Attribution of Human Listeria monocytogenes Infections in England and Wales to Ready-to-Eat Food Sources Placed on the Market: Adaptation of the Hald Salmonella Source Attribution Model

Abstract

Human listeriosis is a rare but serious foodborne disease, with high morbidity and mortality in vulnerable populations (e.g., pregnant women, the elderly, and the immunocompromised). The disease is predominantly caused by the consumption of contaminated ready-to-eat foods. Since 2001, an increase in the number of listeriosis cases has been observed in several European Union countries, including England and Wales, predominantly in the over-60s population. The cause of this selective increased incidence is unknown. The Hald Salmonella Bayesian source attribution model was adapted to determine the potential of this approach to quantify the contribution of different food sources to the burden of human listeriosis in England and Wales from 2004 to 2007. The most important food sources for the overall population were multicomponent foods (sandwiches and prepacked mixed salad vegetables) (23.1%), finfish (16.8%), and beef (15.3%). Attribution of major sources of infection was similar for the elderly population (≥60 years old, multicomponent foods [22.0%], finfish [14.7%], and beef [13.6%]). For pregnancy-associated cases, beef (12.3%), milk and milk products (11.8%), and finfish (11.2%) were more important sources of infection. The adapted model also showed that the serotype 4b was associated with relatively more human infections than that of other serotypes; further, the subtype 4b amplified fragment-length polymorphism V was associated with more pregnancy-associated cases than other subtypes of 4b. This approach of quantifying the contribution of various food sources to human listeriosis provides a useful tool in food safety risk analysis, and underlines the need for further emphasis to be given to the reduction of Listeria monocytogenes in high-risk foods, such as multicomponent foods, which are consumed without any further treatment. The need for targeted dietary advice for the elderly population is also highlighted.

Introduction

L isteria monocytogenes is the causative agent of listeriosis. Human listeriosis is a rare but serious foodborne disease, with high morbidity and mortality in vulnerable populations (e.g., pregnant women and the unborn, newborns, the elderly, and the immunocompromised) (Gillespie et al., 2006; EFSA, 2007), and in the United Kingdom, it is the leading cause of death among foodborne bacterial pathogens (FSA, 2008). Although most cases are foodborne, the epidemiology is complex. The ubiquitous nature of the pathogen in the environment, together with a varied incubation period (1 to >90 days), means that identifying specific food vehicles can be problematic (McLauchlin et al., 2004; Gillespie et al., 2006).

Since 2001, an increase in the number of listeriosis cases has occurred in several European Union countries, including England and Wales, predominantly in persons over 60 years old. The reasons to this increase are not known, but they are independent of recognized outbreaks, sex, season, ethnicity, socioeconomic status, region, or L. monocytogenes serotype and/or amplified fragment-length polymorphism (AFLP) type (Gillespie et al., 2006; Goulet et al., 2008). The increase is also not accounted for by an increasing proportion of the population being elderly (ACMSF, 2009).

Source attribution approaches describe an association between the outcome (the burden of a given disease) and the specific sources or exposures (Batz et al., 2005; Pires et al., 2009). Foodborne disease source attribution is increasingly used to partition foodborne disease to the most important sources and as such support public health and food safety risk management and intervention strategies. The source attribution model developed by Hald et al. (2004) has proved successful in the attribution of human salmonellosis to animal-food sources in Denmark and also to other European Union member states, including the United Kingdom (Pires et al., 2008). The approach taken in the Hald model is the microbial subtyping approach where the relative contributions of different sources to human salmonellosis are inferred from the comparison between Salmonella subtypes in human and infection-source isolates. The increase in listeriosis cases emphasizes the need to better estimate the burden of disease and the presence of this pathogen in the food chain. We adapted and applied the Hald model to L. monocytogenes to determine whether such an approach may provide further insight and quantify the relative importance of food sources to listeriosis cases, including risk population groups such as the elderly.

Materials and Methods

Surveillance of L. monocytogenes in England and Wales

The Health Protection Agency (HPA) Centre for Infections surveils human L. monocytogenes infections in England and Wales as described in detail by Gillespie et al. (2006). Epidemiologic and microbiologic data from this surveillance system are combined and stored in a Microsoft Access database. National and routine microbiological food surveillance in England and Wales is also coordinated by the HPA. Within this food surveillance program, L. monocytogenes isolates obtained from ready-to-eat foods sampled from retail and food service premises have been referred since 2004. Food descriptions and microbiological data are also stored in the Access database.

