3.5.2 Epidemiological data summary and analysis

Of the 33 outbreak reports collected from the published literature and from unpublished data received by FAO and WHO following the call for data, 23 contained sufficient information on the number of people exposed, the number of people that became ill, and the number of organisms in the implicated food to enable calculation of a dose-response relationship. Of the 23 outbreaks, 3 were excluded because the immune status of the persons exposed could not be determined. The remaining 20 outbreaks comprise the database used to calculate a dose-response relationship.

Of the 20 outbreaks in the database, 11 occurred in Japan and 9 occurred in North America. Several serotypes were associated with the outbreaks, including Enteritidis (12 outbreaks), Typhimurium (3), and in single outbreaks, Heidelberg, Cubana, Infantis, Newport and Oranienburg. Several vehicles were implicated, including food (meat, eggs, dairy products and others), water, and a medical dye capsule (carmine dye).

Reports provided by the Ministry of Health and Welfare of Japan (1999) represent a valuable source of information on the real-world dose-response relationship and expand our database of Salmonella pathogenicity considerably. The data in these reports are generated as part of the epidemiological investigations that take place in Japan following any outbreak of foodborne illness. In accordance with a Japanese notification released on March 1997, large-scale cooking facilities that prepare more than 750 meals per day or more than 300 dishes of a single menu at a time are advised to save food for future possible analysis in the event of an outbreak. The notification is also applicable to smaller-scale kitchens with social responsibility, such as those in schools, day care centres and other child-welfare and social-welfare facilities. Thus, 50-g portions of each raw food ingredient and each cooked dish are saved for more than 2 weeks at a temperature below -20ºC. Although this notification is not mandatory, the level of compliance is high, and some of the local governments in Japan also have local regulations that require food saving, but the duration and the storage temperature requirements vary.

The doses, attack rates, serovars and characteristics of the exposed populations derived from the outbreak reports described in the preceding section are summarized in Table 3.14 and Figure 3.10. The analysis of the epidemiological data was intended to serve three purposes:

[1] To determine if there is any epidemiological evidence for greater attack rates in susceptible vs normal populations.

[2] To determine if there is any epidemiological evidence for different attack rates for S. Enteritidis compared with other Salmonella serotypes.

[3] To compare the epidemiological data for dose and attack rate with the estimates generated by the dose-response models.

Table 3.14. Summary of outbreak data.

NOTES: (1) Popn. = population exposed, where N = Normal population and S = Susceptible population. (2) Expected value based on defined uncertainty ranges and distributions.

Figure 3.10. Summary of epidemiological data. Legend numbers indicate outbreak number given in text.

The data in Table 3.14 and Figure 3.10 are coded according to the outbreak number assigned in this document. If additional information related to a specific data point is required, for example the assignment of two data points, the details of the outbreak can be referred to in the previous section. The related assumptions for inclusion, exclusion or multiple data points are certainly issues for discussion and debate, and therefore included in the summary of reported outbreaks.

The data shown in Figure 3.10 appear to reflect our theoretical assumptions regarding the increasing trend in attack rates as dose increases. In addition, although there is a degree of clustering in some of the data points, a dose-response relationship is visually evident.

As noted earlier, some data were excluded from this summary and further analysis. For example, outbreaks numbers 27, 28 and 29 were attributed to S. Enteritidis in a hospital setting, where the exposed population would be expected to be more susceptible. The characteristics of the individuals that were exposed to the food is highly uncertain, so it may in fact be the case that the condition for which they were hospitalized is such that their immunity was not compromised. However, even if they are assumed to have normal susceptibility, these outbreaks were still distinctly different from outbreaks with a similar dose level, if the reported exposures were accurate. Alternative explanations for these data sets are that the individuals served the meal did not actually consume the implicated food, or that concurrent antibiotic therapy prevented the ingested Salmonella from colonization and illness production.

Figure 3.11. Attack rates corresponding to dose for "Normal" and "Susceptible" populations in reported outbreaks.

Susceptible vs Normal Populations

The observed outbreak data were used to gain some insight into the potential differences that may exist between "susceptible" and "normal" populations. The database of quantitative outbreak information collected during the course of this work includes several outbreaks that could be associated with "susceptible" and "normal" populations. Unfortunately, limited data allowed a comparison to be made based only on age. Susceptibility in this analysis was therefore limited to outbreak data for individuals less than 5 years old being classified as "susceptible", with other outbreak data representing a "normal" population. This was the case for all but one of the "susceptible" data points (estimated 85% attack rate, approximately 4.5 log dose), that occurred in a hospital and was attributed to carmine dye capsules. The "susceptible" and "normal" outbreak data were compared on the basis of reported attack rate corresponding with reported dose. Given the potential range in the observed data (dose and attack rate could vary based on the nature of the epidemiological investigation), the comparison was intended to look for overall trends first, and then, if necessary, additional analysis could be done. A plot of dose against attack rate for the "susceptible" and "normal" populations is shown in Figure 3.11.

