ACMSF Minutes: 27 June 2013

Meeting held at 1pm in Aviation House, 125 Kingsway, London WC2B 6NH.


Professor Sarah O’Brien

Dr Bob Adak
Dr Gary Barker
Mr John Bassett
Dr Roy Betts
Mrs Vivianne Buller
Professor John Coia
Mrs Rosie Glazebrook
Professor David McDowell
Mr Paul McMullin
Dr Sally Millership
Mrs Jenny Morris
Mr David Nuttall

Departmental representatives:
Mr Stephen Wyllie (Defra)
Mr Javier Dominguez (FSA)
Ms Maree Barnett (DH)

Officials in attendance:
Mrs Lorna Rowswell

Dr Paul Cook (Scientific Secretary)
Ms Geraldine Hoad (Administrative Secretary)
Dr Sophie Rollinson
Mr Adekunle Adeoye

Invited experts:
Dr Emma Snary (AHVLA)
Dr Paul Gale (AHVLA)
Dr Mike Hutchison (Bristol University)

Elizabeth Andoh-Kesson, BRC
Catherine Cockcroft, Exova
Kaarin Goodburn, Chilled Food Association
Barry Mirhabib, Brakes
Karen Sims, Waitrose
Elizabeth Williamson, Sainsburys
Nicola Wilson, Westward Labs

1. Chair’s introduction

1.1 The Chair welcomed ACMSF Members, members of the public, Dr Emma Snary, Dr Paul Gale and Dr Mike Hutchison to the Committee’s 80th meeting. The Chair also welcomed new ACMSF member Dr Gary Barker to the meeting. Dr Barker had been appointed to provide expertise in risk assessment.

1.2 The Chair passed on the sad news that Prof Doug Georgala, Chair of ACMSF from its inception until 2004 had recently passed away. Several significant issues were considered and reported on during his tenure as Chair including Salmonella in eggs, VTEC and ACMSF reports on foodborne viral infections and microbial antibiotic resistance. The Chair and Committee conveyed their sorrow and thanks to Prof Georgala’s family.

1.3 The Chair informed the Committee that, following discussions with the Social Sciences Research Committee (SSRC), Mrs Joy Dobbs (SSRC Chair) had agreed to become an ex-officio member of ACMSF. It was anticipated this would assist in discussions on areas of mutual interest between the two Committees.

2. Apologies for absence

2.1 Apologies for absence were received from Professor Jim Gray, Professor Rick Holliman, Ms Jenny Hopwood, Ms Ruth Parry (DH representative) and Mrs Joy Dobbs (SSRC ex-officio Member). Mrs Dobbs had provided written comments on papers of interest to the SSRC.

3. Declarations of interest

3.1 The Chair reminded Members of the need to declare any conflicts of interest relating to items on the agenda. Prof Coia declared an interest in relation to item 7 (freezing chicken livers and Campylobacter) as he provided consultancy advice to Tesco. Mr McMullin declared an interest also in relation to item 7 as a poultry veterinarian. Prof McDowell declared an interest in relation to item 7 as he carried out funded research in this area. Members were reminded to raise any other declarations of interest if the need arose during discussions.

4. Minutes of the 79th meeting (ACM/MIN/79)

4.1 The Secretariat was asked to correct paragraph 6.4, 6th bullet as Cephalosporin-based medicines were the subject of the EFSA recommendation rather than ESBLs. Mr Wyllie asked the Secretariat to check that it was Prof Forsythe rather than himself that had been nominated to the antimicrobial resistance subgroup (para 6.6). Subject to these comments the minutes were accepted as an accurate record of the last meeting and the Secretariat was asked to publish them on the ACMSF website.
Action: Secretariat

5. Matters arising (ACM/1105)

5.1 Ms Hoad drew Members’ attention to the summary of actions taken on matters arising from previous meetings. All matters had been actioned.

6. Q fever, raw milk and raw milk products (ACM/1106)

6.1 Dr Rollinson introduced paper ACM/1106, which provided background to the commissioning of a research project on the Q fever risk from unpasteurised milk and milk products. There was a lack of current evidence on this issue and AHVLA were therefore commissioned to assess the Q fever risk to human health from the consumption of contaminated unpasteurised milk and milk products. Members were provided with three outputs from the project, namely a risk profile for Coxiella burnetii in raw milk and milk products, the risk pathways and a draft exposure assessment.

