General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (2/05) Page 1 of 19

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code OZ0322

2. Project title

An evaluation of current animal and human Salmonella research, harmonisation of diagnostic techniques and introduction of novel research tools

3. Contractororganisation(s)

Veterinary Laboratories Agency,Woodham Lane,New Haw,Weybridge,Surrey,KT15 3NB.

54. Total Defra project costs £ 103,682

5. Project: start date................ 01 May 2005

end date................. 30 April 2006

SID 5 (2/05) Page 2 of 19

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.Main objective01 – Literature review and recommendationsTo produce a literature review of recent research related to diagnostic and epidemiological techniques used in animal and human Salmonella investigations and advise on areas for new research.FindingsThis project has sought to provide a rational basis for the harmonisation and enhancement of routine diagnostic and surveillance activities related to Salmonella in the Veterinary Laboratories Agency and the Health Protection Agency.

The scope and limits of the review were agreed at an early planning meeting between Defra and the collaborators and recommendations deriving from the review were agreed between collaborators at a meeting prior to compilation of the final report.

The review has been written as a series of chapters under the following headings:Chapter 1: Veterinary Salmonella surveillance and outbreak investigationsChapter 2: Surveillance of human salmonellosis - methodology; national/international detection, investigation and control of outbreaks; national/international databases; electronic exchange of data; design of surveys.Chapter 3: Early detection systemsChapter 4: Analysis, modelling and risk assessment

4.1 – A Review of the application of spatial analysis and Geographic Information Systems in animal and human Salmonella research4.2 – Mathematical modelling for Salmonella4.3 – Microbial risk assessment

Chapter 5: Salmonella sampling and detection methodsChapter 6: Epidemiological investigations based on strain discrimination and typing

6.1 – Molecular serotyping for Salmonella6.2 – Sub-typing6.3 Antimicrobial resistance typing

Chapter 7: Archiving of SalmonellaChapter 8: Novel techniques

8.1 – DNA-based approaches to study virulence and evolution of Salmonella8.2 – Review of microbial proteomics8.3 – Phenotype screening for Salmonella spp.

SID 5 (2/05) Page 3 of 19

Chapter 9: Methods to study Salmonella pathogenesisChapter 10: Recommendations

Information was derived from a variety of sources including on-line, CD and in-house databases, search engines, specific websites, peer-reviewed journals and articles, textbooks, manuals, technical documents and reports, internal reports, study protocols, Defra study reports, COST Action inventory, Scientific Conference Proceedings and scientific staff. Search criteria are recorded for each chapter.

A study visit was undertaken by three of the authors to the Danish Zoonosis Centre and EpiLab to gather information on salmonella surveillance programmes in Denmark, the DANMAP programme, Vetstat, Salmonella Source Account, and risk-based optimisation of the Danish pork Salmonella programme.

Key recommendations from the review include: Development of shared protocols for statistically structured international, national or local outbreak

investigations and investigating incidents with modelling to inform data collection and experimental design.

Improving standardisation of data collection, collation and databases for storage. Improving access to spatially referenced and denominator data Improving timeliness of data exchange and enhancing links with the Early Detection System for

human salmonella of emerging importance. Developing use of cluster analysis and GIS mapping for detecting outbreaks and focussing risk

factor analyses. Improving use of the iterative nature of microbial risk assessments and using MRA to drive the

direction of research projects. Optimising sampling protocols including use of statistical approaches and novel techniques for

different situations; optimising sample handling and preparation, pre-enrichment and enrichment and developing methods for detecting multiple pathogens.

Development of alternative plating techniques to identify new and emerging serovars and molecular techniques for use in monitoring programmes.

Development of inter laboratory quality assurance schemes for a range of different matrices. Development and harmonisation of sub-typing methods and molecular characterisation and

promotion of techniques amenable to electronic storage and exchange of data.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

MAIN OBJECTIVE01 – Literature review and recommendationsTo produce a literature review of recent research related to diagnostic and epidemiological techniques used in animal and human Salmonella investigations and advise on areas for new research.

BACKGROUNDBoth the Veterinary Laboratories Agency (VLA) and the Health Protection Agency (HPA) conduct routine diagnostic, reference and epidemiological investigations of Salmonella, often working together in outbreak investigations. The two organisations have signed a Memorandum of Understanding (MoU), an associated aim of which is to achieve optimal collaboration and combination of their respective facilities and capabilities including investigation, surveillance and control of zoonoses; surveillance of antimicrobial drug resistance in bacteria from human and animal sources, food and water; investigation, surveillance and control of food-borne infection;

SID 5 (2/05) Page 4 of 19

development of common Standard Operating Procedures (SOPs) for the isolation, identification and typing of zoonotic bacteria and for antimicrobial sensitivity testing and harmonisation of procedures for the surveillance of human and animal infectious disease. This project has sought to provide a rational basis for the harmonisation and enhancement of routine diagnostic and surveillance activities related to Salmonella in the two organisations by reviewing and evaluating the current research and activities in these areas and providing recommendations to Defra for achieving these aims. Recommendations have also been made on areas for new research targeted at the harmonisation of diagnostic and typing techniques, and the introduction of novel research tools.

ACHIEVEMENT OF OBJECTIVE O101/01 – Planning meeting of collaborators and contributorsA planning meeting was held at VLA, Weybridge on 21/06/05 with collaborators from the VLA, the HPA and the Institute for Animal Health to agree the scope and limits of the review and the responsibilities of each contributor. The Defra Project Officer also attended. (Minutes are attached). It was agreed that the review would cover current activities at the VLA and the HPA and would then explore recent research and best practice to define approaches to use in the future.

The review has been written as a series of chapters each with a common format encompassing: Chapter number, title, authors Table of contents Summary Introduction and search criteria Review Key points (summarised in grey shaded boxes) Recommendations (in prioritised order) References

The chapters cover:Chapter 1: Veterinary Salmonella surveillance and outbreak investigationsChapter 2: Surveillance of human salmonellosis - methodology; national/international detection, investigation and

control of outbreaks; national/international databases; electronic exchange of data; design of surveys.Chapter 3: Early detection systemsChapter 4: Analysis, modelling and risk assessment

4.1 – A Review of the application of spatial analysis and Geographic Information Systems in animal and human Salmonella research4.2 – Mathematical modelling for Salmonella4.3 – Microbial risk assessment

Chapter 5: Salmonella sampling and detection methodsChapter 6: Epidemiological investigations based on strain discrimination and typing

6.1 – Molecular serotyping for Salmonella6.2 – Sub-typing6.3 Antimicrobial resistance typing

Chapter 7: Archiving of SalmonellaChapter 8: Novel techniques

8.1 – DNA-based approaches to study virulence and evolution of Salmonella8.2 – Review of microbial proteomics8.3 – Phenotype screening for Salmonella spp.

Chapter 9: Methods to study Salmonella pathogenesisChapter 10: RecommendationsDocuments are appended for:i). Minutes of meetings (for planning on 21.6.05 and recommendations on 27.2.06)ii). Itinerary and report on visit to the Danish Zoonosis Centre and EpiLab

It was agreed that the limits of each section of the review might be different and therefore, the search criteria used would be included at the beginning of each chapter. The agreed common limits were as follows:

Use mainly English language papers but if the summary looks interesting, a limited number of translations may be sought.