Similarly, at other dose intervals there are outbreaks attributed to "normal" populations with attack rates either very similar to or higher than outbreaks involving "susceptible" populations. Given the data that currently exists from outbreaks, there is insufficient evidence to conclude that "susceptible" individuals, as defined in this database, have a higher probability of illness compared with the "normal" population.

It should be noted that, within the database of outbreaks, there are two outbreaks in which a "susceptible" and "normal" population were identified in the same outbreak with differing attack rates. The "susceptible" definition in these cases was again based on an age criteria (<5 years old and >5 years old). In these two outbreaks, shown in Figure 3.12, the attack rate was clearly reported to be higher for the susceptible population compared with the normal population. Taken in isolation, it could be concluded from this information that there is clearly a higher probability of illness for the susceptible population compared with the normal population. However, if we look at the whole picture, we can see other outbreaks involving a "normal" population with higher attack rates at similar doses.

Given the outbreak data that are currently available, it is not possible to conclude that some segments of the population are more susceptible to becoming ill upon exposure to Salmonella than are other segments. Furthermore, it is impossible to derive a quantitative estimate of the increased probability of illness for some segments of the population compared with others. The dose-response relationship for the probability of illness for different segments of the population was therefore assumed to be the same.

The key distinction that needs to be made in this conclusion is that the probability of illness is assumed indistinguishable, given the current data and the susceptible populations defined in the database. It is important to recognize that even if the probability of becoming ill, defined in the dose-response assessment as any degree of gastroenteritis, the severity of the illness may be markedly different for certain segments of the population. To quantify the probabilities of different outcomes, information is needed in the form of quantitative patient follow-up and data on physician visits, hospitalizations, death or other chronic outcomes.

Figure 3.12. Attack rates for two outbreaks in which different populations in the same outbreak were identified.

S. Enteritidis vs other Salmonella serovars

In a similar manner to the comparisons made for susceptible and normal populations, the attack rates in outbreaks associated with S. Enteritidis were compared with outbreaks associated with other Salmonella serovars. This information is summarized in Figure 3.13.

The attack rates observed in outbreaks associated with other Salmonella serovars are indistinguishable from outbreaks associated with S. Enteritidis. At some dose ranges, the highest attack rate reported is for S. Enteritidis, while at others the highest attack rate is for other serovars. Based on this information, S. Enteritidis and other serovars were treated as equivalent for the purposes of the dose-response relationship. It is acknowledged, however, that less virulent strains may infrequently be the cause of foodborne outbreaks and hence would not be captured in this database.

In summary, it was concluded that for the purposes of the current assessment and based upon the existing observed evidence:

(1) a single dose-response relationship for the probability of illness would be used for all members of the population; and

(2) S. Enteritidis and other Salmonella serovars are assumed to have a similar probability of initiating illness at the same dose.

Figure 3.13. Attack rates corresponding to dose for S. Enteritidis and other Salmonella in reported outbreaks.

Comparison of outbreak data with existing Salmonella dose-response models

Three dose-response models for Salmonella exist in the literature. The first (Fazil, 1996) is the beta-Poisson model (Haas, 1983) fitted to the human feeding trial data for Salmonella infection (McCullough & Eisele, 1951a, c, d). The second model was proposed in the US SE RA (USDA-FSIS, 1998) and was based on the use of a surrogate pathogen to describe the dose-response relationship. This model assumed a shift in the dose-response model for "susceptible" and "normal" populations. The third model was introduced in a Salmonella Enteritidis risk assessment done by Health Canada (Health Canada, 2000, but unpublished), which was based on a Weibull dose-response relationship that was updated to reflect selected outbreak information using Bayesian techniques. Similar to the US SE RA model, this one also assumed a higher probability of illness for susceptible populations. The models and their comparison with the outbreak data are shown in Figures 3.14 to 3.16, and discussed in the following sections.

Naive human feeding trial data (beta-Poisson model)

The model suffers from the nature of the feeding trial data (i.e. the subjects used were healthy male volunteers) and may not reflect the population at large. The model also tends to greatly underestimate the probability of illness as observed in the outbreak data (Figure 3.14), even under the extremely conservative assumption that infection, as measured in the dose-response curve, equates to illness.

Figure 3.14. Beta-Poisson dose-response model fitted to naive human feeding trial data compared with reported outbreak data.

Figure 3.15. US SE RA dose-response model compared with reported outbreak data.

Figure 3.16. Health Canada Salmonella Enteritidis dose-response model compared with reported outbreak data.

