Defra 2002 - Risk Assessment: Disposing of catering waste containing meat

RISK ASSESSMENT: USE OF COMPOSTING AND BIOGAS TREATMENT TO DISPOSE OF CATERING WASTE CONTAINING MEAT

Final Report to the Department for Environment, Food and Rural Affairs

Risk Assessment: Use of Composting and Biogas Treatment to Dispose of Catering Waste Containing Meat

Final Report to the Department for Environment, Food and Rural Affairs

Report No: DEFRA

May 2002

Author: Dr Paul Gale

Contract Manager: Dr Paul Gale

Contract No: 12842-0

DEFRA Reference No:

Contract Duration: October 2001 to May 2002

This report has the following distribution: DEFRA – 10 copies

Contract Manager – 2 copies

Any enquiries relating to this report should be referred to the Contract Manager at the following address:

WRc-NSF Ltd, Henley Road, Medmenham, Marlow, Buckinghamshire SL7 2HD. Telephone: + 44 (0) 1491 636500 Fax: + 44 (0) 1491 636501

CONTENTS PAGE LIST OF TABLES VIII LIST OF FIGURES XII EXECUTIVE SUMMARY 18

1. INTRODUCTION 31

1. Objectives 31
2. Project appreciation 32
3. Variation and uncertainty 33
4. General Approach 33

2. SOURCE TERM 35

1. Household consumption of meat 35
2. Meat discarded to the waste bin in household surveys 35
3. Catering Outlets 38
4. Other pathogen sources in household domestic waste 38
5. Poultry 39

3. ANIMAL TISSUES COMPOSITION 40

1. Cattle 40
2. Sheep 41
3. Pigs 41
4. Chickens 43

4. MODELLING THE PATHWAYS 44

1. Growth of pathogenic bacteria in catering waste 44
2. Dilution in the soil 45
3. Decay of pathogens on land 45
4. Receptor Term 48
5. Exposure to grazing cattle, pigs and sheep across England and

Wales. 48

5. AMOUNT OF MATERIAL COMPOSTED IN THE UK 50

1. Total feedstock material 50
2. Types of composted product 50

6. PARTICLE SIZE, TEMPERATURE AND TIME FOR PATHOGEN

DESTRUCTION BY COMPOSTING AND BIOGAS 51

1. Objectives 51
2. Introduction 51
3. Heat inactivation of exotic viruses 51

4. Destruction by composting of bacterial pathogens 56
5. By-pass: Variation and uncertainty in the net pathogen destruction by composting or biogas 57
6. "In-vessel" process 58
7. Windrows 59
8. Destruction of pathogens by thermophilic biogas systems 62
9. Summary of pathogen destructions by composting and biogas 65
10. Particle size and the time/temperature criterion for composting and biogas 66

7. A CREDIT SYSTEM FOR MODELLING THE BARRIERS IN

COMPOSTING AND BIOGAS TREATMENT OF CATERING

WASTE 68

1. Source separation and definitions 68
2. Composting/biogas of source-separated "non-meat" fraction 68
3. Composting the "meat" fraction 69
4. Biogas treatment of the "meat" fraction 69

8. COMPOSTING AND BIOGAS - PUTTING THE BARRIERS

TOGETHER 70

1. Note 70
2. Composting 70
3. Removing the Windrow (2nd barrier) and 18 d stock-piling stages from composting increases the risks by 83-fold 70
4. Biogas 72
5. Storage (18 days) of the "meat" fraction gives a further 5-fold

reduction in the overall risk for the biogas process 73

9. COMPARE WITH OTHER ROUTES OF DISPOSAL –LANDFILL 75

1. Bird control at landfill sites 75
2. Vermin and other pests 75
3. Mud on vehicle tyres 76
4. A quantitative assessment of the pathways out of landfill. 76
5. Conclusions 79

10. CONSIDERATION OF MODES OF BY-PASS FOR COMPOSTED

CATERING WASTE TO LAND 80

1. Implication of "by-pass" 80
2. By-pass of the composting process 82
3. Direct feeding of catering waste to animals as swill - complete by- pass of the whole process 82
4. Direct ingestion of compost by animals – by-pass of decay and

dilution on soil 82

11. A SUMMARY OF KEY ASSUMPTIONS FOR THE RISK

ASSESSMENT 83

1. Barriers 83
2. Policies 84
3. Modelling these policies in the risk assessment 84

4. Dose-response – estimating the risk of infection from the exposure 86

12. BOVINE SPONGIFORM ENCEPATHALOPATHY (BSE) 87

1. Source Term 87
2. Pathways and barriers 88
3. Risk Assessment 93
4. Conclusion 94

13. SCRAPIE 96

1. Source Term 96
2. Pathways 99
3. Predicted risks 100
4. Imported lamb 101

14. FOOT AND MOUTH DISEASE VIRUS 102

1. Epidemiology 102
2. Source Term 102
3. Infectious Dose 108
4. Effect of Cooking 109
5. Decay on soil 109
6. Quantitative risk assessment for FMDV 110
7. Risks from imported material into the UK 111
8. Risks during an FMD outbreak in the UK 114
9. Effect of the time period of the no grazing ban 115
10. Sensitivity analysis – the number of TCID50s comprising a pfu - A 116

