genetic Variability The large genetic variability of microorgan-

isms is the principal reason why so often somesurvive after any unfavorable environmentalchange. Some strains are hypermutable, whichreinforces the potential for survival, and havevery short generation times, with bacterialminutes comparable to human years. As Dr.Lederberg notes, microorganisms are opponentswith whom we cannot race—on their terms.

Environment Environmental factors also contribute to

emergence. Hot, humid climates favor the growthof fungi and the production of mycotoxins. Toborrow an example outside foodborne pathogens,an unusually wet season produced a sharpincrease in the deer mice population and theconsequent outbreak of hantavirus in the FourCorners area of the United States.

Behavior Human actions and behavior directly affect

food safety. People are vectors for disease,traveling far more often, farther, and morerapidly than ever before (3), and moving far moreswiftly than rats, lice, and mosquitoes. Politicalboundaries frequently and perversely act asleaky sieves, letting diseases through unim-peded, while blocking measures for diseaseprevention, control, and treatment (1).

Urbanization Urbanization is a major factor in emergence.

Crowding increases human contact and opportu-nities for transmission. Particularly in develop-ing countries, public health services lag farbehind the rush from farm to city. Cities,especially in industrialized nations, are economicand governmental centers and harbor institu-tions of culture and learning. However, cities arealso massive projects in the intensive monocul-ture of humans. With agriculture, it is entirelypossible to carry out monoculture effectively andproductively—esthetic considerations aside—ifone provides for and tightly controls all essentialinputs and conditions, monitors the processclosely, and is prepared for prompt and effectiveintervention if something goes wrong. Thathardly describes cities anywhere.

Foodborne Illness:Implications for the Future

Richard L. Hall

Address for correspondence: Richard L. Hall, 7004 WellingtonCourt, Baltimore, MD 21212, USA; fax: 410-377-2909.

Many outbreaks of foodborne illness, even those involving newly recognizedpathogens, could have been avoided if certain precautions had been taken. This articlewill draw on existing information to suggest realistic measures that, if implemented, aremost likely to avert or diminish the impact of new foodborne disease outbreaks.


Special Issue

Raw Food Production The effect of changes in raw food production

and harvest practices on opportunities forfoodborne outbreaks have been discussed. Central-ized processing and wide distribution are theprincipal characteristics of the “new scenario,” ascompared to the “old scenario” of local production,home processing, and intrafamily consumption. Aclassic example is the recent Japanese outbreak ofhemorrhagic Escherichia coli.

Denial A behavior that encourages outbreaks is

denying the existence of an epidemic—a practicemore common in developing countries concernedabout the effect of outbreak publicity on touristtrade and exports (4). One of our own attitudes isindifference to outbreaks perceived as common-place and distant (5).

Economics War and economic collapse provide unparal-

leled opportunities for disease outbreaks (e.g.,cholera in central Africa). The infrastructure thatprovides clean water, community medicine,disease surveillance, and food control, evenwhere it exists, is a fragile fabric, easily torn byeconomic, social, and physical disruption.

Technology In spite of their benefits, technologies often

bring new or enlarged risks. This is not anargument for returning to a state of nature. Theinvention of sausage doubtless increased theincidence of botulism; indeed, “botulus” is theLatin word for sausage. Without properprocessing, any modern packaging that excludesoxygen can have the same effect.

Risk FactorsFactors such as age, illness, and medical

treatment increase the risk for foodborneillness. Such increases also result frombehavior that promotes the incidence of otherdiseases (e.g., AIDS).

Failure to Prevent and Control The most common human action that

adversely affects food safety is the avoidable lack ofor failure to use effective prevention and controlmeasures. That failure is why 85% of all outbreaksare traceable, about equally, to mishandling inhomes or in food service establishments.

Interacting FactorsIn much of the developing world, an

interrelated and mutually reinforcing set ofproblems keeps foodborne disease at a high level(Figure). Approximately five million children underthe age of 5 years living in the tropics die each yearof malnutrition and diarrheal disease (6). Palliationis temporary; only economic and technicaldevelopment can break through this net.

The contributing factors already mentionedwould cause us problems, even acting singly.However, they interact, often synergistically.The combination of bacterial genetic variabilityand the ease and frequency of mutation of strainspresent a threat because the process enhancesselection for new and more dangerous pathogens.Acid rain and recycling through ruminants mayhave encouraged the increased environmentaldurability of acid-tolerant E. coli O157:H7. Theincreasing popularity of marinades in thepreparation of foods may have had the same effect.

The globalization of the food trade pullstogether several of these contributing factors.One country’s contaminated water leads toanother country’s contaminated raspberries.Refrigeration and controlled atmosphere canpreserve pathogens, as well as foods, and spreadthem all over the world. The SalmonellaEnteritidis outbreak from contaminated eggs inice cream was a typically broad problem.

Technologic change combines many of thesecontributing factors. The intensive monocultureof plants and animals presents concentrated

Figure. Problems causing foodborne disease indeveloping countries.


Special Issue

opportunities. For example, an outbreak ofsouthern corn blight was attributed to thenarrow genetic base of a popular hybrid corn.Human-guided plant genetics was defeated bythe much more rapid genetic adaptability ofmicroorganisms.

Indiscriminate use of antibiotics and theability of microorganisms to exchange geneticinformation has led to increasing resistance.Minimal processing of food, warmly received bynatural food lovers, is an open invitation toslightly more durable pathogens to take over thefood supply. Centralized processing and massdistribution push us even further in the race Dr.Lederberg reminded us we cannot win. Immuno-suppression, due to disease or medication,combined with the failure to take simple sanitaryor health precautions make the emergence ofpathogens inevitable.

The Role of Public Health MeasuresAchieving optimum public health requires

many measures, each necessary but notsufficient in itself. The most basic measure isclean water, as recent incidents with strawber-ries and other fresh produce washed withcontaminated water illustrate.

Effective food control structures (e.g.,statutory and regulatory frameworks andinspection and enforcement agencies) areessential. A surveillance system has always beenrequired and should be directed to populations athigh risk. Newer methods of typing pathogens,faster response times, and enhanced analyticalsensitivity are making those systems far moreeffective than they have been.

Family health programs and education areequally necessary. Food safety still dependsheavily on how each of us chooses and uses (orabuses) our food, and each of us is not expert.

The Roles of Food ProcessingWe process food for the following main

reasons: nutrition, safety, preservation, distribu-tion, and esthetics. Cooking typically increasesprotein digestibility and destroys antinutrients.Fortification and enrichment add, restore, orincrease essential nutrients. Cooking andcanning destroy pathogenic organisms; cookingand steeping remove natural toxicants, such ascyanide from cassava. Canning, dehydration,salting, smoking, and preservatives, combinedwith proper packaging and storage, decrease

spoilage. Canning, freezing, refrigeration, andcontrolled atmosphere storage provide varietyand year-round availability. Flavors, colors, pre-sale preparation, and stabilizers that preventseparation or crystallization increase consumerappeal and convenience.

Food Processing TechnologiesFood processing technologies reduce our

exposure to dietary pathogens in three ways,summarized by Professor E. M. Foster as “thethree Ks”: keep them out, kill all you can, keepthe rest from growing.

One group of technologies removes ordestroys microorganisms and naturally occur-ring toxicants by careful sanitation; heattreatment (e.g., cooking, retorting, pasteurizing,high temperature/short time treatment, ultra-high temperature treatment, ohmic heating, andother newer techniques); radiation (widely usefuland recently applied to all of our currentmethods); physical separation, (e.g., air separa-tion, gravity tables, visual scanners, and othermethods for reducing “natural and unavoidabledefects” including gross contamination); dehy-dration; and new technologies (e.g., hydrostaticpressure, pulsed light).

A second group of technologies keepscontaminants well below dangerous levels. Theseinclude raw material quality control; carefulsanitation; chemical preservation (i.e., fermenta-tion, bacteriostats, pH control, water activitycontrol, and controlled atmosphere storage);dehydration; and freezing and refrigeration.

The third group of technologies preventsrecontamination through sanitation of thegeneral environment and at the point of service,protective packaging, and proper handling toensure packaging integrity.

Beyond these three groups of technologiesare general principles governing the use of allsuch techniques. All foods require incomingquality control and careful sanitation. For nearlyall foods except fresh produce and dried grains,one technology from each group listed aboveshould be used. All technologies used must beapplied effectively, or none is effective. Thesetechnologies must be used in combination, in asystems approach. Failure to apply properly apreceding technology can cause later technolo-gies to increase microbiologic hazards. Ifoxygen is later excluded from improperlyprocessed food by packaging (canning, foil, an


Special Issue

oil layer), anaerobic pathogens, such asBotulinum, are free to multiply.

Future OptionsFood safety objectives must be established.

There is a clear need for priorities and foremploying the “principle of commensurate effort”(i.e., applying effort to risks in the order of theirprobable impact).

There is also the need for a broad view (1,2).The interaction of several factors, not all ofwhich can be seen in advance, makes a teamapproach essential.

More effort at understanding the evolution ofvirulence (7) could provide us with insights on thegenetic characteristics and the environmentalconditions needed to minimize risk. The 1918-19influenza epidemic can be used as an example;the present search for its genetic makeupsuggests the value of prior knowledge. Knowingor estimating the probability of present or futurevirulence could prepare us for the more effectiveuse of other measures. These include newer,faster, more sensitive methods of detection.

Improved detection would lead to moreeffective prevention and control measures. Theneed to establish priorities suggests here, as incancer research, the value of biomarkers.Biomarkers are genetic or biochemical indicatorsof impending risk in the potential pathogen,preceding clinical or epidemiologic indications.Such biomarkers might indicate potentialvirulence, increased environmental durability, orincreased resistance to heat, cold, unfavorablepH, preservatives, or antibiotics.

We need to apply available knowledge moreeffectively. Most outbreaks of foodbornedisease are due to mishandling food in ways wealready know how to avoid. The points thatfollow are not new, but as our food supplyincreases in complexity and geographic reach,these weapons must grow apace.

Clean water, public sanitation, diseasesurveillance, and food control are more importantthan ever. The value of effective diseasesurveillance has been demonstrated. But whatwe now have and do is not enough. A nationalresponse team must be created. Data should beshared more broadly (electronic linkage oflaboratory and surveillance results) if we are tocope effectively with a global food supply,multistage processing by different firms at

different locations, and national and interna-tional distribution. The effective application ofthese resources requires a Hazard Analysis andCritical Control Points (HACCP) approach (8).

Private health measures must grow ineffectiveness and reach. We must make continu-ing and increasingly effective use of foodprocessing technologies that reduce microbio-logic risks. HACCP cannot be generic or static; itmust be adapted to each specific product andprocessing facility and must be updated withcontinuing feedback on the hazards to beavoided. HACCP, fortified by the public healthmeasures just described and by the results ofthe research discussed at this conference, willenable the food processing industry to meet thechallenges of emerging pathogens.

Even with the steps just described,approximately 85% of all outbreaks occur as aresult of food mishandling in food serviceestablishments or homes. We need to extendHACCP principles to the food service sector aswell. HACCP probably cannot reach intohomes. If we are to communicate better, weneed first to find out what consumers alreadyknow, what they want to know, and what theyneed to know. In short, if we are to educateeffectively, we must have direct evidence ofhow well the information process is working.

Finally, we will never succeed in achievingdesirable consumer risk-management practicesuntil consumers understand the inevitability ofsome risks and minimize them. Food risks havebeen categorized into six groups in decreasingorder of size: microbiologic, nutritional, naturaltoxicant, environmental contaminant, pesti-cide residues, and food additives (9). The take-home message on food safety for all of us, asconsumers, can be summed up in three words:sanitation, variety, moderation. Sanitationdeals effectively with microbiologic risks;variety and moderation deal with nutritionalrisks, minimizing the impact of the fourremaining risks to the extent they are evenpotentially insignificant. Sanitation, variety,and moderation are within our own control asindividual consumers. That is doubtless whywe have done only a mediocre job ofimplementing them. Unless all of us can and dopursue these three objectives effectively, weare all at some unnecessarily increased risk.