US SE RA (beta-Poisson model)

The model uses human feeding trial data for Shigella dysenteriae as a surrogate pathogen, with illness as the measured endpoint in the data. The appropriateness of using Shigella as a surrogate for Salmonella is questionable given the nature of the organisms in relation to infectivity and disease. Compared with the outbreak data (Figure 3.15), and on a purely empirical basis, this curve tends to capture the upper range of the data, but overestimates the probability of illness that is observed in the outbreak data.

Health Canada Salmonella Enteritidis (Weibull-Gamma model)

To date, this model has not been fully documented and lacks transparency. The model uses data from many different bacterial-pathogen-feeding trials and combines this information with key Salmonella outbreak data using Bayesian techniques. Using data from many bacterial-feeding trials and the current lack of transparency regarding their influence is a point of caution. Empirically, the curve describes the outbreak data (Figure 3.16) at the low dose well but tends towards the lower range of response at higher doses.

Dose-response model based on outbreak data

The availability of a reasonably large data set representing real-world observations for the probability of illness upon exposure to Salmonella (outbreak data) allowed a unique opportunity to attempt to develop a dose-response relationship based upon this data. The beta-Poisson model (Equation 1) was used as the mathematical form for the relationship, and this was fitted to the outbreak data.

The maximum likelihood technique was used as the basis for generating the best fitting curve to the data. The fit was optimized using an iterative technique that minimized the deviance statistic, based upon a binomial assumption (Haas, 1983).

The outbreak data have merits as real-world observations of the probability of illness upon exposure to a dose, but there are also some drawbacks in the data. Specifically, it should be recognized that there is a degree of uncertainty in the outbreak data, primarily due to the uncontrolled settings under which the information and data were collected. In some cases, the actual dose ingested can be uncertain, while in other cases the true number of people exposed or ill during the outbreak can be under- or over-estimated.

The uncertainty in the outbreak data set was incorporated into the fitting routine by reviewing the outbreak information and assigning an uncertainty distribution on observed variables that were potentially uncertain. A detailed summary of the assumptions associated with each outbreak and the estimation for the range of uncertainty for each of the variables were described in Section 3.2.2. A summary of the data set, with uncertainty for the variables, is given in Table 3.15.

Table 3.15. Uncertainty ranges assigned to variables in reported outbreak data

Figure 3.17. Dose-response curves generated by fitting to samples from uncertain outbreak observations.

In order to fit the dose-response model to the uncertain outbreak data, the data were re-sampled based on the uncertainty distributions, generating a new data set at each sample. The dose-response model was then fitted to each of the re-sampled data sets. This procedure was repeated approximately 5000 times, generating 5000 dose-response data sets, to which 5000 dose-response curves were fitted. The fitting procedure used (Haas, 1983) places a greater emphasis on fitting the curve through the larger-scale outbreaks compared with the smaller outbreaks. This is primarily a result of the binomial assumption and the greater variance associated with data from a small observation compared with a large one. Figure 3.17 shows an example of the dose-response curves that are generated by fitting to the uncertain data. The observed outbreak data were found to be over-dispersed compared with what would be expected from the binomial assumption inherent in the deviance statistic that is minimized during fitting. As a result, it was not possible to get a statistically significant single "best fitting" curve to the expected value of all the outbreak data points. However, the characterization of the observed outbreak data by the fitted dose-response model was better than that of the other dose-response models described previously. It is important to note that the range of possible responses at any one given dose shown in Figure 3.17 do not represent the statistical confidence bounds of the dose-response fit, but rather the best fit of the beta-Poisson model to different realizations of the observed data, given its uncertainties.

Figure 3.18 shows the comparison between the fitted curves and the expected value for the observed data. The upper bound, lower bound, expected value, 97.5^th percentile and 2.5^th percentile for the dose-response curves fitted to the 5000 data sets are also shown. The fitted dose-response range captures the observed outbreak data quite well, especially at the lower- and mid-dose range. The greater range at the high doses is due to the existence of several large-scale outbreaks at the lower- and mid-dose levels through which the curves attempt to pass, while the two high-dose data points are for relatively small-scale outbreaks that allow greater "elasticity" in the fit.

Figure 3.18. Uncertainty bounds for dose-response curves, compared with expected value for the outbreak data.

Since the fitting procedure generated a dose-response curve for each of the 5000 data sets, there are also 5000 sets of beta-Poisson dose-response parameters (alpha & beta). In order to apply the dose-response relationship in a risk assessment, the ideal approach would be to randomly sample from the set of parameters that are generated, thereby recreating the dose-response curves shown in Figures 3.17 and 3.18. As an alternative, it is also possible to use the upper, lower, expected value, 2.5^th percentile or 97.5^th percentile to represent the uncertainty ranges in the dose-response relationship, as opposed to a full characterization resulting from the sampling of the parameter sets. The parameters that generate dose-response curves that approximate the bounds shown in Figure 3.18 of the dose-response relationship are summarized in Table 3.16.