Confine to zoonotic salmonellas (exclude Pullorum and Gallinarum) Include surveillance information Make use of current reference lists that have been produced for recent VLA and HPA

projects/publications. Build on these. Use recent literature and define extent of search Use key contacts of project participants for advice/up to date and unpublished info. Include information from DANMAP and EpiLab

SID 5 (2/05) Page 5 of 19

01/02 – Conduct searches and collate information on bacteriological and epidemiological techniques applicable to Salmonella investigationsThe searches were carried out according to the outline agreed at the planning meeting and search criteria for each chapter are included below:

Chapter 1: Veterinary Salmonella surveillance and outbreak investigations The recent literature relating to surveillance methods for Salmonella in animals was reviewed in order to produce a series of recommendations for the enhancement of routine surveillance activities for Salmonella in animals carried out by the VLA. The review gave due consideration to the UK Veterinary Surveillance Strategy and its associated information management system, RADAR (Rapid Analysis and Detection of Animal-Related Risks) and included the Salmonella surveillance methodology, monitoring and control of Salmonella in animals in the UK and countries other than the UK, as well as international networks, databases, electronic exchange of data & outbreak detection. There is an unevenness in the information presented from different countries, for example there is more information presented from the Salmonella surveillance system in Denmark, compared with other countries. The reason is that the literature search retrieved more data from Denmark than from other countries. There was no bias in the use of references from different sources.

Search criteria:

Salmonella Typhi, S. Gallinarum and S. Pullorum were excluded from the review, as the first is a recognised human pathogen, while the latter two are almost exclusively animal-related pathogens.

Keywords used: “Salmonella”, “surveillance” Databases searched: Web of Science, MedLine (VLA library version of PubMed), Procite-5 Salmonella

(VLA internal database), Science Direct Time period: 1999-2005 Other sources of information: official websites (listed in bibliography) All languages

Search criteria employed in the review of Outbreak Investigation were as above however additional keywords were employed. These included combinations of “outbreak”, “investigation” and “livestock”. All references on human outbreaks were excluded unless they had been traced to a livestock source.

Chapter 2: Surveillance of human salmonellosis: methodology; national/international detection, investigation and control of outbreaks; national/international databases; electronic exchange of data; design of surveys.The search engines PubMed, Google were used to retrieve relevant reports and peer-reviewed journals. Searches were also made within the HPA website. Search criteria include: Surveillance and Salmonella and HPA; Surveillance and Salmonella and UK (or England & Wales); Salmonella and typing (methods) and HPA; General outbreaks of infectious intestinal disease (surveillance) and England & Wales; (international) outbreak investigation and Salmonella; case-control studies and Salmonella; Cohort studies and Salmonella; nested case-control studies and Salmonella; food studies and HPA and Salmonella; microbiological food surveillance and HPA; international surveillance and Salmonella.

Information was also retrieved from internal reports and study protocols, general epidemiology textbooks and specific sites such as the Enter-net, Salm-gene or Pulse-net Europe websites.

Scientific staff at HPA Centre for Infections, Environmental and Enteric Diseases Department, Information Management and Technology Section and Laboratory of Enteric Pathogens have provided valuable information for the review.

Information specific to S. Typhi, Gallinarum and Pullorum was not included in the review.

Chapter 3: Early detection systemsThe focus of this chapter is the review of early detection systems that have been implemented, world-wide, for the early and rapid detection of increases of Salmonella in both human and animal health surveillance. In addition, focus is paid to the statistical methods used, and the methodological issues that impact on the development of an outbreak detection system for Salmonella for both public and animal health. Before conducting such a review, however, it is important to note the distinction between several concepts associated with applying statistical models to surveillance data. Models may be implemented for forecasting, or aid in early warning or early detection of unwanted health outcomes. Early warning systems aim to model the disease risk based on historical surveillance and contemporary data and forecast the future risk through predictive models. Early detection systems, on the other hand, aim to alert as soon as possible increases in adverse health outcomes as they are occurring. This latter model relies on the use of historical data to estimate what would be expected in the current time period. A decision is made as to whether an outbreak, for example, is occurring by comparing the currently observed number of disease reports to an upper limit or threshold value. These systems are typically automated

SID 5 (2/05) Page 6 of 19

using a single robust statistical algorithm that summarises the main components of the system namely, the data, the prediction element, the alarm signal, and the output.

This literature review is the culmination of a search through the scientific literature and Internet for both published and unpublished early detection systems for Salmonella. The criteria for the literature search are outlined below.

All years, English language only and all countries. Databases searched: Science Direct, Web of Science and Google.

The specific search terms used for each search engine are outlined in the following table.

Table 1: Summary of the search terms used for the various databases or the Internet

Science Direct Web of Science Google Outbreaks and detection

systems Salmonella and

surveillance systems Statistical surveillance Statistical and

surveillance

Outbreak detection systems

Outbreak detection Salmonella and

surveillance and statistical

Statistical surveillance Salmonella and

algorithm Surveillance and

algorithm

Statistical surveillance Salmonella and

surveillance systems Salmonella and statistical

algorithm Salmonella and statistical

algorithm and animal health

Aberration detection methods

Chapter 4: Analysis, modelling and risk assessment4.1 – A Review of the application of spatial analysis and Geographic Information Systems in animal and human Salmonella research

A search was originally carried out through PubMed and Science-Direct using combinations of the keywords “Salmonella”, “spatial”, “space”, “GIS”, “cluster” and “map” to search for appropriate English-language journals. As the journals were scarce, the search was extended to all diseases and not just Salmonella. No date limit was placed on the search due to the limited amount of relevant literature.

4.2 – Mathematical modelling for Salmonella The search criteria for the Web of Knowledge database were the following terms: Salmonella; caecum; disease-transmission; epidemiology; finishing; mathematical models; slaughter. In addition cited references and related records were also searched.. The PubMed database and the internet were also searched, for which the following terms were used: Salmonella, transmission model, poultry/pigs/cattle.

4.3 – Microbial risk assessment (MRA)In many countries, including the UK, MRA is being utilised to address food safety issues, such as Salmonella. Although many countries have developed MRAs for this hazard, here, only the UK MRAs are considered. This is because the risk question and the circumstances under which a MRA is commissioned must always be considered carefully when evaluating and comparing any MRAs. Other countries may have different emphases, different structures and different Salmonella problems, which makes the consideration of these MRAs more difficult and certainly means that comparison to the UK MRAs would be difficult.

To identify the UK MRAs, a literature search was carried out in ProMed. The search terms were: Salmonella, Risk, Assessment. In addition, an inventory of quantitative microbial risk assessments was used, which was co-ordinated by Working Group 3 of COST ACTION 920 (COST920, 2006) and also an existing inventory of risk assessments that was done by the Veterinary Laboratories Agency (VLA) for the Food Standards Agency (Munday & Snary, 2005). MRAs for antimicrobial-resistant Salmonella were included in this review. Although this review only considers in detail the UK Salmonella MRAs, non-UK MRAs and MRAs for other foodborne pathogens have, where appropriate, been referred to. The first UK Salmonella MRA was completed in 1998 (Wooldridge, 1998, Report to EMEA) and, indeed, internationally food safety risk assessment did not start until the late 1990s.