Special Issue

References 1. Farmer P. Social inequalities and emerging infectious

diseases. Emerg Infect Dis 1996;2:259-69. 2. Vidaver AK. Emerging and reemerging infectious

diseases. American Society for Microbiology News1996;62:583.

3. Wilson M. Travel and the emergence of infectiousdiseases. Emerg Infect Dis 1995;1:39-46.

4. Vidaver AK. Emerging and reemerging infectiousdiseases. American Society for Microbiology News1996;62:585.

5. Lederberg J, Shope RE, Oaks SC. Emerging infections:microbial threats to health in the United States.Washington (DC): National Academy Press; 1992.

6. Snyder JD, Merson MH. The magnitude of the globalproblem of acute diarrhoeal disease: a review of activesurveillance data. Bull World Health Organ1996;60(4):605-13.

7. Ewald P. Guarding against the most dangerousemerging pathogens: insights from evolutionarybiology. Emerg Infect Dis 1996;2:245-56.

8. International Committee on Microbiological Safety inFoods. HACCP in microbiological safety and quality.In: Microorganisms in Food. Book 4. Oxford: BlackwellScientific Publications Ltd.; 1988. p. 28-43.

9. Wodicka VO. Wodicka rates food additive hazard aslow. Food Chemical News 1971;12(49):11.


Dispatches

Vero cytotoxin-producing Escherichia coliO157 (O157 VTEC), first associated withoutbreaks of human disease in North America in1982, has since emerged as an important humanpathogen; it causes mild nonbloody diarrhea,hemorrhagic colitis, and hemolytic uremicsyndrome, as well as less common manifestationssuch as thrombotic thrombocytopenic purpurawith neurologic symptoms (1). Laboratory-confirmed infections with O157 VTEC in Englandand Wales increased from fewer than 10 in 1983to 250 in 1990, 792 in 1995, and 660 in 1996(2;3;Laboratory of Enteric Pathogens, unpub.data). The incidence of O157 VTEC infection inScotland relative to its population is up to sixtimes that in England and Wales, but widegeographic variations exist in England, Wales,and Scotland (4,5). The main reservoir for O157VTEC appears to be healthy cattle, althoughrecently organisms have been found in sheep(4,6,7). O157 VTEC infections are usuallyfoodborne, associated with consumption ofundercooked minced beef (most commonly asbeefburgers), unpasteurized milk, and a varietyof other vehicles such as salami, cheese, yogurt,water, salad vegetables, and fruit juice (7,8).Other routes of infection include animal contactand person-to-person spread, both in families andinstitutional settings (4).

O157 VTEC is differentiated by phage typinginto more than 80 phage types (PTs) (9, R.Khakhria, pers. comm.). Polynucleotide DNAprobes identify Vero cytotoxin (VT) genotypes anddivide O157 VTEC into strains with VT1, VT2, orVT1+2 genes. However, certain phage and VTtypes predominate: thus in England and Wales

approximately 50% of O157 VTEC are PT2 andproduce VT2 (2,3). A range of DNA-based methodsis available for further strain discrimination (4),and we have applied some of these to comparestrains associated with outbreaks of O157 VTECinfection in England and Wales in 1995.

Eleven general outbreaks during 1995affected members of more than one household orresidents of an institution (Table), whereas 18general outbreaks were reported during 1992 to1994 (10). Most of the outbreaks in 1995 occurredin late summer and in the community (Table,outbreaks 5, 9, and 10); institutions (Table,outbreaks 1, 4, and 11); catering establishments(Table, outbreaks 3, 6, and 7); or a mixture ofthese (Table, outbreaks 2 and 8). Outbreaksoccurred throughout England and Wales; of the11, four were in the northern region. Theincidence of hemolytic uremic syndrome was 0%to 36% in individual outbreaks and 8% overall.The case-fatality rate was 6%, mainly among theelderly. In six outbreaks, there was epidemiologicevidence for foodborne infection, but O157 VTECwas never isolated from food. Person-to-personspread was probably important in threeoutbreaks. The environmental and epidemio-logic investigations of two of these incidentshave been reported (11, 12).

For most outbreaks, epidemiologic investiga-tion, phage typing of isolates, and further testswith VT1 and VT2 probes (Table) initiallydiscriminated probable outbreak-associated casesfrom sporadic infections at the same time in thesame area. Seven outbreaks were due to strainsof PT2, VT2; two to PT49, VT2; and one to PT1,VT1+VT2. The other outbreak was associated

Vero Cytotoxin-Producing Escherichiacoli O157 Outbreaks in England and

Wales, 1995: Phenotypic Methods andGenotypic Subtyping

Vero cytotoxin-producing Escherichia coli O157 belonging to four phage types(PTs) caused 11 outbreaks of infection in England and Wales in 1995. Outbreak strainsof different PTs were distinguishable by DNA-based methods. Pulsed-field gelelectrophoresis best discriminated among strains belonging to the same PT,distinguishing six of the seven PT2 outbreak strains and both PT49 outbreak strains.


Dispatches

with a VT1+VT2 strain that reacted with thetyping phages but did not conform to a recognizedtype (reacts but does not conform [RDNC]). Inoutbreak 8, the O157 VTEC was resistant tosulphonamides and tetracyclines, whereas theother outbreak strains were sensitive toantimicrobial agents. In relation to communityoutbreaks 2, 5, and 8, several cases were infectedwith O157 VTEC that were similar to theoutbreak strains, but the patients had no knownepidemiologic link with the outbreak. Suchstrains were therefore included in the testsdescribed below to provide evidence for thepossible involvement of these cases. O157 VTECthat hybridize with the VT2 polynucleotide probemay carry different VT2 sequences (13). Strains

from humans commonly possess VT2 or VT2c orboth sequences. Polymerase chain reaction (14)showed that strains from all outbreaks except 2and 3 possessed both types of VT2 sequences(Table). O157 VTEC belonging to PT1, such asthose from outbreak 3, usually carry VT1 andonly the VT2 sequence (13). The presence of theVT2, but not the VT2c, sequence differentiatedthe PT2 strain from outbreak 2 from all the otherPT2-associated outbreaks. This property wasexploited during the course of the outbreak toexclude patients infected with PT2 strains thathad the VT2+VT2c genotype and not theoutbreak VT2 genotype.

The O157 VTEC strains were furtheranalyzed by Southern blot hybridization of Eco RI

Table. Outbreaks of infection with O157 VTEC in England and Wales 1995, properties of outbreak strainsOut- Cases Likelybreak Region/setting (HUS/ Phage VT VT2 RFLP PFGE transmissionNo. Month (ref)a Fatal) type probeb subtypec ϕ32511d XbaΙe of infection 1 Jan Northern/Nursing home 7 2 2 2+2c PT2-A PT2-1 Person-

(0/2) to-person 2 May Wessex/Community; 26 2 2 2 PT2-C PT2-2 Foodborne

hospital (2/0) 3 Jul N.W. Thames/Hotel 5 1 1+2 2 PT1-A PT1-1 Foodborne

(0/0) 4 Jul N. Western/Residential 3 2 2 2+2c PT2-A PT2-1 Person-

home; hospital (1/3) to-person 5 Jul Northern/Community 12 2 2 2+2c PT2-A PT2-4 Foodborne

(11) (0/1) 6 Jul Northern/Restaurant 5 2 2 2+2c PT2-A PT2-1a Foodborne

(1/0) 7 Jul East Anglia/Holiday 4 49 2 2+2c PT49-A PT49-1 Foodborne

camp (0/1) 8 Aug Wales/Day nursery; 49 2 2 2+2c PT2-B PT2-3 Foodborne,

community (2/0) person-to- person

9 Oct W. Midlands/ 11 2 2 2+2c PT2-Avar PT2-1b Foodborne Community (12) (4/0)

10 Oct Various/Community 3 RDNCf 1+2 2+2c RDNC-A RDNC-1 Unknown(0/0)

11 Dec Northern/Day 2 49 2 2+2c PT49-B PT49-2 Unknown nursery (0/0)

HUS=hemolytic uremic syndrome; VT= Vero cytotoxin; RFLP=restriction fragment length polymorphisms; PFGE=pulsed-fieldgel electrophoresisaInvestigation of the epidemiology of two outbreaks has been reported previously (11,12)bDetermined by hybridization with digoxigenin-labeled polynucleotide probes for VT1 and VT2 genes (2,3).cBased on polymerase chain reaction amplification with a sense primer specific for either the VT2 or VT2c sequence and adegenerate antisense primer that would anneal to known VT2 sequences (14).dHybridization with a probe comprising digoxigenin-labeled fragments of the VT2-encoding bacteriophage from strainE32511(15). Patterns were designated according to the phage type of the strain and a letter denoting a unique pattern type. ThePT2-Avar pattern differed from PT2-A by the possession of a single extra hybridizing fragment.eProfiles of XbaI digested genomic DNA. Patterns were designated according to the phage type of the strain and differentiatedby number. Thus patterns PT2-1, PT2-2, PT2-3, and PT2-4 differed from each other by at least three fragment positions. Wherethere were single unique band differences from PT2-1 these were designated PT2-1a, etc.fThe designation RDNC indicates that the strain reacts with the typing phages but does not conform to a currently definedpattern.


Dispatches

restriction enzyme-digested genomic DNA with aprobe comprising digoxigenin-labeled fragmentsof the VT2-encoding bacteriophage ϕ32511 (15).Strains of O157 VTEC give an array of up to 20hybridizing fragments; these restriction frag-ment length polymorphisms (RFLPs) are differ-ent for strains of different PTs and candistinguish between strains of the same PT. TheRFLP patterns of the PT1 and RDNC strainsfrom outbreaks 3 and 10 distinguished them fromeach other and from all the remaining strains.The technique indicated that the PT49 strainsfrom outbreaks 7 and 11 had distinct RFLPpatterns (Table), but it did not differentiate thePT2 strains from outbreaks 1, 4, 5, and 6. Thesestrains gave a single pattern (identical to that ofa PT2 strain reported previously [15]) that wehave found most commonly in strains of this PT(unpub. data). The PT2 strains from outbreaks 2and 8 gave RFLP patterns that were differentfrom this common pattern and from each other(Table). In both instances, the test was used toexclude those cases outside the outbreak.

Pulsed-field gel electrophoresis (PFGE) hasbeen applied widely as a highly discriminatorymethod for fingerprinting bacterial pathogens.The method of Barrett et al. (16), modified asindicated in the Figure, identified between 15and 20 XbaI-generated fragments of the O157VTEC strains (Figure). PFGE patterns of PT1,PT49, and RDNC were reproducible and clearlydistinct from each other; they differed from themost common PT2 pattern by at least sixfragment positions. The PT49 strains associatedwith outbreaks 7 and 11 were distinguished fromeach other by their PFGE profile. Results ofPFGE of O157 VTEC of PT2 from the outbreaksare shown in the Figure. The strains that causedoutbreaks 1 and 4 were indistinguishable by allmethods including PFGE (lanes 2 and 4). Twoother strains, from outbreaks 6 and 9, were verysimilar to these two, differing at only onefragment position (lanes 6 and 8), but thepatterns were distinct and reproducible. The PT2strains from outbreaks 5 and 8 (lanes 5 and 7)differed from those associated with outbreaks 1and 4 by at least 3 fragment positions and weredistinguishable from each other. The strain fromoutbreak 2, which possessed only the VT2 gene,was clearly distinct from all the other PT2 strainsby PFGE (lane 3). Criteria for the interpretationof patterns produced by PFGE have beenpublished by Tenover et al. (17). Results of the

analysis of O157 VTEC by PFGE suggest thesecriteria need modification for closely relatedorganisms such as O157 VTEC (18).