Table 3.16. Beta-Poisson dose-response parameters that generate the approximate bounds shown in Figure 3.18.

Figure 3.19 summarizes all the dose-response models described so far, as well as the outbreak data. It also highlights the expected result of a better characterization of the outbreak data using the current model compared with the alternatives.

Figure 3.19. Comparison of all dose-response models with reported outbreak data.

In dose-response analysis, the critical region is the lower-dose region. These are the doses that are most likely to exist in the real world and this is also the region for which experimental data are mostly non-existent. The outbreak data extend to a much lower dose than is common in experimental feeding trials, and as such may offer a greater degree of confidence in the lower dose approximations generated by the outbreak dose-response model. Table 3.17 and Figures 3.20 and 3.21 summarize the low-dose estimates for the various dose-response models.

Table 3.17. Probability of illness, estimated by alternative dose-response models at selected low-mean-dose values.

Figure 3.20. Comparison of alternative dose-response models in the 0 to 2.0 mean log dose interval

Figure 3.21. Comparison of alternative dose-response models in the -1.0 to 1.0 mean log dose interval.

There is a wide range of estimates generated by the dose-response models. At a dose of 1000 cells, the US SE RA model for the normal population estimates a 65% probability of illness, and an 81% probability for the susceptible population. The Health Canada S. Enteritidis model estimates a 31% probability for susceptible populations and 19% for normal populations, while the outbreak model estimates a probability of 33%. At a dose of 100 cells, the US SE RA model continues to be the most conservative, with estimates ranging from 37% to 64%, while the outbreak model estimates a probability of 13%, lying within the range (10-17%) estimated by the Health Canada S. Enteritidis model. Perhaps the most telling feature of low-dose estimates is the probability of illness estimated by the models upon ingestion of 1 cell. The US SE RA and Health Canada S. Enteritidis models for susceptible populations estimate 9% and 5% probabilities respectively. In the case of the normal population, the Health Canada S. Enteritidis model estimates a higher probability (2.7%) than the USDA model (1.1%). The outbreak model estimates the probability at 0.24%, approximately an order of magnitude lower than the Health Canada model for normal populations.

In conclusion, the dose-response model based upon the observed outbreak data provides an estimate for the probability of illness that is based on real-world data. Given the assumptions associated with some of the other models - surrogate pathogens; infection response with healthy male volunteers; and lack of transparency with non-linear low-dose extrapolation - the outbreak model offers the best current alternative for estimating the probability of illness upon ingestion of a dose of Salmonella.

3.6 Discussion and conclusions

It has been postulated that some strains of S. Enteritidis, particularly the phage types isolated from the increased number of egg-related outbreaks seen in recent years, may be more virulent than other serovars of Salmonella. From the outbreak data used to examine the dose-response relationship, there was no evidence that the likelihood of S. Enteritidis producing illness differed from other serovars. In total, 12 sets of data were evaluated for S. Enteritidis, against 8 sets of data for other serovars. However, increased severity of illness once infected was not evaluated.

It was concluded that there is insufficient evidence in the current outbreak database to conclude that some segments of the population have a higher probability of illness compared with others. There was some indication in two instances, in which two populations potentially exposed to Salmonella in the same outbreak exhibited different attack rates. There is therefore a possibility that the probability of illness upon exposure may be different for some members of the population compared with others. However, in the absence of additional information, the probability of illness could be assumed the same for all members of the population, although the severity of the illness could be potentially different.

This document did not consider a quantitative evaluation of secondary transmission (person-to-person) or chronic outcomes. In addition, the impact of the food matrix was not incorporated into the assessment. These may be considerations for future document development.

The dose-response model fitted to the outbreak data offers a reasonable estimate for the probability of illness upon ingestion of a dose of Salmonella. The model is based on observed real-world data, and as such is not subject to some of the flaws inherent in using purely experimental data. Nevertheless, the current outbreak data also have uncertainties associated with them and some of the outbreak data points required assumptions to be made. Overall, the dose-response model generated in the current exercise can be used for risk assessment purposes, and generates estimates that are consistent with those that have been observed in outbreaks.

3.7 References cited in Chapter 3

Angulo, F.J., & Swerdlow, D.L. 1995. Bacterial enteric infections in persons infected with human immunodeficiency virus. Clinical Infectious Diseases, 21(Suppl. 1): S84-S93.

Angelotti, R., Bailey, G.C., Foter, M.J., & Lewis, K.H. 1961. Salmonella infantis isolated from ham in food poisoning incident. Public Health Reports, 76: 771-776.

