Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 4 APPLICATION OF THE MODEL TO MICROBIOLOGICAL HAZARDS In the report Meat and Poultry Inspection' published in 1985 9 a National Research Council committee stated that Salmonella spp. and Cam~vlobacter ieiuni were the currently recognized major poultry-borne causes ot illness in consumers, but that antemortem and postmortem ~ ~ ~ ~ inadequate to detect these organisms (NRC, 1985b) . That committee recommended that FSIS intensify its efforts to control and eliminate contamination with microorganisms that cause disease in consumers. _ , Inspection methods in effect at that time were _ Specific recommenda- tions to achieve this goal included education of producers, processors, food handlers, and consumers; determination of etiologic agents responsible for gross lesions resulting in condemnation; development of a trace-back capability to determine where pathogenic microorganisms contaminate the poultry between farm and table; and emphasis on the Hazard Analysis Critical Control Point (HACCP) concept in the inspection process (NRC, 1985b). This chapter summarizes a method for identifying and describing risks associated with microbiological contamination of chickens and applies these methods to an evaluation of specific pathogens. The committee made no effort to consider the occupational hazards presented by microorganisms in the workplace (i.e., in chicken slaughtering and processing plants), nor did it evaluate all Possible pathogens associated with chickens. Rather, a model for subsequent refinement, . . ~ ~ ~ · ~ . ~ this chapter is designed to serve as including collection of additional data neecea For risk assessment or microorganisms, and to provide a basis for the development of initial risk-management strategies. METHODOLOGY USED TO DESCRIBE AND IDENTIFY RISKS Hazard Identification and Evaluation General Ancroach. Several basic questions arise in the identification and evaluation of the public health hazard associated with a specific microbial agent. For example, is the microorganism potentially hazardous to health, i.e., is it a pathogen in humans? How does susceptibility vary from person to person or group to group? Are 56
57 all microorganisms in this species, class, or other classification equally hazardous? If the microorganism is a hazard, what disease can it cause? With what frequency does this microorganism cause disease? The question of whether a microorganism is harmful to health has traditionally been approached by linking known disease syndromes with specific microorganisms. The criteria necessary to identify a microorganism as the etiologic agent for a disease were codif fed in the late 19th century as Koch's postulates: the microorganism must always be found in the diseased animal, but not in healthy ones; the organism must be isolated from diseased animals and grown in pure culture away from the animal; the organism isolated in pure culture must initiate and reproduce the disease when reinoculated into susceptible animals; and the organism should be reisolated from the experimentally infected animals. Although documentation of all four criteria is not always possible for human pathogens because of ethical considerations, Koch's postulates provide a framework for determining whether a microorganism is hazardous to human health. In considering the potential health hazard of a microorganism, it is necessary to define the population under consideration. A large number of microorganisms are not pathogenic for normal people but can cause disease and death in susceptible hosts, e.g., persons with underlying diseases that compromise their immunocompetence or who have received chemotherapeutic agents that affect their immune response. For patients whose immune response is totally suppressed (such as patients receiving bone marrow transplants), virtually every microorganism is a potential pathogen, making it difficult to distinguish precisely between pathogens and nonpathogens without some notion of the immune status of the target population. In defining the health hazard caused by a particular microorganism, it is also necessary to understand the organism's virulence factors, or factors that serve as markers for virulence. For many microorganisms, only a small percentage of strains within a species possess a specific plasmid or chromosomally defined factors that permit the organism to cause disease. For example, Escherichia cold is the dominant aerobic bacterium in all normal human stool samples, and most strains of E. cold are harmless to a healthy adult; however, some carry specific virulence plasmids that have been associated with at least four distinct disease syndromes, including toxin-mediated secretory diarrhea and an invasive dysentery syndrome (Levine, 1985~. Similarly, only 1% of environmental Vibr' o parahaemolyticus strains produce a hemolysin, which is correlated with their ability to cause disease (Morris and Black, 19851. For many microorganisms, we still have only a rudimentary understanding of virulence factors. Further research may help to explain apparently contradictory data regarding the pathogenicity of certain species of microorganisms.
58 It is also necessary to understand the disease syndrome and related manifestations . Some microorganisms may only cause mild disease, such as the relatively mild diarrhea seen with many nontoxigenic species of Vibrio; other organisms, such as Salmonella, may also be associated with bacteremia, meningitis, other serious morbidity, and death. Host susceptibility may influence the extent of disease caused by a microorganism. For example, Salmonella bacteremia is more likely to be seen in a patient with sickle-cell disease or an underlying malignancy, whereas meningitis occurs most frequently in neonates. Virulence factors may also influence the severity of disease. For example, strains of Vibrio cholerae that lack the gene for cholera toxin production generally cause only mild disease, whereas fully toxigenic strains can cause rapidly progressive dehydration and circulatory collapse in a previously healthy adult. Finally, one needs to have data on the extent of disease caused by the microorganism (are there 10 cases in the United States per year, or 10,000?) and some index of disease severity, such as the number of hospitalizations or deaths attributable to the microorganism. The overall public health impact of a microorganism that causes a large number of mild cases resulting in few hospitalizations and no deaths may be significantly less than one responsible for only a few cases but with high rates of hospitalization and mortality. Sources of Data. The extent of the human health hazard presented by certain microorganisms must usually be determined by synthesizing data from a variety of sources, including studies in volunteers, animal models, and epidemiological studies. Approximations of disease incidence can be based on epidemiological data and on numbers of cases and deaths reported through state and national reporting systems. For ethical reasons, studies in volunteers are generally limited to healthy adults and to agents that cause only mild disease or for which there is good, effective therapy, but they can provide the best data on pathogenicity, permitting direct testing of Koch's postulates. These studies may also be useful in determining the relationship between dose and infection rates, in characterizing the disease syndrome, and, with an appropriate experimental design, in identifying specific virulence factors. However, they generally cannot be used to develop data on the disease in immunocompromised persons or other high-risk hosts. Many microorganisms are adapted to specific species, so that they cause disease in only one or a few host species. Therefore9 animal models cannot always be used to determine whether a specific microorganism will be pathogenic in humans. Once it has been shown that specific disease manifestations in an animal correlate with disease manifestations in humans, e.g., the correlation between the ability of an E. cold strain to cause keratoconjunctivitis in the guinea pig and its ability to invade enterocytes and cause dysentery in humans (Sereny, 1955), animal models can be useful in characterizing virulence factors or in identifying markers for virulence.
59 Epidemiological studies can provide evidence that an organism is pathogenic by fulfilling one or more of Koch's postulates. For example, studies may demonstrate that a certain microorganism is often found among patients with diarrhea but not among controls, suggesting that the microorganism is respons ible for the diarrhea. Similarly, a study of disease outbreaks may demonstrate that one microorganism is significantly associated with illness among patients and is present in the food or other product incriminated as the cause of the outbreak. Epidemiological studies also permit characterization of microbial disease syndromes and can provide information on susceptibility and disease manifestations in persons other than normal healthy adults. Long-term follow-up may also be possible in such studies, permitting identification of chronic syndromes that may not be apparent in short-term experiments in volunteers. Approximations of incidence can be based on information From epidemiological studies and data reported through national reporting systems. For example, the Centers for Disease Control (CDC) collects national surveillance data on diseases caused by many specific microorganisms, including some of the food-borne bacterial pathogens. However, the number of cases reported through these systems is only a fraction of the actual number. Cases are not reported if a patient does not seek medical attention (e.g., when microorganisms cause relatively mild disease), if the doctor does not order appropriate diagnostic tests (such as a stool culture) for patients that are seen, or if the positive test result is not reported to the appropriate health authorities. By using information from outbreak investigations and other epidemiological studies, it may be possible to obtain a very rough estimate of the actual number of cases that occur for each reported case. For example, it has been estimated that only 1 out of every 20 Shigella infections in the United States is reported (Rosenberg et al., 1977), and possibly as few as 1 out of every 100 Salmonella cases (Aserkoff et al., 1970~. Rates of hospitalization and mortality (which may be more complete) derived from epidemiological studies can also be used to estimate incidence rates in the community. Dose-Resnonse Studies General Approach. For any one person, a certain number of microorganisms is usually necessary to establish an infection; no infection results when fewer microorganisms are present. This threshold value may differ widely from person to person or time to time, depending on such factors as host susceptibility, the method and vehicle with which the inoculum is presented, and the virulence of the microorganism. Nonetheless, it is sometimes possible to estimate the percentage of normal, healthy adults that will become infected after exposure to a viral or bacterial inoculum of known size.
60 When determining the dose response of i nfectious agents, one must also consider the rate of asymptomatic or subclinical infection. Not all persons who become infected with a microorganism develop signs and symptoms of disease . Asymptomatic infections may be quite common, and may be dependent on many factors. For example, only one out of every two or three young adults infected with Epstein-Barr virus will develop s igns and symptoms of infectious mononuclear is; few children infected with the virus will have any symptoms at all. Some infectious agents cause disease by producing toxins that may cause an incremental increase in the severity of illness with increasing toxin levels, rather than a threshold response. The botulinal toxin is one example: food-borne botulism is caused by a preformed neurotoxin present in contaminated food, and disease manifestations and severity correlate directly with the amount of toxin consumed. Sources of Data. Dose-response data on humans are derived primarily from experiments in volunteers. In these studies, the bacterial or viral inoculum can be carefully controlled; however, they are restricted to healthy adults and usually use a defined laboratory strain that may have atypical virulence. Data on the dose response of food-borne pathogens have also been collected in investigations of outbreaks. These are based on estimates of the amount of food consumed and measurements of the concentration of the pathogen present in the implicated food, usually hours or days after the disease outbreak, during which time there may be significant changes in the size of the inoculum. Potential for Human Exposure General Approach. Microorganisms may be transmitted to humans in a variety of ways, including person-to-person transmission, (e.g., through sexual and fecal-oral contact), exposure to air-borne, food-borne, water-borne, and vector-borne pathogens, and contact with contaminated objects. Spread of disease by broiler chickens almost exclusively involves food-borne transmission, although direct contact or air-borne routes may be important in occupational exposures in poultry slaughter and processing plants. Microorganisms may be introduced into food at a number of points. Many of them, including CampYlobacter and Salmonella, are present in the gut flora of animals and may contaminate the surface of the carcass during slaughter or subsequent processing. The actual number of microorganisms present at any one time (and the percentage of carcasses contaminated) may vary widely, depending on temperature, product conditions, and how the product is handled. Most microorganisms are killed during cooking. Infection can occur if the poultry is eaten raw or if parts of it are undercooked. In the kitchen, cooked foods can be contaminated by raw poultry and other foods, by soil, or by food handlers. Once present in a kitchen, these microorganisms may contaminate a variety of foods, including chicken.