6.2 The Chair invited Dr Emma Snary and Dr Paul Gale to present the work undertaken on the project. Dr Snary outlined the information gathered for the risk profile explaining that Q fever was a zoonotic disease caused by C. burnetii, which is endemic in cattle, sheep and goats in the UK. Transmission to humans was mainly through contaminated aerosols from infected animals but there was some strong evidence that contaminated raw milk could also be a source of human infection. Human epidemiology data was presented, including UK outbreak data and Dr Snary highlighted that there had been an increase in the number of reported UK human cases in 2011 (from 55 in 2010 to 112 in 2011). Veterinary epidemiological data was also outlined including data on the seroprevalence of Q fever in UK livestock. It was noted that viable C. burnetii had been detected in commercially sold raw milk but not in raw milk products including cheese. Dr Snary noted that there was very limited enumeration and survival data for C. burnetii in raw milk and milk products. The risk profile concluded that, compared to other routes, milk and milk products were a minor route of infection. It was suggested this could be because the concentration of C. burnetii produced during abortion and present in livestock birth products is much higher than the levels shed in milk and infectivity may be lower through the oral route than the inhalation route.

6.3 Dr Snary summarised the data gaps identified in the project and noted that these meant that a full quantitative risk assessment was not possible due to a lack of data, in particular on survival and concentration of C. burnetii in milk and milk products and dose-response data through the oral route. Dr Snary described the exposure assessment model that had been developed and some draft results from the model, including scenario analysis for a Q fever outbreak. It was noted that exposure is possible and could be occurring frequently but due to the uncertainties associated with the model data, the exposure results are an overestimation, and furthermore the unit of exposure may present a relatively low risk through the oral route. The need for laboratory methods to detect and enumerate viable C. burnetii was highlighted.

6.4 The Chair invited comments and questions on the presentation. The following comments were made in discussions:

  • It was confirmed that Q fever is not a notifiable or reportable disease in animals, although there is a legal obligation to report abortions in cattle under Brucella legislation. In the case of a high abortion rate in a dairy herd it might be expected that a farmer would want to establish a diagnosis and therefore would report abortions.
  • It was noted that human outbreaks appear to be relatively rare and it was queried whether the literature search was restricted to developed countries. Dr Gale clarified that the information and literature found in relation to human outbreaks and cases had tended to focus on the developed world rather than developing world.
  • It was queried whether the apparent increase in Q fever cases in 2011 was a true increase or the effect of changes in diagnosis. It was noted that the gold standard method for human diagnosis is immunofluorescence and demonstration of an increase in antibody level. PHE at Porton offer a PCR test for Q fever. It was suggesting the timing of the Netherlands Q fever outbreak may have a had a bearing on the number of samples submitted for testing in 2011 and further investigation of the number of samples submitted may be valuable. Dr Snary noted that the Defra zoonoses report suggested there may have been recent changes in Q fever diagnosis and this was being followed up with PHE. It was also noted that genotyping of isolates is not undertaken.
  • A Member commented that the outbreak data reviewed came from acute cases of Q fever. The epidemiology of chronic Q fever cases was queried, including whether the two forms of disease were linked to the two development stages of the organism. It was also suggested that better differentiation of forms was important. Dr Gale commented that there were two variants of C. burnetii, small cell and large cell variants. Large cell variants were more fragile and small cell variants were similar to a spore-like form. It was suggested that the organism is less infectious via the oral route than inhalation route because it targets macrophages which are less prevalent in the gut than they are in lung tissue.
  • There was discussion on the 50% Guinea Pig intraperitoneal infectious dose (GP_IP_ID50) unit that was used as the measure of exposure in the exposure assessment and the risk to humans from this via the oral route. Dr Gale suggested that a high number of GP_IP_ID50 units may be needed to make 1 oral infectious dose 50% for humans and this fits with the observed epidemiology of Q fever in humans.
  • It was suggested that dose-response needs to be better elucidated but it was recognised that the time between presentation with symptoms and diagnosis could be quite long so potential infection sources were unlikely to still be available for testing. Dr Gale noted that there was a lack of recent experimental work on infectious dose. A dose-response experiment had been conducted in the 1940s but no quantification of C. burneti was done and therefore the information could not be used to calculate infectious dose.
  • It was suggested that the epidemiological evidence linking consumption of C. burnetii contaminated milk with Q fever illness in humans was fairly weak.
  • A Member asked whether any work was ongoing to establish a better unit of exposure and whether any sensitivity analysis was planned on the model due to the high number of uncertainties. Dr Snary responded that they were not aware of any current work on exposure units and that sensitivity analysis was planned on all the model variables to include the uncertainties.
  • It was suggested that it might be valuable to follow-up regular raw milk drinkers, such as farm workers, to investigate Q fever infection and this could be done via the Gastrointestinal, Emerging and Zoonotic Infections Department at PHE.

