Prof S O’Brien
Mr J Bassett
Dr D Brown
Mrs V Buller
Prof J Coia
Mrs R Glazebrook
Dr R Holliman
Prof T Humphrey
Mr A Kyriakides
Mr P McMullin
Dr S Millership
Mrs J Morris
Prof P Williams
Ms L Redmond (FSA)
Mr S Wyllie (Defra)
Officials in attendance:
Professor Colin Blakemore (Chair of GACS)
Mr Robert Martin (FSA)
Mr Andrew James (FSA)
Miss Lesley Larkin (Defra)
Dr P Cook (Scientific Secretary)
Ms G Hoad (Administrative Secretary)
Mr A Adeoye
Ms S Butler
Dr D Cutts
Invited experts and contractors:
Dr Chris Low (SAC)
Ms Katherine Harris (FERA)
Mr Ian Barber, Watton Produce Ltd
Ms Sally Barber, British Retail Consortium
Mr Steve Batchford, Brakes
Dr Roy Betts, Campden & Chorleywood Food Research Association
Ms Fiona Brookes, Northern Foods
Dr Susan Campbell, Defra
Mr Keneth Chinyama, Food and Drink Federation
Ms Amanda Cryer, British Egg Information Service
Ms Kaarin Goodburn, Chilled Food Association
Ms Audrey Harris, Food and Environment Research Agency
Dr Helen Kendall, Newcastle University
Dr Samantha Kirk, Tesco Stores
Mr Tom Miller, National Consumer Federation
Mr Rick Pendrous, Food Manufacture Magazine
Mr Alan Procter
Dr Helen Rees, ADAS UK Ltd
Mr Bernard Rowe, Consultant
Dr Anna Whyte, FSA Scotland
1. Chair’s Introduction
1.1. The Chair welcomed ACMSF Members and members of the public to the 71st meeting of the Committee. The Chair introduced Ms Geraldine Hoad who had taken over from Dr Lucy Foster as Administrative Secretary and Dr Darren Cutts who had also joined the Secretariat. She thanked Mr Bob Martin for his work covering the role during the interim period. She also welcomed Professor Colin Blakemore, Chair of the FSA’s General Advisory Committee on Science (GACS) who would be presenting item 6, Miss Lesley Larkin (Defra) who would be presenting agenda item 7, Dr Chris Low (Scottish Agricultural College) who would be presenting agenda item 8, and Ms Katherine Holmes (Food and Environment Research Agency) who would assist in presenting agenda item 9.
2. Apologies for absence
2.1. Apologies for absence had been received from Prof Paul Hunter.
3. Declarations of interest
The Chair reminded Members of the need to declare any conflicts of interest relating to items on the agenda, both at the beginning of the meeting and subsequently if the need arose during discussions. Prof Humphrey declared that he provided consultancy to the British Egg Industry Council with reference to an item he wished to raise under Any Other Business. Prof Coia declared he provided consultancy advice to Tesco. Mr McMullin declared an interest in items relating to Campylobacter and Salmonella as he provides a consultancy service to the poultry industry and also advises the British Egg Industry Council. Mr Bassett declared his employer produced many ambient stable foods which might have a bearing on discussions. Mr Kyriakides declared an interest in item 9 as his employer sold chicken products.
4. Minutes of the 70th meeting
4.1. Members approved ACM/MIN/70 as a correct record of the previous meeting with the following amendment: the first sentence of paragraph 8.3 to read “Prof Coia questioned whether being able to identify supershedders on-farm would be the most cost-effective intervention for E. coli control in the food chain . . . “.
5. Matters arising
5.1. The Secretariat had produced a summary of action taken on matters arising from previous meetings (ACM/963). There were no comments on the paper.
6. Work of General Advisory Committee on Science
6.1. Professor Colin Blakemore, Chair of the FSA’s General Advisory Committee on Science (GACS), presented an overview of the role and work of the Committee. GACS provides advice and acts as an independent challenge to the FSA’s Chief Scientist and Board on the Agency’s governance and use of science, it also aids in the development and use of science based evidence throughout the Agency. The Committee comprises the nine Chairs who represent the FSA’s independent scientific advisory committees (SACs) plus six directly elected members.