Armstrong, R.W., Fodor, T., Curlin, G.T., Cohen, A.B., Morris, G.K., Martin, W.T., & Feldman, J. 1970. Epidemic salmonella gastroenteritis due to contaminated imitation ice cream. American Journal of Epidemiology, 91: 300-307.

Banatvala, N., Cramp, A., Jones, I.R., & Feldman, R.A. 1999. Salmonellosis in North Thames (East), UK: associated risk factors. Epidemiology and Infection, 122: 201-207.

Bearson, S., Bearson, B., & Foster, J.W. 1997. Acid stress responses in enterobacteria. FEMS Microbiology Letters, 147: 173-180.

Beck, M.D., Muñoz, J.A., Scrimshaw, N.S. 1957. Studies on the diarrheal diseases in Central America. I. Preliminary findings on the cultural surveys of normal population groups in Guatemala. American Journal of Tropical Medicine and Hygiene, 6: 62-71.

Bellido Blasco, J.B., Gonzalez Cano, J.M., Galiano, J.V., Bernat, S., Arnedo, A., & Gonzalez, M.F. 1998. Factores asociados con casos esporádicos de salmonelosis en niños de 1 a 7 años. Gaceta Sanitaria, 12: 118-125.

Bellido Blasco, J.B., Gonzalez, M.F., Arnedo, P.A., Galiano Arlandis, J.V., Safont, A.L., Herrero, C.C., Criado, J.J., & Mesanza, D.N. 1996. I. Brote de infección alimentaria por Salmonella enteritidis. Posible efecto protector de las bebidas alcohólica. Medicina Cliníca (Barcelona), 107: 641-644.

Blaser, M.J., & Feldman, R.A. 1981. Salmonella bacteraemia: reports to the Centers for Disease Control, 1968-1979. Journal of Infectious Diseases, 143: 743-746.

Blaser, M.J., & Newman, L.S. 1982. A review of human salmonellosis: I. Infective dose. Reviews of Infectious Diseases, 4: 1096-1106.

Boring, J.R., III, Martin, W.T., & Elliott, L.M. 1971. Isolation of Salmonella typhimurium from municipal water, Riverside, California, 1965. American Journal of Epidemiology, 93: 49-54.

Bruch, H.A., Ascoli, W., Scrimshaw, N.S., & Gordon, J.E. 1963. Studies of diarrheal disease in Central America. V. Environmental factors in the origin and transmission of acute diarrheal disease in four Guatemalan villages. American Journal of Tropical Medicine and Hygiene, 12: 567-579.

Buzby, J.C. 2001. Children and microbial foodborne illness. Food Review, 24(2): 34-37.

Cash, R.A., Music, S.I., Libonati, J.P., Snyder, M.J.J., Wenzel, R.P., & Hornick, R.B. 1974. Response of man to infection with Vibrio cholerae. I. Clinical, serologic, and bacteriologic responses to a known inoculum. Journal of Infectious Diseases, 129: 45-52.

Chiu, C.H., Lin, T.Y., & Ou, J.T. 1999. Prevalence of the virulence plasmids of nontyphoid Salmonella in the serovars isolated from humans and their association with bacteremia. Microbiology and Immunology, 43: 899-903.

Coleman, M., & Marks, H. 1998. Topics in Dose-Response Modelling. Journal of Food Protection, 61: 1550-1559.

Cowden, J.M., & Noah, N.D. 1989. Annotation: Salmonellas and eggs. Archives of Disease in Childhood, 64: 1419-1420.

Craven, P.C., Mackel, D.C., Baine, W.B., Barker, W.H., Gangarosa, E.J., Goldfield, M., Rosenfeld, H., Altman, R., Lachapelle, G., Davies, J., & Swanson, R. 1975. International outbreak of S. eastbourne infection traced to contaminated chocolate. The Lancet, 5 April 1975: 788-792.

D’Aoust, J.Y. 1985. Infective dose of Salmonella typhimurium in cheddar cheese. American Journal of Epidemiology, 122: 717-720.

D’Aoust, J.Y. 1991. Pathogenicity of foodborne Salmonella. International Journal of Food Microbiology, 12: 17-40.

D’Aoust, J.Y. 1997. Salmonella species. In: M.P. Doyle, L.R. Beuchat and T.J. Montville (eds). Food microbiology: Fundamentals and frontiers. Washington, DC: American Society for Microbiology Press.

D’Aoust, J.Y., Warburton, D.W., & Sewell, A.M. 1985. S. typhimurium phage type 10 from cheddar cheese implicated in a major Canadian foodborne outbreak. Journal of Food Protection, 48: 1062-1066.

D’Aoust, J.Y., Aris, B.J., Thisdele, P., Durante, A., Brisson, N., Dragon, D., Lachapelle, G., Johnston, M., & Laidely, R. 1975. S. eastbourne outbreak associated with chocolate. Journal Institut Canadien de Technologie Alimentaire, 8: 181-184.