61 Epidemiological studies can often link or attribute outbreaks of food-borne disease to specific products, such as chicken. However 9 the key point in the chain of transmission generally is the kitchen itself: for most food-borne pathogens, highly contaminated meat will not cause disease if properly cooked and handled. Conversely, minimal contamination can lead to maj or disease outbreaks if pathogenic microorganisms are allowed to contaminate an entire kitchen and if food prepared in the kitchen is subsequently mishandled. For many microorganisms, it is possible to measure the level of contamination of a product at each stage of processing. Unfortunately, it is unclear how the level of contamination observed during the processing of such products as raw chicken correlates with the potential for disease transmission. In formal attempts to assess risk, one must recognize the difference between measurable levels of product contamination and disease occurrence. Although the correlation can never be fully resolved, additional data from carefully focused and coordinated epidemiological-microbiological studies could greatly reduce the level of uncertainty. Sources of Data. The U.S. Department of Agriculture and Food and Drug Administration collect some data on levels of product contamination with certain microorganisms at various stages of processing and distribution. More detailed data, and data on additional microorganisms, have been reported in the literature. Data showing a direct link between product contamination and disease in the community (and subsequent disappearance of disease in association with elimination of contamination) have been published for some microorganisms and some products, but there are few data available on raw chicken or its major pathogens (Bryan, 1980a; CDC, 1983a,b; Horwitz and Gangarosa, 1976~. The CDC collects data on food-borne disease outbreaks through its national food-borne disease surveillance system. As in other national reporting systems and for individual cases of disease, however, there is undoubtedly significant underreporting and an emphasis on more serious illnesses. For example, almost all outbreaks of botulism are reported, whereas outbreaks of other illnesses involving only a few persons with relatively mild symptoms frequently are not investigated or reported. Data on routes of transmission can be obtained from outbreak reports and from case-control studies of specific pathogens. In properly designed case-control studies, it may be possible to determine the percentage of reported cases due to one microorganism that can be attributed to broiler chickens, i.e., attributable risk. To completely characterize the risk associated with microbial contamination of chicken, there must be sufficient data from which to predict changes in disease occurrence that would result from changes in levels of product contamination at each of several steps from production to the consumer's table. Such data are likely to show an
62 increasing correlation between the potential for diseases to occur and microbial contamination as one moves from production to the ingestion of poultry products. The microbial load on the cooked, ready-to-eat product is the only reliable indicator of disease potential, and the further removed from this point data are obtained, the less reliable they are. Characterization Of Risk There are sufficient data to clearly establish many microorganisms as a hazard to human health and to estimate (very roughly) the risk of illness associated with eating contaminated chicken. It is also possible to identify where microorganisms may be introduced during processing (see Figures 3-1 and 3-2 in Chapter 3) and to quantitate microbial contamination at those points. These control points, and their identification as part of an HACCP system, have been discussed in detail in a previous report of the National Research Council (NRC, 1985a) . Microorganisms contaminating chicken during processing may die off or reproduce rapidly before reaching the consumer, depending on environmental and product conditions Thus, levels of contamination on the dinner plate may have little relation to data obtained at the time of slaughter or in the kitchen prior to cooking. Therefore, measurement of the risk posed by microbial contamination of a specific food item e, such as chicken, must involve identification and enumeration of cases of illness within an exposure group in addition to sampling during production. APPLICATION OF METHODS TO INDIVIDUAL AND GROUPS OF POULTRY-BORNE PATHOGENS Many pathogenic microorganisms have been associated with the production of chickens or with eating chicken. Following are some examples, which are broken down into three general categories: 1. Bacteria known to be pathogenic in humans that are carried on or transmitted by broiler chickens at retail: CampYlobacter jejuni, Salmonella spp. (excluding S. typhi) Yersinia enterocolitica 2. Bacteria known to be pathogenic in humans that have been associated with eating chicken: Bacillus cereus Clostridium botulinum Clostridium perfringens Shigella sppe Staphylococcus aureus
63 Contamination by these human pathogens probably occurs during processing and preparation. 3. Microorganisms known to be pathogenic in chickens that are of questionable pathogenic significance for healthy humans when carried on or transmitted by broiler chickens at retail: Bacteria: Haemophilus gallinarum Pasteurella multocida Escherichia cold MYcobacterium avium Mycop ~ asma spp. Clostridium colinum ChlamYd~a psittaci Viruses and associated diseases: Picornavirus (encephalomyelitis) Alphavirus (equine encephalitis) Herpesvirus (Marek's disease, infective laryngotracheitis) Par~yxovirus (Newcastle disease) Adenovirus (adenoviral tracheitis) Poxvirus (fowl pox) Coronavirus (infectious bronchitis) Orthomyxovirus (avian influenza) Reovirus (viral arthritis Fungi: Aspergillus spp. DactYlar~a spp. Parasites: CrYptoscoridium _ For some of these microorganisms, their significance as human pathogens and information on their transmission by broiler chickens are discussed in the following paragraphs. Known Human Pathogens Carried on or Transmitted by Broiler Chickens at Retail CampYlobacter jejuni Hazard Identification and Evaluation. The gram-negative bacterium CampYlobacter ieiuni has recently been implicated as one of the most common bacterial causes of- gastroenter~t~s among both children and adults (Pai et al., 1979; Skirrow, 1977) . Studies ~n the United States, Canada, and Europe have demonstrated that C. jejuni is significantly more common in stool samples from patients with diarrhea than in samcles from healthY controls (Blaser et al., 1983a). in a clinical al., 1983b). _ , , _ _ ~ _ _ _ Ingestion of C. iejuni by volunteers has resulted diarrhea! syndrome (Black et al., 1983; Blaser et
64 C. jejuni is pathogenic for normal, healthy adults (Black et al., - 1983; Blaser and Reller, 1981~. Illness may be more severe in elderly or debilitated patients, and the rare deaths from C. jejuni infection have occurred mostly in that group (Blaser and Reller, 1981~. Virulence factors and pathogenic properties of C. jeJuni include invasiveness, enterotoxin production, and cytotoxin production (Kl~pstein et al., 1985~. Some investigators have suggested that diarrhea-assoc~ated strains are more likely to have one or more of these properties than strains isolated from asymptomatic individuals (Klipstein et al., 1985; Ruiz-Palacios et al., 1983), but these results have not been confirmed by some other investigators (Mathan et al., 1984). Predominant symptoms among patients identified by positive stool culture include diarrhea, abdominal pain, malaise, fever, nausea, and vomiting (Blaser and Reller, 1981; SKCDPH, 1984) . Up to 80% of patients have fever, and approximately 50t give a history of bloody diarrhea (Blaser et al ., 1983a; SKCDPH, 1984) . Abdominal pain is severe enough to mimic acute appendicitis in about 19 of cases (Skirrow, 1977~. Complications rarely include toxic megacolon, gastrointestinal hemorrhage, convulsions (in children with high fevers), a typhoid-like syndrome, meningitis, reactive arthritis (in men with the HLA-B27 histocompatibility antigen), urinary tract infect) ons , and cholecystitis (Blaser and Reller, 1981~. Illness is reported most frequently among children less than a year old, and among young adults 20 to 29 years of age (Blaser et al., 1983b ; CDSC, 1981; Finch and Riley' 1984; Riley and Finch, 1985)0 Illnesses vary widely in severity, ranging from mild diarrhea lasting less than 24 hours to severe bloody diarrhea with abdominal pain and fever lasting several weeks. Some investigators have reported that most patients recover in less than a week (Blaser and Reller, 1981; Butzler and Skirrow, 1979), but a recent study of 225 culture-confirmed cases indicates that the average duration of illness is 13.5 days (SKCDPH, 1984~. In that study, 15 (6.7~) of the 225 patients were hospitalized up to 11 days as a result of their infection, most commonly (13 of the 15) for 5 days or less (SKCDPH, 1984). Since the late 1970s 9 most clinical microbiology laboratories in the United Kingdom have included cultures for Campylobacter as part of their routine evaluation of stool specimens (CDSC, 1981~. In 1981 there were about 12,500 reported isolations of Campylobacter, giving an annual isolation rate of approximately 28 per 100,000 population (Blaser et al., 1983b; CDSC, 1981~. Some clinical laboratories in the United States still do not routinely culture stool samples for Campylobacter, and nationwide surveillance for Campylobacter, which began In January 1983, is still not uniformly applied. During the first year , 42 states reported 8, 282 isolates of C . j e juni (Riley and Finch, 1985 ~ O In all 33 states that participated for at least 1l months in 1983, the number of Campylobacter isolates exceeded the
65 number of Shigella isolates reported. In four states, there were more Campylobacter isolates than Salmonella isolates. Until national surveillance data are more complete, the best estimates of CampYlobacter incidence in the United States will continue to come from laboratories serving defined population groups. In the metropolitan area of Denver, Colorado, investigators found that the annual incidence of reported Campylobacter infections was 17.2 per 100,000 people (Blaser et al. , 1983b). In the Seattle-King County Department of Public Health (SKCDPH) study (which did not include active recruitment of patients with diarrhea), the annual incidence was estimated to be 100 per 100,000 (SKCDPH, 1984~. For children less than 1 year or age, rates were 332.9 per 100,000 for males and 194.9 per 100,000 for females. The figures in both of the above studies are based on reported isolation of Campylobacter from stool samples. One study in England projected a 1.1% (1,100 per 100,000) mean annual rate of Campylobacter infection in a defined population (Kendall and Tanner, 1982) or approximately 40 times the reported annual rate of isolations ~ 28 per 100, 000) . Blaser et al. (1983b) have estimated that in developed countries such as the United States, the incidence of Camped obacter infections, whether symptomatic or not, is about 1% to 2% per year (1,000-2,000 per 100,0009. The SKCDPH data (with a relatively high incidence rate) suggest that the annual incidence of hospitalized cases would be approximately 7 per 100,000 if all hospitalized cases in the study population were correctly identified. Data are inadequate to calculate mortality rates (Blaser et al., 1983b). Dose-Response Studies. For campylobanteriosis, information on dose response is based on a few studies involving small numbers of volunteers and is difficult to interpret. In volunteer studies with C. jejuni strain A3004, for example, 5 of 10 healthy adult volunteers ingesting 8 x 102 colony-formin~ units (cfu) became infected, and one of the 5 became ill. At 8 x 10 cfu, 6 of 10 were infected, but none were ill. The rate of infection increased with increase" in the size of the inoculum until 100% were infected at a dose of 10 cfu. At 9 x 104 cfu, the rate of illness was 46%; however, as the size of inoculum was increased, the rate of illness at each dose level ranged from only 9% to 22% (Black et al., 1983~. These data, in which the dose varied 100,000-fold, show that risk is far from linear with dose, perhaps because of marked variability in susceptibility. Epidemiological studies of both person-to-person and water-borne transmission of C. jejuni suggest that relatively small inocula are capable of causing disease (Blaser et al., 1983b). No outbreaks have been reported where it has been possible to clearly associate infection rate with a known inoculum.
66 Potential for Human Exposure. The prevalence of C . j ejuni infection in live broiler chickens is highly variable. Microbiological surveys of live broilers have demonstrated that grow- out premises are inconsistently infected and that infection in one poultry housing unit does not mean that other housing units on the scone premises contain infected birds (Smitherman et al., 1984~0 Within infected flocks, however, infection rates tend to be high, and fecal samples6from infected chickens tend to be heavily contaminated (e.g., 10 cfu/g of feces) (Grant et al., 1980; Smitherman et al., 1984)e Ce iejuni has been found at many points during slaughter and processing, and a significant proportion of the broiler chicken carcasses available for retail sale carry the microorganism (Table 4-1~. Persons eating raw or rare chickens contaminated with C. iejuni may ingest sufficient numbers of microorganisms to become infected. CampYlobacter can also enter the kitchen via contaminated broiler carcasses and subsequently cross-contaminate other foods. In one study of commercial and institutional kitchens, Hutchinson et al. (1983) found that the hands of workers and kitchen work surfaces often became contaminated with CampYlobacter during the preparation of raw chicken. This is similar to the environmental contamination that occurs during the preparation of chicken contaminated by Escherichia cold (De Wit et al., 1979~. Several studies have shown an epidemiological link between chicken and gastroenteritis caused by C. Unit. Raw chicken eaten by military recruits during a training exercise was implicated in an outbreak in the Netherlands (Brouwer et al., 1979), and ingestion of poorly cooked chicken was associated with illness at a barbecue in Colorado (Blaser et al., 1983b) . In another study in Colorado, sporadic it lness was associated with the handling of raw chickens in home kitchens (Hopkins and Scott, 1983~. Consumption of chicken was associated with sporadic illness in Sweden (Norkrans and Svedham, 1982) and in urban areas of the Federal Republic of Germany (Kiss, 1983~. In the SKCDPH (1984) study, significantly more (p = 0.00003) patients with C. iejuni-caused enteritis had eaten chicken than had controls. The relative risk for chicken consumption was 2.4 (95% confidence interval [CI], 1.6-3.6~ ; for rare or raw chicken, it was 7~6 (95% CI, 2.1-27.6~; and for cooked chicken, it was 2.3 (95% CI, 1.5-3.51. The method of Walter (1975) was used to estimate that 48.2% of C. jejuni cases were attributable to eating chicken. The link - between chicken and human illness was reinforced by finding similarities in antibiograms, plasmid content, and serotypes of C. jejuni isolates taken from poultry at retail and from human cases (SKCDPH, 1984). Characterization of Risk. The studies cited above indicate that C. iejuni is clearly pathogenic for humans, resulting in as many as 1, 000 to 2, 000 cases per 100, 000 people annually in the United States (Blaser et al., 1983b). Epidemiological studies suggest that at least 48% of Camp~lobacter cases are attributable to chicken (SKCDPH, 1984~.