6.5 Ms Dobbs provided written comments and suggested a number of areas where the SSRC could provide assistance if required, such as on studies to look at the number of people consuming raw milk, amount consumed, type of products consumed and storage of products. Ms Dobbs suggested the SSRC would want to keep a watching brief on this issue as the work progressed and should also consider whether to include questions on consumption of raw milk and milk products in the next Food and You survey.

6.6 The Chair reminded members that they had been asked to comment on:

  • The evidence that Q fever can be transmitted by unpasteurised milk and milk products.
  • The data gaps identified in the risk pathway document, in particular the most significant gaps to address in terms of assisting risk assessment.
  • The exposure assessment approach taken.
  • Any other approach that would help is assessing the level of risk.

6.7 The Chair summarised the Committee’s discussions, noting that the link between Q fever and unpasteurised milk and milk products could be considered not proven. However, it was agreed that unpasteurised milk was one of the less significant of the known infection routes. There were many data gaps that would hinder the risk assessment and most of these need to be addressed. However, improving laboratory methods, human diagnoses and follow-up of raw milk drinkers, using a sentinel approach were important. It was also important to clarify if the recent increase in human cases was an artefact related to changes in diagnostics or a true increase. The exposure assessment approach used seemed pragmatic in the circumstances and there were no particular improvements recommended by the Committee on the approach taken. The report should be explicit in stating the literature search criteria and in explaining which milk products were included and excluded from the scope and why.

7. Freezing chicken livers and Campylobacter (ACM/1107)

7.1 Item 7 was introduced by Mrs Rowswell who reminded the Committee that there had been an increase in reported human outbreaks of Campylobacter associated with chicken liver pâté or parfait in recent years, thought to be linked to the undercooking of livers. The Agency had commissioned research to look at interventions that may help in reducing the risk from these products, including a project on the effect of freezing livers on Campylobacter numbers. Members were asked to consider the study findings and comment on:
- Whether application of the findings could contribute to reducing the risk of campylobacteriosis from imperfectly cooked chicken livers or chicken liver products?
- Whether there are other factors (ingredients, treatments) which could help reduce Campylobacter contamination in chicken livers with or without freezing?
- Possible further research arising from these findings.

7.2 Dr Mike Hutchison was invited to present the findings of the study. A small number of samples of frozen and fresh raw livers were collected from retail premises. The mean Campylobacter counts in frozen livers were found to be significantly lower than those measured in fresh livers. Livers were collected from final clearance flocks in slaughterhouses to increase the probability that they were contaminated with campylobacters. The livers were subjected to different freezing treatments by varying the rates and durations of freezing and the number of freeze treatments. The rates of freezing were chosen to represent the worst-case conditions achievable in a domestic freezer towards the end of its working life and the best-case conditions in a catering freezer. Freezing to either 15°C or 25°C was attempted. The freezer set to -15°C was only able to freeze the livers to -11°C after 24 hours. Freezing to -11°C or -25°C for 24 hours was found to significantly reduce the numbers of campylobacters on the livers compared with unfrozen livers. The freezer set to -15°C was able to lower the temperature of the livers to -15°C by 48h. Freezing for one week at -15°C gave a further significant reduction in Campylobacter numbers compared with the 24h freeze. Two freezes to -25°C gave the biggest reduction in Campylobacter numbers (up to three logs).