All L. monocytogenes isolates from patient cases and food referred to HPA Centre for Infections are confirmed by polymerase chain reaction (McLauchlin, 1997). Isolates are characterized into subtypes by serotyping and AFLP analysis (Guerra et al., 2002; Doumith et al., 2004).

Data included in the model

Data that were used as input for the model included only data collected as part of the L. monocytogenes surveillance programs described above and represented the years 2004–2007 (Table 1).

Table 1.

Data Included in the Adapted Hald Model: Number of Listeria monocytogenes Isolates Found in Sporadic Cases of Human Listeriosis and in Ready-to-Eat Foods Placed on the Market in England and Wales, 2004–2007

		Reported human Listeria monocytogenes cases			Ready-to-eat food sources
Type	Listeria monocytogenes sero-AFLP subtype	Sporadic cases	Pregnancy-associated cases	Cases ≥60 years old	Beef	Pork/ham	Tongue	Other red meats ^a	Chicken	Turkey	Game bird	Finfish	Shellfish	Milk/milk products	Salad vegetables/fruit	Multicomponent foods ^b	Other foods ^c
1	1/2a IX	53	3	37	16	114	27	2	19	16	4	30	7	17	8	195	2
2	1/2a VII	44	7	25	8	23	8	1	2	5	1	4	1	35	24	92	15
3	1/2c VII	19	1	12	17	45	14	4	7	6	0	2	0	23	3	69	4
4	4b I	143	18	108	10	21	20	0	7	1	2	8	4	27	4	49	6
5	4b IV	91	7	62	17	35	3	0	3	3	0	3	5	3	5	36	1
6	1/2b II	40	2	31	9	18	13	0	9	5	0	2	0	4	1	31	3
7	1/2b IV	15	3	12	1	8	1	0	5	2	2	3	1	47	0	16	1
8	1/2a III	18	2	9	3	28	1	0	5	1	0	2	1	5	6	8	1
9	1/2a XIV	19	2	11	1	4	0	0	0	0	0	25	0	4	3	5	3
10	1/2a VI	3	0	2	0	4	2	0	1	0	0	0	0	11	9	8	0
11	4b V	73	23	37	3	3	0	0	0	1	0	2	0	0	0	11	1
12	1/2a II	2	0	2	0	1	0	0	0	0	0	1	0	2	1	7	0
13	1/2a XI	7	0	6	0	1	0	0	0	0	0	0	1	0	2	7	0
14	4b II	1	0	1	0	1	0	0	1	0	0	0	0	0	0	1	1
15	4b XV	5	1	3	0	0	0	0	0	0	0	1	0	1	1	0	1
16	1/2b IX	1	0	1	1	0	0	0	0	1	0	0	0	0	0	1	0
17	4b VII	1	0	1	0	1	0	0	0	0	0	0	0	0	1	1	0
18	1/2a XV	1	0	0	0	0	0	0	0	0	0	0	0	1	0	1	0
19	1/2b XV	1	0	1	0	0	0	0	0	0	0	0	0	1	1	0	0
20	1/2a XII	1	0	1	0	1	0	0	0	0	0	0	0	0	0	0	0
21	Others^d	26	1	19	0	0	0	0	0	0	0	0	0	0	0	0	0
Total		564	70	381	86	308	89	7	59	41	9	83	20	181	69	538	39

Pates, chopped livers, and fermented meats.

Sandwiches and prepacked mixed vegetable salads.

Desserts and cooked rice.

Included subtypes that caused human listeriosis but were not isolated from the investigated food sources.

AFLP, amplified fragment-length polymorphism.

Food descriptions were classified according to the food type (poultry, red meat, etc.) and the main ingredient (chicken, pork, etc.). In total, 1529 isolates from 13 ready-to-eat food sources were included in the analysis: beef, pork/ham, tongue, other red meat products (pates, chopped livers, and fermented meats), chicken, turkey, game bird, finfish, shellfish, milk/milk products, salad vegetables/fruit, multicomponent foods (e.g., sandwiches and prepacked mixed vegetable salads), and other foods (e.g., desserts and cooked rice) (Table 1).