67 TABLE 4-1. Rates of Salmonella, Campylobacter, and Yersinia Contamination of Chicken Carcasses Part of Chicken Sampled Microorganism and Stage of Procedure CampYlobacter Chicken livers Mechanical deboning of chicken Chicken livers from giblet chiller Wings ready for packaging dings on arrival at supermarket Chicken; supermarket shelf Salmonella Yersinia No . Sampled Percent Reference 40 30 Stern et al., 40 12.5 1984 36 69.4 Wempe et al ., 1983 36 66.7 94 82.9 Kinde et al., 1983 862 Chicken; supermarket shelf Whole, eviscerated chickens; chill tank exit in 1967 in 1979 Whole, eviscerated chickens; chill tank exit 862 597 28.6 601 36.9 Whole, eviscerated chickens; 171 20.5 chill tank exit Chicken; supermarket shelf 862 1.3 23.1 Harris et al., 1986b 4.3 Harris et al., 1986b Green et al., 1982 11.6 Campbell et al., 1983 Surkiewicz et al., 1969 Harris et al., 1986b Species of Campylobacter are present on a high percentage of chickens at the time of retail sale. Any reduction in the frequency or intensity of contamination may result in a corresponding reduction in the frequency of illness, but the commmittee found no studies that clearly support or refute this hypothesis. Plasmid-mediated resistance to antimicrobial agents has been reported for isolates of C. j ej uni
68 isolated from poultry (Bradbury and Munroe, 1985; Harris et al., 1986a,b), and could compromise therapy in infected humans if resistance influences response to a clinically relevant drug. Nontyphoidal Salmonella Species Hazard Identification and Evaluation. Salmonella species are among the most common bacterial causes of gastroenteritis in humans. They are found significantly more often in stools of persons with diarrhea than in samples from asymptomatic persons. In studies in volunteers, isolates from the stools of diarrhea patients cause diarrhea! disease when fed to humans (McCullough and Eisele, 1951~. Normal, healthy adults are susceptible to infection, although illness tends to be more severe among the very young, the very old, or patients with some type of underlying immunosuppression. Certain serotypes or serogroups are characteristically more virulent than othersO Putative virulence plasmids have been identified for certain serotypes (Helmuth et al., 1985), but there does not appear to be a single virulence plasmid common to multiple serotypes. Symptoms usually appear within 6 to 48 hours after ingestion of contaminated food or water. They often begin with nausea and vomiting, followed by diarrhea, abdominal pain, and fever (Blaser et ale, 1983a; Hook, 1985; Hornick, 1983~. In some cases, the stools may be bloody. Temperature elevations as high as 38° to 39°C are common, as are chills. Abdominal cramps occur in approximately two-thirds of the patients. In severe cases, fluid and electrolyte losses may be profound enough to cause hypovolemic shock. Toxic megacolon has been described as a rare complication (Mandal and Mani, 1976~. Transient bacteremia occurs in less than 5% of adults with gastroenteritis, but at a somewhat higher rate in children and persons with major underlying diseases (Hook, 1985; Hyams et al., 1980~. Localized suppurative infections (bronchopneumonia, emphysema, endocarditis, pericarditis, arteritis9 pyelonephritis, osteomyelitis, arthritis) develop in about 10% of patients with bacteremia (Hook, 1985; Hornick, 1983~. Nontyphoidal Salmonella species, particularly S. cholerae-suis g can cause fever and sustained bacteremia without manifestations of enterocolitis (Saphra and Wassermann, 1954~. In the majority of uncomplicated cases, fever lasts for less than 2 days. Diarrhea usually lasts less than 7 days, but occasionally persists for as long as several weeks (Hook, 1985~. In the SKCDPH study , mean length of illness was 10 . 25 days (SKCDPH, 1984~. Illness is more severe in children, in the elderly, and in patients who have had a gastrectomy or gastroenterostomy, or who have achlorhydria, sickle-cell anemias or other conditions that impair resistance to infection. In food-borne outbreaks, hospitalization rates have ranged from O to more than 50% . The CDC (1981) reported that 8% of 2,356 cases in 51 outbreaks reported in 1978 through the CDC Foodborne Disease
69 Surveillance system were hospitalized. In the SKCI)PH (1984) study, 3 (4.1~) of 72 outpatients with positive cultures were later hospitalized. In Massachusetts, a case-fatality rate of 1.4% was found in a study of 2,600 confirmed cases occurring between 1940 and 1955. A higher mortality rate was seen in older age groups, debilitated patients, and patients with bacteremia (MacCready et al., 1957~. A much lower case-fatality rate of 0.1% was observed among 16,172 patients in 261 food-borne outbreaks reported to CDC between 1972 and 1978 (Sours and Smith, 1980). Salmonellosis has been a reportable disease in the United States since 1942, and data have been collected by CDC through its Salmonella surveillance system since 1963. The incidence of reported cases was approximately 8 per 100,000 in 1963; in 1983, there were approximately 19 cases per 100,000 (CDC, 1984a) . Although there are differences by serotype, age-specific rates tend to peak between 2 and 4 months of age at approximately 300 cases per 100,000 and decline rapidly thereafter. After age 5, rates for all serotypes combined remain relatively constant at approximately 10 cases per 100,000 per year (CDC, 1981~. As described for CampYlobacter, it is likely that reported cases represent only a small fraction of the total number of infections that occur. From data on outbreaks, it has been estimated that only one in 75 to 100 cases of salmonellosis in the United States is reported (Aserkoff et al., 1970; Hauschild and Bryan, 1980~. These types of data indicate that the incidence of actual infections ranges from 1,SOO to 2,000 cases per 100,000 population, roughly the same as rates estimated for Campvlobacter infections (Blaser et al., 1983b). In the SKCDPH (1984) study, there were 1.4 hospitalized cases of salmonellosis per 100,000 members of the population annually. Similar figures are obtained from reported cases and the percentages of hospitalization reported in studies of outbreaks. The reported number of cases and estimated mortality rates indicate that mortality would range from 0.02 to 0.2 deaths per 100,000 population per year. Dose-Response Studies. Studies in volunteers have suggested that an inoculum of 10 cfu or more is necessary to produce disease (McCullough and Eisele, 1951~. However, data from outbreak investigations suggest that the infective dose in nature may be as low as 10 to 10 cfu (Craven et al., 1975; Silverstolpe et al., 1961~. In outbreaks associated with low infective doses, the vehicles of transmission were foods with a high buffering capacity or a high fat content, which may have protected the microorganism from the lethal action of gastric acid (Gangarosa, 1978b). Potential for Human Exposure. Two relatively host-specific Salmonella serotypes, S. pullorum and S. gallinarum, cause acute or chronic diseases in chickens. S. pullorum, the etiologic agent of pullorum disease in poultry, causes an acute systemic disease that
70 produces a high mortality rate in chicks less than 3 weeks of age, but most often causes a chronic localized disease or may be asymptomatic in older chickens. These two serotypes rarely cause disease in humans. Infection of chickens by other serotypes of Salmonella most often results in chronic intestinal carriage. These infections occur in poultry worldwide, but the flock incidence is extremely variable (Lee, 1977~. Nationwide surveys of market-ready broilers conducted in 1967 and 1979 both demonstrated extreme variability in the contamination rates among various lots of birds, with an incidence range of 7 . 5% to 73. 7% in 1967 and 2. 5% to 87. 5% in 1979 (Green et al., 1982) . Persons eating inadequately cooked chicken infected with Salmonella may ingest sufficient numbers of microorganisms to become infected. The presence of Salmonella on chi cken carcasses may also lead to contamination of the kitchen environment and other foods. Characterization of Risk. Salmonella clearly causes disease in humans . Proj ections from available data suggest that there are between 1,500 to 2,000 cases per 100,000 members of the population per year. Data on attributable risk are not as complete as those for C O j ej uni, but numerous epidemiological studies document a major role for broiler chickens as a vehicle for salmonellosis in humans. Salmonella is present on chickens, and many studies have documented the points at which contamination occurs during production and processing. At the time of retail sale, a high percentage of chickens bear Salmonella (CDC, 1977, 1983b, 1984b, 1985~. A reduction in the frequency or degree of contamination of chicken with Salmonella would logically result in a reduction in human disease; however, as for C. jejuni, there are no quantitative data linking contamination levels with disease occurrence. Thus, it is not known with certainty whether a reduction in Salmonella on broiler chickens would have any appreciable influence on the incidence of salmonellosis. In the absence of such data, it is not possible to do a quantitative assessment of the public health risk posed by specific levels of Salmonella on chicken carcasses . Yers inia enterocolitica :> Hazard Identification and Evaluation. Yersinia enterocolitica is . recognized as a significant cause of acute enteritis and enterocolitis (Bottone, 1977~. It may also cause mesenteric adenitis, hepatosplenic abscesses, or septicemia, and may initiate a variety of autoimmune processes, including erythema nodosum and polyarthritis (Anonymous, 1984; Bottone, 1977~. There is some evidence implicating Y. enterocolitica in Reiter's syndrome, carditis, glomerulonephritis, Grave's disease, and Hashimoto's thyroiditis (Anonymous, 1984; Larsen, 1980~. Host factors appear to play a major role in pathogenicity and in the expression of yersiniosis: diarrhea is seen more frequently in young children, mesenteric adenitis in older children and adolescents, and autoimmune phenomena in women. Arthritis correlates with occurrence of the HLA B27 histocompatibility antigen.
71 Illness has been associated with the isolation of Y. enterocolitica from clinical specimens in several outbreak investigations (Black et al., 1978; Shayegani et al., 1983; Tacket et al., 1985~. At least one case-control study showed an association between the occurrence of diarrhea and isolation of the microorganism (Marks et al., 1980~. Chronic Yersinia infection can be established by history and serologic evidence of prior infection (Anonymous, 1984~. There are no good data on Y. enterocolitica infections from studies in volunteers. Given the invasive nature of the microorganism and the potential seriousness of infections, it is unlikely that such data could be obtained at this time. The pathogenicity of Y. enterocolitica appears to depend on several factors, including the presence of the O serogroup and invasive strains of plasmids ranging in molecular mass from 40 to 48 megadaltons (Gemski et al., 1980; Portnoy et al., 1981~. Disease-associated Y. enterocolitica strains isolated in Europe have generally belonged to . . serogroups 0:3 and O:9 (de Groote et al., 1982~. In the United States serogroup 0:8 is the most common serogroup isolated, but this pattern may be changing (Bottone, 1983~. The invasive strains of plasmids encode, among other properties, the production of specific outer membrane proteins, calcium dependence, and adherence to HEp-2 tissue culture cells (Heesemann et al., 1984~; other proposed virulence factors, including invasion of HeLa cells and enterotoxin production, appear to be chromosomally mediated (Kay et al., 1983; Schiemann and Devenish, 1982~. At this time no single factor appears to be sufficient to differentiate pathogenic from nonpathogenic strains (Kay et al., 1983~. Laboratories in the United States that routinely culture all stool samples report isolation rates for Y. enterocolitica in stools, blood, and other samples that are only about 3.8% to 33.6% of the rate for Salmonella (Marymont et al., 1982; Snyder et al., 1982~. Studies in Europe suggest that stool isolates account for more than 95% of total Yersinia isolates. Extrapolating from the previously cited data on Salmonella (approximately 19 reported cases per 100,000 per year), and assuming that there are 5% as many Yersinia isolates as Salmonella isolates, the incidence of diagnosed Yersinia infections would be approximately one per 100,000 members of the population per year. In the absence of good data relating total infection rates to diagnosed infections, it is not possible to predict the actual number of infections that may occur. Up to 30% of isolates of Y. enterocolitica are obtained from hospitalized patients, and as much as 10% of reported infections may be associated with a syndrome of mesenteric adenitis mimicking appendicitis (de Groote et al., 1982; Snyder et al., 1982; Tacket et al., 1985~. These data suggest that hospitalization rates are proportionately much higher than those resulting from Salmonella infections (1-2~. Data are inadequate to calculate mortality rates accurately. Dose-Response Studies. Because of the potential seriousness of yersiniosis, studies of infectious doses have not been performed on
72 large numbers of volunteers Og Enterocolitis and fever resulted in one person who ingested 3.5 x 10 organisms. The symptoms lasted for 4 weeks (Szita et al., 1973~. In epidemiological studies, the necessity of using the standard microbiological enrichment procedures for contaminated specimens makes it difficult to estimate levels of contamination from epidemiologically linked fOodse Potential for Human Exposure. Investigators who have specifically looked for colonization of Y. enterocolitica have identified the microorganism in a small percentage of broiler chickens available for retail sale. Y. enterocolitica has also been isolated from a variety of sources, including wild and domestic animals and water (Vantrappen et al., 1982) as well as poultry, beef, pork, fish, oysters, raw milk, and pasteurized milk (Lee, 1977; Morris and Feeley, 1976; Rechtman et al., 1985; Schiemann and Toma, 1978; Stengel, 1985)0 Characterization of Risk. The widespread presence of Yersinia in the environment and the diversity of food vehicles potentially involved in food-borne outbreaks suggest that broiler chickens may not be a major source of yersiniosis in humans, despite the microorganism's presence on chickens in retail stores. Better data are needed on the roles of various foodstuffs, including poultry, in the epidemiology of . . . yerslnlos IS o Known Human Pathogens Associated with the Ingestion of Chicken CampYlobacter, Salmonella, and Yersinia are not the only pathogens associated with eating chicken. Generally, however, raw chicken does not appear to be a major channel through which other pathogenic microorganisms enter the kitchen. Spore-forming microorganisms such as Bacillus cereus, Clostridium botulinum, and C. perfringens appear to be ubiquitous in nature. Although they may be present on the surface of chickens, they are as likely, or more likely, to be brought into the kitchen in soil or on vegetables, grains, or other farm products. Shigella, which is infectious at very low doses, tends to be transmitted by direct fecal-oral contact. For example, it may be transferred directly onto food by food handlers whose hands have been contaminated with feces containing the microorganism. Isolates of Shigella boYdii have infrequently been obtained from poultry feces (Fagerberg and Quarles, 1982~. The direct role of consumer-ready chicken in human shigellosis is unclear, but is probably negligible. Staphylococci tend to be transferred to food by persons whose hands are infected with staphylococci or by people who are chronic nasal or rectal carriers of the microorganism. Two species in this group, B. cereus and C. botulinum, are described in detail in the following paragraphs. A full formal risk assessment will require similar analyses of other pathogens in this category.