From a review of the literature 12 UK risk assessments for Salmonella have been identified. Rather than review each MRA individually, themes of interest have been identified and discussed. From within each theme, some recommendations have been identified and these are summarised.

SID 5 (2/05) Page 7 of 19

Chapter 5: Salmonella sampling and detection methods For the preparation of this chapter a large number of scientific articles, books and technical documents have been reviewed. Key books and chapters:1 Pathogen populations on poultry farms (2005). In: Food safety control in the poultry industry., Woodhead Publishing, 1st Edition Ed. by G.C.Mead2 Salmonella in domestic animals, 2000, CAB International, 1st Edition. Edited by C. Wray and A. Wray.3 Salmonella enterica Serovar Enteritidis in humans and animals, Epidemiology, pathogenesis and control, 1st Ed., Iowa State University Press, 1999, Ed. by A. Saeed4 Salmonella isolation methods in the UK. Rob Davies. 1995, In: Proceedings of Symposium on the diagnosis of Salmonella infections, Reno, Nevada.5 Manual of Diagnostic tests and vaccines for terrestrial animals, 5th Edition, 2004, OIE.6 Microbiological food safety in animal agriculture. 1st Ed, 2003. Ed. By Torrence, Iowa State Press.7 Veterinary Epidemiologic Research, 2003, Ian Dohoo, W. Martyn, H. Stryhn

Technical documents:ISO documents1 ISO 6579:2000 (Incl. Annex D) and ISO 16140:20032 ISO 7218: Microbiology of food and animal feeding stuffs - General rules for microbiological

examinations. International Organisation for Standardization, Geneve, Switzerland3 ISO 6887-1: Microbiology of food and animal feeding stuffs - Preparation of test samples, initial

suspension and decimal dilutions for microbiological examinations - Part 1: General rules for the preparation of the initial suspension and decimal dilutions. ISO, Geneva, Switzerland

3 Isolation and identification of Salmonella from meat, poultry and egg products, USDA4 FSA carcass swabbing methods http://www.ukmeat.org5 EU protocols – EU layer survey protocol – Technical specification, SANCO/34/2004 Rev3, July 2004

Peer-reviewed papers. Pubmed search terms :

Salmonella AND swab (from 2000) Salmonella AND tissues AND isolation (from 2000) Salmonella AND hides Salmonella AND carcass AND poultrySalmonella AND carcass NOT (Salmonella AND carcass AND poultrySalmonella AND poultry AND sampling (from 2000Salmonella AND poultry AND samplingSalmonella AND poultry AND faecesSalmonella AND poultry AND poolSalmonella AND poultry AND litterSalmonella AND dustSalmonella AND drag swabSalmonella AND socksSalmonella AND bootsSalmonella AND dustSalmonella AND sampling AND abattoirSalmonella AND hatcherySalmonella AND sampling AND eggsSalmonella AND sampling AND miceSalmonella AND sampling AND ratsSalmonella AND poultry houses AND ratsSalmonella AND feed AND samplingSalmonella AND millSalmonella AND pre-enrichmentSalmonella AND enrichmentSalmonella AND universal pre-enrichmentSalmonella AND ferrioxamineSalmonella AND Delayed secondary enrichmentSalmonella AND SPRINTSalmonella AND RambachSalmonella AND mix-ELISASalmonella AND rapid methodsSalmonella AND PCR AND fecesSalmonella AND BAXSalmonella AND VIDAsSalmonella AND SPRINT

SID 5 (2/05) Page 8 of 19

Salmonella AND MalthusSalmonella AND Rabit

Chapter 6: Epidemiological investigations based on strain discrimination and typing6.1 – Molecular serotyping for Salmonella

The following bibliographical searches were performed for this review:Date 25 August 2005.Search engine PUBMEDSalmonella & molecular & serotyping 207 hitsSalmonella & DNA & serotyping 242 hitsSalmonella & PCR & serotyping 136 hitsSalmonella & array & serotyping 8 hitsSalmonella & MLST 4 hits

When these references were combined, and repeated entries discarded, a database of 368 references was created in Endnote. This list was screened visually based on the abstract information. A final list of 33 references was created for the purpose of this review. Also, a small number of selected references have been included for general remarks about Salmonella nomenclature, antigen composition, and microarray technologies.

6.2 – Sub-typingMedline search using limits “last two years (2003-2005)”Search criteria “salmonella + molecular typing” and specific searches for each new molecular typing technique discussed.

6.3 - Antimicrobial resistance typingThe following bibliographical searches were performed for this review:Date 12 May 2006.Search engine PUBMEDSalmonella & antimicrobial & resistance & typing 113 hitsSalmonella & antibiotic & resistance & typing 255 hits

The titles of these references were screened visually and selected abstracts examined; only two papers reviewed the role of antimicrobial susceptibility testing in the epidemiological typing of Salmonella; the remaining articles were mostly prevalence studies relating to specific regions, cases or time periods. Two papers commented on ‘phage typing in relation to antimicrobial resistance.

Chapter 7: Archiving of Salmonella The information quoted within this chapter derives primarily from internet- based searches using the following scientific search facility: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi The search criteria were based on the following keywords or strings of keywords.Storage methods for microorganisms Archiving of microorganismsLow temperature storage methods Survival of microorganism after storageArchiving of salmonella Survival rate of microbes after storageEffects of long-term preservation on microorganisms

Survival rate of salmonella after storage

Chapter 8: Novel techniques8.1 – DNA-based approaches to study virulence and evolution of Salmonella

The literature survey for this review was performed using search terms such as: “Salmonella + microarrays + comparative genomics”; “Salmonella + comparative genomics”; “Salmonella + evolution + genomics”; “Salmonella + subtractive hybridization”; “Salmonella + SNPs” for entries for the last 20 years in PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?DB=pubmed). The search resulted in several hundred entries of which only prominent key papers were selected for this review.

8.2 – Review of microbial proteomicsThe literature survey for this section of the review was performed using search terms such as: “Salmonella + proteomics”; “Microbial proteomics”; “Bacteria proteomics”; “Bacteria MALDI-ToF”; for entries for the last 10 years in PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?DB=pubmed). The search resulted in up to 100 enteries of which only prominent key papers were selected for this review.

Chapter 9: Methods to study Salmonella pathogenesis A literature survey was performed by interrogating entries for the last 20 years in PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?DB=pubmed) using the search terms ‘Salmonella + pathogenesis + method (1501 entries)’ and ‘Salmonella + virulence + method (133 entries)’.

01/03 – Meeting to discuss recommendations for areas of new researchA meeting was held at the VLA, Weybridge on 27.2.06 with collaborators from VLA, HPA and IAH to discuss recommendations. Minutes are attached. Authors presented an outline of each section of the review and their key

SID 5 (2/05) Page 9 of 19

recommendations. Collaborators commented on each section and discussed additions and amendments. Recommendations were discussed and a process for prioritising and circulating a complete list for comment was agreed. Recommendations are given in each chapter and a prioritised list is given in Chapter 10.

PEER-REVIEWExperts from the HPA and the VLA were identified as potential peer-reviewers for each section of the project and are listed in the meeting minutes of 27.2.06. All sections received comments which were incorporated into the final versions submitted with this report.