6.2. Examining the way in which the Agency uses scientific based evidence, GACS found there to be good practice in terms of the FSA’s commitment to science and the use of this evidence to inform policy. The Agency’s approach to openness and the Agency’s use of expert Committees was also commended. However, the Agency needed to ensure that thorough scientific evaluation of the impact of policies was performed. Better clarity on the relationship between risk assessment and risk management was also required with improved communication and interaction between SACs and a wider network of experts. Further, to achieve this there was a need for engagement and improved interaction between science and policy. There was concern that the distinct separation between scientific risk assessment and policy risk management drawn up after the Philips Report created artificial boundaries. It was also considered important that as policies change in light of risk assessments, such information is fed back to committees in case the science has to be re-assessed. GACS has set up a working group on risk assessment and risk management to examine how governance and good practice have worked.
6.3. Prof Blakemore highlighted the importance of horizon scanning (HS) in ensuring that the Agency is in a good position to achieve its strategic objectives. As part of this it was noted that GACS had held a HS workshop in June 2009 to discuss the implications of trends in food production. In addition, GACS was looking into ways of developing and including expertise to augment the knowledge already in place on the SACs. This work helps GACS input into the approach, priorities and principles of the Agency’s scientific evidence base and strategy.
6.4. Prof Blakemore asked Members for their views on how the role of the SACs could be improved including better lines of communication. It was noted that one of the members of ACMSF had attended the recent meeting organised by the FSA and Government Office for Science for lay members of SACs and this had provided a useful means of dialogue. Members highlighted the need for more interaction between Committees, for example where social science issues needed to be included in risk assessments. Members agreed that following risk assessment there needed to be an iterative process to ensure that the Committee was informed about evidence that might impact on their assessments and advice to the Agency. Similarly, although ACMSF does not make risk management decisions, the Committee should be aware of management options which may also impact on advice. Indeed ACMSF often presents risk management options for the FSA to consider.
7. UK-wide survey of the prevalence of Johne’s disease in UK dairy herds
7.1. The Chair introduced the item stating that in 2000 (ACM/MIN/38) the Committee considered the risk to human health from Mycobacterium avium subspecies paratuberculosis (MAP) in cows’ milk. ACMSF concluded that the risk to human health was neither proven nor disproven and did not recommend any change in the advice regarding the consumption of cows’ milk. Miss Lesley Larkin (Defra) was invited to present data on a survey of MAP, the cause of Johne’s disease, in the UK dairy herd.
7.2. Miss Larkin stated that MAP infection occurs most frequently in calfhood although the disease may not be manifest for many years. Cattle may also be infectious before they show signs of disease with some never showing clinical signs. There is currently no effective treatment for infected animals. The survey estimated the UK wide prevalence of MAP infected dairy herds using randomly selected herds, geographically stratified to reflect general distribution of dairy herds, in the 4 countries of the UK. A total of 136 herds participated in the study with 13,691 samples collected, these included blood, faeces, environmental and bulked milk. Although samples were generally representative of UK herds they were more weighted towards farms with a prior history of Johne’s disease as these were more likely to participate in the study.
7.3. ELISA results showed that for the 136 herds tested a total of 86 (65.4%) had one or more animal seropositive for MAP and out of 13,691 samples, 340 (2.5%) were found to be seropositive. Liquid culture results showed that from 13,691 faecal samples (combined into 2791 pools) 36 (26.5%) herds had one or more positive pool. From the 804 environmental samples collected 25 (18.4%) herds were found to have positive samples. Only herds with positive pooled faecal samples had positive environmental samples. Four PCR methods were used to analyse 135 bulk milk samples for MAP resulting in the identification of 40 (29.6%) herds with positive samples. Culture tests, HEYM, Middlebrook & BACTEC were negative for all bulk milk samples.