7.3 The Chair invited comments from Members on the research, the following points were discussed:

  • It was clarified that chest freezers and upright freezers were used in the study. The different types of freezer froze the livers at different rates. Both the surface area to volume ratio of livers and rate of freezing affected Campylobacter viability.
  • The length of time taken to reach the desired low temperature (i.e. the rate of freezing) was important since that affected Campylobacter viability. A rapid rate of freezing was more important than the lowest temperature reached.
  • There are two bactericidal mechanisms operating during freezing. When freezing rates were low, patches of frozen water formed in the extracellular water. As that ice formed, it expelled dissolved ions into the unfrozen water, increasing its osmotic potential. As the ionic strength of the extracellular water became more concentrated, it began to remove water from the cytoplasms of the Campylobacter cells. When freezing was rapid, ice crystal formation was the main method for Campylobacter population decline.
  • Dr Hutchison responded that the rate of freezing was a critical parameter that was much faster in the -25°C freezer where the damage to cells was probably due to ice crystal formation.
  • The study showed that freezing could be a useful intervention to reduce Campylobacter but there were still several important questions on the effect of variables such as rate of freezing and time between defrosting and re-freezing. It was agreed that recommendations to freeze livers might lead to some confusion for consumers and caterers if they were advised to freeze livers twice as the general advice is not to refreeze foods once defrosted.
  • In response to a question, Dr Hutchison clarified that no work was done in this project on any organoleptic changes to the livers as a consequence of freezing. Dr Hutchison stated his recollection was that, after two freezes, the livers produced more exudate but they weren’t ‘mushy’. Mrs Rowswell added that FSA were in the process of commissioning further work to look at the production of pâté and parfait and sensory testing to investigate organoleptic issues.
  • Members were informed that in the last two years there had been a decrease in the number of Campylobacter outbreaks associated with chicken liver pâté, with 15 outbreaks in 2011, 6 in 2012 and one to date in 2013. It was suggested that some of prosecutions that had taken place in relation to catering premises and chicken liver pâté might have had an effect on catering practices.
  • It was queried whether the use of essential oils in pâté/parfait preparation had been considered. It was clarified that the project scope was only to assess the implications of freezing.
  • It was noted that a risk assessment model created by EFSA for general chicken meat reported that a two-log reduction in Campylobacter numbers would give a significant reduction in human foodborne disease cases. The size of the reduction achieved by freezing livers was an important and useful piece of work. Were Campylobacter levels contaminating livers reduced by freezing, there would also be an additional potential benefit of reduced cross-contamination from the livers during preparation in kitchens.

7.4 Ms Dobbs provided some written comments on this item which queried the factors underlying the increase in outbreaks linked to chicken liver pâté and whether this was due to increased consumption of chicken liver pâté, increased food poisoning risk per serving or greater propensity to undercook livers. If the cause was related to behaviour changes then research to analyse and understand that behaviour might identify interventions to modify it.

7.5 The Chair summarised that freezing could be considered a risk reduction measure and FSA should consider appropriate messaging to ensure that consumers/caterers were not confused over freezing advice. The research had raised some questions over variables, such as the rate of freezing, time between freezes, and effect on organoleptic properties. The use of essential oils in pâté and parfait preparation could also be considered.

8. Infant formula (ACM/1108)

8.1 Dr Paul Cook presented paper ACM/1108 explaining that the Department of Health were seeking further advice from the FSA on safe preparation of powdered infant formula (PIF) with respect to microbiological risks. The current UK recommendation was to prepare standard PIF using hot water (>70°C or above). The Committees’ views on the relative risks of different preparation, storage and feeding scenarios were sought.