Although phage typing and polynucleotideprobes for VT1 and VT2 genes rapidlycharacterized strains from the outbreaks in1995, DNA-based methods were valuable indistinguishing outbreak cases from outlyingones. VT2 gene subtyping was rapid but notvery discriminatory, whereas RFLP and PFGEtechniques differentiated strains but weretime-consuming. The highest level of discrimi-

Figure. Pulsed-field gel electrophoresis of XbaI-digested genomic DNA of O157 VTEC PT2 isolatedfrom outbreaks in 1995. Digests were separated on 1%agarose for 42h at a voltage gradient of 5.6 volts per cmwith a pulse ramp time of 5 to 50 sec. Lane 1 containeda phage lambda DNA 48.5 kb ladder (Sigma). Lanes 2to 8 contained digests of PT2 strains from outbreaks asfollows: lane2, outbreak 1; lane3, outbreak 2; lane4,outbreak 4; lane5, outbreak 5; lane6, outbreak 6;lane7, outbreak 8; lane8, outbreak 9.


Dispatches

nation (by PFGE) distinguished certain O157VTEC that appeared identical by all othermethods. Some of the differences detected byPFGE were minor, but for each outbreak, theywere reproducible with all the strains exam-ined in this study. The epidemiologic orevolutionary significance of these minorvariations is difficult to evaluate.

Investigation of the epidemiology of out-breaks of O157 VTEC infection requires acombined use of typing and fingerprintingmethods in a hierarchic manner consistent withpractical and economic constraints. In outbreakswell defined by epidemiologic studies, phagetyping and the identification of VT1 and VT2genes, including VT2 subtyping, are likely to besufficient. In outbreaks less clearly definedepidemiologically, DNA-based methods mayassist in identifying those strains not associatedwith the outbreak; this is particularly helpfulwhen the outbreak is due to a common phagetype. DNA-based methods have been useful inlinking human infections with associated foods(15) and animals (19).

Although PFGE gives a high level ofdiscrimination between closely related O157VTEC, it has certain disadvantages. It is time-consuming and may not be suitable for rapididentification of large numbers of strains. Werecommend the use of a combination of phagetyping, VT typing, and PFGE to provide gooddiscrimination of O157 VTEC strains inepidemiologic investigations.

AcknowledgmentsWe thank laboratories in England and Wales for sending

clinical samples and bacterial isolates and colleagues in theLaboratory of Enteric Pathogens for technical assistance.

Geraldine A. Willshaw,* Henry R. Smith,*Thomas Cheasty,* Patrick G. Wall,†

and Bernard Rowe**Central Public Health Laboratory, London, United

Kingdom; †Communicable Disease SurveillanceCentre, London, United Kingdom

References 1. Karmali MA. Infection by Verocytotoxin-producing

Escherichia coli. Clin Microbiol Rev 1989;2:15-38. 2. Thomas A, Chart H, Cheasty T, Smith HR, Frost JA,

Rowe B. Vero cytotoxin-producing Escherichia coli,particularly serogroup O157, associated with humaninfections in the United Kingdom: 1989 to 1991.Epidemiol Infect 1993;110:591-600.

3. Thomas A, Cheasty T, Frost JA, Chart H, Smith HR,Rowe B. Vero cytotoxin-producing Escherichia coli,particularly serogroup O157, associated with humaninfections in England and Wales: 1992-4. EpidemiolInfect 1996;117:1-10.

4. Advisory Committee on the Microbiological Safety ofFood. Report on Verocytotoxin-producing Escherichiacoli. London: HMSO; 1995.

5. Sharp JCM, Reilly WJ, Coia JE, Curnow J, Synge BA.Escherichia coli O157 infection in Scotland: anepidemiological overview, 1984-94. PHLS MicrobiologyDigest 1995;12:134-40.

6. Chapman PA, Siddons CA, Harkin MA. Sheep as apotential source of verocytotoxin-producing Escherichiacoli O157. Vet Rec 1996;138:23-4.

7. Griffin PM, Tauxe RV. The epidemiology of infectionscaused by Escherichia coli O157:H7, otherenterohemorrhagic E. coli, and the associated hemolyticuremic syndrome. Epidemiol Rev 1991;13:60-98.

8. Neill MA, Tarr PI, Taylor DN, Trofa AF. Escherichiacoli. In: Hui YH, Gorham JR, Murrell KD, Oliver DO,editors. Foodborne disease handbook 1994. New York:Marcel Dekker; 1994. p. 169-213.

9. Khakhria R, Duck D, Lior H. Extended phage-typingscheme for Escherichia coli O157:H7. Epidemiol Infect1990;105:511-20.

10. Wall PG, McDonnell RJ, Adak GK, Cheasty T, SmithHR, Rowe B. General outbreaks of Vero cytotoxin-producing Escherichia coli O157 in England and Walesfrom 1992-1994. Commun Dis Rep CDR Rev1996;6:R26-33.

11. Stevenson J, Hanson S. Outbreak of Escherichia coli O157phage type 2 infection associated with eating precookedmeats. Commun Dis Rep CDR Rev 1996;6:R116-8.

12. Gammie AJ, Mortimer PR, Hatch L, Brierley AFM,Chada N, Walters JB. Outbreak of Vero cytotoxin-producing Escherichia coli O157 associated withcooked ham from a single source. PHLS MicrobiologyDigest 1996;13:142-5.

13. Thomas A, Smith HR, Rowe B. Use of digoxigenin-labelled oligonucleotide DNA probes for VT2 and VT2human variant genes to differentiate Vero cytotoxin-producing Escherichia coli strains of serogroup O157. JClin Microbiol 1993;31:1700-3.

14. Thomas A, Cheasty T, Chart H, Rowe B. Isolationof Vero cytotoxin-producing Escherichia coliserotypes O9ab:H- and O101:H- carrying VT2variant gene sequences from a patient withhaemolytic uraemic syndrome. Eur J MicrobiolInfect Dis 1994;13:1074-6.

15. Willshaw GA, Thirlwell J, Jones AP, Parry S, SalmonRL, Hickey M. Vero cytotoxin-producing Escherichiacoli O157 in beefburgers linked to an outbreak ofdiarrhoea, haemorrhagic colitis and haemolyticuraemic syndrome in Britain. Letters in AppliedMicrobiology 1994;19:304-7.

16. Barrett TJ, Lior H, Green JH, Khakhria R, Wells JG,Bell BP, et al. Laboratory investigation of a multistatefood-borne outbreak of Escherichia coli O157:H7 byusing pulsed-field gel electrophoresis and phagetyping. J Clin Microbiol 1994;32:3013-7.


Dispatches

17. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA,Murray BE, Persing DH, Swaminathan B. Interpretingchromosomal DNA restriction patterns produced bypulsed-field gel electrophoresis: criteria for bacterialstrain typing. J Clin Microbiol 1995;33:2233-9.

18. Bender JB, Hedberg CW, Besser JM, Boxrud DJ,Macdonald KL, Osterholm MT. Surveillance forEscherichia coli O157 infections in Minnesota bymolecular subtyping. N Engl J Med 1997;337:388-94.

19. Trevena WB, Willshaw GA, Cheasty T, Wray C,Gallagher J. Vero cytotoxin-producing E. coli O157infection associated with farms. Lancet 1996;347:60-1.


Dispatches

Long considered a veterinary disease,cryptosporidiosis has emerged as an importantinfectious disease in humans (1). In immunocom-petent persons, the disease is usually self-limiting; however, in the immunocompromised,it is frequently chronic, more severe, andsometimes fatal. Cryptosporidiosis is one of themajor secondary diagnoses in people with AIDSand is associated with a twofold greater hazard ofdeath than other AIDS-defining diagnoses (2).

A number of major waterborne outbreaks ofcryptosporidiosis have occurred in urban settings(3); however, the disease also occurs sporadically.Since most Cryptosporidium parvum infectionsare self-limiting and symptomatically similar toother diarrheal diseases, the disease may oftenbe undiagnosed or misdiagnosed in the absence ofa recognized outbreak. Consequently, the actualincidence of cryptosporidiosis and the relativeimportance of each of its many modes oftransmission are largely unknown. For thesereasons, laboratory tools are needed forquantitative and qualitative environmentalsampling and for strain analysis ofCryptosporidium isolates. These tools would beextremely valuable for source identification andoutbreak investigations, for correlation withclinically important phenotypes, and for deter-mining risk factors in nonepidemic settings.

A number of new nucleic acid-basedapproaches have been developed for detection,diagnosis, and typing of C. parvum, among thempolymerase chain reaction (PCR)-based teststhat focus either on random amplification of DNApolymorphisms or on specific polymorphicgenetic loci (4-7). These PCR-based tests suggestthe existence of strain variation and thepossibility of two distinct transmission cyclesamong C. parvum isolates that infect humans. Inthis study, we examined genetic polymorphismamong C. parvum isolates from human andnonhuman sources to identify strain-specificmarkers that could be correlated with epidemio-logically important phenotypes.

Analytic Approach

Parasite IsolatesThirty-nine isolates were examined from

stool samples positive for C. parvum: 17 wereobtained from humans or calves duringoutbreaks in the United States and Canada(Table 1); one was a calf isolate (Iowa calf)routinely passaged in neonatal Holstein calves inour laboratories; and 21 were obtained fromcattle from Georgia, Alabama, Ohio, Oklahoma,Kansas, Iowa, Idaho, Utah, and Washington. Allsamples were collected and placed directly into a

Genetic Polymorphism AmongCryptosporidium parvum Isolates:Evidence of Two Distinct Human

Transmission Cycles

We report the results of molecular analysis of 39 isolates of Cryptosporidiumparvum from human and bovine sources in nine human outbreaks and from bovinesources from a wide geographic distribution. All 39 isolates could be divided into eitherof two genotypes, on the basis of genetic polymorphism observed at thethrombospondin-related adhesion protein (TRAP-C2) locus. Genotype 1 was observedonly in isolates from humans. Genotype 2, however, was seen in calf isolates and inisolates from a subset of human patients who reported direct exposure to infected cattleor consumed items thought to be contaminated with cattle feces. Furthermore,experimental infection studies showed that genotype 2 isolates were infective to mice orcalves under routine laboratory conditions, whereas genotype 1 isolates were not.These results support the occurrence of two distinct transmission cycles of C. parvum inhumans.


Dispatches

2.5% potassium dichromate solution and werestored at 4°C. Oocysts were purified by usingdiscontinuous sucrose and Percoll or cesiumchloride gradients (8,9).

Isolation of Genomic DNAParasite DNA was isolated as described by

Kim et al. (10). Briefly, oocysts were ruptured byusing five freeze-thaw cycles (dry ice ethanolbath and 65°C) in a lysis buffer (120 mM NaCl, 10mM EDTA, 25 mM Tris pH 7.5, 1% Sarkosyl)containing proteinase K. The samples wereincubated for 1 hour at 55°C to inactivatenucleases. Then DNA was extracted with phenol/chloroform/isoamyl alcohol (25:24:1) and chloro-form/isoamyl alcohol (24:1), precipitated withabsolute ethanol, washed with 70% ethanol,and resuspended in TE buffer (10 mM Tris pH8.0, 1 mM EDTA).

PCR Amplification and Sequencingand Analysis

The gene fragment of interest, a 369-bpregion of the thrombospondin-related adhesiveprotein (TRAP-C2) of C. parvum, was amplifiedwith the following primers: 5'-CAT ATT CCCTGT CCC TTG AGT TGT-3' and 5'-TGG ACAACC CAA ATG CAG AC-3', which correspond,respectively, to positions 812 to 835 on the codingstrand and positions 1,161 to 1,180 on thenegative strand, of GenBank sequence X77586.The reactions were performed with Perkin-Elmer(Perkin-Elmer Corporation, Foster City, CA)PCR reagents, including 1X PCR buffer, 2.5 mM

MgCl2, 0.2 mM each dNTP, 0.4 mM of eachspecific primer, and 2.5 U of Taq DNApolymerase. After a 1-minute hot start at 94°C,the reactions went through 35 to 40 cycles ofdenaturing at 94°C for 30 seconds, annealing at45°C for 30 seconds, and extension at 72°C for 1minute, followed by a 72°C incubation forstrand completion.