Since the purpose of the model was to quantify the relative importance of ready-to-eat food sources to listeriosis cases within risk groups, we categorized cases as pregnancy-associated (all maternal–fetal patients and neonatal patients; a mother–baby pair was considered as one case-patient) and nonpregnancy-associated (patient >1 month of age). Nonpregnancy case-patients were categorized further into those who were over 60 years or not. Three scenarios were investigated, two of them corresponding to risk groups, and one for the overall population. In total, 564 sporadic cases of listeriosis were reported in the overall population from 2004 to 2007 (Table 1). Among these cases, 538 were caused by 20 L. monocytogenes sero-AFLP subtypes, which were isolated from different food sources accounted for in the analysis. These 20 L. monocytogenes sero-AFLP subtypes were included in this scenario analysis, along with a “type” labeled as “others” that included subtypes (1/2a XIII, 1/2a XIX, 1/2b VI, 1/2b XVII, 4b VI, and 4b VIII) that caused human infection but were not isolated from the investigated food sources. In the scenario describing the risk group constituted by pregnant women, 70 reported cases, caused by 12 different L. monocytogenes subtypes (including “others”) (Table 1), were attributed to the food sources. The third scenario attributed human cases of listeriosis among the elderly population (≥60 years old), and utilized 381 cases caused by 20 types (including “others”) (Table 1).

Adaptation of the Hald Salmonella Bayesian risk attribution model

The model developed by Hald and colleagues compares the number of sporadic human cases of salmonellosis by different subtypes with their distribution in different food sources, weighted by the amount of food consumed. The full portrayal and discussion of the assumptions and statistics behind the Salmonella Bayesian risk attribution model is described by Hald et al. (2004). The principle of our adapted model is to compare the number of human sporadic cases caused by different L. monocytogenes subtypes with the prevalence of the subtypes isolated from different 13 food sources. The differences in the abilities of the L. monocytogenes subtypes to cause human disease (q_i ) and of the food sources to act as a vehicle of infection (a_j ) are accounted for by prior distributions. The Hald model also incorporates a food consumption weighting factor. In our adapted model, differences in consumption patterns of the investigated foods in the population were accounted for in a_j (as there were no readily available food consumption estimates). Mullner et al. (2009) and Whalstrom et al. (2008) have, however, previously shown that the exclusion of food consumption weights in the model when applied to Campylobacter and Salmonella in New Zealand and Salmonella in Sweden had no major consequences, but noted that the a_j became less comparable. For example, in the adapted model a high a_j could reflect a high exposure to the food source and not necessarily a high ability of the individual food source to act as a vehicle for disease.

Posterior distributions for the q_i and a_j factors were estimated by fitting the adapted model to the observed number of cases per L. monocytogenes subtype using a Markov Chain Monte Carlo simulation. The equation used to estimate the number of human cases per food source and L. monocytogenes subtype was defined as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes {10} {9} {7} {6} \begin{document} \begin{align*} \lambda_{ji} = p_{ji} \times q_i \times a_j, \end{align*} \end {document}

where λ_ji is the expected number of cases per year of type i from source j, p_ji is the prevalence of subtype i in source j, q_i is the subtype-related factor for subtype i, and a_j is the food-source-related factor for source j.

The model used the assumptions that all L. monocytogenes serotypes have different q_i , and AFLP groups within the same serotypes were assumed to have the same q_i. As the q_i values are all relative, we fixed the q prior estimate for serotype 1/2a as 1. All other q estimates are relative to this one. Additionally for the pregnant-associated cases scenario, the q_i prior distribution for subtype 4b V was defined as an independent prior distribution. The assumption was based on the evidence of a higher likelihood of infection in pregnant women with subtype 4b V (Gillespie and McLauchlin, 2007). No similarities between the a_j for each source were assumed.

The model was set to calculate the expected number of cases in each risk group per L. monocytogenes subtype (λ_ji ) according to the above equation. The observed data or number of human cases of type i was then linked with the prior distribution by assuming that the observed cases per L. monocytogenes subtype were Poisson distributed with a parameter value equal to the expected number of cases. The model was set up in the WinBUGS software (Gilks et al., 1994). Five independent Markov chains of 20,000 iterations each were run, and convergence was monitored as described by Gelman and Rubin (1992) and Brooks and Gelman (1998).