73 Bacillus cereus Hazard Identification and Evaluation. Bacillus cereus is an aerobic, spore-forming, gram-positive rod that has recently been implicated as a cause of gastroenteritis. High counts of B. cereus have been detected in food implicated in disease outbreaks and, less often, in stool samples from ill patients. There are no data showing an association between B. cereus and cases of sporadic diarrhea, although up to 14% of asymptomatic persons in the general population may have the microorganism in their stool (Ghosh, 1978~. There have been limited volunteer experiments with this microorganism. Hauge (1955) inoculated vanilla sauce with a strain of B. cereus isolated from an outbreak of food poisoning in Norway, and ingested about 200 ml of the sauce containing approximately 101° microorganisms; 13 hours later he experienced severe abdominal pain, diarrhea, and rectal tenesmus, which persisted for 8 hours. In a previous experiment, four of six other volunteers who drank 155 to 270 ml of vanilla sauce containing 30 to 60 million B. cereus per milliliter experienced similar symptoms (Gilbert, 1979; Hauge, 1950~. Two clinical syndromes are associated with B. cereus: a short incubation vomiting syndrome (similar to that seen with staphylococcal food poisoning) associated with a heat-stable emetic toxin and a diarrhea! syndrome (similar to that seen with Clostridium p~erfringens) associated with production of a heat-labile, cyclic AMP-mediated enterotoxin (Terranova and Blake, 19781. B. cereus has also been implicated as a cause of bacteremia, pneumonia, and meningitis, and has been isolated from serious wound infections in otherwise healthy patients as well as from immunocompromised patients (Turnbull et al., 1977, 1979). B. cereus has been widely recognized in Europe as an important cause of food-borne disease. Between 1960 and 1966, for example, B. cereus was the third most commonly reported cause of food-borne disease outbreaks in Hungary (Goepfert et al., 19721. Such outbreaks have been identified infrequently in the United States, however. Of approxi- mately 1,800 food-borne disease outbreaks of known etiology reported to CDC between 1966 and 1979, only 19 were attributed to B. cereus, including 10 outbreaks of the diarrhea! syndrome and 9 of the emetic syndrome (Morris, 1981~. These numbers probably reflect severe underrecognition of B. cereus outbreaks. Most laboratories in the United States are not familiar with the microorganism and do not attempt to isolate it routinely from stool and food samples. Dose-Response Studies. There are inadequate data to Clearly establish an infectious dose for B. cereus. Doses of 10- ~ have caused disease in rhesus monkeys and volunteers. Counts higher than 105 per gram of food have been associated with disease in outbreak investigations (Morris, 1981~. Potential for Human Exposure. B. cereus is ubiquitous in soil and can be isolated from virtually all farm produce, including vegetables
74 (51.2% of samples in one study [Nygren, 1962]), grain (91.2% of uncooked rice samples [Gilbert and Parry, 1977~), and milk and dairy products (including, in one study, 86.7% of bottled pasteurized milk samples [Gilbert, 1979 ; Ionescu et al ., 1966 ~ ~ . It is also common ~ n dried and processed foods, including spices (72.39s of samples [Nygren, 1962~) and products such as chocolate pudding powder (469s of samples tested [Nygren, 1962 ] ~ and dried chicken (47 . 8% of samples [Nygren, 19623~. Outbreaks of the vomiting syndrome have almost all been associated with fried rice; conditions inside rice warmers used in Chinese restaurants appear to be ideal for germination 9 growth, and production of the emetic toxin (Gilbert et al., 1974~. Chicken has been implicated as the vehicle of B. cereus transmiss ion in at least one outbreak of diarrhea! disease in the United States (Midura et al., 1970) and in several other outbreaks reported in the literature, but there are no good data on rates of isolation from raw chicken (Gilbert, 1979~. Data are not adequate to determine whether contamination of chickens before or after retail sale played a role in these outbreaks, and there are no data to suggest that chickens play a more significant role than other foods in the transmission of B. cereus. Characterization of Risk. The wide distribution of B. cereus in the environment and in various foods suggests that contaminated broiler chickens are not a major factor in the spread of this microorganism. An accurate assessment of the public health risk attributable to broiler chickens would require additional knowledge of the epidemiological-characteristics of disease caused by food-borne B. cereusO Clostridium botul~num Hazard Identification and Evaluation Investigators have demonstrated a clear association between the presence of botulinal toxin in food and human illness. Group I, II, and possibly IV (types A, B. E, F. and G) toxins produced by Clostridium botulinum are responsible for the necrologic signs and symptoms observed in human cases of botulism (Morris and Hatheway, 1983 ~ . No well- documented studies of food-borne botulism in volunteers have been conducted. When botulinal toxin is used to treat strabismus in humans, however, it ~ s known to cause localized paralysis after injection into muscle (Scott, 19811. Botulism patients typically have blurred vision as well as difficulty speaking and swallowing. During the 24 to 72 hours after onset, there may also be a progressive descending motor paralysis, resulting, in the most severe cases, in total loss of motor function, including respiratory function (Hughes et al., 1981; Morris and Hatheway, 1983~. Recovery is usually complete in patients who receive adequate supportive care, including mechanical respiratory support. The average duration of respiratory support is 1 month, but has been provided for as long as 7 months (Morris and Hatheway , 1983 ~ O
75 Between l9SO and 1979, approximately 25 cases of food-borne botulism were reported in the United States each year (CDC, 1979~. Virtually all these patients were hospitalized. The case-fatality rate for botulism is currently less than 109~, down from approximately 70% between 1910 and 1919. This improvement is attributable primarily to the widespread availability of long-term respiratory support (Morris and Hatheway, 1983~. Dose-Response Studies. In contrast to most other food-borne diseases, botulism is caused by the presence of toxin in food, rather than by the ingestion of the microorganisms themselves. Consequently, disease severity is directly related to the amount of toxin present in the food. A dose of 10 g of purified botulinal toxin is probably sufficient to cause disease in a human; doses between 10-6 and 10- g can be a lethal (Morris and Hatheway, 1983)0 Potential for Human Exposure. C. botulinum spores are widely distributed in the environment and may be isolated from soils and marine sediments, the surfaces of vegetables and fruit, and fish and other seafood. To germinate, spores require anaerobic conditions (restricted oxygen and sufficiently low Eh [redox potential]), adequate nutrients, low acidity (pH > 4.6), sufficient availability of water (low solute concentrations; Aw twater activity] 0.93), suitable temperature, and lack of inhibiting substances. If spores are present on food maintained under such conditions (e.g., in improperly home-canned food), they will germinate and toxin will be released as vegetative cells lyse. The toxin is heat labile and is destroyed by cooking (Morris and Hatheway, 1983~. Chicken or chicken-containing products were implicated as the vehicle of infection in 4 of the 190 outbreaks of food-borne botulism of known etiology reported between 1950 and 1977 (CDC, 1979~. These included two outbreaks resulting from the consumption of commercially prepared chicken pot pie (one from type A toxin and one from a toxin of an undetermined type), one outbreak involving commercially prepared chicken livers (type A toxin), and one involving home-canned chicken soup (type B toxin). Given the heat-lability of the toxin, mishandling of the product (and subsequent toxin production) most likely occurred after initial cooking. Although chicken carcasses may occasionally carry C. botulinum spores, there are no data to suggest that this has any significant effect on the occurrence of disease in the human population. Chickens are themselves susceptible to botulism. Type C toxin (produced by group III C. botulinum growing in decaying vegetation or decomposing carcasses) causes limberneck (a disease characterized by diffuse motor paralysis) in chickens, turkeys, and several species of wild birds. Apparently, type C toxin can also affect mammals, and there are at least two poorly documented reports of type C disease in humans (Hariharan and Mitchell, 1977~. There are no data to suggest that humans have acquired type C disease by eating affected birds; even if affected birds were eaten, cooking should destroy any toxin present in the tissue.
76 Characterization of Risk. The risk to humans presented by preformed botulinal toxin in raw chicken appears to be negligible. Botulinal spores may be carried on raw chicken, but given the wide distribution of spores in the environment, there is no evidence that This is associated with an increased risk of illness. Microorganisms; Known to he Pathogen; c in (10 skins: That Are of Questionable Significance as Food-Borne Pathogens Transmitted by, Broiler Chickens Many other microorganisms, including bacteria, viruses, and parasites, are common causes of disease in poultry but play an uncertain role in human disease. Host-specific~ty appears to preclude many of these microorganisms from contributing to the burden of food-borne disease in humans. Other microorganisms, such as ChlamYdia psittaci, may cause disease in humans, but are transmitted through routes other than oral. Still others, such as Mycobacterium avium, are shared by humans and poultry, but there is no evidence that broiler chickens serve as a vehicle for transmitting this microorganism to humans . This section provides examples of the information necessary to assess the risk to public health posed by the microorganisms in this group. Inasmuch as new human enter pathogens continue to be identified, a more thorough and continuing review of avian pathogens should be a part of a complete public health risk assessment of broiler chicken inspection. Bacteria O Haemophilus gal~inarum causes infectious coryza in poultry, and some strains kill mice when injected intraperitoneally, but it does not appear to be pathogenic in humans (McGaughey, 1932~. Infectious coryza is relatively common among the diseases of broilers, but the affinity of the agent for upper airway tissues, which are removed during slaughter and processing, should keep contamination low even in severely affected birds. Pasteurella multocida causes fowl cholera in domestic poultry O Most avian strains belong to capsular type A, but some are type D; both of these antigenic types can cause respiratory disease in humans (Carter, 1962, 1972; Smith, 1955~. Although outbreaks of fowl cholera can occur in flocks, P. multocida is transmitted to humans mainly by inhalation of infectious aerosols emanating from coughing poultry or livestock. Thus, it presents an occupational risk of respiratory disease rather than a risk of food-borne infection to the consumers Infection of broilers by Escher~ch~a cold is an important cause of several clinical syndromes. E. cold Is also an important cause of disease in humans, different serotypes typically being identified with different pathogenic mechanisms (Gangarosa, 1978a; Gangarosa and Merson, 1977; Guerrant, 1980)0 There is some overlap between O group serotypes isolated from diseased poultry and those characterized as enteropathogenic, enterotoxigenic, and enteroinvasive for humans, but
77 data are inadequate to determine whether E. cold strains from poultry are important contributors to food-borne disease (Glantz, 1971; Gross, 1983~. Better definitions of the extent to which E. cold causes enteric disease in humans and the role of contaminated poultry as a vehicle are needed. Clostridi~ colinum is associated with ulcerative enteritis in young chickens. This is not particularly common in the broiler industry, and no evidence was found to suggest a pathogenic role for C. £olinum in humans (Borriella, 1985~. Species of Mycoplasma are found mainly in the oral cavity, the upper respiratory tract, and distal genital tract of humans and birds, and are host-specific (Hayflick, 1969~. Therefore, although M. gallisepticum is a major contributor to chonic respiratory disease in chickens and M. sYnoviae causes both respiratory disease and infectious synovitis in poultry, current knowledge does not suggest that they pose a health hazard to consumers of broiler chickens. Nontuberculous mycobacteria, including those in the MYcobacterium avium complex, cause chronic pulmonary disease, lymphadenitis, other soft tissue infections, and bone and joint disease in humans (Wolinsky, 1979~. Immunosuppressed individuals and those with underlying chronic lung disease are especially susceptible (D~msker and Bottone, 1985; Good, 1985~. Infection in poultry is caused by serotypes 1, 2, and 3 (Thoen et al., 1981), but clinically evident disease is restricted almost entirely to old laying hens. Therefore, in countries where extensive long-term hen holding is not practiced, e.g., in the United States, human infections with M. avium complex microorganisms are due primarily to strains other than the avian serotypes 1, 2, and 3 (Meissner and Anz, 1977; Schaefer, 1968~. For this group of microorganisms, the sources and vehicles of human infections need further investigation In many cases, they appear to be of environmental origin (e.g., from water, sand, and sawdust) (Meissner and Anz, 1977; Schaefer, 1968~. In humans, Chlamydia psittaci can cause infections ranging from asymptomatic to severe systemic disease. Most uncompromised people who become ill have only a flu-like syndrome, but infections in the elderly can be life-threatening (Kuritsky et al., 1984~. Of the 100 to 150 cases of psittacosis in humans reported in the United States each year, most are associated with caged pet birds (Potter and Kaufmann, 19791. Headache, chills, and fever are the most commonly reported symptoms. Roentgenographic evidence of pnewmonitis is found commonly in patients with little clinical evidence of pulmonary lesions (Kuritsky et al., 1984~. Chickens are susceptible to infection by C. E~sittac~, but such infections are rare. Nearly all poultry- associated cases of psittacosis in humans appear to result from occupational exposure to infected turkeys during slaughtering (Anderson et al., 1978; Durfee et al., 1975; Filstein et al., 1981; Hines et al., 1957; Meyer and Eddie, 1942 ~ . Where is no evidence that market- ready chickens present a risk for psittacosis in humans.