An aliquot of each PCR product wasexamined by agarose gel electrophoresis; theremaining PCR product was purified by using theWizard PCR Prep Kit (Promega Corporation,Madison, WI). Purified PCR fragments weresequenced directly on an ABI 377 automatedsequencer by fluorescent cycle sequencing usingdye-terminator chemistry with AmpliTaq FS(Perkin-Elmer-Applied Biosystems) according tothe manufacturer’s recommended procedure.The same primer sets used initially for PCR wereused again for sequencing, diluted to aconcentration of 10 pMoles in the finalsequencing reaction. Downstream analysis ofsequence data was accomplished by using theSequence Navigator program (Perkin–Elmer-Applied Biosystems). Multiple sequence align-ments were performed by using the Pileupprogram (11). Animal isolates were manuallysequenced by the dideoxy chain-terminationmethod (12), using the Sequenase Version 2.0 kit(J.T. Baker, Phillipsburg, NJ) with the sequences ofa few isolates confirmed by automated sequencing.

Experimental Infection StudiesPurified oocysts ranged in age from 1 to 6

months at the time of inoculation of cell culturesor animals. This age range is well within thestorage time that maintains oocyst viability andinfectivity (e.g., laboratory-passaged isolates are40% to 50% viable after storage for 6 months).Approximately 106 oocysts were administeredorally to 2-day-old calves or to 4- to 6-day-oldBALB/c or SCID mice by using establishedprocedures (8,13). Beginning at day 5, stools werecollected and examined daily by light micros-copy or by immunofluorescent flow cytometryfor C. parvum oocysts (14).

Findings

Sequence Determination and AnalysisA single specific band of 369 bp, correspond-

ing to bases 812 to 1180 of the 1.1 kb C. parvumTRAP-C2 gene (GenBank accession number

Table 1. Cryptosporidium parvum isolates examined

ImplicatedIsolate source Host Ref.

Maine,1993 apple cider human 16Wisconsin, 1993 drinking water human 3Wisconsin, 1996 drinking water human *Georgia (day person-tohuman *

care), 1995 personGeorgia (water- recreational human 17

park), 1995 waterFlorida, 1995 drinking water human 18British drinking water human 15

Columbia, 1996Texas, 1996 unknown human *Pennsylvania, bovine contact human, *

1997 calf

*Reference is this article.


Dispatches

X77586) was amplified from 39 different isolates(Figure 1). Although the sequence similarity wasvery high among all gene fragments, multiplealignments showed two primary genotypes.These genotypes could be established on the basisof nucleotide substitutions at five independentpositions, three being silent changes and theother two resulting in amino acid changes(Figure 2). Of the five changes, four weretransitions, and one was a transversion.

Genotype 1 included human isolates fromWisconsin, Georgia, Florida, and Texas. Geno-type 2 contained human isolates from Maine andBritish Columbia, human and calf isolates fromPennsylvania, the Iowa calf laboratory strain,and 21 bovine isolates from various locationsaround the country. Two human isolates (onefrom Florida and the Texas isolate) appeared torepresent a variant of genotype 1. In both cases,they shared the first four positions with the othergenotype 1 isolates. The fifth position, however,was the same as that of genotype 2 isolates.

Experimental Infection StudiesThe results of experimental infection studies

are shown in Table 2. The genotype 2 isolatesfrom human outbreaks in Maine and Pennsylva-

nia and from a calf in Iowa all readily infect bothmice and calves. The genotype 2 isolate fromBritish Columbia was also reported to beinfective to immunosuppressed C57BL/6 mice(15). None of the genotype 1 isolates fromhumans—from Wisconsin, Florida, a Georgiaday-care facility, and a Georgia water park—could be established in either mouse or calf. Asingle sample, the Georgia day-care isolate, wasexamined for its ability to infect a neonatal pig.This isolate caused a brief moderate infection in aneonatal pig host (data not shown) but not incalves or mice. One of the Wisconsin isolates andthe Georgia day-care isolate were tested for theirinfectivity to MDCK cell cultures; both success-fully infected this cell line (data not shown).

ConclusionsAll isolates examined in this study could be

grouped easily into two distinct genotypesdefined by nucleotide substitutions at fivepositions within the TRAP-C2 locus, withgenotype 1 containing a variant at the fifthposition that was represented by two isolates. Allisolates in genotype 1 were from human stool.The isolates in genotype 2, however, were fromboth human and bovine sources. In the limitednumber of isolates that were tested inexperimental infection studies, all genotype 2isolates could be established readily in mice andcalves. None of the genotype 1 isolates, however,could be shown to be infective to either of thesehosts. The genotype and experimental infectiondata suggest the possibility of two distinctpopulations of C. parvum cycling in humans. Onepopulation appears to involve zoonotic transmis-sion from calf-to-human with subsequenthuman-to-human and human-to-calf transmis-sion. The other population appears to involve ananthroponotic transmission cycle, exclusively inhumans. This hypothesis is consistent with thedata from the epidemiologic investigations fromwhich the isolates were obtained.

Genotype 2 characteristics were identified inhuman isolates from the Maine 1993, BritishColumbia 1996, and Pennsylvania 1997 out-breaks, and in all isolates from bovine sources.Both the Maine and Pennsylvania outbreakscould be directly linked to a calf source of C.parvum. The Maine outbreak was associatedwith contaminated apple cider (16). Interest-ingly, C. parvum oocysts were isolated directlyfrom apple cider, the press used for preparing the

Figure 1. Alignment of TRAP-C2 nucleotide positionsthat show polymorphism among Cryptosporidiumparvum isolates from human and nonhuman sources.Published calf sequence refers to Genbank accessionnumber X77586. Other bovine (n=21) refers to 21samples (from Georgia, Alabama, Ohio, Oklahoma,Kansas, Iowa, Idaho, Utah, and Washington) thathad the same genotype.

Position: 15 42 64 111 244Isolate:

Milwaukee 93 /1 G C T C T Milwaukee 93 /2 G C T C T Milwaukee 93 /3 G C T C T Milwaukee 96 G C T C T Georgia-DC 95 G C T C T Georgia-WP 95 /1 G C T C T Georgia-WP 95 /2 G C T C T Florida 95 /1 G C T C T Florida 95 /2 G C T C T Florida 95 /3 G C T C T Florida 95 /4 G C T C TFlorida 95 /5 G C T C CTexas 96 G C T C C

Maine Cider 93 A T G T CB.C. Canada 96 A T G T C Pennsylvania 97 (H) A T G T C Pennsylvania 97 (C) A T G T C Iowa calf (CDC) A T G T C Published calf A T G T C Other bovine (n=21) A T G T C

HumanIsolates

CalfIsolates

Genotype 1

Genotype 2


Dispatches

cider, and a calf stool specimen from the farmthat supplied the apples.

The Pennsylvania focus involved threefamilies that together purchased three youngcalves that subsequently developed scours. Ninemembers of three families had diarrhea, and twowere hospitalized. C. parvum oocysts were isolatedfrom two calves and five humans; all isolatesexamined demonstrated the genotype 2 pattern.

The British Columbia isolate came from ahuman patient infected in an outbreak(approximately 2,000 cases) that occurred inthe small rural community of Cranbrook (15).During the outbreak investigation,Cryptosporidium oocysts were identified inhuman fecal specimens, in cattle manurespecimens found near the watershed, and inwater samples from the reservoir intake.

Figure 2. DNA and putative amino acid sequences of Cryptosporidium parvum TRAP-C2 genotypes 1 and 2.


Dispatches

Of the genotype 1 isolates examined,epidemiologic investigations were conducted forthe Georgia water park 1995, Florida 1995, andWisconsin 1993 outbreaks. In the Georgia waterpark outbreak, approximately 2,900 persons metthe case definition for clinical cryptosporidiosis.In a sample of these patients, the following riskfactors were evaluated in telephone interviews:swimming in lakes or pools, exposure to day careor to persons with diarrhea, contact with younganimals, drinking water from various sources,chronic illness, and water park attendance. Theonly factor independently associated with diar-rheal disease was water park attendance (17).

The Florida 1995 outbreak occurred at a daycamp in central Florida and had approximately70 cases (18). Risk factors examined includedparticipating in camp activities, eating lunchesprovided at the camp, and drinking water fromvarious specified sources. C. parvum oocystswere observed in the stools of 16 persons and inwater from an outside tap. Fecal contamination(of unknown origin) of the tap was the suspectedsource of the outbreak. Five specimens wereexamined from this outbreak, all of whichbelonged to the genotype 1 grouping; onedisplayed additional polymorphism at nucleotideposition 244 (Figure 1). The Texas 1995 isolateshowed this same polymorphism, which we thinkis most accurately described as a subset ofgenotype 1.

The Wisconsin 1993 outbreak, which affectedmore than 403,000 people, is the largestwaterborne disease outbreak ever recorded in theUnited States. Four isolates were examined,three isolated during the original outbreak and a

fourth isolated in 1996 from an AIDS patientwith a chronic infection who had initially beeninfected in the 1993 outbreak.

During the Wisconsin outbreak, possiblesources of contamination of Lake Michigan withCryptosporidium oocysts included cattle alongtwo rivers that fed Milwaukee Harbor, slaughter-houses, and human feces (3,19). The genotypicand experimental infection data from the fourisolates we examined suggest a human ratherthan bovine source. However, these resultscome from the analysis of only four samplesfrom a massive outbreak, and the degree towhich these samples are representative of theentire outbreak remains uncertain.

All genotype 2 isolates examined in thisstudy came from persons that had direct links orpotential exposure to C. parvum from an infectedanimal. All samples tested in experimentalinfection studies were also infective to both miceand calves. In the genotype 1 isolates, however,while the initial source of the cases was neverdirectly determined experimentally, no con-firmed links to bovine sources were found, butexposure to water contaminated with humanfeces could have occurred. Furthermore, of theisolates tested in experimental infection studies,none could successfully infect laboratory ani-mals. These results lead us to suggest thepossibility of a second transmission cycle that isanthroponotic and maintained through person-to-person contact or through human sewagecontamination of the water supply.

The observations reported here with respectto genotypic variation among C. parvum isolatesfrom humans and animals are very similar tothose reported by other groups. These studiesgenerally reported one allozyme pattern orgenotype associated with human isolates and asecond genotype or allozyme in bovine samplesand a subset of human samples. The specificgenes or regions examined differed in each studybut included electromorphs of phosphoglucomu-tase and hexokinase (20), random amplifiedpolymorphic DNA (RAPD) analysis of anunspecified region (4), a repetitive DNA sequence(6), the 18S rRNA gene and adjacent internaltranscribed spacer (ITS) region I (5), theCryptosporidium oocyst wall protein (COWP) locus(7), and the dihydrofolate reductase-thymidylatesynthase (DHFR-TS) gene (21). At least twoadditional studies suggest the possibilities of twotransmission cycles, on the basis of epidemiologic or

Table 2. Experimental infection studies withCryptosporidium parvum isolates from various sources

ExperimentalIsolate host InfectionMaine, 1993 mouse +

calf +Wisconsin, 1993 mouse -

calf -Georgia (day care), 1995 mouse -

calf NDGeorgia (water park), 1995 mouse -

calf -Florida, 1995 mouse -

calf NDIowa (bovine), 1984 mouse +

calf +ND = not done


Dispatches

experimental observations (22,23).The TRAP-C2 protein is a member of a class

of proteins present in all apicomplexansexamined to date (24-26). This protein isassociated with the cell surface and micronemalcomplex of these parasites and is thought to beinvolved in surface attachment; consequently,changes in this protein could affect attachmentspecificity and the resultant host range. Werethis the case, such a mechanism could explainwhy the host range of one genotype might bedifferent from that of a second genotype,resulting in distinct transmission cycles. In thetwo isolates that make up the variant of genotype1, the T-to-C transition results in a change in theamino acid sequence from tyrosine to histidine. Ifvariations in this protein affect host preference,the histidine-to-tyrosine change would have to beinconsequential with respect to protein functionand host specificity. Observations in Plasmo-dium suggest that the WCSP motif in the TRAPgene is the functional domain involved insurface attachment; however, the polymor-phism we observed in C. parvum did notinvolve this region. Additional studies areneeded to clarify the relationship, if any, ofpolymorphism in this gene to host range.