Results and Discussion

Adaptation of the Hald et al. (2004) model to quantify the contribution of ready-to-eat food sources to sporadic cases of listeriosis in England and Wales presented a valid means to assess the relative risk from different foods, the results of which are presented in Table 2. For the overall population, the risk model for L. monocytogenes attributed 551 human cases (97.7%), a close match to the observed number of 564 cases; 13 cases (2.3%) could not be associated with any of the food sources. The major sources of infection were multicomponent foods (23.0%), finfish (16.8%), and beef products (15.3%). Similarly in the over-60s population group, the major sources of infection were found to be multicomponent foods (22.0%), finfish (14.7%), and beef (13.6%). Consumption of multicomponent foods—particularly, supplied prepacked sandwiches in hospitals—has been shown to be associated with recent outbreaks of listeriosis in England and Wales (Little et al., 2007; HPA, 2008). However, in the pregnancy-associated cases group, beef (12.3%), milk and milk products (11.8%), and finfish (11.2%) were ranked as important major sources of infection. Multicomponent foods (i.e., sandwiches), sliced meats such as beef, fish, and milk and milk products are considered high-risk foods with regard to L. monocytogenes and thus listeriosis as they have potential for contamination with this pathogen, support its growth to high numbers, can be sold with a long shelf-life under refrigerated conditions, and are consumed as a ready-to-eat food product (Sim et al., 2002; Little et al., 2007, 2009; ACMSF, 2009; HPA, 2008; Wagner and McLauchlin, 2008).

Table 2.

Estimated Major Ready-to-Eat Food Sources of Cases of Human Listeriosis in England and Wales, 2004–2007

	Estimated no. of cases					Percentage of sporadic cases
Source	Mean	SD	Median	95% CI		Mean	95% CI
Overall population (n = 564)
Beef	86	53.8	82	5	200	15.3	0.8	35.5
Pork/ham	30	19.4	27	4	78	5.4	0.7	13.7
Tongue	13	11.8	10	0	44	2.3	0.1	7.7
Other meat products	7	6.5	5	0	24	1.2	0.0	4.2
Chicken	11	9.9	8	0	37	1.9	0.1	6.5
Turkey	56	42.1	48	2	155	10.0	0.4	27.4
Game bird	7	6.8	5	0	25	1.3	0.0	4.5
Finfish	95	27.6	93	44	153	16.8	7.9	27.0
Shellfish	11	10.3	8	0	38	2.0	0.1	6.8
Milk/milk products	12	10.4	9	0	39	2.1	0.1	6.9
Salad vegetables/fruit	33	16.6	32	6	70	5.9	1.0	12.3
Multicomponent foods	130	56.5	128	27	246	23.0	4.8	43.6
Other foods	59	30.3	57	7	122	10.4	1.2	21.7
Unknown source	13	23.5	13	− 34	58	2.3	− 6.0	10.3
Pregnancy-associated cases (n = 70)
Beef	9	7.1	7	0.3	26.4	12.3	0.4	37.6
Pork/Ham	5	4.4	3	0.1	16.4	6.8	0.2	23.5
Tongue	5	4.1	4	0.1	15.2	6.7	0.2	21.7
Other meat products	2	1.8	1	0.1	6.7	2.7	0.1	9.6
Chicken	4	3.7	3	0.1	13.8	5.8	0.2	19.7
Turkey	5	5.0	4	0.2	18.5	7.8	0.2	26.5
Game bird	4	3.8	3	0.1	13.9	6.1	0.2	19.9
Finfish	8	6.2	6	0.3	23.3	11.2	0.4	33.3
Shellfish	5	4.2	4	0.1	15.4	6.8	0.2	22.0
Milk/Milk products	8	6.1	7	0.4	22.9	11.8	0.5	32.7
Salad vegetables/fruit	6	5.1	5	0.2	19.0	9.2	0.3	27.1
Multicomponent foods	6	5.6	4	0.2	20.8	8.7	0.2	29.7
Other foods	15	10.0	13	0.7	37.5	21.3	1.1	53.5
Unknown source	0	9.1	0	− 30.7	4.8	0	− 43.9	6.9
Elderly (≥60 years old) cases (n = 381)
Beef	52	33.8	47	3	126	13.6	0.7	33.0
Pork/ham	26	16.8	23	4	67	6.9	0.9	17.6
Tongue	21	15.9	18	1	59	5.6	0.2	15.5
Other meat products	6	5.4	4	0	20	1.6	0.0	5.3
Chicken	12	10.5	9	0	39	3.0	0.1	10.2
Turkey	27	21.8	21	1	81	7.0	0.2	21.3
Game bird	8	7.6	6	0	28	2.2	0.1	7.4
Finfish	56	20.6	55	20	101	14.7	5.2	26.4
Shellfish	16	13.1	12	1	49	4.1	0.1	12.7
Milk/milk products	10	8.5	7	0	32	2.6	0.1	8.3
Salad vegetables/fruit	25	13.1	23	4	54	6.5	1.0	14.2
Multicomponent food	84	43.4	82	9	174	22.0	2.3	45.5
Other foods	33	20.8	30	2	80	8.6	0.6	20.9
Unknown source	6	19.3	6	− 33	43	1.6	− 8.6	11.3

SD, standard deviation; CI, Bayesian credibility interval.