78 Avian Viruses. There are numerous reports on infection of humans by the paramyxovirus that causes Newcastle disease in poultry. One hundred cases of Newcastle disease virus (NDV) infections in humans were reported frown 1943 to 1958 (Chang, 1981), but there have been fewer reports in the more recent medical literature. Of the three types of NDV, two (lentogenic and velogenic viruses) have been reported to cause infection in humans (Dardiri et al. 9 1962; Miller and Yates, 1971; Reagan et al. 5 1956~. The most common symptom of these infections is a unilateral conjunctivitis, which lasts for 3 to 4 days ~ Bilateral conj unctivitis, chills, malaise, headache, and fever have also been reported. Newcastle disease in humans is considered to be an occupational hazard for those who have close contact with poultry and for laboratory personnel who work with the virus. There is no evidence that ingestion of meat contaminated with NDV causes infection of humans. Strains of influenza A viruses infect humans and a large variety of birds, including chickens (Murphy and Webster, 1985; Wood et al., 1985) . The common antigenic components of these strains (e . go, Asian H2N2) suggest that avian species may have been involved in the origin of the Asian virus that affects humans (Kaplan, 1980; Nerome et al., 1984~. The avian influenza outbreak in chickens that occurred in Pennsylvania and surrounding states from October 1983 through February 1984 was associated with HSN2 influenza A viruses. People in direct contact with infected birds were shown to carry the avian virus over a short period, but there was no evidence that they developed infections (Bean et al., 1985~. The avian retroviruses, such as avian leukosis virus (ALV), cause a variety of neoplastic diseases in chickens (Crittenden, 1976; Penner et al., 1974~. The chick embryos used in the production of yellow fever vaccine during World War II were contaminated with ALV, but thus far, no association has been found between immunization with vaccines containing oncogenic ALV and leukemia, lymphoma, or other cancers in humans (Richman et al., 1972; Waters et al., 1972~. Researchers have found no etiologic relationship between avian and human leukemia (Solomon and Purchase, 1969), no evidence of ALV group-specific antibodies in human serum (Roth and Dougherty, 1971), and no association between cancers in humans and avian myeloblastosis virus (Hehlmann et al., 1972~. There is also no evidence that ALV and related viruses are infectious or carcinogenic in humans. Arboviral encephal~tides in mammals (including humans) and birds are caused by alphaviruses. These viruses are transmitted by a mosquito vector. Thus, although both chickens and humans are susceptible to infection by both Eastern and Western equine encephalitis viruses, meat from infected broiler chickens would not present a food-borne hazard to consumers. The medical literature provides no evidence that Marek's disease herpesvirus (MDV) is pathogenic for humans. Relatively high titers of
79 anti-MDV antibody were found in 64 samples of human sera by Naito et al. (1970~. In a similar study, however, investigators found no significant anti-MDV antibody titers in 205 serum samples from persons who worked with the virus in laboratories or who had contact with infected chickens (Sharma et al., 1973~. Other serologic studies to obtain evidence of MDV infections in humans have produced negative results, both in healthy individuals and in cancer patients (Makari, 1973; Purchase and Witter, 1986~. The committee found no evidence of human infection by avian reoviruses, poxviruses, adenoviruses, or avian infectious bronchitis virus. Although avian infectious laryngotracheitis virus is not known to infect humans, a related herpesv~rus has been reported to cause subacute myelo-opticoneuropathy in humans in Japan (Biggs, 1982; Inoue, 1973, 1975; Inoue and Nishibe, 1973; Nishimura and Tobe, 1973; Roizman and Batterson, 1985~. Parasites. Since its recognition as a human pathogen in 1976, Cryptosporidium has been identified as an important cause of diarrhea! disease worldwide (Navin, 1985~. In humans, cryptosporidiosis is generally manifested as a short-term cholera-like diarrhea! illness in immunocompetent people or as a prolonged life-threatening illness in immunodeficient patients (Current, 1985~. Among immunocompetent people, the infection appears to be more common In children than in adults. In poultry, Cryptosporidium has been associated with mild intestinal disease and severe respiratory tract disease. Although early studies documented the potential for bird-to-hu~an transmission (Current, 1985), most Cryptosporidium infections in humans are probably not acquired directly from infected birds (Navin, 1985~. The frequency with which these cases are due to person-to-person, water-borne, or food-borne transmission is unknown. USING THE RISK MODEL TO DEVELOP PROGRAMS AND STRATEGIES The need to reduce the public health risk associated with microbial contamination of foods, including poultry products, has been clearly established. Reports to the CDC during the past 10 years indicated that outbreaks of illnesses attributable to the ingestion of chicken have included infections and intoxications due to Salmonella spp., Campylobacter jejuni, Shigella spp., Staphylococcus aureus, Clostridium perfring;ens, Bacillus cereus, and Clostridium botulinum. Archer and Kvenberg ( 198S ~ estimated that millions of cases of food-borne illnesses costing billions of dollars occur each year. It has also been welt documented that various known and potential pathogens can contaminate the poultry product at different stages of production, from transovarial infection of eggs by Salmonella (Faddoul and Fellows, 1966) to contamination of carcasses or parts by env~ro~mental Clostridium perfringens during further processing or preparation for consumption (Bryan, 1980b). Managing the risks produced in such divergent ways by microorganisms with such diverse characteristics requires careful application of the precepts of formal risk assessment and risk management.
80 Analysis of the data with the tools of risk assessment would permit FSIS to val idate the assumptions made during the preceding risk assessment, to determine the relative effectiveness of alternative risk management tools, and to establish priorities for each activity in the subsequent program period. Regular reevaluation of priorities would permit FSIS to take advantage of emerging technologies to manage new public health risks. Through continual reassessment, it would be possible to identify and replace inspection methods that had outlived their usefulness without disrupting program continuity As discussed in Chapter 3, before a rational program to manage the risk of poultry-borne disease can be developed, one must first have data derived from a comprehensive risk assessment. The activities necessary for the collection of such data are discussed below along with proposed risk-management options. Risk Assessment Activity 1. Identification of Potenti al Pathogens on Broiler Chickens. Bacteria, fungi, protozoa, and viruses may be associated with chickens, either producing disease or living as commensals. These microorganisms can infect poultry during production or can become contaminants during slaughter, processing, and further handling O Microorganisms present in or on chickens have been described by several groups (NRC, 1969, 1985b) and in the first section of this report. To supplement this information, it would be desirable to collect and analyze end product contamination data, i . e ., what microorganisms does the consumer encounter at the time broiler chickens are purchased or consigned? Such data can be obtained in part from reviews of the scientific literature. For example, Todd (1980) reviewed the world literature published during the 1960s and 1970s and found extremely variable rates of contamination of fresh and frozen chicken by Salmonella. This author also described other microbial contaminants of market-ready poultry. Numerous reports of infectious diseases of poultry appear annually in journals of veterinary medicine and poultry science. Each year, for example, Avian Diseases, the journal of the American Association of Avian Pathologists, contains diagnostic summaries from selected laboratories that diagnose poultry disease. From 1981 to 19839 these summaries included more than 30 infectious diseases and conditions that should be considered in a complete risk assessment. Journals covering food science and microbiology also contain reports on microbial contaminants of poultry and edible poultry products. A second and equally important source of data is the careful analysis of PSIS microbiological surveys. The data should be supple mented by the collection of additional specific critical data, such as determination of the level of microbial contamination in condemned birds--something recommended in a previous NRC report (NRC, 1985a). These comparative data on microbial contamination of passed and condemned carcasses are needed for evaluation of inspection procedures.
81 Data can also be obtained from field studies of stores and homes. Although ESIS authority beyond processing is limited, field studies critical to its major mission, i.e., to ensure ~ use of inspection to protect the public health. These studies should include a quantitative component so that microorganisms present only intermittently on a small percentage of chickens (or present only at very low levels) can be distinguished from microorganisms found persistently on a high percentage of carcasses, and contamination levels can be compared over time, between and within geographic areas, from flock to flock, and in other ways related either to human disease or to possible control measures. These data will permit the development of a comprehensive list of microorganisms present on broiler chickens and will indicate what further risk-assessment data are needed. Quantitative data are necessary in the risk-assessment process itself. homes. are the sound and effective This information should be continually updated as new reports are published and results are obtained from a well-designed FSIS surveillance system. Careful review of the data should lead to the identification of critical gaps in the data base and to the initiation of new studies to provide the needed information. Data from the rapidly maturing field of avian microbiology should be continually integrated into FSIS risk- assessment processes . Activity 2: Collection of Data Necessary for Risk Assessment. Formal risk assessments should be conducted for each potentially hazardous microorganism by using the format outlined in this chapter. The resulting data should be as accurate and comprehensive as possible to support a formal determination about the risk-management strategy that should be applied. The data may indicate that some microogranisms do not require risk management if they are judged to present minimal risk, if they are not pathogenic in humans, or if their presence in broiler chickens is not important in the epidemiology of human diseases. Other microorganisms (e.g., Salmonella and Campylobacter) may be judged to be important publi c health risks, because data support an important role for poultry as a vehicle of food-borne disease in consumers. These microorganisms will be considered in the further development of risk-management strategies. For many microorganisms known to be pathogenic in chickens, there are only limited data on the risk to human health. For microorganisms that are known human pathogens, such as Yers~nia spp., more careful assessment of the potential for human exposure is needed; for example, evaluation of data on product contamination levels and acquisition of data on attributable risk (i.e., what percentage of cases can be attributed to chickens) similar to the study of Campylobacter conducted by the Seattle-King County Department of Public Health for FDA (SKCDPH, 1984).
82 . By using the data developed in this activity, one should be able to identify and establish priorities for managing high-risk microorganisms that are on or in broiler chickens at retail, are known to be human pathogens, and are transmitted through food. These priorities should guide FSIS in its management of microbial risks to public health. In particular, they should carry implications regarding the kinds and intensities of inspection that would be most appropriate. For example, what portion of available resources should be devoted to microbiological inspection? How much does organoleptic inspection reduce microbial risks? Would a sampling program, combined with an FSIS and industry shift from detection to prevention, provide enhanced protection? These and related questions will require the best possible quantitative risk assessments. Thus, the priority list will have substantial importance. From the data reviewed in this chapter, it is likely that the list would include Salmonella, Campylobacter, and possibly Yersinia O Activity 3: Determination of Maj or Control Points . As a first step toward risk management, efforts should be made to determine the major control points for each high-risk microorganism by using the flow diagrams given in Chapter 30 In general, the control points can be grouped into two maj or categories: those that influence the level of microbial contamination at the time of retail sale and those that dictate levels at the time the product is consumed ~ the so - called dinner plate county. It would clearly be advantageous if microbial contamination could be totally eliminated before retail sale or before consumption, but complete control at either stage is unlikely. For example, total elimination of microorganisms such as Salmonella from chicken carcasses does not appear to be likely at present, nor is it likely that mishandling of chickens by consumers can ever be completely prevented. If complete control cannot be achieved, the problem then becomes one of assessing the relative importance of each of these areas (or each group of control points) as determinants of disease. While studies have demonstrated a great deal of fluctuation in the levels of microbial contamination during processing, contamination at the time of sale can be directly measured and should be susceptible to modification by using an HACCP system explicitly designed for the control of each microorganism within the framework of production practices in each slaughter facility . The HACCP systems would undoubtedly have some effect on disease incidence, but the limited studies reviewed in this chapter do not show a clear relationship between reduced contamination levels and the occurrence of illness. Numerous studies have documented both the expense and the frustration of trying to control microbial contamination of poultry and suggest that very strict guidelines for limiting microbial contamination during processing may not provide commensurate health benefits.