STUDY VISIT TO DENMARKThe aims of the study visit to the Danish Zoonoses Centre and EpiLab were discussed.

Two veterinarians and one scientist from the Centre for Epidemiology and Risk Analysis and the Food and Environmental Safety Departments at the VLA visited Denmark on 9-10 March 2006. Visitors and hosts presented information on their work at their respective Institutes. The itinerary for the visit is attached as is a brief report on the outcome. Information from the visit has been incorporated into the review.

01/04 – Draft final literature review and recommendations for final reportReviews of each aspect will be collated and recommendations written for submission in the final report to Defra.

REVIEW AND RECOMMENDATIONS

The key findings from the chapters are summarised below.

Chapter 1: Veterinary Salmonella surveillance and outbreak investigations Salmonella surveillance in GB is supported by legislation, which makes mandatory the reporting of Salmonella isolations from statutory species, their environment and animal feeding stuffs. The role of VLA in Salmonella surveillance in animals is receipt and distribution of Salmonella reports, laboratory testing, data management, analysis and reporting and investigation of Salmonella incidents. Salmonella surveillance data provided to and collated by the VLA include currently numerator data only, but it is one of RADAR’s aims to collect and collate data from various sources, including denominator data and enable access of researchers to such data. Isolation methods according to ISO 6579:2002 are used for most samples tested for Salmonella. Serotyping and ‘phage typing is performed with the same techniques (‘phage typing with the HPA Colindale scheme and serotyping with the Kauffmann-White scheme) used in Salmonella surveillance in humans. Every isolate from one Salmonella incident report is tested in vitro for its sensitivity against a panel of 16 antimicrobials using a disk diffusion technique on Oxoid agar with antibiotic disks. Standardised definitions are used for the reporting and an e-mail alert system, which is in place for Salmonella from animals in GB. On farm investigations are carried out when isolates of higher public and/or animal health importance are reported. A standardised protocol to support these investigations is planned to be developed in the near future. A voluntary monitoring and control plan for Salmonella in finisher pig herds supplying pigs to quality assured abattoirs (Zoonoses Action Plan scheme) is in place in GB. A similar Salmonella monitoring programme for slaughter pigs in N. Ireland applies to all pig herds and slaughterhouses. Statutory monitoring of breeding flocks of domestic fowl for the serovars Salmonella Enteritidis, Hadar, Virchow, Infantis and Typhimurium is required and control measures in the event of an infection are rigorous. Flocks infected with S. Enteritidis or S. Typhimurium are slaughtered. Farmers with flocks infected with other serovars of public health significance are provided with advice on control of infection and are required to implement a farm action plan. The recent literature related to the Salmonella monitoring and control systems in other countries within Europe, such as Italy, Spain, France, Austria, Denmark, Norway and Finland, but also for Canada, was also reviewed in order to seek examples of best practice. Furthermore, international networks which enable data exchange, outbreak detection and Salmonella prevention and control, such as SALM-NET, MedVetNet, Global Salm-Surv and EPIET.

The need for use of mathematical modelling for evaluation and review of the surveillance system for Salmonella in GB was considered to be of top priority, as the result of the review, in order for routine surveillance activities to be enhanced. Other proposals included the provision of industry incentives to improve Salmonella control and the improvement of targeted Salmonella surveillance, according to species, production type, production level and clinical illness. The other recommendations, which resulted from this part of the review are included in the section of “Cross-cutting recommendations” in the epidemiology part of the report.

Salmonella is considered endemic in many sectors of the livestock industry and control, therefore, is generally focussed on reducing disease to predefined levels rather than all out eradication. Thus, when considering an outbreak of Salmonella in livestock, the term usually refers to the isolation of a rare or emerging strain or a strain that poses a significant threat to human or animal health. Under the Zoonoses Order 1989 all UK isolations of Salmonella from animals, birds, their carcases, products or surroundings and animal feed stuffs must be reported, there is no mandatory obligation to investigate reports of Salmonella in pigs or cattle in the UK unless they pose a significant threat to public health, including investigation of new and emerging strains in order to avoid these

SID 5 (2/05) Page 10 of 19

strains becoming established in the national livestock population (for example MDR S. Newport and MDR S. Java). Investigations may take the form of visits, laboratory examinations, or telephone/verbal enquiries and often the nature of a visit is left to the discretion of the officer. For investigations of Salmonella in livestock that are associated with human disease outbreaks, the medical or environmental health authorities may request the assistance of Animal Health (AH - formerly the State Veterinary Service) in investigating an outbreak where food of animal origin is suspected as a source. There is at present no standardised on-farm protocol for investigating outbreaks of Salmonella in humans traced to an animal source.

An important component of outbreak investigation is the collection epidemiological information about the disease to enhance our understanding about spread and control. Data collected from Salmonella incidents in GB are are entered into the VLA networked Farmfile database which contains baseline data on all VLA submissions and enables better linkage of relevant data to improve the quality and timeliness of information and allow enhanced epidemiological analysis of the data. In the UK, the information management system RADAR (Rapid Analysis and Detection of Animal Related Risks) provides researchers with access to veterinary surveillance data but lacks denominator data on livestock populations other than the cattle and poultry populations.

Recommendations specific to this section of the review included improved systems for recording and tracing of movements of animals and animal products to enable food-borne outbreaks in humans to be traced back to specific farms, and of less importance the increased analysis of epidemiological information gathered during Salmonella outbreak investigations.

Chapter 2: Humans: surveillance methodology; national/international detection, investigation and control of outbreaks; national/international databases; electronic exchange of data; design of surveys.

This Chapter describes the surveillance and control activities for human salmonellosis undertaken at national level by the Health Protection Agency (HPA), including:

National reporting mechanisms for cases of human salmonellosis and databases held at HPA Standard laboratory tests performed for strain isolation and identification Enhanced surveillance including general outbreaks of Infectious intestinal diseases surveillance and food

studies. Outbreak detection mechanisms and investigation of outbreaks including study design Processes for disseminating information on surveillance and control activities

The HPA is also involved in a number of international projects, relating to human salmonellosis. The international network for enteric infections caused by salmonella and VTEC O157 (Enter-net), is coordinated from the HPA and involves the participation of 34 countries worldwide. International surveillance has been enhanced since Salm-gene, a 3- year research project involving the participation of nine countries in Europe including England & Wales and Scotland. That study (2001-2004) was established to assess the added value of using molecular subtyping for the detection of outbreaks and involved the standardisation of PFGE subtyping and analysis in numerous countries. Salm-gene has in addition to enhancing international surveillance of human salmonellosis, laid out the foundation for MedVetNet (2004-present), which is an EU-funded surveillance project integrating veterinary, medical and food science. Through the HPA, England participates in the PulseNet Europe network, which is one of the workpackages of MedVetNet. PulseNet EU builds on the experience of PulseNet USA, which is a national surveillance network based on PFGE for all human salmonellosis. Other programmes the HPA participates in include (a) the European Programme for Intervention Epidemiology Training, which is coordinated by the European Centre for Disease Prevention and Control. This programme provides training and practical experience in intervention epidemiology with the aim of developing a European network of intervention epidemiologists (b) WHO global Salm-surv, establish to enhance laboratory based surveillance of Salmonella worldwide (c) the Early Warning and Response System, which aims to alert public health authorities in Member States and the Commission on international outbreaks.