Davis, R.C. 1981. Salmonella sepsis in infancy. American Journal of Diseases of Children, 135: 1096-1099.

Delarocque-Astagneau, E., Desenclos, J.C., Bouvet, P., Grimont, P.A. 1998. Risk factors for the occurrence of sporadic Salmonella enterica serotype enteritidis infections in children in France: a national case-control study. Epidemiology and Infection, 121: 561-567.

Fazil, A.M. 1996. A quantitative risk assessment model for salmonella. Drexel University, Philadelphia PA. [Dissertation].

Fontaine, R.E., Arnon, S., Martin, W.T., Vernon, T.M. Jr, Gangarosa, E.J., Farmer, J.J. III, Moran, A.B., Silliker, J.H., & Decker, D.L. 1978. Raw hamburger: an interstate common source of human salmonellosis. American Journal of Epidemiology, 107: 36-45.

Fontaine, R.E., Cohen, M.L., Martin, W.T., Vernon, T.M. 1980. Epidemic salmonellosis from cheddar cheese: Surveillance and Prevention. American Journal of Epidemiology, 111: 247-253.

George, R.H. Small infectous dose of Salmonella. The Lancet, 24 May 1976: 1130.

Glynn, J.R., & Palmer, S.R. 1992. Incubation period, severity of disease, and infecting dose: evidence from a Salmonella outbreak. American Journal of Epidemiology, 136: 1369-1377.

Greenwood, M.H., & Hooper, W.L. 1983. Chocolate bars contaminated with S. napoli: an infectivity study. British Medical Journal, 286: 1394.

Haas, C.N. 1983. Estimation of risk due to low doses of microorganisms: a comparison of alternative methodologies. American Journal of Epidemiology, 118: 573-582.

Haas, C.N., Rose, J.B., Gerba, C., Regli, S. 1993. Risk assessment of virus in drinking water. Risk Analysis, 13: 545-552.

Health Canada. [2000]. Risk assessment model for Salmonella Enteritidis. Unpublished Health Canada document.

Hennessy, T.W., Hedberg, C.W., Slutsker, L., White, K.E., Besser-Wiek, J.M., Moen, M.E., Feldman, J., Coleman, W.W., Edmonson, L.M., MacDonald, K.L., & Osterholm, M.T. 1996. A national outbreak of Salmonella enteritidis infections from ice cream. New England Journal of Medicine, 334: 1281-1286.

Hormaeche, E., Peluffo, C.A., & Aleppo, P.L. 1939. Nuevo contrabucion al estudio etiologico de las "Diarreas infantiles de Verano". Archivas Uruguayos de Medicina, Cirurgia y Especialiadades, 9: 113-162.

Hornick, R.B., Geisman, S.E., Woodward, T.E., DuPont, H.L., Dawkins, A.T., & Snyder, M.J. 1970. Typhoid fever: pathogenesis and immunologic control. New England Journal of Medicine, 283: 686-691.

Jaykus, L.A., Morales, R.A., & Cowen, P. 1997. Development of a risk assessment model for the evaluation of HACCP-based quality assurance programs for human infection from Salmonella Enteritidis: Preliminary estimates. Epidemiolgie et Santé Animale, 31-32: 06.11.

Kapperud, G., Lassen, J., & Hasseltvedt, V. 1998. Salmonella infections in Norway: descriptive epidemiology and a case-control study. Epidemiology and Infection, 121: 569-577.

Kapperud, G., Stenwig, H., & Lassen, J. 1998. Epidemiology of Salmonella typhimurium O:4-12 infection in Norway: evidence of transmission from an avian wildlife reservoir. American Journal of Epidemiology, 147: 774-782.

Kass, P.H., Farver, T.B., Beaumont, J.J., Genigeorgis, C., & Stevens, F. 1992. Disease determinants of sporadic salmonellosis in four northern California counties. A case-control study of older children and adults. Annuals of Epidemiology, 2: 683-696.

Khan, M.A. 1995. HLA-B27 and its subtypes in world populations. Current Opinion in Rheumatology, 7: 263-269.

Khan, M.A. 1996. Epidemiology of HLA-B27 and arthritis. Clinical Rheumatology, 15 (Suppl. 1): 10-12.

Kourany, M., & Vasquez, M.A. 1969. Housing and certain socioenvironmental factors and prevalence of enteropathogenic bacteria among infants with diarrheal disease in Panama. American Journal of Tropical Medicine and Hygiene, 18: 936-941.

Lang, D.J., Kunz, L.J., Martin, A.R., Schroeder, S.A., & Thomson, L.A. 1967. Carmine as a source of noscomial salmonellosis. New England Journal of Medicine, 276: 829-832.