83 There is a need for additional quantitative data relating disease incidence to microbial contamination at the time of retail sale. Unfortunately, however, the development of such data is not easy. One approach, which has already been applied, is based on the use of a large health maintenance organization in conjunction with a population-based surveillance system (SKCDPH, 1984~. An alternative approach, which offers the advantage of providing a broader-based population sample and a quantifiable risk, would be based on a system of sentinel county health departments in counties where broiler chickens are supplied almost entirely from a few identified processing plants. The county would require a public health infrastructure that includes centralized laboratory facilities and prompt reporting to the epidemiologist in charge; microbiological competence for isolating and identifying specified pathogens; and the ability to perform epidemiological studies adequately so that risk factors can be identified for each infection found. Data on disease incidence in these counties could then be correlated with FSIS microbiological survey data, thereby permitting direct correlation of illness rates and microbial contamination levels. These sentinel counties could also be used to test the relative impact of changes in inspection strategies and the efficacy of control progams for specific microorganisms. Risk Management As outlined in the 198S NRC report on meat and poultry inspection (NRC, 1985b), programs to limit microbial contamination during processing of broiler chickens are best considered as part of an HACCP system. The success of such programs can be measured directly by end-product sampling; however, there is also a need for data of the type described above if changes in microbial contamination are to be related to the occurrence of disease in the community (i.e., to public health). There has been considerable debate concerning the advisability and feasibility of classifying raw foods of animal origin as either acceptable or unacceptable on the basis of microbiological criteria. To justify the establishment of regulatory standards, such as mandatory limits on the frequency or numbers of specified pathogenic microorganisms, statistically defensible background data and a clearly defined need are necessary. There are clear deficiencies in the data base on poultry-borne disease in the United States. Therefore, this committee concurs with the earlier NRC committee (NRC, 1985a) that microbiological standards for pathogens in raw poultry are inappropriate at this time. The committee does, however, encourage FSIS and the industry to begin exploring strategies to decrease the levels of fecal contamination of edible poultry tissue to improve the overall microbial quality while new data are being generated. The production of Salmonella-free poultry could serve as the first line of defense against salmonellosis (Lipton, 1983~. Programs to eradicate S. gallinarum and S. pulloru~ in poultry have been largely
84 successful, but attempts to eliminate other serotypes in poultry have been less so, largely because of the multiplicity of sources, the complexity of transmission, and the lack of coordination among groups responsible for control at different points. Control oF microorganisms that are not host-adapted will be similar for Salmonella serotypes and other microorganisms. In defining critical control points, consideration should be given to a number of factors, including facility design; isolation of broilers; controlled access to poultry houses, feed, water, litter, breeding stock, and eggs; cleaning and disinfection; and availability of veterinary diagnosis for sick or dead birds. Careful control of these points has been shown to effect large reductions in the prevalence of salmonellosis in turkeys (Campbell et al., 1982)0 In slaughter and processing facilities, clean equipment and good sanitation are essential to prevent contamination of edible poultry tissues by intestinal flora and environmental microorganisms. Many investigators have attempted to measure the levels of contaminants on poultry carcasses at various points in the slaughter or processing chain, most often for Salmonella, and have shown that microbial loads are established early in the slaughter process (Campbell et al., 19841. Carcasses are chilled in a common cold water bath, but if the chillers are operated properly, there is little opportunity for cross~contamination, as demonstrated by the Commission of European Communities (CEC, 1976), which reported that reduction of contamination before chilling was more important than the type of chilling procedures used. Other possible sources of contamination and cross-contamination of carcasses during slaughter and processing should be similarly studied to determine where and when contamination is introduced or exacerbated so that the principles of the HACCP system can be applied to reduce the microbial load on the market-ready broiler. Consumer handling of broiler chickens after retail purchase plays a major role in determining whether microorganisms present at the time of purchase will cause human disease. The principles of good food handling are relatively simple: chicken should always be thoroughly cooked, efforts should be made to minimize cross-contamination within the kitchen, and food should be kept either warm, at least to 140°F (60°C), or cold, below 40°F (4.44°C), to minimize multiplication of potential food-borne pathogens. A range of programs could be developed, or current programs expanded, to educate consumers about the importance of good food-handling practices. This might include increased emphasis on school education programs, development of adult consumer education materials, dissemination of information on public or commercial television or radio, or the development of product inserts describing proper handling techniques for poultry. Given the frequency of food-borne disease outbreaks associated with restaurants and institutions, efforts should also be directed toward reenforcing the importance of correct food-handling practices in those establishments.
85 It is much more difficult to assess the final outcome of educational programs than to assess programs to limit microbial contamination during processing. However, it would be possible to pretest the public health impact of major programs in all these areas by using a cou~unity-based surveillance system, as described above. The proposed educational programs for food handlers or homemakers could be instituted in such counties, and their subsequent impact on disease incidence could be measured directly. IMPLICATIONS OF THE PROPOSED RISK MODEL FOR MICROBIAL CONTAMINATION FOR THE CURRENT FSIS INSPECTION PROGRAM The committee believes that the present system of inspection does very little to protect the public against microbial hazards in broiler chickens. The issue of whether bird-by-bird inspection offers greater public health protection than would inspection of a sample has not been m`1~-~!~!~ ;~ ~ A-; {S^~-^llC C!~; ^ - taxi Pi rat mom - A' Likewise, microbiological differences between passed and condemned carcasses have not been demonstrated. Furthermore, the public health impact of current inspection procedures has not been defined. In the absence of such data, the committee recommends that FSIS begin now to lay the groundwork for a shift of resources from its present inspection strategy to a program that is more likely to have a substantial impact on human diseases and that FSIS make such a shift as soon as possible. Further development of quantitative health risk assessment data, which must include a comprehensive evaluation of all pathogens on chickens, will be an essential tool in this change. .~ Whatever public health-based inspection system evolves from the risk-assessment procedure, systems for monitoring compliance with critical control point parameters and effects on product quality and public health would have to be designed. Compliance could be determined by physical measurement of contamination at critical control points and by analysis of microbial contamination, if any microbiological standards or guidelines were established. Surveillance to determine microbial contamination of the end products would provide a measure of the overall program's success in reducing the prevalence of human pathogens on market-ready poultry and would permit the tracking of trends in product contamination. Because of the influence of food handling on the occurrence of food-borne disease, however, end-product surveillance alone can not be used to judge the public health impact of the poultry inspection program. Through community-based surveillance systems (described above), FSIS could define the role of poultry products in food-borne disease in humans, establish trends to define the benefits of reduced carcass contamination, and test the relative impact of various inspection strategies.
86 REFERENCES Anderson, D. C., P. Ao Stoesz, and A. F. KaufmannO 1978. Psittacosis outbreak in employees of a turkey-processing plant. Am. J. Ep idemiol . 107: 140 - 148 . Anonymous . 1984e Yersiniosis today. Lancet 1: 84-85 . Archer, D. L., and J. E. Kvenberg. 1985. Incidence and cost of foodborne diarrhea! disease in the United S tates. J. Food Protect. 48 887-894. Aserkoff, B., S. A. Schroeder, and P. S. Brachman. 1970. Salmonellosis in the United States--a five -year review. Am. J. Epidemiol. 92:13O24. Bean, W. J. a Y. Kawaoka, J. M. Wood, J. E. Pearson and R. G. Webster. 1985. Characterization of virulent and avirulent A/chicken/ Pennsylvania/83 influenza A viruses: Potential role of defective interfering RNAs in nature. J. Viral. 54: 151-160. Biggs, P. M. 1982. The epidemiology of avian herpesviruses in veterinary medicine. Dev. Bio] . Stand. 52: 3-11. Black, R. E ., R. J . Jackson, T . Tsai, M. Medvesky, M. Shayegani, J . C O Feeley, K. Ie E. MacLeod, and A. M. Wakelee. 1978. Epidemic Yersinia enterocolitica infection due to contaminated chocolate milk. N O Engl . J . Med. 298 :76 - 79 O Black , R. E ., M. M. Levine , M. J . Blaser , M . L. Clements, and T . P . Hughes . 1983 . Studies of Campylobacter j e juni infection in volunteers. P. 13 in Ao D. Pearson, M. B. Skirrow, B. Rowe, J. R. Davies, and D. M. Jones, eds. Campylobacter TI: Proceedings of the Second International Workshop on Campylobacter Infections held in Brussels, September 6 - 9, 1983 . Public Health Laboratory Service, London. Blaser, M. J . , and L. B. Reller. 1981. Campylobacter enteritis O N. Engl. J. Med. 305: 1444-1452. Blaser, Me J., J. G. Wells, R. A. Feldman, R. Ao Pollard, and J. R. Allen. 1983a. Campylobacter enteritis in the United States: A multicenter study. Ann. Intern. Med. 98:360-36S. Blaser, M. J., D. N. Taylor, and R. A. Feldman. 1983b Epidemiology of Campylobacter jejuni infections O EpidemiolO Rev. 5:lS7-176.
87 Borriella, S. P., ed. 1985. Clostridia in Gastrointestinal Disease. CRC Press, Boca Raton, Fla. 248 pp. Bottone, E. J. 1977. Yersinia enterocolitica: A panoramic view of a charismatic microorganism. CRC Crit. Rev. Microbiol. 5:211-241. Bottone, E. J. 1983. Current trends of Yersinia enterocolitica isolates in the New York City area. J. Clin. Microbiol. 17:63-67. Bradbury, W. C., and D. L. G. Munroe. 19850 Occurrence of plasmids and antibiotic resistance among Campylobacter Jejune and Campylobacter cold isolated from healthy and diarrheic animals. J. Clin. Microbiol. 22:339-346. Brouwer, R., M. J. A. Mertens, T. H. Siem, and J. Katchaki. 1979. An explos ive outbreak of Campylobacter enteritis in soldiers. Antonie van Leeuwenhoek 45:517-5190 Bryan, F. L. 1980a. Foodborne diseases in the United States associated with meat and poultry. J. Food Protect. 43:140-150. Bryan, F. L. 1980b. Poultry and poultry meat products. Pp. 410-458 in Microbial Ecology of Foods, Vol. II: Food Commodities, by the International Commission of Microbiological Specifications for Foods. Academic Press, New York. Butzler, J. P., and M. B. Skirrow. 1979. Campylobacter enteritis. Clin. Gastroenterol. 8:737-765. Campbell, D. F., S. S. Green, C. S. Custer, and R. W. Johnston. 1982. Incidence of Salmonella in fresh dressed turkeys raised under Salmonella-controlled and uncontrolled environments. Poult. Sci. 61:1962-1967. Campbell, D. F., R. W. Johnston, G. S. Campbell, D. McClain, and J. F. Macaluso. 1983. The microbiology of raw, eviscerated chickens: A ten year comparison. Poult. Sci. 62:437-444. Campbell, D. F., R. W. Johnston, M. W. Wheeler, K. V. Nagaraja, C. D. Szymansaki, and B. S. Pomeroy. 1984. Effects of the evisceration and cooling processes on the incidence of Salmonella in fresh dressed turkeys grown under Salmonella-controlled and uncontrolled environments. Poult. Sci. 63:1069-1072. Carter, G. R. 1962. Animal serotypes of Pasteurella multocida from human infections. Can. J. Public Health 53:158-161. Carter, G. R. 1972. Simpli£ied identification of somatic varieties of Pasteurella multocida causing fowl cholera. Avian Dis. 16:1109-1114.
88 .~ CDC (Center for Disease Control) . 1977. Salmonella Surveillance, Annual Summary 1976 . HEW Publ . No . (CDC) 78 - 8219 . Center for D-' sease Control, Public Health Service, UO S . Department of Health, Education, and Welfare, Atlanta. 23 ppe CDC (Center for Disease Control) ~ 1979 . Botulism i n the United States, 1899~1977. Handbook for Epidemiologists, Clinicians, and Laboratory Workers. Center for Disease Control, Public Health Service9 u.s'. Department of Health, Education, and Welfare, Atlanta. 41 pp. CDC (Centers for Disease Control). 1981. Salmonella Surveillance, Annual Summary 1978. HES Publ. No. (CDC) 81-8219. Centers for Disease Control, Public Health Service, U.S. Department of Health and Human Services, Atlanta. 25 ppO CDC (Centers for Disease Control). 1983a. Foodborne Disease Surveillance, Annual Summary 1980. HHS Publ. No. (CDC) 83-8185. Centers for Disease Control, Public Health Service, U.S. Department of Health and Human Services, Atlanta. 32 pp. CDC (Centers for Disease Control). 1983b. Foodborne Disease Surveillance, Annual Summary 1981. HHS Publ. No. (CDC) 83-8185. Centers for Disease Control, Public Health Service, U.S. Department of Health and Human Services, Atlanta. 41 pp. (Centers for Disease Control). 1984a. Annual summary 19830 Reported morbidity and mortality in the United States. Morbid. Mortal. Weekly Rep e 32 (annual suppl. ~ . CDC (Centers for Disease Control). 1984b. Notifiable diseases- Reported cases by geographic division and area, United States, 1983 . Morbid. Mortal . Weekly Rep @ 33: 542- 547 . CDC (Centers for Disease Control). 1985. Foodborne Disease Survey ance g Annual Summary 1982 . HHS Publ . No . (CDC) 85 - 8185 . Centers for Disease Control, Public Health Service, UOS. Department of Health and Human Services 9 Atlanta. 38 pp. CDSC (Communicable Disease Surveillance Centre). 1981. Review of Campylobacter reports to CDSC 1977-80. Communicable Dis. Rep. 12:3-4. CEC (Commission of European Communities). 1976. Evaluation of the Hygienic Problems Related to the Chilling of Poultry Carcasses. Information on Agriculture Series No. 22. European Economic Community, BrusselsO 110 pp. Chang, Pa We 1981. Newcastle disease. Pp. 261-274 in G. W. Beran, ed. CRC Handbook Series in Zoonoses, Section B: Viral Zoonoses, Vo1. II. CRC Press, Boca Raton, Fla.