8.2 Dr Cook outlined the main microbiological hazards associated with infant formulae. These were Cronobacter spp, which was mainly associated with sporadic cases and Salmonella enterica which was generally associated with outbreaks. Cronobacter was the primary organism of concern. Infection with Cronobacter is rare but serious. Between 1992 and 2012 there were 36 reports of isolation of Cronobacter from blood or CSF from infants less than 12 months in England and Wales. The source of these infections was not known. Dr Cook highlighted that, despite what many care-givers believe, PIF is not a sterile product and controlling Cronobacter during manufacture is challenging and contamination may occur at very low levels. It was noted that feeds are not always consumed straight after preparation and may be stored for some time and there is nothing intrinsic in the powdered formula that would prevent growth of Cronobacter or other bacteria if reconstituted formula was not stored at appropriate temperatures. Dr Cook noted that a risk assessment model for Cronobacter in PIF has been developed by FAO/WHO and this on-line model was used to explore the effect of varying parameters such as preparation temperature, storage time and feeding duration on the risk from Cronobacter in PIF. The model gives a relative measure of risk reduction as its output. Members were asked to comment on:

  • The information provided regarding microbiological hazards associated with powdered infant formula, its preparation and use.
  • The risk reduction achieved by different preparation scenarios using the FAO/WHO model and their relative importance.
  • The conclusion that the reconstitution of PIF with water at 70°C can make a significant contribution to risk reduction from i) intrinsic contamination of PIF and, ii) extrinsic contamination arising from equipment and the preparation environment.

8.3 The Committee covered the following points in discussions:

  • It was noted that the output is exposure reduction rather than risk reduction. Exposure and risk are correlated but the increments may be smaller. It was suggested that this may reduce the utility of comparisons on exposure risk between scenarios.
  • Dr Cook clarified there was no dose response information available for infants and Cronobacter and it was difficult to know what levels infants were exposed to when illness was identified as the reconstituted formula was often no longer available for testing.
  • Using a risk ratio approach generates greater uncertainty as there is uncertainty in both the numerator and denominator. This makes the output difficult to interpret except at very low and high temperatures.
  • The recommendation to use water at 70°C may create confusion as many people don’t know what this means in practice and how to achieve it. It may be more helpful to recommend use of boiling water. It was clarified however, that because of concerns with handling boiling water and the effect on vitamins use of boiling water to make up PIF is not currently advised. Current advice aims is ensure that PIF is reconstituted with freshly boiled water which is no lower than 70°C.
  • It was suggested that when access to freshly boiled water is not available e.g. on long journeys, people follow different practices. Some make up formula and transport it for later use and some transport boiled water for making up formula as needed, the risks are different for these different scenarios. It was suggested that this is an area where social science may have a role in looking at consumer behaviour. Dr Cook noted that some research had been done to look at this, in terms of temperature tracking and behaviour. It was also suggested that the most important thing is probably controlling growth rather than the kill, so making formula up with water at ambient temperature is probably less of a risk than making formula up with water at 50°C and storing it before feeding.
  • It was suggested that advice should be that PIF is made up with hot water and if it has to be stored it should be chilled as soon as possible. Advising consumers that PIF is microbiologically unstable and should be treated as other foods in terms of temperature control could be helpful and simple advice.
  • Based on the figures in the paper it was noted that making up with water at 70°C significantly reduced viable Cronobacter numbers.
  • In relation to extrinsic contamination it was noted that if freshly boiled water was used any contamination on the equipment should be destroyed.
  • It was suggested that where advice on long held practices is changed the effect on micro-organisms other than the target organism should also be considered.

8.4 In summarising the Chair noted that some Members had concerns over certain aspects of the model used but the important factor was controlling growth of bacteria in the formula. Growth could be controlled by making up PIF with freshly boiled water which would reduce any contamination present in the PIF or on the associated equipment.

9. Epidemiology of Foodborne Infections Group (EFIG) (ACM/1109)

9.1 The Chair invited Dr Paul Cook to update Members on the outcome of the EFIG meeting which took place on 3 June 2013. Dr Cook reported on trends in Salmonella prevalence in animals from data collected during 2012. There had been a reduction in reports of Salmonella in cattle, sheep and ducks compared with 2011 and data from pigs was comparable with 2011. Trends in laboratory reports for Salmonella, Campylobacter, Listeria monocytogenes and E.coli O157 in humans were also reported. Numbers of cases of Salmonella infection have continued to fall whilst Campylobacter cases are continuing to rise although case numbers in Scotland, Wales and N. Ireland appear to be levelling off. Overall, the number of outbreaks reported in 2012 declined by 35%. EFIG were also updated on the FSA’s recent foodborne viruses and Campylobacter workshops and on PHE food surveillance studies.