Le Bacq, F., Louwagie, B., Verhaegen, J. 1994. Salmonella typhimurium and Salmonella enteritidis: changing epidemiology from 1973 until 1992. European Journal of Epidemiology, 10: 367-371.

Lee, L.A., Puhr, N.D., Maloney, E.K., Bean, N.H., & Tauxe, R.V. 1994. Increase in antimicrobial-resistant Salmonella infections in the United States, 1989-1990. Journal of Infectious Diseases, 170: 128-134.

Levine, M.M., & DuPont, H.L. 1973. Pathogenesis of Shigella dysenteriae 1 (Shiga) dysentery. Journal of Infectious Diseases, 127: 261-270.

Levy, M., Fletcher, M., Moody, M., and 27 others. 1996. Outbreaks of Salmonella serotype enteritidis infection associated with consumption of raw shell eggs - United States, 1994-1995. Morbidity and Mortality Weekly Report, 45: 737-742.

Lipson, A. 1976. Infecting dose of salmonella. The Lancet, 1 May 1976: 969.

Mackenzie, C.R., & Livingstone, D.J. 1968. Salmonellae in fish and foods. South African Medical Journal, 42: 999-1003.

McCullough, N.B., & Eisele, C.W. 1951a. Experimental human salmonellosis. I. Pathogenicity of strains of Salmonella Meleagridis and Salmonella anatum obtained from spray-dried whole egg. Journal of Infectious Diseases, 88: 278-289.

McCullough, N.B., & Eisele, C.W. 1951b. Experimental human salmonellosis. II. Immunity studies following experimental illness with Salmonella Meleagridis and Salmonella anatum. Journal of Immunology, 66: 595-608.

McCullough, N.B., & Eisele, C.W. 1951c. Experimental human salmonellosis. III. Pathogenicity of strains of Salmonella Newport, Salmonella derby, and Salmonella Bareilly obtained from spray dried whole egg. Journal of Infectious Diseases, 89: 209-213.

McCullough, N.B., & Eisele, C.W. 1951d. Experimental human salmonellosis. IV. Pathogenicity of strains of Salmonella pullorum obtained from spray-dried whole egg. Journal of Infectious Diseases, 89: 259-266.

Ministry of Health and Welfare, Japan. 1999. Research on micriobiological risk assessment. Prepared by Susumu Kumagai and Shigeki Yamamoto, in: Report of Grants for Health Science of the Ministry of Health and Welfare, Japan.

Morales, R.A., Jaykus, L.A., & Cowen, P. 1995. Preliminary risk assessments for the transmission of Salmonella Enteritidis in shell eggs. Abstract. Proceedings 76^th Conference of Research Workers in Animal Diseases. Chicago, Illinois, 13-14 November 1995.

Mossel, D.A., & Oei, H.Y. 1975. Letter: Person to-person transmission of enteric bacterial infection. Lancet, 29 March 1975 1: 751.

Murray, M.J. 1986. Salmonella: virulence factors and enteric salmonellosis. Journal of the American Veterinary Medical Association, 189: 145-147.

Narain, J.P., & Lofgren, J.P. 1989. Epidemic of restaurant-associated illness due to S. Newport. Southern Medical Journal, 82: 837-840.

Nguyen, B.M., Lanata, C.F., Black, R.E., Gil, A.I., Karnell, A., & Wretlind, B. 1998. Age-related prevalence of Shigella and Salmonella antibodies and their association with diarrhoeal diseases in Peruvian children. Scandinavian Journal of Infectious Diseases, 30: 159-164.

Olsen, S.J., Bishop, R., Brenner, F.W., Roels, T.H., Bean, N., Tauxe, R.V., & Slutsker, L. 2001. The changing epidemiology of Salmonella: Trends in serotypes isolated from humans in the United States, 1987-1997. The Journal of Infectious Diseases, 183: 753-761.

Pavia, A.T., Shipman, L.D., Wells, J.G., Puhr, N.D., Smith, J.D., McKinley, T.W., & Tauxe, R.V. 1990. Epidemiologic evidence that prior antimicrobial exposure decreases resistance to infection by antimicrobial-sensitive Salmonella. Journal of Infectious Diseases, 161: 255-260.

Regli, S., Rose, J.B., Haas, C.N., Gerba, C. 1991. Modelling risk for pathogens in drinking water. Journal American Water Works Association, 83: 76-84.

Reitler, R., Yarom, D., & Seligmann, R. 1960. The enhancing effect of staphylococcal enterotoxin on Salmonella infection. The Medical Officer, 104: 181.

Riley, L.W., Cohen, M.L., Seals, J.E., Blaser, M.J., Birkness, K.A., Hargrett, N.T., Martin, S.M., Feldman, R.A. 1984. Importance of host factors in human salmonellosis caused by multiresistant strains of Salmonella. Journal of Infectious Diseases, 149: 878-883.