15. CLASSICAL SWINE FEVER VIRUS (HOG FEVER) 117

1. Source Term 117
2. Sensitivity analysis – uncertainty over viral titres in blood 119
3. Oral ID50 for pigs 119
4. Survival in food 119
5. Decay in soil 120
6. Risks of CSF from imported material into the UK 122
7. Risks during a CSF outbreak in the UK 124
8. Effect of the time period of the no grazing ban 124

16. SWINE VESICULAR DISEASE 126

1. Source Term 126
2. Survival in meat 126
3. Dose-response of SVDV in pigs 128
4. Assume no decay in the soil 129
5. Risks of SVD from imported pig meat into the UK 129

17. AFRICAN SWINE FEVER 131

1. Source Term 132
2. Oral ID50 for pigs 132
3. Survival in foods 133
4. Decay in soil 134
5. Risks of ASF from imported material into the UK 135

6. Effect of the time period of the grazing ban 138
7. Other routes of infection 138

18. NEWCASTLE DISEASE 139
    1. A quantitative risk assessment 139
19. PROTOZOAN PARASITES SUCH TOXOPLASMA 141

1. Cryptosporidium 141
2. Toxoplamsa gondii 141

20. ENDEMIC FAECAL BACTERIAL PATHOGENS – SALMONELLAS,

E. COLI O157 AND CAMPYLOBACTERS 147

1. Source Term 147
2. Regrowth of bacterial pathogens on discarded food. 152
3. Routes of exposure for endemic bacterial pathogens through composted catering waste 153
4. A quantitative model for E. coli O157 in catering waste 153
5. Quantitative Risk Assessment for Campylobacter 157
6. Risk assessment for salmonellas 158

21. TRICHINAE (T. SPIRALIS) 161

1. Epidemiology 161
2. Risk assessment 162

22. CLOSTRIDIUM BOTULINUM (BOTULISM) 164

1. Types of botulism 164
2. Infectious and lethal doses 165
3. Heat resistance of C. botulinum spores 166
4. Growth of C. botulinum spores in composting and in food 167
5. Survival in soil. 167
6. Number and frequency of spores in food. 167
7. Risk assessment for infant botulism – a "What-if?" scenario 168
8. Note on C. botulinum spore densities in soil 168
9. Risk assessment for cattle 168
10. Conclusion 168

23. PLANT PATHOGENS 170
24. OTHER FORMS OF COMPOSTING 173

1. Home-composting 173
2. Vermiculture 173

25. ADDRESSING THE SPECIFIC OBJECTIVES 174

1. Maintain the current ban on the use of composting and Biogas to dispose of catering waste containing meat 174
2. Adopt the new EC rules 174
3. Adopt specific UK standards 174

26. REFERENCES 176

LIST OF TABLES

Table 2.1 Household consumption of meats by households in UK

(Source DEFRA) 35 Table 2.2 Household waste survey (summer). 250 households sampled. 36 Table 2.3 Household waste survey (winter). 156 households sampled. 36 Table 2.4 Summary of total kitchen waste discarded weekly by

households from three studies 37 Table 2.5 Summary of results of survey by WRc-NSF of 39 domestic

kitchens. Note, percentages are only estimates and were not

weighed. 37 Table 2.6 Breakdown of meat distribution to catering outlets in UK (The

Foodservice market meat monitor, 2001). Tonnes purchased

per week by sector. 38 Table 3.1 Cattle by-products (data from MLC). Organs discarded as SRM

in italics. 40 Table 3.2 Sheep by-products – weights in lambs. Organs discarded as

SRM in italics. For sheep tissues multiply by a factor of 1.6.

Data from MLC. 41 Table 3.3 Pig by-products (Data from MLC) 42 Table 4.1 Summary of parameters for decay of pathogens in sewage

sludge after application to soil. 47 Table 4.2 Total tillage and grass land in England and Wales (Anon 1997) 48 Table 4.3 Total number of cattle, pigs and sheep in England and Wales

(Anon 1997). 49 Table 5.1 Quantity and proportion of composted material 50 Table 6.1 Net destruction of FMDV over 10 minutes. Data from Turner et

al. (2000). Data averaged for medium and pig slurry counts. 52 Table 6.2 Net destruction of CSFV over 10 minutes. Data from Turner et

al. (2000). Assumes starting titre for experiments at 60C and

65C is 7.0 log10. Data averaged for medium and pig slurry

counts. 53 Table 6.3 Net destruction of Aujeszky's disease virus (ADV) over 15

minutes. Data from Turner et al. (2000). Data averaged for

medium counts (and not pig slurry). 54 Table 6.4 The effect of heat on survival of SVD virus in milk (from

Herniman et al 1973). 55 Table 6.5 The effect of heat on inactivation of SVD virus in slurry (from

Herniman et al. 1973). 55 Table 6.6 Effect of Windrow (at laboratory scale) on bacterial pathogens

spiked into sewage sludge. Data from Horan and Lowe (2001). 56 Table 6.7 Effect of "within-batch" and "between-batch" variation (e.g. from

short-circuiting and dead spaces in a digester) on the net

destruction of pathogens. 57 Table 6.8 Fate of E. coli in bags placed in situ in aerated static piles

(Data from Horan and Lowe 2001). 58

Table 6.9 Percentage cross sectional area of windrows reaching certain

temperatures throughout green organic processing (Joshua et

al. 1998). 60 Table 6.10 Log10 counts of pathogen remaining in a windrow after N turns.