Although the risk of listeriosis is higher when the food is contaminated with high levels of the organism (McLauchlin et al., 2004; EFSA, 2007), the infectious dose of the pathogen is also dependent on the type of food product, strain virulence, and host susceptibility; therefore, relatively lower doses may still cause infection in high-risk populations (McLauchlin et al., 2004). The food-source-related factor (a_j ) in the model may account for characteristics of the food item that influence its ability to act as a vehicle for exposure to the pathogen and infection (e.g., general differences in the pathogen load, food characteristics influencing growth behavior, or preparation procedures), and/or food consumption patterns. The highest value for the food-source-related factor (a_j ) in the overall population and in both risk population groups was found for “other red meats,” which include meats generally on the market from delicatessen stores (overall population 1.4, 95% Bayesian credibility intervals [CI] 0.04–5.0; pregnancy-associated cases 0.5, 95% CI 0.01–1.7; elderly cases 1.2, 95% CI 0.03–4.3) (Fig. 1).

FIG. 1.

Estimated values and 95% credibility intervals of the ready-to-eat food-source-related factors (a_j ).

Meats, including delicatessen meats and pates, have been shown to be important sources of L. monocytogenes in the United Kingdom and elsewhere (Lianou and Sofos, 2007; Gormley et al., 2009; Little et al., 2009). Ready-to-eat food may be contaminated at point of consumption because contamination was not eliminated entirely earlier in the production chain or as a result of postprocess cross-contamination. The U.S. microbiological risk assessment categorized delicatessen meats at the highest level of risk of causing serious illness and death associated with consumption of ready-to-eat foods potentially contaminated with L. monocytogenes, on both a serving and per annum basis (FDA, 2003). Continuous efforts should be made to improve food safety by adequate cooking and appropriate hygienic measures to avoid contamination, and through use of appropriate refrigeration temperatures to prevent growth of L. monocytogenes in these products.

The subtype-related factor (q_i ) represents differences in the survivability of the L. monocytogenes subtypes along the food chain, and differences in their virulence and pathogenicity to humans. The estimated values for the L. monocytogenes subtype-related factors (q_i ) suggest that the ability of serotype 4b to survive food processing and/or cause disease in the overall population, pregnancy-associated cases, and the elderly group is higher than that of other serotypes (Fig. 2). Moreover, in the pregnancy-associated group, subtype 4b V was estimated to have a higher ability to cause disease; the estimated q_i value for this subtype was 63.9 (95% CI 24.6–147.7); the remaining AFLP types within serotype 4b were estimated to have a q_i value of 3.3 (95% CI 1.7–5.8), followed by types 1/2b (q_i = 1.1, 95% CI 0.3–2.6) and 1/2c (q_i = 0.4, 95% CI 0.04–1.1) (Fig. 2). In concordance with these results, the L. monocytogenes serotypes most often causing infection in England and Wales from 2000 to 2008 are 4b (47%), 1/2a (18%), and 1/2b (8%) (HPA, unpublished data), and among pregnancy-associated cases subtype 4b V is predominant (Gillespie and McLauchlin, 2007).

FIG. 2.

Estimated values and 95% credibility intervals of the Listeria monocytogenes serotype–related factors (q). As the q_i values are all relative, the q prior estimate for serotype 1/2a was fixed as 1.

Although monitoring of convergence of the adapted model provided satisfactory results, our risk attribution results presented wide 95% CI. The wide CIs may reflect the epidemiology of L. monocytogenes and the resulting uncertainty on dose response. Dose–response relationships are not well understood for human L. monocytogenes infection. These relationships comprise three major components, each with major uncertainties. The three components are the environment (the food matrix and storage conditions that allow contaminating Listeria to grow to hazardous levels), the pathogen (virulence), and the host (susceptibility and immune status) (McLauchlin et al., 2004). The relative contributions of host and environmental factors in acquiring listeriosis require further study.