7.4. Due to the different tests applied to samples and the different sample matrices (individual, pooled and bulked) a model was developed to combine and assess the overall prevalence while taking account of sampling, sensitivity and specificity issues. The model assumed that one true positive implied the herd was positive and that all tests would need to be negative for herd to be considered negative. The model also used “prior” sensitivity and specificity estimates to feed into Bayesian probabilistic estimates which were then updated depending on the output of the model. The prevalence of UK dairy herds infected with MAP was estimated to be 34.7% (95% CI 27.6% - 42.5%). The full report on the surveillance of Johne’s Disease was published on the Defra website on the 30th November 2009.
7.5. Noted in discussion:-
- Discrepancies between the ELISA, PCR and culture growth results - there was particular concern as to why growth was not observed in bulk milk samples. Miss Larkin stated that research looking at this had not been performed.
- The ability to make a full assessment of the item was questioned since a paper had not been submitted for review. This would have been useful particularly in light of the depth and intricacies of the model. Mr Wyllie explained that the full report had only just been published and Defra had made the judgment that it was better to bring the item to ACMSF in December, hot off the press, rather than defer the discussion until the spring.
- More information was required for the Committee to determine the validity of the model. However, it was acknowledged that this model was only providing a prevalence estimate and not a full risk assessment. The use of sensitivity and specificity estimates in the model was questioned as it was not clear how these were derived. There was concern that if these were incorrect then the number of false positives and negatives would not be known. In response Miss Larkin stated that the model used “prior” sensitivity and specificity estimates based on data from literature. This was fed into the model which then produced UK specificity and sensitivity results. Sensitivity analysis was also performed using the model to examine the impact of different levels of test sensitivity and specificity.
7.6. In summary the Chair stated the cause of Crohn’s disease remains unknown and that the Committee needed to see more data and clarification on the testing methods used. A summary paper describing clearly the sampling methods and computer model was requested. The Chair added that the complex nature of the model raised the question of whether the Committee would require additional statistical expertise for discussions of this sort in the future.
8. E.coli O157: Defra and FSA research with relevance to Super-shedders
8.1. The Chair introduced the item stating that at ACMSF 70 (September 2009), in response to a FSA request for the ACMSF to advise on the public enquiry recommendation that super-shedders should be explored as a mechanism for spreading E. coli O157, the Committee suggested that a review should be undertaken to examine the latest information available. Dr Chris Low (Scottish Agricultural College) was invited to present a review of recent research.
8.2. Dr Low opened the presentation by stating that it was important to recognise that E.coli O157 was not just a foodborne risk but also an environmental risk. A brief description of E.coli pathotypes was provided highlighting that verocytotoxigenic E.coli (VTEC) O157 produces verotoxin (VT) which is detectable using Vero cells. Scotland was noted as having a higher rate of E.coli O157 infection in humans than the rest of the UK with 200 to 250 cases per year from both outbreaks and sporadic cases. Sporadic cases accounted for approximately 20% of all cases. Regional differences were also observed with a 2004 survey showing the Grampian area to have the highest rate of culture and serum positives at almost 1 in 10,000 of the population.
8.3. Dr Low added that a SEERAD prevalence study of E.coli O157 based on 925 Scottish farms had shown 23% of farms and 8% of cattle to be positive. A follow-up study by IPRAVE on 481 farms provided similar results with comparable numbers of farms found positive. However, it was noted that a longitudinal study had shown that individual farms commonly changed their status between positive and negative. The IPRAVE study additionally showed that 89.7% of the isolates were VT2 positive while 8.9% were positive for both VT1 and VT2. All isolates were positive for intimin.
8.4. E.coli O157 was shown to have a preference in attaching itself to the terminal portion of the rectum and that after ingestion of E.coli O157, terminal rectal colonisation occurred after 3 days. Colonisation also occurred on direct application of E.coli O157 to the terminal rectum. It was stated that cattle with E.coli colonisation at the rectum were associated with a longer duration of shedding and a high level of faecal excretion, these were described as “super-shedders”. However, it was noted that the status of super-shedders is transient. It was found that the protein Tir, an intimin receptor, was active in the binding of E.coli to the host cells and the removal of Tir led to reduced shedding in cattle.