Chapter 3: Early detection systemsEarly detection systems aim to alert as soon as possible increases in adverse health outcomes as they are occurring. This is achieved by using historical data to estimate what would be expected in the current time period. A decision is made as to whether an outbreak, for example, is occurring by comparing the currently observed number of disease reports to an upper limit or threshold value. These systems are typically automated using a single robust statistical algorithm that summarises the main components of the system namely, the data, the prediction element, the alarm signal, and the output.

Early detection systems have been implemented, world-wide, for the early and rapid detection of increases of Salmonella and other pathogens in both human and animal health surveillance. More specifically, systems have been implemented in Australia, England and Wales, and the United States. These and other systems have used a variety of statistical approaches including time series, regression analysis and one-sided Cumulative Sum. The majority of systems use five years historical data to provide an estimate of what is expected to occur in the current time period, either a week or month, and this forms the basis for calculating the threshold value. Furthermore, the systems vary in terms of their assumptions regarding the surveillance data (either Poisson or

SID 5 (2/05) Page 11 of 19

Normally distributed), whether seasonality, past outbreaks and temporal trends are accounted for and the number of pathogens included within the model.

In addition to the currently applied methods, there are a wide range of other statistical techniques that can be used for prospective analysis of surveillance data, including Salmonella. These include the Zero-inflated Poisson model, Scan statistics, sets technique and Hidden Markov Chain models. Each of these methods have different underlying assumptions, disadvantages and advantages which impacts on when they are most appropriate to use and what their output will inform users about the data. Indeed, no specific method appears to be more beneficial than any other in terms of its application for a Salmonella early detection system and the most appropriate method to use would depend upon the criteria of the system and the characteristics of the data used. Many of these characteristics or attributes are common to both public and animal health surveillance data (i.e. definition of outbreak, reporting delays, presence of past outbreaks, trends and seasonality, coverage and submission rate). However, additional issues specific to Salmonella animal health data include the epidemiological unit, reason for sampling and increased number of species represented in the data. Consequently, in applying a statistical model to animal health surveillance data, further considerations need to be accounted for in determining which statistical method is most appropriate.

Based on the review, it is apparent there are numerous data issues that infringe on the ability to develop an early detection system for Salmonella in both public and animal health surveillance. Not all of these issues can be addressed within a single algorithm. For example, the submission rate is known to vary over time in both public and animal health. However without accurate denominator data, an underlying assumption is often made to address this issue, namely that the submission rate is constant over time and representative of the underlying affected population. These and other assumptions made need to be borne in mind as well as a general understanding of the data when interpreting the outputs from any early detection system. Given this, a detection system is an additional tool for enhanced surveillance and to be of benefit an underlying network of good communication is required to ensure that local knowledge of specific increases in Salmonella are communicated between public and/or animal health officials.

Chapter 4: Analysis, modelling and risk assessment4.1 – A Review of the application of spatial analysis and Geographic Information Systems in animal and human Salmonella research

A literature review of the use of spatial analysis and GIS in infectious diseases of humans and animals showed that many spatial techniques were designed in the 1950’s for use in botany. The techniques were adapted to be used for human epidemiological analysis and only recently have these been adapted once again to try and analyse animal information realistically.

The most common uses of spatial statistics are for mapping and cluster analysis. The mapping of disease rates or occurrence is an invaluable tool in presenting findings, although there are many inherent problems. The denominator, the level of reporting in a particular area and missing data, all significantly alter the interpretation of a map. Methods have been designed to ‘smooth’ data using either prior data or weighting the data from neighbouring areas. Smoothing the data removes the fluctuations of results, especially in low population areas or when dealing with rare diseases, but utilises statistically complex methods and can create more severe adjustments than one would expect. Modelling techniques have also been used to smooth the data and adjust for confounders. Maps and models have also been used to target risk analysis and to monitor surveillance data, alerting any significant increases in rates, or to show the spread of an outbreak.

There are many types of cluster analysis test, which look for groupings of heterogeneous occurrences. Some tests look at spatial or temporal situations or both, some can incorporate denominator data, changes in time or confounder information. A mixture of tests and techniques need to be used to suit the data and the hypothesis being tested. The Scan statistic is very useful in showing where the most likely clusters are, and can incorporate the distribution of the population at risk. The other space-time interaction tests are more useful in calculating whether there is clustering, but not indicating where. The Knox test has the advantage of indicating the smallest distance and time frame with which the data is clustered. Kulldorf’s scan statistic, K-nearest neighbour and the Knox test allow clusters to be analysed without knowledge about the population at risk which would be beneficial for Salmonella surveillance projects where ‘control’ information is incomplete.

The recommendations are that the HPA and the VLA improve their in-house skills and, where appropriate, add spatial analysis components to future project designs and proposals. Cluster analysis and improved mapping of data should be used routinely, and with improved spatial data, the modelling of spatial factors should take place.

4.2 – Mathematical modelling for Salmonella Existing mathematical models representing the transmission of Salmonella within-herds and between-herds were reviewed to determine areas for further work. Models have been developed for pigs, cattle and poultry, and form key parts of risk assessments and investigations into effective control measures. For within-herd models, a key area for further work for models applicable to all species is on refinement of parameter estimates of transmission rates and incubation periods, since these estimates are main drivers of the outputs of the models. Therefore, it is

SID 5 (2/05) Page 12 of 19

vital that modelling is used to inform experimental design at an early stage. The resulting data from current or new studies could be combined with existing data and state of the art data analysis methods applied to refine parameter estimates. For between herd transmission of Salmonella, there was little existing work in the literature, with the only published work focusing on control measures for Salmonella in broilers. A key area for further work is the application of network analysis techniques to recently collected data to explore the potential for Salmonella transmission through the poultry industry, and the evaluation of the effectiveness of current control measures. Similar methods could also be applied to pig movement data, which is available.

4.3 – Microbial risk assessmentMicrobial risk assessment (MRA) is a scientific tool that can be used to evaluate the level of exposure and the subsequent risk to human health relating to a specific hazard, such as Salmonella. Since the late 1990s such techniques are becoming increasingly used within the area of food safety because they facilitate scientific investigations of food related risks including quantification of uncertainty and prioritisation of control strategies. Although many countries have developed MRAs for this hazard, here, only the UK MRAs will be considered. This is because the risk question and the circumstances under which a MRA is commissioned must always be considered carefully when evaluating and comparing any MRAs. Other countries may have different emphases, different structures and different Salmonella problems, which makes the consideration of these MRAs more difficult and certainly means that comparison to the UK MRAs would be highly complex.

In this review, a total 12 UK Salmonella MRA (4 qualitative; 8 quantitative) were identified and reviewed. Common themes of interest were identified and discussed, such as methods used, data gaps, modelling the pre-harvest phase and making the most of MRA. As a consequence of these discussions, 7 key recommendations have been put forward; these are:

Further investigation of mathematical methodologies and/or IT development to fully utilise the data available, to better characterise uncertainty and to decrease model runtime.