The conclusion that two transmission cyclesexist for C. parvum is now supported by theresults of independent groups, using markers atsix different genetic loci. This conclusion, if valid,may have important implications for theprevention and control of cryptosporidiosis inurban settings. Cattle have been the mostcommonly implicated source of water contamina-tion in outbreaks outside the United States butnot conclusively within the United States.Measures for preventing water contaminationhave in some cases included the removal of cattlefrom watershed areas in or around municipali-ties. If, however, sewer overflows and inadequatesewage treatment are the primary source ofwater contamination in urban settings whereanthroponotic cycles are being maintained,focusing solely on cattle could fail to eliminate avery important source of infection.

The results of this study suggest the need to1) combine the typing approaches of variousgroups into a multilocus approach for genetictyping of C. parvum that would result in areliable and robust method for strain typing, 2)apply multilocus typing to a large number of C.

parvum isolates both from epidemic and isolatedcases and from a large geographic distribution todetermine the prevalence of these two genotypesand their quantitative importance as indicatorsof specific risk factors, and 3) identify additionalgenetic loci that will allow more precisedetermination of strain variation and linkage ofgenotypic variation to specific clinical andepidemiologically important outcomes.

AcknowledgmentsWe thank Thomas R. Navin, David G. Addiss, and

Dennis D. Juranek, Centers for Disease Control andPrevention (CDC), for critical reading of the manuscript;Gisela Withers, Alfred L. Loban, Susan Yeager, BarbaraFisher, and Marshall P. Deasy, Pennsylvania Department ofHealth, Charles Brummit, St. Lukes Medical Center,Milwaukee, WI, and Michael J. Beach, CDC, for providingsamples; Brian Holloway and staff, CDC, for providing DNAoligonucleotide primers; Robert W. Ryder, Yale UniversitySchool of Medicine, and Mark L. Wilson, The University ofMichigan, for help in coordinating the study; and KimberleyB. Donaldson, Donghyun Hahn, Kimberly Y. Won, ChunfuYang, and Lillian Escalante, CDC, for technical support incompleting the work.

Michael M. Peng,* Lihua Xiao,† Amanda R.Freeman,† Michael J. Arrowood,† Ananias A.

Escalante,† André C. Weltman,‡Corinne S.L. Ong,¶ William R. Mac Kenzie,†

Altaf A. Lal,† and Charles B. Beard†*The University of Michigan, Ann Arbor, Michigan,USA; †Centers for Disease Control and Prevention,

Atlanta, Georgia, USA; ‡Pennsylvania Department ofHealth, Harrisburg, Pennsylvania, USA;

and ¶University of British Columbia,British Columbia, Canada

References 1. Guerrant RL. Cryptosporidiosis: an emerging, highly

infectious threat. Emerg Infect Dis 1997;3:51-7. 2. Colford JM, Tager IB, Hirozawa AM, Lemp GF, Aragon

T, Petersen C. Cryptosporidiosis among patientsinfected with human immunodeficiency virus. Am JEpidemiol 1996;144:807-16.

3. MacKenzie WR, Hoxie NJ, Proctor ME, Gradus MS,Blair KA, Peterson DE, et al. A massive outbreak inMilwaukee of Cryptosporidium infection transmittedthrough the public water supply. N Engl J Med1994;331:161-7.

4. Morgan UM, Constantine CC, O’Donoghue P, MeloniBP, O’Brien PA, Thompson RCA. Molecularcharacterization of Cryptosporidium isolates fromhumans and other animals using random amplifiedpolymorphic DNA analysis. Am J Trop Med Hyg1995;52:559-64.

5. Carraway M, Tzipori S, Widmer G. Identification ofgenetic heterogeneity in the Cryptosporidium


Dispatches

parvum ribosomal repeat. Appl Environ Microbiol1996;62:712-6.

6. Bonnin A, Fourmaux MN, Dubremetz JF, Nelson RG,Gobet P, Harly G, et al. Genotyping human and bovineisolates of Cryptosporidium parvum by polymerasechain reaction-restriction fragment lengthpolymorphism analysis of a repetitive DNA sequence.FEMS Microbiol Lett 1996;137:207-11.

7. Spano F, Putignani L, McLauchlin J, Casemore DP,Crisanti A. PCR-RFLP analysis of the Cryptosporidiumoocyst wall protein (COWP) gene discriminatesbetween C. wrairi and C. parvum, and between C.parvum isolates of human and animal origin. FEMSMicrobiol Lett 1997;150:209-17.

8. Arrowood MJ, Sterling CR. Isolation of Cryptosporidiumoocysts and sporozoites using discontinuous sucrose andisopycnic Percoll gradients. J Parasitol 1987;73:314-9.

9. Arrowood MJ, Donaldson K. Improved purificationmethods for calf-derived Cryptosporidium parvumoocysts using discontinuous sucrose and cesiumchloride gradients. J Eukaryot Microbiol 1996;43:S89.

10. Kim K, Gooze L, Petersen C, Gut J, Nelson RG.Isolation, sequence and molecular karyotype analysisof the actin gene of Cryptosporidium parvum. MolBiochem Parasitol 1992;50:105-14.

11. Wisconsin Package [Computer program]. Version 9.0.Madison (WI): Genetics Computer Group; 1996.

12. Sanger F, Miklen S, Coulson AR. DNA sequencing withchain-terminating inhibitors. Proc Natl Acad Sci U S A1977;74:5463-7.

13. Mead JR, Arrowood MJ, Sidwell RW, Healey MC.Chronic Cryptosporidium parvum infections incongenitally immunodeficient SCID and nude mice. JInfect Dis 1991;163:1297-304.

14. Arrowood MJ, Hurd MR, Mead JR. A new method forevaluating experimental cryptosporidial parasite loadsusing immunofluorescent flow cytometry. J Parasitol1995;81:404-9.

15. Ong CSL, Pearce M, Eisler D, Goh SH, King AS, BowieWR, and Isaac-Renton JL. An outbreak ofcryptosporidiosis in southeastern British Columbia,Canada. In: Proceedings of the Symposium onWaterborne Cryptosporidium American Water WorksAssociation. In press 1997.

16. Millard PS, Gensheimer KF, Addiss DG, Sosin DM,Beckett GA, Houck-Jankoski A, Hudson A. Anoutbreak of crytposporidiosis from fresh-pressedapple cider. JAMA 1994;272:1592-6.

17. Beach MJ, McNeil M, Arrowood M, Kreckman L, CraigA, Donaldson K, et al. Cryptosporidium in a waterpark: largest U.S. recreational waterborne outbreak[abstract]. 45th Annual Epidemic Intelligence ServiceConference, Atlanta. 1996.

18. Centers for Disease Control and Prevention. Outbreakof cryptosporidiosis at a day camp—Florida, July-August 1995. MMWR Morb Mortal Wkly Rep1996;45:442-4.

19. Addiss DG, Mac Kenzie WR, Hoxie NJ, Gradus MS,Blair KA, Proctor JE, et al. Epidemiologic features andimplications of the Milwaukee cryptosporidiosisoutbreak. In: Betts WB, Casemore D, Fricker C, SmithH, Watkins J, editors. Protozoan parasites and water.Cambridge, United Kingdom: Royal Society ofChemistry; 1995. p. 19-25.

20. Awad-El-Kariem FM, Robinson HA, Dyson DA, EvansD, Wright S, Fox MT, McDonald V. Differentiationbetween human and animal strains of Cryptosporidiumparvum using isoenzyme typing. Parasitology1995;110:129-32.

21. Vasquez JR, Gooze L, Kim K, Gut J, Petersen C, NelsonRG. Potential antifolate resistance determinants andgenotypic variation in the bifunctional dihydrofolatereductase-thymidylate synthase gene from human andbovine isolates of Cryptosporidium parvum. MolBiochem Parasitol 1996;79:153-65.

22. Pozio E, Gomez Morales MA, Barbieri FM, La Rosa G.Cryptosporidium: different behaviour in calves ofisolates of human origin. Trans R Soc Trop Med Hyg1992;86:636-8.

23. Hojlyng N, Molbak K, Jepsen S. Cryptosporidiosis inhuman beings is not primarily a zoonosis. J Infect1985;11:270-2.

24. Robson KJH, Hall JRS, Davies LC, Crisanti A, HillAVS, Wellems TE. Polymorphism of the TRAP gene ofPlasmodium falciparum. Proc R Soc Lond B Biol Sci1990;242:205-16.

25. Muller HM, Scarselli E, Crisanti A. Thrombospondinrelated anonymous protein (TRAP) of Plasmodiumfalciparum in parasite-host cell interactions.Parassitologia 1993;35 (suppl):69-72.

26. Templeton TJ, Kaslow DC. Cloning and cross-speciescomparison of the thrombospondin-related anonymousprotein (TRAP) gene from Plasmodium knowlesi,Plasmodium vivax and Plasmodium gallinaceum. MolBiochem Parasitol 1997;84:13-24.


Commentary

Irradiation Pasteurization of SolidFoods: Taking Food Safety to the NextLevel

In the 19th century, milk from diseased cattleproduced in unsanitary surroundings anddistributed under filthy conditions to anincreasingly urbanized population sickened andkilled consumers by the thousands (1,2). Wideacceptance of the germ theory and the sanitaryawakening that followed led to vast improve-ments in animal health and hygiene, and thesafety of dairy products improved substantially.Dairy farmers in northern Europe discoveredthat heating milk fed to their calves furtherreduced the risk for tuberculosis in their herds,and after overcoming concerns that the newthermal pasteurization technology would corruptthe dairy industry, destroy the nutritional valueof milk, and lead to serious public healthproblems, the same level of protection was offeredto human consumers a few decades later (3,4).More recently, thermal pasteurization has beensuggested for eliminating low level contamina-tion of juice by foodborne pathogens (5,6).However, for the safety of solid foods that enterkitchens as raw agricultural commodities,including meat, poultry, and seafood, wecontinue to rely solely on animal health programsand sanitation. Therefore, as we approach the21st century, preventable illness and deathcaused by vegetative bacterial and parasiticfoodborne pathogens remain substantial publichealth problems (7,8).

Irradiation pasteurization of solid foods withlow doses of gamma rays, X-rays, and electronswill effectively control vegetative bacterial andparasitic foodborne pathogens (9-11). Publicconcerns, similar to those raised against thermalpasteurization of milk, have been advanced inopposition to irradiation pasteurization, and ithas been claimed that if we but paid moreattention to sanitation and proper cooking, theseproducts could be safely consumed withoutintroducing new technologies.

Perhaps. However, the residual risk forinfection that remains after state-of-practicesanitation during production, harvest, process-ing, distribution, and preparation yields anunacceptable level of illness and death. Inaddition, the admonition to properly cook worksonly if culturally acceptable food preferences do

not include undercooked and raw foods.Increased interest (encouraged by the publichealth and nutrition community) in freshproduce as part of a high fiber, low fat diet,further reduces the effectiveness of propercooking as a disease control strategy (12).

Recent outbreaks of foodborne illness associ-ated with undercooked meat and uncooked freshproduce, and the emergence of the previouslyunrecognized foodborne hazards that spawnedthe conference whose proceedings are reported inthis issue of Emerging Infectious Diseases, havestimulated interest in methods of pasteurizingsolid food without altering its raw appearanceand characteristics. Research is under way on avariety of promising approaches, includingpulsed energy, bright light, high pressure, andother nonthermal technologies, but few are readyfor immediate application (13,14; Fed Reg61:42381-83, 1997). Irradiation pasteurization,on the other hand, is a well-established processwith clearly documented safety and efficacy thatcan be put into widespread use as quickly asfacilities can be sited and built (15).