Another limitation of our source attribution model was the small size of the dataset with 564 human cases compared with 41,264 Salmonella cases from England and Wales used in the Hald model (Gordon L. Nichols, HPA, pers. comm.). The original Hald Salmonella model itself also utilized a greater number of cases (3268). A similar observation in CIs was also noted in the New Zealand adaptation of the Hald model to Campylobacter where 481 cases were included (Mullner et al., 2009). Increasing the number of Listeria cases in the model may improve precision of the estimates, but compared to Campylobacter and Salmonella in England and Wales, listeriosis is a rare disease with currently around 200 cases reported per annum. The food isolate data included in the model were obtained from food surveys and investigations carried out during 2004–2007. It is conceivable that there could be some foods that contribute a risk for listeriosis which are not represented in this dataset and whose effect will be underestimated as a result. The rationale for the development of a European surveillance network and community food surveys for listeriosis (Denny and McLauchlin, 2008; EFSA, 2009) would, however, present an opportunity to collate pan European data on listeriosis cases, prevalence and levels of L. monocytogenes in different foods, and on L. monocytogenes subtypes both from cases and food sources that could be analyzed using this modified Hald model. Such an approach could provide an improved estimate of the burden of disease and the presence of this pathogen in the food chain.

Conclusions

The modified Hald model for L. monocytogenes provides a valid and useful approach for the attribution of the burden of listeriosis to the specific food sources, and for the prioritization of effective food safety interventions. Further emphasis needs to be given to the reduction of L. monocytogenes in high-risk foods, such as multicomponent foods (prepacked sandwiches and prepacked mixed vegetable salads), which are consumed without any further treatment. High-quality ingredients, hygienic manufacture, appropriate shelf life, and correct refrigerated storage are critical to minimize the potential for L. monocytogenes to be present at the point of consumption at levels hazardous to health. Although listeriosis continues to occur in association with pregnancy, it is now predominantly an infection of the older population. Intervention measures including dietary advice were previously predominantly targeted toward the pregnant woman. However, because of the change in the age distribution of cases, together with a demographic shift in the population as a whole, intervention measures now will therefore be most effective when targeted toward the elderly population.

Footnotes

Acknowledgments

This work was supported in part by a grant from the Med-Vet-Net Short Term Mission/Scientific Exchanges. Med-Vet-Net is a European Network of Excellence for Zoonoses research, working for control and prevention of zoonoses and foodborne diseases. The authors thank Tine Hald, Technical University of Denmark, from Workpackage 28 in Med-Vet-Net for support.

Disclosure Statement

No competing financial interests exist.

References

[ACMSF] Advisory Committee on the Microbiological Safety of Food. Ad hoc group on vulnerable groups. Report on the increased incidence of listeriosis in the UK. July 2009. www.food.gov.uk/multimedia/pdfs/committee/acmsflisteria.pdf. 2009 December 10(Online.)

Batz

, Doyle

, Morris

Jr.

et al. Attributing illness to food. Emerg Infect Dis, 2005; 11:993–999.

Brooks

, Gelman

. Alternative methods for monitoring convergence of iterative simulations. J Comput Graph Stat, 1998; 7:434–455.

Denny

, McLauchlin

. Human Listeria monocytogenes infections in Europe—an opportunity for improved European surveillance. Euro Surveill 2008;13:pii:8082. www.eurosurveillance.org/ViewArticle.aspx?ArticleId=8082. 2009 December 10(Online.)

Doumith

, Buchrieser

, Glaser

et al. Differentiation of the major Listeria monocytogenes serovars by multiplex PCR. J Clin Microbiol, 2004; 42:3819–3822.

[EFSA] European Food Safety Authority. Report of task force on zoonoses data collection on proposed technical specifications for a survey on Listeria monocytogenes in selected categories of ready-to-eat food at retail in the EU. EFSA J, 2009; 2616:1–66.

[EFSA] European Food Safety Authority. Scientific opinion of the panel on biological hazards on a request from the European Commission on request for updating the former SCVPH opinion on Listeria monocytogenes in ready-to-eat foods and the related risk for human illness. EFSA J, 2007; 599:1–42.

[FDA] Food, and Drug Administration. U.S. Department of Agriculture, U.S. Centers for Disease Control and Prevention. Quantitative assessment of relative risk to public health from foodborne Listeria monocytogenes among selected categories of ready-to-eat foods, September 2003. www.foodsafety.gov/∼dms/lmr2-toc.html. 2009 December 10(Online.)