Consideration of EU and third world (non-EU) trade in MRA Higher utilisation of the iterative nature of MRA. Generation of data identified by data gaps within the

previous MRA, which is then updated. Inclusion of pre-harvest models in farm-to-consumption MRA to allow investigation on the impact of on-farm

controls on public health. Adopt a holistic approach to MRA, where multiple hazards (including animal health pathogens) and controls

can be considered. Establish MRA methods/tools that can deliver quality estimates of risk in a short period of time Use of MRA to ‘drive’ research projects. Entire project focused around the MRA, which is fully integrated into

the project. Available budget for not only re-running the model with the new data but for other improvements for the MRA.

In conclusion, MRA has made significant progress since its emergence in the area of veterinary public health; however its use is not yet being optimised in the UK and the recommendations suggested above should enable MRA to help it fulfil its potential to support evidence-based policy making.

Chapter 5: Salmonella sampling and detection methods The detection of infection or contamination with Salmonella is done in a variety of scenarios that may involve, for example, living populations, the environment or livestock premises. In each case, the suitability of a methodology depends on a range of factors including the type of matrix, accompanying micro-flora and presence of inhibitory substances. The choice of statistically representative samples should also be determined in environmental sampling. Monitoring Salmonella in the end products of the food chain usually requires the development of rapid techniques, or specific refinements over traditional isolation methods. For detection of Salmonella in poultry houses and feedmills, environmental sampling is preferred, but each situation requires development of its own method. Traditional culture methods are well established in many laboratories, but may be labour intensive due to the necessity of checking for multiple false positive colonies. New chromogenic media should be investigated using a range of field samples. New ‘rapid’ methods should be validated by comparison with existing, well established methodologies. A reasonable level of standardisation has been achieved with some cultural isolation techniques (e.g. ISO 6579:2002), but new rapid molecular technologies (e.g. PCR) still require a great deal of standardisation across laboratories to be used in routine monitoring for the analysis of veterinary samples. The main priorities for research are: 1. To determine sensitivity of and optimise sampling and testing methods for establishing bacterial load throughout the food chain. 2. To harmonise, develop and evaluate rapid combined molecular detection/typing methods throughout the food chain in collaboration with commercial companies who are leading in this area.

Chapter 6: Epidemiological investigations based on strain discrimination and typingNumerous subtyping methods have been described for salmonella over the last two years, however there have been few published reports comparing these to established methods and few descriptions of application to both clinical and veterinary samples. It would seem that existing methods especially serotyping are still a requirement, and PFGE is still the molecular method of choice for subtyping in many laboratories.

SID 5 (2/05) Page 13 of 19

For many new molecular methods a more detailed knowledge of the genome would be beneficial, therefore ongoing efforts for whole genome sequencing may help in choosing genes, especially for methods such as MLST and VNTR where specific information regarding the genetic diversity may benefit further method development. The development of reliable Salmonella serotype-specific probes as suitable molecular markers through a thorough phylogenetic analysis of this genus is still required, and in progress in several laboratories. Molecular serotyping could be an alternative to conventional serotyping if such specific regions could be located to allow identification of at least the most relevant serotypes.To facilitate improved collaborations between the Veterinary Laboratories Agency and the Health Protection Agency in order to promote better surveillance, there needs to be a drive towards the appropriate testing and selection of techniques, which offer improved discrimination, ease of use, data sharing and which also offer substantial potential advantages to subtyping already in use.

6.3 - Antimicrobial resistance typingThe patterns of antimicrobial susceptibility of Salmonella isolates to panels of antimicrobials can provide a useful means of sub-dividing strains occurring within a serotype; in some circumstances susceptibility to single antimicrobial compounds provides useful strain data. Such patterns have been used for many years to provide information on the epidemiology of strains with a particular resistance phenotype, the emergence of new (or previously unrecognised) strains and the development of new types of resistance in existing strains. Susceptibility tests are commonly performed for therapeutic purposes and secondarily provide data which is useful for epidemiological purposes. A number of published methods are available for the detection of the antimicrobial resistance phenotype in Salmonella and these are generally robust, reproducible and relatively cheap. When choosing antimicrobial susceptibility testing methods considerations may include the comparability of medical and veterinary results within a geographical area, or the comparison of medical or veterinary results between different geographical areas. Because of the multiplicity of susceptibility testing methods in use in Europe, these considerations will strongly influence the test methods selected. The stability of antimicrobial resistance characteristics in a given strain may be important; resistance plasmids or genes may occasionally be lost or alternatively, at the level of individual genes, non-functional or not expressed. Identification of specific resistance genes at the molecular level can also assist in discriminating between strains of Salmonella, particularly in situations where several different resistance genes can confer resistance to the same antimicrobial. The resistance genes that are detected in Salmonella isolates of human and animal origin are commonly identical; in these situations, the flanking genetic sequences, organisation of the resistance genes or the occurrence of associated mobile genetic elements can be very useful in tracing the evolution and epidemiology of strains. HPA and VLA have already completed ring trials in which the susceptibility of Salmonella strains was compared. On the whole, the results have shown a very good correlation between the laboratories concerned for most antimicrobials. For some types of resistance, for example resistance to third generation cephalosporins or to fluoroquinolones, then the full significance of the resistance to animal and public health cannot properly be assessed without molecular characterisation of the resistant strain. Antimicrobial susceptibility results may be analysed in conjunction with other types of strain-typing data and often provide an excellent preliminary filter for the detection of strains of interest.

Chapter 7: Archiving of Salmonella The aim of this archiving review is to present a comprehensive overview of current literature relating to advocated storage conditions and the recognised effects of long term storage on the quality of isolate archives. Additionally, the review seeks to increase awareness of current microbiological archiving issues within the scientific community, with specific reference to Salmonella. The considerations and commitments of setting up such a resource are considered along with the implications for the scientific community of not acknowledging the necessity for such facilities. The document also details several critical issues that would need to be considered prior to establishment of such a facility (WFCC Standards Committee, 1990). The two main recommendations are the establishment of a formal, centralised microbiological archive facility and assessment of new preservation approaches or the development of optimal procedures for preservation of sensitive groups of isolates.

Chapter 8: Novel techniques8.1 – DNA-based approaches to study virulence and evolution of Salmonella

Salmonella are human and animal pathogens responsible for enteric and typhoid-like disease worldwide. The traits possessed by S. enterica strains capable of zoonotic or epidemic transmission are poorly characterized, as are the differences in their virulence determinants and mechanisms which allow colonization of differing hosts. Advances in molecular tools such as sequencing have helped to study the genome of such pathogens in detail, although this is limited only to the sequenced organism. Techniques such as microarray have been a useful adjunct to establish the evolution of these pathogens, the core genome, and genomic similarities and differences which exist within different serovars and populations of field and clinical isolates. Virulence and other differentiating genes, identified using comparative studies, can in future be included in miniaturised arrays that are suitable for high through-put use. The use of such techniques in routine diagnostics would provide a wealth of information on the genetic variation that exists within clinical isolates, which may be correlated to disease. However, the limitation of microarrays is that it is based on information gathered only from genes present in sequenced strain(s). Although this information is ever increasing due to greater numbers of strains being sequenced, there will always be a gap between sequence information and the natural variability of strains.

SID 5 (2/05) Page 14 of 19

Therefore, subtractive hybridisation and diversity arrays, which give information about genes unique to one organism, are useful in complementing microarray data. Polymorphism at the gene level can also be useful to distinguish between gene functionality and be used as diagnostic markers. Techniques such as pyrosequencing and dHPLC have been used as high throughput techniques for rapid identification of single nucleotide polymorphism.