Rejnmark, L., Stoustrup, O., Christensen, I., & Hansen, A. 1997. Impact of infecting dose on severity of disease in an outbreak of food-borne Salmonella Enteritidis. Scandinavian Journal of Infectious Diseases, 29: 37-40.

Ryder, R.W., Merson, M.H., Pollard, R.A., & Gangarosa, E.J. 1976. From the Center for Disease Control: salmonellosis in the United States, 1968-1974. Journal of Infectious Diseases, 133: 483-486.

Schliessmann, D.J., Atchley, F.O., Wilcomb, M.J., & Welch, S.F. 1958. Relation of environmental factors to the occurrence of enteric disease in areas of eastern Kentucky. Public Health Service Monographs, 54.

Schmid, H., Burnens, A.P., Baumgartner, A., & Oberreich, J. 1996. Risk factors for sporadic salmonellosis in Switzerland. European Journal of Clinical Microbiology and Infectious Diseases, 15: 725-732.

Slauch, J., Taylor, R., & Maloy, S. 1997. Survival in a cruel world: how Vibrio cholerae and Salmonella respond to an unwilling host. Genes and Development, 11: 1761-1774.

Smith, J. 2002. Campylobacter jejuni infection during pregnancy: Long-term consequences of associated bacteraemia, Guillain BarrJ Syndrome, and reactive arthritis. Journal of Food Protection, 65: 696-708.

Smith, P.D., Lane, H.C., Gill, V.J., Manischewitz, J.F., Quinnan, G.V., Fauci, A.S., & Masur, H. 1988. Intestinal infections in patients with the acquired immunodeficiency syndrome (AIDS). Etiology and response to therapy. Annals of Internal Medicine, 108: 328-333.

Sprinz, H., Gangarosa, E.J., Williams, M., Hornick, R.B., & Woodward, T.B. 1966. Histopathology of the upper small intestines in typhoid fever. American Journal of Digestive Diseases, 11: 615-624.

Sprong, R.C., Hulstein, M.F., & Van der Meer, M.R. 1999. High intake of milk fat inhibits intestinal colonization of Listeria but not of Salmonella in rats. Journal of Nutrition, 129: 1382-1389.

Tacconelli, E., Tumbarello, M., Ventura, G., Leone, F., Cauda, R., & Ortona, L. 1998. Risk factors, nutritional status, and quality of life in HIV-infected patients with enteric salmonellosis. Italian Journal of Gastroenterology and Hepatology, 30: 167-172.

Taylor, D.N., Bopp, C., Birkness, K., & Cohen, M.L. 1984. An outbreak of salmonellosis associated with a fatality in a healthy child. American Journal of Epidemiology, 119: 907-912.

Telzak, E.E., Greenberg, M.S., Budnick, L.D., Singh, T., & Blum, S. 1991. Diabetes mellitus: A newly described risk factor for infection from Salmonella enteritidis. Journal of Infectious Diseases, 164: 538-541.

Teunis, P.F.M., Van der Heijden, O.G., Van der Giessen, J.W.B., & Havelaar, A.H. 1996. The dose response relation in human volunteers for gastro-intestinal pathogens. National Institute of Public Health and the Environment, Bilthoven, The Netherlands. RIVM Report No. 284550002.

Travers, K., & Barza, M. 2002. Morbidity of infections caused by antimicrobial-resistant bacteria. Clinical Infectious Diseases, 34(Suppl. 3): S131-S134.

USDA-FSIS. 1998. Salmonella Enteritidis Risk Assessment. Shell Eggs and Egg Products. Final Report. Prepared for FSIS by the Salmonella Enteritidis Risk Assessment Team. 268 pp. Available on Internet as PDF document at: www.fsis.usda.gov/ophs/risk/contents.htm.

Varela, G., & Olarte, J. 1942. Infecion experimental del Hombre con Salmonella anatum. Medicina Revista Mexicana, 22: 57-58.

Vought, K.J., & Tatini, S.R. 1998. S. enteritidis contamination of ice cream associated with a 1994 multistate outbreak. Journal of Food Protection, 61: 5-10.

Waterman, S.R., & Small, P.L. 1998. Acid-sensitive enteric pathogens are protected from killing under extremely acidic conditions of pH 2.5 when they are inoculated onto certain solid food sources. Applied Environmental Microbiology, 64: 3882-3886.

Wong, S.S., Yuen, K.Y., Yam, W.C., Lee, T.Y., & Chau, P.Y. 1994. Changing epidemiology of human salmonellosis in Hong Kong, 1982-93. Epidemiology and Infection, 113: 425-434.

Woodward, W.E. 1980. Volunteer studies of typhoid fever and vaccines. Transactions of the Royal Society of Tropical Medicine and Hygiene, 74: 553-556.