Good practice guidelines and Hazard Analy-sis and Critical Control Points programs(HACCP) can result in raw meat, poultry,seafood, and produce with sufficiently low levelsof pathogen contamination that irradiation dosesas low as 1 to 3 kGy yield adequate margins ofsafety for common foodborne pathogens such asCampylobacter, Cryptosporidium, Escherichiacoli, Listeria, Salmonella, and Toxoplasma (9).No other control for Campylobacter contamina-tion of poultry meat is apparent, and otherapproaches to ground beef safety have proveninadequate to prevent intermittent low levelcontamination with E. coli O157:H7. Likewise, noother solutions are immediately available tocontrol pathogen contamination of produceintended for raw consumption, and irradiationdoses used appear adequate for the bacterial andparasitic pathogens involved in recent outbreaks(Donald Thayer, pers. comm.)

The food industry appears reluctant to fullyembrace irradiation pasteurization despite theobvious and painful failure of alternativeapproaches to prevent foodborne infections.Much of this reluctance stems from theperception that consumers reject the process andwill refuse to buy irradiated food. Indeed, surveyshave shown considerable consumer confusionand ignorance about food irradiation (16), and


Commentary

reports on public antipathy toward thingsradioactive abound. However, consumer surveysalso demonstrate profound and growing publicconcern about microbial food safety, anddecreasing concerns about the safety of irradi-ated food (17). Knowing as little about it as theyappear to, approximately half of the consumerssurveyed have expressed willingness to tryirradiated food if it will decrease their risk forillness (16). In addition, when educated aboutfood irradiation, 90% of survey participantsexpressed interest in purchasing irradiatedfoods; sampling such food increased interest to99% (18).

Irradiation pasteurization is not the cure forall food safety ills. Pasteurization of any sort is nomatch for bad sanitation and substandardpractices, and irradiation pasteurization can beoverwhelmed by large numbers of pathogens.Just as obviously, foods produced and processedunder appropriate conditions that are thenproperly packaged and irradiated are subject topostpasteurization contamination. In addition,the doses of irradiation used to pasteurize freshmeat and poultry are not sufficient to killbacterial spores. Thus, if anaerobic packaging isthe method used to protect irradiated foods frompostpasteurization contamination, Clostridiumbotulinum could pose a risk if the cold chain isdisrupted.

Vibrio infections associated with consump-tion of raw molluscan shellfish can be preventedwith irradiation pasteurization, but the Norwalk-like viruses also frequently associated with rawshellfish appear to be more radioresistant thanvegetative bacterial pathogens. Levels of irradia-tion an order of magnitude greater thanpasteurizing doses for meat and poultry also arenecessary to inactivate hepatitis A virus. Toreduce the risk for foodborne hepatitis A andNorwalk virus infections, it will be necessary toreduce the level of exposure of food to humanfeces. This is true regardless of whether or notfoods are to be pasteurized. Although irradiationpasteurization will not eliminate all seafood-borne pathogens, it will reduce the potential ofseafood to cause illness. Seafood HACCP andadvances in viral diagnostics and environmentalvirology will help ensure that prepasteurizationconditions are sufficient to yield seafoodappropriate for irradiation pasteurization (19,20).Just as thermal pasteurization works well for

liquid foods like milk and juice, but not for solidfoods for which raw characteristics are desired,irradiation pasteurization works well for meat,poultry, seafood, and soft fruit, but wilts leafyvegetables and sprouts. That irradiation pasteur-ization does not work for every food and everypathogen is poor justification for not applying itfor those food/pathogen combinations for which ithas been shown to work so well.

Consumer surveys have demonstrated publicconcerns over worker and environmental safetythat have also contributed to the reluctance ofsome to build and use food irradiation facilities.These concerns are appropriate and addressable.Because food irradiation and irradiation steril-ization of nonfood items like medical supplies areso well established, proper facilities design andoperating characteristics are well known. Therelatively short half-life of Cobalt 60 and itsinsolubility in water reduce environmentalconcerns, which can be eliminated altogether byusing electricity-generated X-rays and electronbeams instead of a radioactive source. Propereducation and training has protected employeesof irradiation sterilization facilities; employees offood irradiation facilities should not be qualita-tively different from other employees in similarindustries.

Thus, a broadly applicable solution to manyof our food safety problems exists and has existedfor a number of decades. It is disappointing thatthe public health community has been so silentfor so long on this issue. Faced with the liability ofmarketing hazardous foods, it is puzzling why thefood industry has not stepped into the vacuumcreated by this lack of leadership from publichealth. Presentations at the Conference onEmerging Foodborne Pathogens make it clearthat new foodborne hazards are being stacked ontop of old, unresolved food safety problems—broadly applicable solutions are desperatelyneeded. Just as thermal pasteurization of milkprotected us from E. coli O157:H7 before we knewit was in raw milk, irradiation pasteurization canprotect us from tomorrow’s emerging foodbornepathogen.

Michael T. Osterholm* and Morris E. Potter† *Minnesota Department of Health, Minneapolis,

Minnesota, USA; and †Centers for Disease Controland Prevention, Atlanta, Georgia, USA


Commentary

References 1. Magruder GL. Further observations on the milk

supply of Washington, D.C. JAMA 1910;55:581-9. 2. Kelley ER, Osborn SH. Further evidence as to the

relative importance of milk infection in thetransmission of certain communicable diseases ofman. Am J Pub Health 1920;10:66-73.

3. Steele JH, Engel RE. Radiation processing of food. JAm Vet Med Assoc 1992;201:1522-9.

4. Potter ME, Kaufmann AF, Blake PA, Feldman RA.Unpasteurized milk: hazards of a health fetish.JAMA 1984;252:2048-52.

5. Besser RE, Lett SM, Weber JT, Doyle MP, BarrettTJ, Wells JG, Griffin PM. An outbreak of diarrheaand hemolytic uremic syndrome from Escherichiacoli O157:H7 in fresh-pressed apple cider. JAMA1993;269:2217-20.

6. Millard PS, Gensheimer KF, Addiss DG, Sosin DM,Beckett GA, Houk-Jankowski A, Hudson A. Anoutbreak of cryptosporidiosis from fresh-pressedapple cider. JAMA 1994;272:1592-6.

7. Foegeding PM, Roberts T. Foodborne pathogens:risks and consequences. Ames (IA): Council forAgricultural Science and Technology; 1994; TaskForce Report No. 122.

8. Tauxe RV. Emerging foodborne disease: an evolvingpublic health challenge. Emerg Infect Dis 1997;3:425-34.

9. Monk JD, Beuchat LR, Doyle MP. Irradiationinactivation of food-borne microorganisms. Journalof Food Protection 1995;58:197-208.

10. Radomyski T, Murano EA, Olson DG, Murano PS.Elimination of pathogens of significance in food bylow-dose irradiation: a review. Journal of FoodProtection 1994;57:73-86.

11. Lagunas-Solar MC. Radiation processing of foods: anoverview of scientific principles and current status.Journal of Food Protection 1995;58:186-92.

12. Foerster SB, Kizer KW, DiSogra LK, Bal DG, KriegBF, Bunch KL. California’s 5 a day—for betterhealth campaign: an innovative population-basedeffort to effect large-scale dietary change. A JPrevent Med 1995;11:124-31.

13. Paul P, Chawla SP, Thomas P, Kesavan PC. Effect ofhigh hydrostatic pressure, gamma-irradiation andcombination treatments on the microbiologicalquality of lamb meat during chilled storage. Journalof Food Safety 1997;16:263-71.

14. Shigeshia T, Ohmori T, Saito A, Taji S, Hayashi R.Effect of high hydrostatic pressure on characteristicsof pork slurries and inactivation of microorganismsassociated with meat and meat products. Int J FoodMicrobiol 1991;12:207-16.

15. Thayer DW, Josephson ES, Brynjolfsson A, GiddingsGG. Radiation pasteurization of food. Ames (IA):Council for Agricultural Science and Technology;1996; Issue Paper No. 7.

16. Resurreccion AVA, Galvez FCF, Fletcher SM, MisraSK. Consumer attitudes toward irradiated food:results of a new study. Journal of Food Protection1995;58:193-6.

17. Bruhn CM. Consumer attitudes and marketresponse to irradiated food. Journal of FoodProtection 1995;58:175-81.

18. Bruhn CM. Consumer concerns: motivating toaction. Emerg Infect Dis 1997;3:511-15.

19. Jaykus LA. Epidemiology and detection as optionsfor control of viral and parasitic foodborne disease.Emerg Infect Dis 1997;3:529-39.

20. Garrett ES, Lima dos Santos C, Jahncke ML. Public,animal, and environmental health implications ofaquaculture. Emerg Infect Dis 1997;3:453-57.


Letters

Non-O157 Shiga Toxin–ProducingEscherichia coli Infections in Europe

To the Editor: Shiga toxin–producing Escherichiacoli (STEC) infections are an important cause ofsevere human disease. Although most infectionsare caused by strains of serogroup O157, STECpathogenic to humans may belong to otherserogroups usually referred to as non-O157 STEC.

Recently, Tarr et al. (1) and Acheson et al. (2)described infections attributable to STEC O103and expressed concern that non-O157 STEC maypose an underestimated threat to public health inthe United States. In fact, non-O157 STEC isoften overlooked in clinical microbiology labora-tories because the toxigenic phenotype is notexploited to identify such pathogens. Rather,most laboratories use sorbitol MacConkey agarand serotyping (which cannot detect most non-O157 STEC) to identify E. coli O157:H7.

Since the end of the 1980s, non-O157 STECinfections have caused as many as 10% to 30% ofsporadic cases of hemolytic uremic syndrome(HUS) in Germany (3), Italy (4), and the UnitedKingdom (5). Moreover, HUS outbreaks havebeen associated with STEC O111:H- in Italy (6)and France (7).

During 1996, we observed a sudden increasein infections attributable to STEC O103 and O26in Germany and Italy. In our laboratory inGermany, E. coli O103:H2 was not identifiedamong 345 non-O157 STEC isolated between1985 and 1995 but represented 12 (18.2%) of 66 ofthe non-O157 STEC isolated during 1996. HUSdeveloped in two infected patients.

Among cases reported to Italy’s nationwideHUS surveillance system from 1988 to 1995,evidence of infection with STEC O103 or O26 wasfound in two (1.5%) and nine (6.6%) of 135 cases,respectively. Since 1996, infection with STECO103 and O26 has been diagnosed in three (11%)and nine (33%) of 27 HUS cases, respectively.

These observations indicate that identifica-tion of non-O157 STEC should be considered byclinical laboratories. Immunoenzymatic tests(based on either toxin antibodies or receptors)that detect Shiga toxins produced by fecalbacterial isolates or present in stool specimensare now available (8,9). Use of these tests shouldbe considered in analyzing the stools of patientswith HUS, bloody diarrhea, or painful nonbloodydiarrhea, if classic microbiologic analysis fails to

yield E. coli O157:H7 or another standardenteric pathogen, such as Campylobacter,Salmonella, or Shigella.

The sudden appearance or increase of rarenon-O157 STEC in our populations is worrisome.Most non-O157 STEC, as well as the sorbitolfermenting O157:H- strains (10) associatedwith HUS in several European countries,would be missed by laboratories using standardmicrobiologic detection methods, such assorbitol MacConkey agar screening. Because ofthe considerable clinical and epidemiologicurgency, clinical microbiologists and physi-cians should seek out these such pathogens inappropriate clinical situations.