9.2 The Chair invited Dr Bob Adak to give a presentation on surveillance data and denominators for human Infectious Intestinal Disease. Dr Adak explained that it was the current international industry standard to use population data as the denominator for routine national or regional laboratory surveillance. This allowed comparison of data in an international and cross-disciplinary context in a way that was transparent and consistent and was the approach used in the Chief Medical Officer’s annual report. Any change in approach would need to show demonstrable benefits and the requirements for the output of any new approach were outlined. Currently laboratory reports are used as a proxy for illness in the community, in an ideal world the number of people that became infected after exposure would be more informative i.e. the number of symptomatic and asymptomatic cases. A number of alternative denominators for human surveillance were outlined and it was highlighted that the feasibility of collecting centralised data on specimen throughput from clinical laboratories was being investigated. Use of the number of specimens tested for certain pathogens from PHE sentinel laboratories was discussed. Using specimens tested as a denominator might allow for adjustments to account for changes in sampling and testing that might affect laboratory report ascertainment. Several possible scenarios using this approach were outlined.

9.3 The Chair invited questions and comments. It was suggested that Defra may want to review their processes as one of the frustrations was having data presented in different ways in the same booklet and being misinterpreted. It was noted that one of the big differences between animal and human surveillance is that a lot of foodborne zoonosis don’t cause disease in animals and the only way to find out if animals are carrying the organism is to do routine surveillance on healthy animals.

10. Committee sub-groups

10.1 Dr Roy Betts provided a report on the fourth, fifth, sixth and seventh meetings of the Ad Hoc Group on Raw, Rare and Low Temperature Cooked Foods. The group had considered low temperature cooked foods and sous vide at their first four meetings and looked at trying to define low temperature cooking. Meetings five to seven had considered raw and rare foods and the range of factors that could mediate the increased risk from these foods. The group aim to have a draft final report for the October 2013 ACMSF meeting.

10.2 Prof O’Brien provided a report on the twelfth meeting of the Ad Hoc Group on Foodborne Viral Infections. The group had met to finalise the drafting of their report and agree the wording of their recommendations. The report will be presented to the October 2013 ACMSF meeting.

10.3 Prof McDowell updated the Committee on the setting up of the new ACMSF Working Group on Antimicrobial Resistance. The Committee had agreed at their January meeting to look at antimicrobial resistance in the foodchain in more detail. Prof McDowell mentioned the ACMSF members that had been nominated to the Group and also noted that a number of external experts had been invited to join. A teleconference was arranged for July to discuss the terms of reference for the group with a follow-up meeting in September. It was anticipated that the group would be established as a working group rather than an ad hoc group as a continued overview of the issue was necessary.

11. General Advisory Committee on Science (GACS)

11.1 The Chair informed Members that the minutes of the March 2013 GACS meeting were available on the GACS website.

12. Dates of future meetings (ACM/1110)

12.1 Members were asked to note the dates of the final 2013 ACMSF meeting on 3 October and the 2014 meetings on 30 January, 26 June and 2 October.

13. Any Other Business

13.1 Dr Cook informed Members that the FSA is planning to look at risk ranking of certain ready-to-eat foods, with respect to Listeria monocytogenes and the Committee will be invited to comment on this work.

13.2 The Chair informed Members that the revised WRAP reports, on composting and anaerobic digestates have been submitted to FSA. The reports have been amended to address comments, including those from ACMSF. The Secretariat will shortly be arranging a meeting of the ACMSF subgroup that looked at these reports to review the revised documents.

13.3 Ms Dobbs provided written comments on information paper ACM/1113, the forward work plan, highlighting that the forthcoming items on foodborne viral infections and risk–ranking of ready-to-eat foods had most potential for input from a social science perspective.

Public questions and answers

The Chair drew formal proceedings to a close and invited questions and comments from the public in relation to ACMSF and risk assessment. There were no questions.