8.5. Dr Low stated that in support of the super-shedder model, prevalence studies had revealed that observed surveillance results did not fit those predicted by modelling based on chance events. This showed that infection was not random as chance events would have led to fewer negative farms. Remodelling the data assuming animals exist with high transmission rates (super-shedders) led to accurate fitting of the observed data. In addition, the modelling suggested that 4% of cattle had a 50 times higher transmission rate than the average with shedding periods of approximately 18 days. Importantly, the number of infections that would occur from a single super-shedder ‘R0’ was 1.9. Therefore super-shedders were said to be responsible for the continued infection of herds. However, it was noted that reducing shedding to below 104 cfu/g would bring the R0 value below 1 and therefore lead to the eventual eradication of infection.
8.6. Noted in discussion:-
- Immune responses to E.coli O157 and the possibility of vaccinating cattle to reduce the incidence of infection. In response Dr Low stated that the UK does have experimental vaccines but these are not used on cattle herds. However, the vaccines did appear to reduce the likelihood of infection and the amount of infectivity.
- MHC sequencing to determine any association with the immune system. It was reported that this had not been done.
- Prioritising routes of transmission when considering interventions.
- It was noted that in order to mitigate infectivity it was important to prioritise what the routes of infection were when looking at interventions, particularly as environmental routes appeared to be the main cause.
- Investigation of the length of time E.coli remains in the soil and the type of soil matrix. Dr Low added that rather than soil, it is believed that in the Grampian region the main route of infection was through private water supplies.
- Investigation of other pathotypes of VTEC. It was noted that there was a focus on VTEC O157 while other pathotypes were not investigated for their impact although it was acknowledged that approximately 99% of all cases in Scotland were E.coli O157.
8.7. In summary the Chair noted that there was a need for cross government working on the management of E.coli O157 as transmission was not solely through food. More work was also required on the definition of “super-shedder”. The cost effectiveness and benefit of interventions should be investigated and in particular more information on vaccine development is required. Finally, it was important to remember that other strains of VTEC must not be excluded.
9. Campylobacter and Salmonella in retail chicken survey
9.1. Mr Andrew James (FSA), Ms Geraldine Hoad (FSA) and Ms Katherine Holmes from (FERA) were invited to the table. Ms Hoad gave a presentation of the findings of a recently published UK-wide survey of Campylobacter and Salmonella in chicken. It was highlighted that Campylobacter was currently the most common bacterial cause of foodborne illness in the UK. In 2005 the FSA’s strategic plan had contained a target to reduce the prevalence (assumed to be 70%) of Campylobacter in chicken at retail by half by 2010. The survey performed between May 2007 and September 2008 was designed to measure progress towards this target.
9.2. It was noted that for Campylobacter, the prevalence results were based on combined data for both the presence/absence and enumeration methods used in the survey. The prevalence of Campylobacter in chicken was found to be 65.2% for 927 samples taken at retail and 76.1% for 416 samples of whole UK-origin chicken. Campylobacter jejuni accounted for 52.9% of isolates and Campylobacter coli 47.1%. The prevalence of Campylobacter was higher in chilled chicken (47.6%) than frozen (13.6%) with lower levels also observed in the frozen chicken. Campylobacter prevalence was also found to be higher in free-range and organic flocks than for housed birds. Unfortunately due to the change in approach used to determine prevelence in this survey, it was impossible to compare these data directly with an earlier 2005 study.
9.3. The prevalence of Salmonella was 6.6% which compared well to the 2001 figure of 5.7%. Thirty different serotypes were isolated during the survey. All isolates of Salmonella Java PT Colindale were found to be multi-drug resistant, the first time this had been observed in the UK. Salmonella prevalence was 5.7% in UK-origin chicken and 11.3% in non-UK chicken. Prevalence was higher at 11.7% in frozen chicken than chilled at 5.9%. In conclusion it was stated that a significant proportion of fresh chicken in the UK was contaminated with Campylobacter and addressing this risk remains a priority for the Agency.