Alfredo Caprioli,* Alberto E. Tozzi,*Gianfranco Rizzoni,† and Helge Karch‡*Istituto Superiore di Sanità, Rome, Italy;

†Ospedale Pediatrico Bambino Gesù, Rome, Italy;‡Universitat Wuerzburg, Germany

References 1. Tarr PI, Fouser LS, Stapleton AE, Wilson RA, Kim HH,

Vary JC, Clausen CR. Hemolytic-uremic syndrome in asix-year-old girl after a urinary tract infection withShiga-toxin-producing Escherichia coli O103:H2. NEngl J Med 1996;335:635-8.

2. Acheson DWK, Wolf LE, Park CH. Escherichia coli andthe hemolytic-uremic syndrome. N Engl J Med1997;336:515.

3. Bitzan M, Moebius E, Ludwig K, Muller-Wiefel DE,Heesemann J, Karch H. High incidence of serumantibodies to Escherichia coli O157 lipopolysaccharidein children with the hemolytic uremic syndrome. JPediatr 1991;119:380-5.

4. Caprioli A, Luzzi I, Rosmini F, Pasquini P, CirrincioneR, Gianviti A, et al. Hemolytic-uremic syndrome andvero-cytotoxin-producing Escherichia coli infection inItaly. J Infect Dis 1992;166:154-8.

5. Kleanthous H, Smith HR, Scotland SM, Gross RJ,Rowe B, Taylor CM, Milford DV. Haemolytic-uraemicsyndromes in the British Isles, 1985-8: association withverocytotoxin producing Escherichia coli. Part 2:microbiological aspects. Arch Dis Child 1990;65:722-7.

6. Caprioli A, Luzzi I, Rosmini F, Resti C, Edefonti A,Perfumo F, et al. Communitywide outbreak ofhemolytic-uremic syndrome associated with non-O157verocytotoxin-producing Escherichia coli. J Infect Dis1994;169:208-11.

7. Boudaillez B, Berquin P, Mariani-Kurkdjian P, Bef D,Cuvelier B, Capek J, et al. Possible person-to-persontransmission of Escherichia coli O111-associatedhemolytic uremic syndrome. Pediatr Nephrol1997;11:36-9.

8. Kehl KS, Havens P, Behnke CE, Acheson DWK.Evaluation of the premier EHEC assay for detectionof Shiga-toxin-producing Escherichia coli. J ClinMicrobiol 1997;35:2051-4.


Letters

9. Basta M, Karmali M, Lingwood C. Sensitive receptor-specified enzyme-linked immunosorbent assay forEscherichia coli verocytotoxin. J Clin Microbiol1989;27:1617-22.

10. Gunzer F, Bohm H, Russmann H, Bitzan M, Aleksic S,Karch H. Molecular detection of sorbitol-fermentingEscherichia coli O157 in patients with hemolytic-uremic syndrome. J Clin Microbiol 1992;30:1807-10.

The Taxonomy of Cyclospora

To the Editor: In the article by N.J. Pieniazek andB.L. Herwaldt (1) on the rRNA gene of Cyclosporacayetanensis, the authors suggest that Cyclosporashould be placed in the genus Eimeria becausethe rRNA genes of the two genera have similarsequences. The article refers to Norman D.Levine’s chapter on the Apicomplexa in theIllustrated Guide to the Protozoa (2). Regretta-bly, the authors failed to read the whole chapterand to recognize that the initial characteristicsfor placing the oocyst of a coccidium in its propergenus are the number of sporocysts and then thenumber of sporozoites in each sporocyst. Thegenus Eimeria has four sporocysts and twosporozoites in each sporocyst. The genusCyclospora has two sporocysts, each of which hastwo sporozoites.

The original taxonomists (3) of C. cayetanensisrecognized that it should be placed in thetaxonomic family Eimeriidae, close to Eimeria,but they adhered to the traditional designationfor genera of coccidia. Pieniazek and Herwaldtshould be cognizant of the rules of zoologicnomenclature as well as the fact that certainmorphologic characteristics of protists haveserved us well for many decades and continue tobe useful. There are serious consequences tochanging the classification of an organism, and itshould not be thought that one can make such achange casually. I encourage the editors ofEmerging Infectious Diseases to seek the adviceof those who understand what should be donewith respect to the classification and nomencla-ture of organisms.

William C. MarquardtColorado State University, Fort Collins,

Colorado, USA

References 1. Pieniazek NJ, Herwaldt BL. Reevaluating the

molecular taxonomy: Is human-associated Cyclosporaa mammalian Eimeria species? Emerg Infect Dis1997;3:381-3.

2. Levine ND. Apicomplexa. In: Lee JJ, Hutner SH, BoyceEC, editors. An illustrated guide to the protozoa.Lawrence (KS): Society of Protozoologists; 1985. p.322-74.

3. Ortega YR, Gilman RH, Steling CR. A new coccidianparasite (Apicomplexa:Eimeriidae) from humans. JParasitol 1994;80:625-9.

Reply to W.C. Marquardt: Dr. Marquardt’sadvocacy for reliance on morphologic characteris-tics even if phylogenetic data become availablethat lead to a different conclusion runs counter tothat expressed in an article he coauthored, whichsupports the importance of molecular data (1).The introduction of that paper states thefollowing:

“Early systematists relied largely on lightmicroscopic structures and life cycle patterns toseparate protozoa taxonomically....Apicomplexans display enormous variations inlife cycle patterns, physiology, cytology, andbiochemistry. There is no consensus on whichcharacteristics should be relied upon to inferphylogenetic relationships. Developmental andultrastructural features have been used to inferevolutionary relationships among representativegenera in the class Sporozoea. However,comparisons of phenotypic characters arequalitative and lack objective quantitativeassessment to infer genetic relationships.Sequence similarities between proteins or geneswhich share a common evolutionary history canbe used to infer quantitative phylogeneticrelationships. The small subunit (16S-like)rRNAs and their coding regions are especiallyuseful for estimating the extent of geneticrelatedness over broad evolutionary ranges.”

That paper concludes with the statement that“ribosomal RNA sequence analyses of otherapicomplexans are required in order to test thevalidity of relationships inferred from structuresand life cycle patterns.” Similarly, we concluded ourpaper as follows: “Reports based on morphologicfeatures alone may suffer from poor resolution offeatures needed for classification of closely related


Letters

organisms. To improve our understanding of thetaxonomy of human-associated Cyclospora, mo-lecular evaluation of isolates of additionalCyclospora and Eimeria species, especially othermammalian species, is needed.”

Norman J. Pieniazek and Barbara L. HerwaldtCenters for Disease Control and Prevention,

Atlanta, Georgia, USAReferences 1. Gajadhar AA, Marquardt WC, Hall R, Gunderson J,

Ariztia-Carmona EV, Sogin ML. Ribosomal RNAsequences of Sarcocystis muris, Theileria annulataand Crypthecodinium cohnii reveal evolutionaryrelationships among apicomplexans, dinoflagellates,and ciliates. Mol Biochem Parasitol 1991;45:147-54.


News and Notes

Foodborne Diseases Active SurveillanceNetwork (FoodNet)

The Foodborne Diseases Active SurveillanceNetwork (FoodNet) is the foodborne diseasecomponent of the Emerging Infections Program(EIP) of the Centers for Disease Control andPrevention (CDC). A collaborative project ofCDC, the seven EIP sites, the U.S. Department ofAgriculture (USDA), and the U.S. Food and DrugAdministration (FDA), FoodNet consists of activesurveillance for foodborne diseases and relatedepidemiologic studies designed to help publichealth officials better understand the epidemiol-ogy of foodborne diseases in the United States.FoodNet was established in 1995 in five locations:Minnesota, Oregon, and selected counties inGeorgia, California, and Connecticut. The totalpopulation of these sites, or catchment areas, is14.7 million, or 6% of the population of the UnitedStates. FoodNet was expanded to selectedcounties in Maryland and New York in 1997. Thegoals of FoodNet are to describe the epidemiologyof new and reemerging bacterial, parasitic, andviral foodborne pathogens; estimate the fre-quency and severity of foodborne diseases thatoccur in the United States per year; anddetermine how much foodborne illness resultsfrom eating specific foods, such as meat, poultry,and eggs.

Foodborne diseases are common; an esti-mated 6 to 33 million cases occur each year in theUnited States. Although most of these infectionscause mild illness, severe infections and seriouscomplications do occur. The public healthchallenges of foodborne diseases are changingrapidly; in recent years, new and reemergingfoodborne pathogens have been described, andchanges in food production have led to new foodsafety concerns. Foodborne diseases have beenassociated with many different foods, includingsome previously thought to be safe, such as eggsand fruit juice, both of which have transmittedSalmonella during recent outbreaks. Publichealth officials in the seven EIP sites aremonitoring foodborne diseases, conducting epide-miologic and laboratory studies of these diseases,and responding to new challenges from thesediseases. Information gained through thisnetwork will lead to new interventions andprevention strategies for addressing the publichealth problem of foodborne diseases.

Current “passive” surveillance systems relyupon reporting of foodborne diseases by clinicalmicrobiology laboratories to state health depart-ments, which in turn report to CDC. Althoughfoodborne diseases are extremely common, only afraction of them are routinely reported to CDCthrough these surveillance systems. Inadequatereporting results from a complex chain of eventsthat must occur before a case is reported, and abreak at any linkage along the chain results in acase not being reported (Figure). FoodNet is an“active” surveillance system, meaning publichealth officials frequently contact microbiologylaboratory directors to find new cases offoodborne diseases and report these caseselectronically to CDC. In addition, FoodNet isdesigned to monitor each of the events that occursalong the foodborne diseases pyramid andthereby allow more accurate and preciseestimates and interpretation of the prevalence offoodborne diseases over time. Because mostfoodborne infections cause diarrheal illness,FoodNet focuses these efforts on persons whohave a diarrheal illness.

Figure. The prevalence of illness pyramid. Passivesurveillance data represent only the tip of the iceberg. For abacterial infection to be included in the passive surveillancesystem, it must pass through the following steps: a personbecomes ill with a diarrheal disease, the patient must go to adoctor, the doctor must order a bacterial stool culture, theassigned microbiology laboratory must culture for thisorganism and report the infection to the state healthdepartment, and the state health department in turn mustreport the infection to CDC. This passive surveillancesystem is the means by which the number of cases offoodborne illness is currently determined at CDC; if any stepdoes not occur, foodborne illness is not reported. FoodNet isdesigned to collect information along each step of thispyramid.


News and Notes

FoodNet Components

Active Laboratory-Based SurveillanceThe core of FoodNet is population-based

active surveillance at over 300 clinical microbiol-ogy laboratories that test stool samples in theseven participating sites. In active surveillance,the laboratories in the catchment areas arecontacted regularly by collaborative FoodNetinvestigators to collect information on alllaboratory-confirmed cases of diarrheal illness.Since January 1996, information has beencollected on every laboratory-diagnosed case ofSalmonella, Shigella, Campylobacter, Escheri-chia coli O157, Listeria, Yersinia, and Vibrioinfection among residents of the catchment areasof the five original sites; this information istransmitted electronically to CDC. The result is acomprehensive and timely database of foodborneillness in a well-defined population.

Survey of Clinical LaboratoriesIn October 1995, collaborative FoodNet

investigators conducted a baseline laboratorysurvey of all microbiology laboratories in the fiveoriginal catchment areas to determine whichpathogens are included in routine bacterial stoolcultures, which tests must be specificallyrequested by the physician, and what specifictechniques are used to isolate the pathogens. Abaseline survey will be conducted in the two newsites, and a follow-up survey to assess any recentchanges in laboratory practices was conducted inthe original sites in 1997. Practices in clinicallaboratories have been found to vary; somelaboratories look for a wider variety of bacteriathan others. The methods used to collect andexamine specimens are being investigatedbecause these can influence whether thelaboratory finds disease-causing bacteria.