8.2 – Review of microbial proteomicsThe rapidly emerging post-genomic technology of proteomics has been reviewed with specific applications for Salmonella typing down to the strain level. Proteomics measures protein expression and is a necessary level of analysis as several transcriptomic studies have shown a poor correlation with mRNA. Early proteomics studies were based on analysis by 2D-gel electrophoresis but various limitations with this technology have driven the development of automated HPLC-mass spectrometry procedures. Whole proteome analysis of enteric bacteria including Escherichia coli and Salmonella enterica have been successfully achieved by HPLC-mass spectrometry. The main utility of such research analyses is the detection of biomarkers for the development of rapid phenotypic tests. Clinical laboratory methods, based on Matrix Assisted Laser Desorption Ionisation Time of Flight (MALDI-ToF) mass spectrometry of intact bacterial cells, have been developed for microbial speciation and have potential for the rapid detection of certain types of antibiotic resistant microbes (e.g. MRSA) and multiplexed phage typing. This approach, termed 'intact cell MALDI' (ICM), is highly amenable to a high throughput environment of a clinical laboratory and the Health Protection Agency at Colindale has been instrumental in developing this technology. The sensitivity of ICM is around 100,000 organisms and may be further enhanced by phage amplification.

8.3 – Phenotype screening for Salmonella spp. It is important to understand the phenotypes and behaviours of bacteria. However, traditional biochemical methods are only suitable to study a small number of traits of bacteria. A new technique, Phenotype Microarray (PM) has been used to study a larger number (~2000) of traits of bacteria in microtitre plates. This method has been successfully used to identify phenotype differences among closely related Salmonella isolates. It has potential in the identification of phenotype markers; however, the high cost prevents it from being used on a large number of isolates.

Chapter 9: Methods to study Salmonella pathogenesis Chapter 9 reviews methods used to study Salmonella pathogenesis, with an emphasis on approaches that may inform the development of novel diagnostic tools or strategies for control of Salmonella in humans and farmed animals that are not reliant on antibiotic use. A pressing need exists to define the molecular mechanisms underlying the ability of Salmonella enterica to colonise the intestines of humans and food-producing animals, invade epithelia, induce enteritis and translocate to systemic sites. Furthermore, the traits possessed by S. enterica strains capable of zoonotic or epidemic transmission, as wells as strains differing in virulence and host range, are not understood.

The review considers models to study Salmonella pathogenesis and explores the relative merits of studies in food-producing animals versus surrogate rodent or cell-based assays. Approaches to study the function of Salmonella genes are described (including targeted, random and signature-tagged mutagenesis, novel microarray-based methods and cloning of virulence factors by gain-of-function), together with methods to detect the expression of Salmonella genes (transciptomics, proteomics, reporter fusions and methods to identify in vivo-induced genes). The value of DNA-based approaches to dissect the virulence and evolution of Salmonella by analysis of the repertoire, sequence and conservation of bacterial genes is reviewed separately under the section on Novel Techniques.

Owing to constraints of space, only key methods are listed together with their value to DEFRA and other stakeholders, gaps in knowledge and recommendations for future research on Salmonella pathogenesis, where appropriate.

Chapter 10: recommendationsRecommendations from each main chapter of the review have been prioritised as follows:

Priority A – those areas for immediate developmentPriority B – those areas which are important but not a top priorityPriority C – those areas for development after A and B have been addressed

Priority A recommendations are summarised here.

SID 5 (2/05) Page 15 of 19

Chapters 1-4 (Epidemiology, Surveillance, Early Detection Systems, Analysis, Modelling & Risk Assessment Collect data: library of spatially referenced information and access to denominator data such as livestock

populations. Existing data from databases could be cross-referenced and cleaned to produce lists of site locations, map references and layout shape maps.

Harmonise both phenotypic and molecular Standard Operating Procedures (SOPs) for the isolation, identification and typing of Salmonella across veterinary and human surveillance, nationally and internationally to allow data to be shared.

Develop SOPs between the HPA and VLA for investigating incidents, such as an unexpected increase in reports of a particular strain of Salmonella

Develop SOPs for collaboration between VLA and HPA and other appropriate agencies during international, national or local outbreak investigations. For example, improve liaison with local and health authorities in trace back of human outbreaks/ development and agreement of protocol for such investigations; ensure regular contact with the Food Standards Agency in relation to possible food-related outbreaks where a product withdrawal may be necessary.

Improve timeliness of data on Salmonella surveillance exchange between VLA and HPA. Currently this is done annually through the Salmonella in Livestock Production Annual report and the UK Zoonoses Report.

Access to denominator data, such as details of animal populations and design of appropriate databases to store it in. The outputs of veterinary surveillance will thus improve. Encourage improved standardized data collection, collation and incorporation of denominator data in analysis (e.g. early detection system) and dissemination of surveillance information to motivate surveillance participation.

Statistically structured epidemiological investigations (including outbreak investigation protocols) and statistically valid samples. Also, enhanced surveillance of common Salmonella serotypes by routine use of molecular typing. This can aid in distinguishing between sporadic infections and outbreaks.

Harmonisation with the Early Detection System used for Salmonella in the human health sector with regards to development of early warning system for emergence of multi-drug resistant Salmonella epidemic strains. Enhancements in the Salmonella Early detection system of emerging serovars in animals in liaison with HPA

Modelling to inform data collection and experimental design at the planning stage.

Chapter 1: Veterinary Salmonella surveillance and outbreak investigations Use of mathematical modelling for evaluation and review of the surveillance system for Salmonella in GB.

Chapter 3: Early detection systems Increase harmonisation and communication between HPA and VLA on early detection outputs and

methods Research into the inclusion of a denominator within the system to address the issue of changing

submission rates over time; this would include research into identifying the most appropriate denominator to use

Examine ways to encourage increased submission of samples from domestic livestock for clinical illness associated with Salmonella; this will ensure that the underlying ‘sick’ population is better represented in the system

Further research into the application of statistical methods to data with many zeros in the historical dataset in order to detect slow increases, over time, of a potentially new emerging serotype; This recommendation will be addressed within Defra project ED1039.

Research into the best method for adjusting for reporting delays.

Chapter 4: Analysis, modelling and risk assessment The use of techniques to detect outbreaks and monitor new and emerging strains should be implemented

into study proposals, such as the Salmonella Newport surveillance project. Use cluster analysis as a routine tool for detecting outbreaks, and testing specific hypotheses. When

combined with GIS mapping, cluster analysis could be used to explore disease occurrence data and ‘focus’ risk factor analysis e.g. clusters linked to particular feedmills

To meet these needs, the departments at the VLA and HPA need to improve in-house spatial skills to help implement these concepts into project design and to carry out spatial-temporal analyses.

Apply statistical models to the estimation of test sensitivity for the current diagnostic tests. Work is underway to estimate the sensitivity of pooled and individual faecal samples for detection of salmonella in pigs (OZ0323) and these methods could be extended for use for other tests/other species.

Higher utilisation of the iterative nature of MRA. Generation of data identified by data gaps within the previous MRA, which is then updated. This is already being done to some extent but primarily at the farm level.