9.4. Members asked whether the country of origin of the chickens was known and whether it was more likely that Campylobacter was from the UK or overseas. Ms Hoad responded that 50% of the Salmonella Java PT Colindale isolates were from non-UK samples. However, analysis of contamination origin was not investigated. A Member added that it was unfortunate that results could not be compared with previous surveys and it was noted that consistency in methodology should be considered for the next survey. The Chair added that while Campylobacter contamination remains high it was right that the Agency continues to give this work priority.
10. Campylobacter Workshop
10.1. Ms Liz Redmond (FSA) provided feedback on a joint FSA/Defra/BBSRC Campylobacter workshop that had taken place in October. It was noted that the meeting, which focused on Campylobacter in poultry, included stakeholders including producers, meat industry and retailers. Research needs and priorities, the need for a vaccine control method and the need for joined up funding were issues identified at the meeting. The BBSRC are in the process of compiling the output of the workshop which will be reported at the next ACMSF meeting. The Committee noted that it was important to maintain the impetus of the workshop to ensure timely commissioning of research. Ms Redmond responded that a meeting to discuss the way forward was scheduled for 15 December 2009.
11.1. The Chair introduced the item reminding Members that the ACMSF had previously considered “openness” and the way in which the Committee operates with particular attention given to subgroups. A paper was drafted by the secretariat setting out options for future subgroup meetings for the Committee to discuss. Mr Bob Martin (FSA) was invited to present the paper. Mr Martin stated that the original discussion on openness stemmed from a FSA review of Scientific Committees which recommended that Committees should conduct as much of their business in open session as possible. Further, the FSA Board supported this position but added clear criteria must be in place should it be felt items must be held in reserved business.
11.2. Opening the discussion Members welcomed the move to more openness and stated that the proposed outline provided enough flexibility for the subgroups in terms of how data is handled. However, they raised concern about the application of the Freedom of Information (FOI) Act on data which stakeholders may request to be kept confidential. It was noted that it would be difficult for the Agency to agree that documents would be kept confidential due to the interpretation of the FOI Act and what is in the public interest. A clear definition and protocol would be required to provide stakeholders assurances that confidentially would not be challenged under the FOI Act. Members asked whether the FSA had consulted those who would submit sensitive information whether they would be happy to work to these conditions. Mr Martin responded stating that there had not been a consultation as this was a provisional document seeking views of the Committee. However, each item submitted by stakeholders would be reviewed on a case by case basis in terms of FOI issues. Members added that subcommittees from other advisory bodies should be consulted to see how they deal with FOI issues. SEAC was highlighted as a Committee with good practice in this area.
11.3. Members added that the Agency should recognise that open meetings could impede the spontaneity of discussion in the early stages of a subgroup’s work when ideas and thoughts were still in their formative stage. Members could also lose their willingness to hypothesise on the interpretation of data in case this was misunderstood. Members agreed that a move to openness had to be managed well and suggested that there may be a requirement to have a pre-discussion before the open meeting to ensure that all members’ thoughts were recorded. The Chair highlighted the FSA Board as a group which adopts this method of working.
11.4. In summary the Chair stated that in principle the Committee acknowledged and appreciated the FSA’s move to more openness. However, it was thought that there needs to be more consideration given to the practicalities.