Survey of PhysiciansTo obtain information on physician stool-

culturing practices, collaborative FoodNet inves-tigators mailed a survey questionnaire to 5,000physicians during 1996. Analysis of these data isongoing. Because laboratories test stool speci-mens from a patient only upon the request of aphysician or other health-care provider, it isimportant to measure how often and under whatcircumstances physicians order these tests. Aschanges occur in the way health care is providedin the United States, stool-culturing practices

may also change. The practices of physicians whosend stool samples to laboratories within thecatchment areas will be monitored by surveysand validation studies.

Survey of the PopulationCollaborative FoodNet investigators contact

randomly selected residents of a catchment areaand ask whether the person has had a recentdiarrheal illness, whether the person soughttreatment for the illness, and whether the personhad consumed certain foods known to havecaused outbreaks of foodborne illness. During1996, 750 residents of the catchment areas wereinterviewed by telephone each month (9,000/year). Because many who become ill withdiarrhea do not see a physician, little is knownabout the number of cases of diarrhea in thegeneral population and how often persons withdiarrhea seek medical care. The populationsurvey is an essential part of active surveillancefor foodborne illness because it allows for anestimate of the population who seeks medicalcare when affected by diarrheal illness.

Case-Control StudiesIn 1996, the FoodNet began case-control

studies of E. coli O157 and Salmonella serogroupB and D infections. More than 60% of Salmonellainfections in the United States are caused byserogroup B and D Salmonella. These large case-control studies will provide new and more preciseinformation about which food items or otherexposures may cause these diseases. To allow themost precise classification of the isolates from thepatients in these studies, the Salmonella and E.coli O157:H7 laboratory specimens from thesepatients are sent from FoodNet sites to CDC forfurther study, including antibiotic resistancetesting, phage typing, and molecular subtyping.

AccomplishmentsSince becoming operational on January 1,

1996, FoodNet has tracked the rates of foodbornediseases. Even in the first year of data collection,numerous interesting patterns and outbreakswere detected. Surprisingly high isolation ratesfor Y. enterocolitica in Georgia and Campy-lobacter in California were detected. An outbreakof Salmonella infections caused by contaminatedalfalfa sprouts was detected in Oregon. Twooutbreaks of E. coli O157:H7 infections weredetected in Connecticut, one due to lettuce and


News and Notes

one to apple cider. FoodNet has also provided theinfrastructure for conducting active surveillancefor new and reemerging diseases. When anassociation between bovine spongiform encepha-lopathy in cattle and variant–Creutzfeldt-Jakobdisease in humans was suspected in the UnitedKingdom, EIP personnel conducted surveillancefor this rare human disease. EIP personnel alsocollaborated in the investigation of a multistateoutbreak of Cyclospora infections associated withconsumption of raspberries from Guatemala.

Future FoodNet ProjectsSeveral projects in 1997 will focus on

Campylobacter, including a case-control study todetermine the risk factors for infection anddetermination of antibiotic resistance patternsamong Campylobacter strains.

Collaborative FoodNet investigators will estab-lish active surveillance for hemolytic uremicsyndrome, a serious complication of E. coli O157:H7infection.

FoodNet Working Group*

*Frederick Angulo, David Swerdlow, Robert Tauxe,Patricia Griffin, Drew Voetsch, Thomas Boyce, SudhaReddy, Mary Evans, Sam Yang (CDC); Duc Vugia, BenWerner, Kevin Reilly (California Department of HealthServices); Sue Shallow, Gretchen Rothrock, Pam Daily,Felicia Chi (California EIP); Paul Blake, Jane Koehler

(Georgia Department of Human Resources); MonicaFarley, Wendy Baughman, Molly Bardsley, Suzanne

Segler, Shama Desai (Georgia EIP); James Hadler, PatMshar (Connecticut Department of Public Health);

Ruthanne Marcus, Terry Fiorentino (Connecticut EIP);Michael Osterholm, Craig Hedberg, Jeff Bender, JulieHogan, Valerie Deneen, Heidi Kassenborg (MinnesotaDepartment of Health); Paul Cieslak, John Townes,Beletshachew Shiferaw, Maureen Cassidy, Theresa

McGivern, Regina Stanton (Oregon Health Division);Diane Dwyer, Peggy Pass (Maryland Department ofHealth and Mental Hygiene); and Dale Morse, Julia

Kiehlbauch, Hwa-Gan Chang, Cathy Stone (New YorkDepartment of Health); I.Kaye Wachsmuth, Jill

Hollingsworth, Peggy Nunnery, Art Baker, PhyllisSparling (USDA-FSIS); Ken Falci, Bing Garthright, Sean

Altekruse (FDA-CFSAN).

The Centers for Disease Control andPrevention, the Council of State and TerritorialEpidemiologists, the American Society forMicrobiology, and the National Foundation forCDC, along with more than 50 agencies andorganizations are cosponsoring the 1998 Interna-tional Conference on Emerging InfectiousDiseases. This conference will offer newopportunities for education, collaboration, andpartnership with colleagues worldwide to explorethe most current research, surveillance, andprevention and control programs addressing allaspects of emerging infectious diseases. Atten-dance is limited to 2,500 participants.

The meeting will consist of general andplenary sessions, symposia, and roundtableswith invited speakers, presentations on emerg-ing infections activities, oral and posterpresentations based on submission of an acceptedabstract, and exhibits. Conference topics willinclude work on surveillance, epidemiology,research, communication, training, and preven-tion and control of emerging infectious diseases,as well as emergency preparedness and response.

Abstracts should address new, reemerging,and drug-resistant infectious diseases that affecthuman health. Deadline for abstract submissionis October 31, 1997. Information on latersubmission of abstracts for consideration as latebreakers is included in the program materials.For additional information, access announce-ments at www.cdc.gov/ncidod/EID/eid.htm or theNational Center for Infectious Diseases(www.cdc.gov/ncidod/whatsnew.htm), send an e-mail to [email protected], or call 202-942-9248. Proceedings of the conference willbe published in the Emerging InfectiousDiseases journal.


News and Notes

Errata

Vol. 3, No. 1In “Fluoroquinolone Resistance in Neisseria

gonorrhoeae” by J.S. Knapp et al., in the Table, p. 35,under the column R of Equivalent MIC (mg/ml), and onthe row Ofloxacin, 400 mg subrow Ofx, 5, which citedreference 6, the number should be ≥2.0, not ≥1.0. Thereference for this criterion (reference 6 on p. 3) shouldread “Detection of quinolone-resistant Neisseriagonorrhoeae.”

Vol. 3, No. 3In “Host Genes and HIV: The Role of the Chemokine

Receptor Gene CCR5 and Its Allele (∆CCR5) by J.M.McNicholl et al., the following figure should have beenprinted as Figure 2, p. 264. The figure legend is correct.

In “Emerging Foodborne Diseases” by S.F.Altekruse et al., in Table 1, p. 285, the number ofEscherichia coli O157:H7 cases per year in the UnitedStates should be 25 x 103, not 725 x 103. The legend toFigure 2, p. 288 should read “Percentage of U.S.population over 65 years of age, 1900-2040 (projected).”

In “Molecular Epidemiologic Investigations ofMycoplasma gallisepticum Conjunctivitis in Songbirdsby Random Amplified Polymorphic DNA Analyses” byD.H. Ley et al., in Figure 1, p. 376, the photographs of thehouse finch and American goldfinch were inadvertentlyreversed.

We apologize to our readers for these errors.

Instructions to AuthorsManuscripts should be prepared according to the

“Uniform Requirements for Manuscripts Submitted toBiomedical Journals” [Ann Intern Med 1997:126[1]:36-47].

Begin each of the following sections on a new page andin this order: title page, abstract, text, acknowledgments,references, each table, figure legends, and figures. On the titlepage, give complete information about each author. Givecurrent mailing address for correspondence (include faxnumber and e-mail address). Follow Uniform Requirementsstyle for references. Consult List of Journals Indexed in IndexMedicus for accepted journal abbreviations. Tables andfigures should be numbered separately (each beginning with1) in the order of mention in the text. Double-space everything,including the title page, abstract, references, tables, and figurelegends. Italicize scientific names of organisms from speciesnames all the way up, except for vernacular names (virusesthat have not really been speciated, such as coxsackievirusand hepatitis B; bacterial organisms, such as pseudomonads,salmonellae, and brucellae).

All articles are reviewed by independent reviewers. TheEditor reserves the right to edit articles for clarity and tomodify the format to fit the publication style of EmergingInfectious Diseases.

Documents sent in hardcopy should also be sent ondiskette, or by e-mail. Acceptable electronic formats for textare ASCII, WordPerfect, AmiPro, DisplayWrite, MSWord,MultiMate, Office Writer, WordStar, or Xywrite. Sendgraphics documents in Corel Draw, Harvard Graphics,Freelance, or save as .TIF (TIFF), .GIF (CompuServe), .WMF(Windows Metafile), .EPS (Encapsulated Postscript), or.CGM (Computer Graphics Metafile). The preferred font forgraphics files is Helvetica. If possible, convert Macintoshfiles into one of the suggested formats. Submit photographsas glossy, camera-ready photographic prints.

Send all manuscripts and correspondence to theEditor, Emerging Infectious Diseases, National Center forInfectious Diseases, Centers for Disease Control andPrevention, 1600 Clifton Road, Mailstop C-12, Atlanta,GA 30333, USA, or by e-mail to [email protected].

Perspectives: Contributions to the Perspectives sectionshould provide insightful analysis and commentary aboutnew and reemerging infectious diseases or related issues.Perspectives may also address factors known to influencethe emergence of infectious diseases, including microbialadaptation and change; human demographics andbehavior; technology and industry; economic developmentand land use; international travel and commerce; and thebreakdown of public health measures. Articles should beapproximately 3,500 words and should include references,not to exceed 40. Use of additional subheadings in the mainbody of the text is recommended. If detailed methods areincluded, a separate section on experimental proceduresshould immediately follow the body of the text. Photographsand illustrations are encouraged. Provide a short abstract(150 words) and a brief biographical sketch.

Synopses: Submit concise reviews of infectious diseases orclosely related topics. Preference will be given to reviews ofemerging and reemerging infectious diseases; however,timely updates of other diseases or topics are also welcome.Synopses should be approximately 3,500 words and shouldinclude references, not to exceed 40. Use of subheadings inthe main body of the text is recommended. If detailedmethods are included, a separate section on experimentalprocedures should immediately follow the body of the text.Photographs and illustrations are encouraged. Provide ashort abstract (150 words) and a brief biographical sketch.

Dispatches: Provide brief updates on trends in infectiousdiseases or infectious disease research. Include descrip-tions of new methods for detecting, characterizing, orsubtyping emerging or reemerging pathogens. Develop-ments in antimicrobial drugs, vaccines, or infectiousdisease prevention or elimination programs are appropri-ate. Case reports are also welcome. Dispatches (1,000 to1,500 words of text) should not be divided into sections.Provide a short abstract (50 words); references, not toexceed 10; and figures or illustrations, not to exceed two.

Editorial Policy and Call for ArticlesEmerging Infectious Diseases is a peer-reviewed journal established expressly to promote the recognition of new and

reemerging infectious diseases around the world and improve the understanding of factors involved in disease emergence,prevention, and elimination.

The journal has an international scope and is intended for professionals in infectious diseases and related sciences. Wewelcome contributions from infectious disease specialists in academia, industry, clinical practice, and public health, as wellas from specialists in economics, demography, sociology, and other disciplines. Inquiries about the suitability of proposedarticles may be directed to the Editor at 404-639-3967 (tel), 404-639-3075 (fax), or [email protected] (e-mail).

Emerging Infectious Diseases is published in English and features three types of articles: Perspectives, Synopses, andDispatches. The purpose and requirements of each type of article are described in detail below. To expedite publication ofinformation, we post journal articles on the Internet as soon as they are cleared and edited.

Spanish translations of some articles can be accessed at ftp://fcv.medvet.unlp.edu.ar/pub/EID. Articles by authorsfrom non-English-speaking countries can be made simultaneously available in English and in the author’s native language(electronic version of the journal only). Articles published in this way are translated from English into the author’s nativelanguage and appear in the same issue of the journal.