89 Craven, P. C., W. B. Baine, D. C. Mackel, W. H. Barker, E. J. Gangarosa, M. Goldfield, H..Rosenfeld, R. Altman, G. Lachapelle, J. W. Davies, and R. C. Swanson. 1975. International outbreak of Salmonella eastbourne infection traced to contaminated chocolate. Lancet 1:788-793. Crittenden, L. B. 1976. The epidemiology of avian lymphoid leukosis. Cancer Res. 36:570-573. Current, W. L. 1985. Cryptosporidiosis . J . Am. Vet. Med. Assoc. 187:1334-1338. Damsker, B., and E. J . Bottone. 1985 . Mycobacterium avium-Mycobacterium intracellulare from the intestinal tracts of patients with the acquired immunodeficiency syndrome: Concepts regarding acquisition and pathogenes~s. J. Infect. Dis. 151:179-181. Dardiri , A. H., A. ~ . Yates , and T . D . Fl anagan. 1962 . The reaction to infection with the B1 strain of Newcastle disease virus in man. Am. J . Vet. Res . 23: 918- 921. de Groote, G., J e Vandepitte, and G. Wauters. 1982. Surveillance of human Yersinia enterocolit~ca infections in Belgium: 1963-1978. J. Infect. 4:189-197. De Wit , J . C ., G . Broekhuizen, and E . H . Kampelmacher . 1979. Cross - contamination during the preparation of frozen chickens in the kitchen. J . Hyg. 83: 27 - 32 . Durfee, P. T., M. M. Pullen, R. W. Currier II, and R. L. Parker. 1975. Human psittacosis associated with commercial processing ot turkeys . J . Am. Vet . Med. Assoc . 167: 804- 808 . Faddoul, G. P., and G. W. Fellows . 1966 . A five-year survey of the incidence of salmonellae in avian spec~es . Avian Dis . 10: 296 - 304 . Fagerberg, D. J ., and C . L. Quarles . 1982 . Final Report, Phases I, II, III, 1978-1981. Data Base for Drug Resistant Bacteria for Animals. Prepared for the U.S. Food and Drug Administration. FDA Contract No. 223-77-7032. Animal Sciences Department, Colorado State University, Fort Collins, Colo. t1590 pp.] Fenner, F., B. R. McAuslan, C. A. Mims, 3. Sambrook, and D. O. White. 1974. Viral oncogenesis: RNA viruses. Pp. 508-542 in The Biology of Anima1 Viruses, 2nd ed. Academic Press, New York. Filstein, M. R., A. B. Ley, M. S. Vernon, K. A. Gaffney, and L. T. Glickman. 1981. Epidemic of psittacosis in a college of veterinary med~cine. J. Am. Vet. Med. Assoc. 179: 569-574.
go Finch, M. J.9 and L. W. Riley. 1984. CampYlobacter infections in the United States: Results of an 11-state surveillance. Arch. Intern. Med. 144:1610-1612. Gangarosa, E. J. 1978a. Epidemiology of Escherichia cold in the United States. JO Infect. Dis. 137:634-638. Gangarosa, E. J. 1978b. What have we learned from 15 years ot Salmonella surveillance? Pp. 9-28 in the Proceedings of the National Salmonellosis Seminar held in Washington, D.C., January 10-11, 1978. U.S. Animal Health Association, Richmond, Va. Gangarosa, E. J., and M. H. Merson. 1977. Epidemiologic assessment of the relevance of the so-called enteropathogenic serogroups of Escherichia cold in diarrhea. N. Engl. J. Med. 296:1210-1213. Gemski, P., J. R. Lazere, and T. Casey. 1980. Plasmid associated with pathogenicity and calcium dependency of Yersinia enterocolitica. Infect. Immun. 27:682-685. Ghosts A. C. 1978. Prevalence of Bacillus cereus in the faeces of healthy adults. J. Hyg. 80:233~236. Gilbert, R. J . 1979 . Bacillus cereus gastroenteritis . Pp . 495 - S18 in H. Riemann and F. L. Bryan, eds. Food-borne Infections and Intoxications 9 2nd ed. Academic Press, New York. Gilbert, R. J., and JO M. Parry. 1977. Serotypes of Bacillus cereus from outbreaks o£ food po~soning and from routine foods. J. Hyg. 78:69-74. Gilbert, R. J ., M. F. Stringer , and T . C . Peace . 1974. The survival and growth of Bacillus cereus in boiled and fried rice in relation to outbreaks of food poisoning. J. Hyg. 73:433-444. Glantz, P. J. 1971. Serotypes of Escherichia cold associated with colibacillos is in neonatal animals. Ann. N. Y. Acad ~ Sci . 176:67 ~ 79. Goepfert, J. M., W. M. Spira, and H. U. Kim. 1972. Bacillus cereus: Food poison~ng organism. A rev~ew. J. Milk Food Technol. 35:213-227. Good, R. C. 1985. Opportunistic pathogens in the genus Mycobacterium. Annul Rev. Microbiol. 39:347-369. Grant, I. H., N. Je Richardson, and V. D. Bokkenheuser. 1980. Broiler chickens as potential sources of CampYlobacter infections in humans. J. Clin. Microbiol. 11:508-510.
91 Green, S . S ., A. B . Moran, R. W. Johnston, P . Uhler , and J . Chiu. 1982. The incidence of Salmonella species and serotypes in young whole chicken carcasses in 1979 as compared with 1967. Poult. Sci. 61:288-293. Gross, R. J. 1983. Escherichia cold diarrhoea. J. Infect. 7:177-192. Guerrant, R. L. 1980. Yet another pathogenic mechanism for Escherichia cold diarrhea. N. Engl. J. Med. 302:113-115. Hariharan, H., and W. R. Mitchell. 1977. Type C botulism: The agent, host spectrum and environment. Vet. Bull. 47:95-103. Harris, N. V., N. S. Weiss, and C. M. Nolan. 1986a. The role of poultry and meats in the etiology of Campylobacter jejuni/coli enteritis . Am. J . Public Health 76: 407-411. Harris, N. V. , D. Thompson, D. C. Martin, and C. M. Nolan. 1986b. A survey of Campylobacter and other bacterial contaminants of pre-market chicken and retail poultry and meats, King County, Washington. Am. J. Public Health 76:401-406. Hauge, S. 1950. Matforgiftninger fremkalt av Bacillus cereus. Nord. Hyg . Tidskr . 31: 189 - 206 . Hauge, S. 1955. Food poisoning caused by aerobic spore-forming bacilli. J. Appl. Bacterial. 18:591-595. Hauschild, A. H. W., and F. L. Bryan. 1980. Estimate of cases of food- and waterborne illness in Canada and the United States. J. Food Protect. 43:435-440, 446. Hayflick, L., ed. 1969. The Mycoplasmatales and the L-Phase of Bacteria. Appleton-Century-Crofts, New York. 750 pp. Heesemann, J., B. Algermissen, and R. Laufs. 1984. Genetically manipulated virulence of Yersinia enterocolitica. Infect. Immun. 46:105-110. Hehlmann, R., D. Kufe, and S. Spiegelman. 1972. Viral-related RNA in Hodgkins' disease and other human lymphomas. Proc. Natl. Acad. Sci. U.S.A. 69:1727-1731. Helmuth, R., R. Stephan, C. Bunge, B. Hoog, A. Steinbeck, and E. Bulling. 1985. Epidemiology of virulence-associated plasmids and outer membrane protein patterns within seven common Salmonella serotypes. Infect. Immun. 48:175-182.
92 Hines, M. P ., P . M. Page , N . Hirschberg , and L. G . Maddry . 1957 Ornithosis and leptospirosis survey of a chicken and turkey processing plant and textile mill in North Carolina. Vet. Med. 52: 337 - 33S, 356 . Hook , E O W . 1985 . Salmonella species (including typhoid fever). Pp. 125601269 in G. L. Mandell, R. G. Douglas, Jr., and J. E. Bennett, eds. Principles and Practice of Infectious Diseases, 2nd ed. Wiley, New York. Hopkins, R. S., and A. S. Scott. 1983. Handling raw chicken as a source for sporadic Campylobacter j ej unit infections O J . Infect . Dis . 148:770. Hornick, R. B. 1983. Nontyphoidal salmonellosis. Pp. 655-661 in P. D. Hoeprich, ed. Infectious Diseases: A Modern Treatise of Infectious Processes, 3rd ed. Harper & Row, Philadelphia. Horwitz, M. A., and E. J. Gangarosa. 1976. Foodborne disease outbreaks traced to poultry, United States, 1966-l 974. J. Milk Food Technol . 39: 859 - 863 . Hughes, J . M. 9 J . R . Blumenthal, M. H . Merson, G. L. Lombard, V . R . Dowell, and E. J. Gangarosa. 1981. Clinical features of type-A and type-B food°borne botulism Ann. Intern. Med. 95:442-445. Hutchinson? D. N., H. C. Dawkins, and F. J. Bolton. 1983. Do campylobacters contaminate food preparation areas? Pp. 169-170 in A. D. Pearson, M. B. Skirrow, B. Rowe, J. R. Davies, and D. M. Jones, eds. C~mpylobacter II: Proceedings of the Second International Workshop on Campylobacter Infections held in Brussels, September 6-9, 1983. Public Health Laboratory Service, London. Hyams, J . S ., W. A. Durbin, R. J O Grand, and D. A. Gol dmann. 1980. Salmonella bacteremia in the first year of life J. Pediatre 96:57 - 59 . Inoue, Y. K. 1973 . Avian infectious laryngotracheitis and S .M. O .N . in Japan. Lancet 1:1009. Inoue, Y. 1975. An avian-related new herpesvirus infection in man--subacute myelo-optico-neuropathy (SMON). Proge Med. Virol. 21: 32-42 ~ Inoue, Y. K., and YO Nishibe. 1973. Serological relationship between S . M. O . N . virus and avian infectious laryngotracheitis virus. Lancet 1: 776 - 777 .
93 Ionescu, G ., C . Ienistea, and C . Ionescu. 1966 . Freeventa B . cereus in laptele crud si in laptele pasteurizat. Microbiol., Parazitol., Epidemiol. 11: 423-430 . Kaplan, M. M. 1980. Some epidemiological and virological relationships between human and animal influenza. Comp. Immunol Microbiol . Infect. Dis . 3: 19 - 24 0 Kay, B. A., K. Wachsmuth, P. Gemski, J. C. Feeley, T. J. Quan, and D. J. Brenner. 1983. Virulence and phenotypic characterization of Yersinia enterocolitica isolated from humans in the United States O J. Clin. Microbiol. 17:128-138. Kendall, E. J. C., and E. I. Tanner. 1982. in general practice. J. Hyg. 88:155-163. Campylobacter enteritis Kinde, H., C. A. Genigeorgis, and M. Pappaioanou. 1983. Prevalence of Campylobacter iejuni in chicken wings. Appl. Environ. Microbiol. 45: 1116 - 1118 . Kist, M. 1983. Campylobacter enteritis in an industrial country: Epidemiologic features in urban and rural areas. P. 140 in A. D. Pearson, M. B. Skirrow, B. Rowe, J. R. Dav~es, and D. M. Jones, eds. Campylobacter II: Proceedings of the Second International Workshop on Campylobacter Infections held in Brussels, September 6-9, 19830 Public Health Laboratory Service, London. Klipstein, F. A., R. F. Engert, H. Short, and E. A. Schenk. 1985. Pathogenic properties of Campylobacter j ejuni: Assay and correlation with clinical manifestations. Infect. Immun. 50:43-490 Kuritsky, J . N., G . P . Schmid, M. E. Potter , D. C . Anderson, and A. F. Kaufmann. 1984. Psittacosis--a diagnostic challenge. J. Occup. Med. 26:731-733. Larsen, J. H. 1980. Yersinia enterocolitica infections and rheumatic diseases. Scand. J. Rheumatol. 9:129-137. Lee, W. H. 1977. An assessment of Yersinia enterocolitica and its presence in foods. J. Food Protect. 40:486-489. Levine, M. M. 19850 Escherichia cold infections. N. Engl. J. Med. 313:445-447. Linton, A. H., ed. 1983. Guidelines on Prevention and Control of Salmonellosis. World Health Organization, Geneva. 128 pp. MacCready, R. A., J. P. Reardon, and I. Saphra. 1957. Salmonellosis in Massachusetts; a sixteen-year experience. N. Engl. J. Med. 256:1121-1128.