12. Epidemiology of Foodborne Infections Group (EFIG)
12.1. Dr Paul Cook (FSA) introduced the item reminding members that EFIG was an inter-departmental group which coordinates the surveillance of animal and human infection with the aim of identifying the actions required to ensure the microbiological safety of food. A brief overview of the November 2009 EFIG meeting was provided. Key items highlighted were the January 2009 to June 2009 figures for Salmonella in livestock which were very similar to the same period last year. In the same period the incidents of Salmonella in chickens fell by 54%. This was attributed to a fall in S. Enteritidis and S.Typhimurium. These figures excluded isolates from vaccine strains. Salmonella 4,5,12:i was reported in UK pigs for the first time and further work is being done to characterise this strain. In humans there was a continuation of the downward trend for non-typhoidal salmonellas during January 2009 to September 2009, again attributed to reductions in S. Enteritidis. Both Listeria monocytogenes and VTEC O157 reports in humans were shown to be relatively constant over the last 4-5 years although in 2009 there had been a slight increase compared with the same period in 2008 for both organisms particularly in pregnancy associated cases of listeriosis. Campylobacter infections continued to show an upward trend with three outbreaks reported in England and Wales associated with chicken liver pâté.
12.2. Members raised concern regarding the reporting of data, in particular Salmonella incidents. It was important that denominator data was recorded so that the significance of any changes could be measured and interpreted. In many instances only positives were reported and no record of the total number tests undertaken.
13. Codex Committee Food Hygiene Meeting (CCFH)
13.1. Dr Paul Cook provided the committee with a brief summary of the 41st CCFH meeting. Key points included the submission of documents on Risk Analysis Principles and Procedures Applied by CCFH, leafy green vegetables and Vibrio spp. in seafood, to the Codex Alimentarius Commission for adoption. Progress was reported on the development of Guidelines for the Control of Campylobacter and Salmonella in chicken meat, although issues with divergent opinions regarding the inclusion of chemical decontamination methods were highlighted. CCFH discussed progress with work on foodborne viruses and safety considerations regarding a proposed draft standard for smoked fish, smoked flavoured and smoke-dried fish with respect to Clostridium botulinum. CCFH also agreed to take forward new items of work including the revision of the Principles for the Establishment and Application of Microbiological Criteria for Foods and revision of the Code of Hygiene Practice for Natural Mineral Waters. In summer 2010 the Food and Agriculture Organization plan to launch a web based risk management tool which will support the implementation of the Codex guidelines for the control of Campylobacter and Salmonella in chicken meat.
14. Committee sub-groups
14.1. Prof Tom Humphrey, Chair of the Ad Hoc Group on Vulnerable Groups (AGVG) and the Surveillance Working Group (SWG) reported that these groups were to meet on the 9th and 10th December. However, it was noted that the AGVG would discuss a position paper on the current situation of Toxoplasmosis in the UK and review a paper on Toxoplasma surveillance sampling. The SWG would be discussing the FSA survey on the microbiological contamination of fresh red meats and a draft commentary paper on the methods used in the surveillance of Campylobacter in poultry. A report of these two meetings will be provided at the next ACMSF.
15. Dates of Future Meetings
15.1. Members were reminded that ACMSF meetings for 2010 would be held on the: 25th March, 3rd June, 23rd September and 2nd December.
16.1. Professor Humphrey raised concern that recent FSA advice on eggs for caterers had appeared not to include recommendations from the ACMSF. It was felt that the advice concentrated too much on contamination from egg shells rather than the risks from liquid egg which was a good growth medium for Salmonella. It was added that the Agency extolled vaccination methods and that the British Egg Industry Council (BEIC) believed the FSA was not doing enough to support this. Professor Humphrey added that although recognising that this was a risk management issue, it may be more beneficial that different lines of communication are used in the future to ensure that the information is directed to the right people. He concluded that it was important that the advice for caterers to use reputable suppliers was given more prominence to reinforce this message.
16.2. Ms Liz Redmond responded that the questions raised were risk management issues and as such ACMSF was not the forum for discussion. However, the FSA had taken on board what the Committee had recommended and key aspects were included in the advice. The FSA noted the concerns raised and would consider these in future versions. Ms Redmond added that it was important to understand that in developing the Agency’s advice there was a need to consider the wider risk management implications and the limitations of focusing on one control method such as vaccination when other methods were available. There was also no EU requirement to identify vaccinated or unvaccinated eggs and therefore no way for caterers to identify such eggs from across the EU. As such the FSA has focused on egg handling methods as an important factor in cross contamination where caterers should handle all eggs as if contaminated.