[FSA] Food Standards Agency. Annual report of the chief scientist 2006/7, London, UK, 2008. www.food.gov.uk/multimedia/pdfs/publication/chiefscientist1107.pdf. 2009 December 10(Online.)

10.

Gelman

, Rubin

. Inference from iterative simulations using multiple sequences (with discussion) Stat Sci, 1992; 7:457–511.

11.

Gilks

, Thomas

, Spiegelhalter

. A language program for complex Bayesian modelling. Statistician, 1994; 43:169–178.

12.

Gillespie

, McLauchlin

. Advisory Committee on the Microbiological Safety of Food, Update on Listeriosis in England and Wales, December 2007. www.food.gov.uk/multimedia/pdfs/committee/879listeria.pdf. 2009 December 10(Online.)

13.

Gillespie

, McLauchlin

, Grant

et al. Changing pattern of human listeriosis in England and Wales, 2001–2004. Emerg Infect Dis, 2006; 12:1361–1366.

14.

Gormley

, Little

, Grant

et al. The microbiological safety of ready-to-eat specialty meats from markets, specialist food shops: a UK wide study with a focus on Salmonella and Listeria monocytogenes. Food Microbiol, 2009, 10.1016/j.fm.2009.10.009(Online.)

15.

Goulet

, Hedberg

, Le Monnier

et al. Increasing incidence of listeriosis in France and other European countries. Emerg Infect Dis, 2008; 14:734–740.

16.

Guerra

, Bernardo

, McLauchlin

. Amplified fragment length polymorphism (AFLP) analysis of Listeria monocytogenes. Syst Appl Microbiol, 2002; 25:456–461.

17.

Hald

, Vose

, Wegener

et al. A Bayesian approach to quantify the contribution of animal-food sources to human salmonellosis. Risk Anal, 2004; 24:255–269.

18.

[HPA] Health Protection Agency. Human listeriosis linked to hospital sandwiches: implications for procurement and storage. HPR Wkly, 2008; 2. www.hpa.org.uk/hpr/archives/2008/hpr3508.pdf. 2009 December 10(Online.)

19.

Lianou

, Sofos

. A review of the incidence and transmission of Listeria monocytogenes in ready-to-eat products in retail and food service environments. J Food Prot, 2007; 70:2171–2198.

20.

Little

, Barret

, Grant

et al. Microbiological safety of sandwiches from hospitals and other health care establishments in the United Kingdom with a focus on Listeria monocytogenes and other Listeria species. J Food Prot, 2007; 71:309–318.

21.

Little

, Sagoo

, Gillespie

et al. Prevalence and level of Listeria monocytogenes and other Listeria species in selected retail ready-to-eat foods in the UK. J Food Prot, 2009; 72:1869–1877.

22.

McLauchlin

. The identification of Listeria species. Int J Food Microbiol, 1997; 38:77–81.

23.

McLauchlin

, Mitchell

, Smerdon

et al. Listeria monocytogenes and listeriosis. A review of hazard characterization for use in microbiological risk assessment of foods. Int J Food Microbiol, 2004; 92:15–33.

24.

Mullner

, Jones

, Noble

et al. Source attribution of food-borne zoonoses in New Zealand: a modified Hald model. Risk Anal, 2009; 29:970–984.

25.

Pires

, Evers

, van Pelt

et al. Attributing the human disease burden of foodborne infections to specific sources. Foodborne Pathog Dis, 2009; 6:417–424.

26.

Pires

, Kasbohrer

, Spitznagel

et al. Salmonella source attribution in different European countries. Food Micro 2008, Aberdeen, UK. http://www.foodmicro2008.org/fm08/resources/374/1143/pdf/FM2008_0211.pdf. 2009 December 10.

27.

Sim

, Hood

, Finnie

et al. Series of incidents of Listeria monocytogenes non-invasive febrile gastroenteritis involving ready-to-eat meats. Lett Appl Microbiol, 2002; 35:409–413.

28.

Wagner

, McLauchlin

. Biology. Handbook of Listeria monocytogenes. Dongyou

. Boca Raton, Florida: CRC Press, 2008; 3–25.

29.

Whalstrom

, Andersson

, Plum-Forsell

, Pires

. Source attribution of reported human Salmonella cases in Sweden. Med-Vet-Net 4th Annual Scientific Meeting, 2008 St. Malo, France.