Inclusion of pre-harvest models in farm-to-consumption MRA to allow investigation on the impact of on-farm controls on public health. Already being done for Salmonella in pigs but could be extended to other food-animals.

SID 5 (2/05) Page 16 of 19

Use of MRA to ‘drive’ research projects. Entire project focused around the MRA, which is fully integrated into the project. Available budget for not only re-running the model with the new data but for other improvements for the MRA.

Chapter 5: Salmonella sampling and detection methods To develop optimal but practical and economic standard sampling programmes for the various stages of

the food chain and to determine their sensitivity and relevance in terms of the epidemiology of the organism and its public health risk.

To develop improved external quality assessment (EQA) for controlling test quality in Salmonella–approved laboratories by means of standardised realistic ring- trial samples and to identify deficits in testing carried out by commercial laboratories.

To investigate sentinel process and waste monitoring surveillance in focal operations which are involved in processing large volumes of potentially contaminated material such as feedmills, abattoirs, livestock markets, sewage plants etc. to detect new and emerging strains.

To evaluate real time PCR and other new rapid methods (e.g. BAX) for the detection and enumeration of Salmonella, especially for test and release of feed and breeding stock.

Chapter 6: Epidemiological investigations based on strain discrimination and typing

Reliable markers exist for most of the commonly encountered serotypes, if a strategy for molecular serotyping could be developed it would enable faster identification, as well as considerable savings (by means of a possible rationalisation) in comparison to conventional serotyping. Serotype-specific PCR is a valuable alternative

Needs to be a drive towards the testing of subtyping methods on both clinical and veterinary samples; most of the literature describes new methodologies in either a clinical or veterinary context

Improved harmonisation and standardisation of methods needs to be achieved in the HPA and VLA, including all aspects such as SOPs and the data analysis

The promotion of subtyping techniques, which are amenable to electronic storage. e.g. MLST or VNTR, following suitable evaluation. This would facilitate better collaborations and instigate the development of shared databases

Continue to perform susceptibility tests on isolates from “new incidents” of Salmonella infection as a preliminary filter to identify isolates of particular importance by:

- resistance pattern (eg S. Typhimurium DT 104)- unusual resistances (eg MDR S. Newport resistant to third generation cephalosporins).

(This is particularly important in cases where follow-up action may be required, because susceptibility testing is generally performed relatively quickly after primary isolation.)

In some cases, quantitative susceptibility monitoring by determination of MIC may be required, for example to determine the degree of resistance to fluoroquinolones.

For detailed characterisation of selected isolates, determination of MIC by microtitre broth microdilution using plates pre-loaded with antimicrobials, in conjunction with molecular investigation using gene profiling microarrays, is likely to prove useful.

Targeted surveillance could be based on emerging problems reported outside the UK to detect new and emerging conditions; this could include sentinel surveillance of imported products.

Denominator data is important in relating the emergence of “new” resistant strains to the intensity of sampling. The system of defining incidents partly addresses this (VLA limits “incident” to isolations occurring within 30 days of primary isolation).

Where organisms occur at low prevalence it is not always possible to collect sufficient numbers of samples for statistical considerations to be valid. In this case, for important serotypes, the available results should be reported with appropriate caveats.

The core sample types collected during on-farm investigations should be similar in order facilitate comparison between different investigations.

The emergence of epidemic resistant strains of Salmonella should be compared with the general mathematical models of resistance spread, to see how well these models also apply to this situation.

Horizon-scanning for new epidemic serotypes is currently undertaken by the HPA and VLA for the Defra Antimicrobial Resistance Co-ordination Group on an ad hoc basis. It is important that horizon-scanning is maintained.

Chapter 7: Archiving of Salmonella

Establish a formal, centralised microbiological archive facilityo Strict guidelines need to be established to define the contents of the archive. The guidelines should be based

on the validated importance of the material/ isolates and sound expert judgment for their retention. o Start-up costs and maintenance costs must be agreed prior to inception.

SID 5 (2/05) Page 17 of 19

o Well-designed database implementation is required to ensure good searchability. Well-designed databases will facilitate good traceability.

o The envisioned life-span for the archive must be defined.o Importantly, we need to consider the potential user of material/ isolates archives and identify their

requirements, stakeholder input pre-inception is very important. Assess new preservation approaches or the development of optimal procedures for preservation of sensitive

groups of isolates.

Chapter 8: Novel techniques

The routine use of existing validated antimicrobial miniaturised diagnostic microarrays on Salmonella field and clinical isolates submitted to the VLA Enteric Bacteriology Reference Unit. This will result in a database of epidemiological information which would be useful in understanding the diversity of antimicrobial resistance determinants which exists within Salmonella isolates from different serovars in the UK. This information can then be correlated to host animal, farming methods and other niche specific adaptations which may have occurred in these pathogens, and ultimately be useful for risk analysis and control measures.

PROPOSALS FOR FUTURE WORKA key aim of this work was to provide a rational basis for the harmonisation and enhancement of routine diagnostic and surveillance activities related to Salmonella in the VLA and the HPA. The recommendations indicate the work required to achieve these aims. It is hoped that those developments identified as priority A will attract funding for implementation.

The EU Zoonoses Directive 2003/99 and Regulation 2160/2003 sets out requirements for monitoring and control of zoonotic organisms particularly Salmonella and currently a series of surveys is underway to determine prevalences of Salmonella in different livestock sectors and to set targets for future improvements. These improvements will be brought about through National Control Programmes. At the time of writing this review, the impact of the EU Zoonoses Regulation could not be fully addressed as control programmes had not been finalised. It is further recommended, therefore, that the recommendations in this report are subjected to a limited review later in the year in the light of the EU Zoonoses Regulation and UK National Control Plans.

The scope for publications from this work has been considered and the following are being pursued:

From chapter 5, a review of sampling techniques. From chapters 8 and 9, a review article or book chapter relating to application of novel tools to inform our

understanding of Salmonella pathogenesis and discussion of future perspectives. Report in the CDR Weekly

In addition, Defra will be asked whether it wishes to have an editorial in the Veterinary Record to coincide with publication of the report on the Defra website.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

PUBLICATIONS FROM THE REVIEW

Bacteriological detection of Salmonella Enteritidis in eggs: A review by J.J. Carrique-Mas & R.H. Davies. Scientific and Technical Review of the OIE. Volume 27 (3), December 2008

Sampling and bacteriological detection of Salmonella in poultry and poultry premises: A review by J.J. Carrique-Mas & R.H. Davies. Scientific and Technical Review of the OIE, Volume 27 (3), December 2008

Muna Anjum (who co-authored Chapter 8) has written a chapter on Salmonella - Revealing the mosaic nature of Salmonella genomes, in: "Salmonella pathogenesis, molecular typing and epidemiology" for Horizon Press edited by Mikael Rhen, Duncan Maskell, Pietro Mastroeni and John Threlfall. The book includes a section on “Use of DNA microarrays for diagnostics” which formed part of her review. John Threlfall has also co-authored a chapter in this book on various aspects of antimicrobial resistance, many of which incorporated various facets discussed in Chapter 6 of the report.

SID 5 (2/05) Page 18 of 19

SID 5 (2/05) Page 19 of 19