94 Makari, J. G. 1973. Association between Marek's herpesvirus and human cancer . I . Detection of cross - reacting antigens between chicken tumors and human tumors. Oncology 28:177-183. Mandal, B. K. 9 and V. Mani. 1976. Colonic involvement in salmonellosis. Lancet 1: 887 - 888 . Marks, M. I., C. H. Pai, L. Lafleur, L. Lackman, and 0. H~mmerberg. 1980. Yers~nia enterocolitica gastroenteritis: A prospective study of clinical, bacteriologic, and epidemiologic features. J . Pediatr 0 9 6: 26 - 31. Marymont, J. H., Jr., K. K. Durfee, H. Alexander, and J. P. Smith. 1982. Yersinia enterocolitica in Kansas: Attempted recovery from 1, 212 patients. Am. J. Clin. Pathol. 77: 753-755. Mathan , V . I ., I) . P . Raj an , F . A . Kl Ups te in , and R . F . Engert . 1 9 84 . Enterotoxigenic Campylobacter j ej uni among children in South India. Lancet 2:9810 McCullough, N. B ., and C . W. E:isele . 1951. Experimental human salmonellosis; pathogenicity of strains of Salmonella newport, Salmonella derby and Salmonella bareilly obtained from spray- dried whole egg. J. Infect. Dis. 89: 209-213. McGaughey, C . A. 1932 . Organisms of the B O influenzas group in fowls O J. Comp. Pathol. Ther. 45: 58-66 0 Meissner, G., and W. Anz. 1977. Sources of Mycobacterium avium complex infection resulting in human diseases. Am. Rev. Respir Dis. 116:1057-1064. Meyer, K. F., and B. Eddie. 1942. Spontaneous ornithosis (psittacosis) in chickens the cause of a human infection. Proc. Soc . Exp . Biol . Med. 49: 522- 525 . Midura, T., M. Gerber, R. Wood, and A. R. Leonard. 1970. Outbreak of food poisoning caused by Bacillus cereus. Public Health Rep. 85:45-48. Miller, L. T., and V. J. Yates. 1971. Reactions of human sera to avian adenovirues and Newcastle disease virus. Avian Dis 15:781-788. Morris, G. K., and J. CO Feeley. 1976. Yersinia enterocolitica: A review of its role in food hygiene. Bull. W.H.O. 54:79-85. Morris, J . G., Jr. 1981. Bacillus cereus food poisoning. Arch O Intern. Med. 141: 7110
as Morris, J. G., Jr., and R. E. Black. 1985. Cholera and other vibrioses in the United States. N. Engl. J. Med. 312:343-350. Morris, J. G., Jr., and C. L. Hatheway. 1983. Botulism. Pp. 1115-1124 in P. D . Hoeprich , ed. Infectious Diseases : A Modern Treatise of Infectious Processes, 3rd ed. Harper & Row, Philadelphia Murphy, B. R., and R. G. Webster. 1985. Influenza viruses. Pp. 1179-1239 in B. N. Fields, D. M. Knipe, R. M. Chanock, J. L. Melnick, B. Roizman, and R. E. Shope, eds. Virology. Raven Press, New York. Naito, M., K. Ono, S. Tanabe, T. Doi, and S. Kato. 1970. Detection in chicken and human sera of antibody against herpes type virus from a chicken with Marek's disease and EB virus demonstrated by the indirect immunofluorescence test. liken J. 13:205-212. Navin, T. R. 1985. Cryptosporidiosis in h~,rnans: Review of recent epidemiologic studies . Eur . J . Epidemiol . 1: 77 - 83 . Nerome , K., Y. Yoshioka , C . A. Torres , A. Oya , P . Bachmann, K. Ottis, and R. G. Webster. 1984. Persistence of Q strain of H2N2 influenza virus in avian species: Antigenic, biological and genetic analys is of avian and human H2N2 viruses . Arch . Clairol . 81: 239 - 250 . Nishimura , C ., and T . Tobe . 1973 . Antigenicity of S . M. O . N . - associated virus closely related to avian infectious laryngotracheitis virus. Lancet 2:1445. Norkrans, G., and A. Svedham. 1982. Epidemiological aspects of CampYlobacter jejuni enteritis. J. Hyg. 89 :163-170. NRC (National Research Council). 1969. An Evaluation of the Salmonella Problem. Report of the Committee on Salmonella, Division of Biology and Agriculture e National Academy of Sciences, Washington, D.C. 207 pp. NRC (National Research Council). 1985a. An Evaluation of the Role of Microbiological Criteria in Foods and Food Ingredients. Report of the Subcommittee on Microbiological Criteria, Committee on Food Protection, Food and Nutrition Board. National Academy Press, Washington, D.C. 436 pp. NRC (National Research Council). 1985b. Meat and Poultry Inspection: The Scientific Basis of the Nation's Program. Report of the Committee on the Scientific Basis of the Nation's Meat and Poultry Inspection Program, Food and Nutrition Board. National Academy Press, Washington, D.C. 209 pp.
96 Nygren, B. 1962. Phospholipase C-producing bacteria and food poisoning. An experimental study on Clostridium perfringens and Bacillus cereus. Acta Pathol. Microbial. Scand., Suppl. 160:1-88. Pai, C. H., S. Sorger, L. Lackman, R. E. Sinai, and M. I. Marks. 1979. C~pylobacter gastroenteritis in children. J. Pediatr. 94: 589 - 591. Portnoy , D . A., S . L. Moser ey, and S . Falkow. 1981 e Characterization of plasmids and plasmid-associated determinants of Yersinia enterocolitica pathogenesis. Infect. Immun. 31:775-782. Potter, M. E., and A. F. Kaufmann. 1979. Psittacosis in humans in the United States 9 1975-1977 . J . Infect. Dis . 140 :131-134. Purchase, H. G., and R. L. Witter. 1986. Public health concerns from human exposure to oncogenic avian herpesviruses. J. Am. Vet. Med. Assoc. 189: 1430-1436. Reagan, R. L., S. C. Chang, F. S. Yancey, and A. L. Brueckner. 1956. Isolation of Newcastle disease virus from man with confirmation by electron microscopy. J. Am. Vet. Med. Assoc. 129:79-80. Rechtman, J., L. Lapeyrie, and H. H. Mollaret. 1985. Contribution 1'ecologie de Yersinia enterocolitica et especes apparentees: Recherche dans des denrees alimentaires. Med. Mall Infect. 15:130-134. Richman, A. V., C. G. Aulisio, W. G. Jahnes 9 and N. M. Tauraso. 1972. Avian leukosis antibody response in individuals given chicken embryo derived vaccines. Proc. Soc. Exp. Biol. Med. 139: 235-237. Riley, L. W., and M. J. Finch. 198S. Results of the first year of national surveillance of C~mpylobacter infections in the United States. J. Infect. Dis. 151:956-959. Roizman9 B., and W. BattersonO 1985. Herpesviruses and their reclication. Pc. 497-525 in B. N. Fields, Do M. Knipe, R. M D Chanock, JO L. Melnick, B. Roizman, and Ro BO Shope, eus. Virology. Raven Press, New York. Rosenberg, M. L., E . J . Gangarosa , R. A. Pollard, O . Brolnitsky , and J. S. Marr. 1977. Shigella surveillance in the United states9 1975. J. Infect. Diso 136:458-460. Roth, F. K., and R. M. Dougherty. 1971. Search for group-specific antibodies of avian leukosis virus in human leukemic sera. J. Natl. Cancer Inst. 46:1357-1359.
97 Ruiz-Palacios, G. M., J. Torres, N. I. Torres, E. Esc~milla, B. R. Ruiz-Palacios, and J. Todays. 1983. Cholera-like enterotoxin produced by Campylobacter jejuni. Lancet 2:250-2S3. Saphra, I., and M. Wassermann. 1954. Salmonella cholerae suds: A clinical and epidemiological evaluation of 329 infections identified between 1940 and 1954 in the New York Salmonella Center. Am. J. Med. Sci. 228:525-533. Schaefer, W. B. 1968. Incidence of the serotypes of Mvcobacterium avium and atypical mycobacteria in human and animal diseases. Am. Rev. Respir. Dis. 97:18-23 Schiemann, D. A., and J. A. Devenish. 1982. Relationship of HeLa cell infectivity to biochemical, serological, and virulence characteristics of Yersinia enterocolitica. Infect. Immun. 35:497-506. Schiemann, D. A., and S. Toma. 1978. Isolation of Yersinia enterocolitica from raw milk. Appl. Environ. Microbiol. 35:54-58. Scott, A. B. 1981. Botulinum toxin injection of eye muscles to correct strabismus. Trans. Am. Ophthalmol. Soc. 79: 734-770. Sereny, B. 1955. Experimental Shigella keratoconjunctivitis ; a preliminary report. Acta Microbiol. Acad. Sci. Hung. 2: 293-296. Sharma, J. M., R. L. Witter, B. R. Burmester, and JO C. Landon. 1973. Public health implications of Marek's disease virus and herpesvirus of turkeys. Studies on human and subhuman primates. J . Natl. Cancer Inst. 51 :1123-1128 . Shayegani, M., D. Morse, I. DeForge, T. Root, L. M. Parsons, and P. S . Maupin. 1983. Microbiology of a major foodborne outbreak of gastroenteritis caused by Yersinia enterocolitica serogroup 0:8. J. Clin. Microbiol. 17: 35-40. Silverstolpe , L., U. Plazikowski , J . Kj ellander , and G. Vahlne. 1961. An epidemic among infants caused by Salmonella muenchen. J. Appl. Bacterial. 24:134-142. SKCDPH (Seattle-King County Department of Public Health) . 1984 . Surveillance of the Flow of Salmonella and Campylobacter in a Community. Prepared for the Bureau of Veterinary Medicine, U.S. Food and Drug Administration. Contract No. 223-81-7041. Communicable Disease Control Section, Seattle -King County Department of Public Health, Seattle. [250 pp. ~ Skirrow, M. B. 1977. Campylobacter enteritis: A "new" disease . Br . Med. J. 2:9-ll.
98 Smith, J. E. 1955. Studies on Pasteurella septica. I. The occurrence in the nose and tonsils of dogs. J. Comp. Pathol. 65:239-245. Smitherman, R. E., C. A. Genigeorgis, and T. B. Farver. 1984. Preliminary observations on the occurrence of Campylobacter Jejune at four California chicken ranches. J. Food Protect. 47:293-298. Snyder, J. D., E. Christenson, and R. A. Feldman. 1982. Human Yersinia enterocolitica infections in Wisconsin: Clinical, laboratory and epidemiologic features. Am. J. Med. 72:768-774. Solomon J. J., and H. G. Purchase. 1969. A search for avian leukosis virus and antiviral activity in the blood of leukemic and nonleukemic adults and children. J . Natl . Cancer Inst. 42:29 - 33 e Sours, H. E., and D. G. Smith. 1980. Outbreaks of foodborne disease in the United States, 1972-1978. J. Infect. Dis . 142 :122-125. Stengel, G. 1985. Yersinia enterocolitica. Occurrence and importance in foods O Fleischwirtschaft 65:1493-1495. Stern, N. J., S. S. Green, N. Thaker, D. J. Krout, and J. Chiu. 1984. Recovery of CampYlobacter jejuni from fresh and frozen meat and poultry collected at slaughter. J . Food Protect . 47: 372- 374 O Surkiewicz, B . F ., R. W. Johnston, A. B . Moran, and G . W. Krumm. 1969. A bacteriological survey of chicken eviscerating plantse Food Technol. 23:1066-1069. Szita, J., M. Kali, and B. Redey. 1973. Incidence of Yersinia enterocolitica infection in Hungary. Pp. 106-110 in A. Grumbach, ed. Contributions to Microbiology and Immunology, VolO 2r Karger, New York. Tacket, C. O., J. Ballard, N. Harr's, J. Allard, C. Nolan, T. Quan, and M. L. Cohen O 1985. An outbreak of Yersinia enterocolitica infections caused by contaminated tofu (soybean curd). Am. J. Ep idemiol . 121: 705 - 711 . Terranova, W., and P. A. Blake. 1978. Baci flus cereus food poisoning. N. Engl. J . Med. 298: 143-144. Thoen, C . O ., A . G . Karlson 9 and E . M. Himes . 1981 . Mycobacterial infections in animals. Rev. Infect. Dis. 3: 960-972. Todd, E. C. D. 1980. Poultry-associated foodborne disease--its occurrence, cost, sources and prevention. J. Food Protect. 43:129-139.
99 Turnbull, P. C. B., T. A. French, and E. G. Dowsett. 19770 Severe systemic and pyogenic infections with Bacillus cereus. Br. Med. J 1:1628-1629. Turnbull, P. C. B., K. Jorgensen, J. M. Kramer, R. J. Gilbert, and J. M. Parry. 1979. Severe clinical conditions associated with Bacillus _ereus and the apparent involvement of exotoxins. J. Clin. Pathol. 32:289-293. Vantrappen, G., K. Geboes, and E. Ponette. Med. Clin. North Am. 66:639-653. 1982. Yersinia enteritis. Walter, S. D. 1975. The distribution of Levin's measure of attributable risk. Biometrika 62:371-374. Waters, T. D., P. S. Anderson, Jr., G. W. Beebe, and R. W. Miller. 1972. Yellow fever vaccination, avian leukosis virus, and cancer risk in man. Science 177:76-77. . Wempe, J. M., C. A. Genigeorgis, T. B. Farver, and H. I. Yusufu. 1983. Prevalence of CampYlobacter jejuni in two California chicken processing plants. Appl. Environ. Microbial. 45:355-359. Wolinsky, E. 1979 . Nontuberculous mycobacteria and associated diseases O Am. Rev. Respir. Dis. 119:107-159. Wood, J. M., R. G. Webster, and V. F. Nettles. 1985. A/chicken/Pennsylvania/83 (HSN2) influenza virus. 29:198-207. Host range of Avian Dis.