Assumes there are 1,000 (3-log) counts in the windrow at t = 0.

Model allows for different degrees of destruction () in the high

temperature zone and assumes that 80% of the material is in

the high temperature zone (fh = 0.8 in Equation 2). 61 Table 6.11 T90 values for Selected Microorganisms in Biogas and Slurry

Systems. Data from Bendixen (1999). 62 Table 6.12 Average counts of FS in four thermophilic biogas plants (Table

IV from Bendixen 1999) 62 Table 6.13 Daily removals of faecal streptococci between the receiver

tanks and the digestion tank from Table VI of Bendixen (1999).

The conditions for this plant (no. VIII) were thermophilic

digestion tanks (53C) with mean guaranteed retention time

(MGRT) of 5 h and hydraulic retention time (HRT) of 19 days. 63 Table 6.14 Daily removals of faecal streptococci between the receiver

tanks and the storage tank from Table VI of Bendixen (1999).

The conditions for this plant (no. VIII) were thermophilic

digestion tanks (53C) with mean guaranteed retention time

(MGRT) of 5 h and hydraulic retention time (HRT) of 19 days. 64 Table 6.15 Faecal streptococci counts in biogas plant VIII after

improvement of operational procedures to eliminate minor

irregularities causing contamination with raw material. 65 Table 6.16 Net removal ratios for the processes and storage 65 Table 6.17 Estimated heat transfer times into spherical compost particles

(Table 5.5 from Haug 1993). 66 Table 7.1 A credit system for the barriers for composting the "non-meat"

fraction. 68 Table 7.2 A credit system for the barriers for biogas treatment of the

"non-meat" fraction. 68 Table 7.3 A credit system for the barriers for composting the "meat"

fraction. 69 Table 7.4 A credit system for the barriers for biogas treatment of catering

waste containing meat. 69 Table 9.1 A qualitative comparison for control points for pathogens in

uncooked meat in catering waste 77 Table 9.2 A credit system (representing log-reductions) for the barriers

for composting and landfilling 77 Table 10.1 Effect of "within-batch" and "between-batch" variation (e.g. from

short-circuiting and dead spaces in a digester) on the net

destruction of pathogens. 80 Table 11.1 A credit system for the barriers 84 Table 12.1 BSE in EU countries. Data for 2001 from DEFRA 88 Table 12.2 Fate of DRG in cattle carcasses (Data taken for Figure 3.1 of

DNV (1997). 89

Table 12.3 Fate of CNS in under 30 month cattle slaughtered at UK

abattoirs 89 Table 12.4 Fate of CNS in carcasses of under 30 month cattle slaughtered

abroad and imported into the UK 92 Table 12.5 Summary of risks of BSE transmission to cattle in

England/Wales from application of composted catering waste. 94 Table 13.1 Infectivity titres (bioassay in mice) in tissues from up to 9

Suffolk sheep (34-57 months old). Data from Anon (1994). 96 Table 13.2 Distribution of scrapie infectivity in ovine tissues with age

(taken from DNV, 2002). 97 Table 13.3 Arithmetic mean scrapie infectivity titres (intracerebral ID50 g-1)

estimated for ovine tissues with age of animal 97 Table 13.4 Weights and utilisation of ovine tissues in food. 98 Table 13.5 Predicted ovine oral ID50 in food per infected animal. 99 Table 14.1 pH values of imported meat and offal determined in a London

cold-storage warehouse (from Henderson & Brooksby, 1948) 103

Table 14.2 Foot and mouth disease viral titres in tissues of 62 pigs two

days after experimental infection. Data taken from Farez and

Morley (1997). 105 Table 14.3 FMD loadings in an infected pig 106 Table 14.4 Summary of ID50s for FMD. 109 Table 14.5 Likelihood of exposure of domestic swine to exotic disease

agents through uncooked swill in the USA. 112 Table 14.6 Summary of predicted FMD risks from composting of catering

waste from the illegal importation of 10,000 FMD-infected

"bone-in" porcine carcasses – assumes no time interval period. 112 Table 14.7 Summary of predicted numbers of FMD cases from composted

catering waste assuming the illegal importation of 10,000 FMD-

infected "bone-in" porcine carcasses. Numbers based on

0.52% of UK herd grazing on land to which compost has been

applied – assumes 1 year ban with 1% of animals spending 7

days accidentally grazing on that land. 113

Table 14.8 Summary of predicted number of FMD cases from composted

catering waste from the importation of 2.28 x 105 tonnes of

boneless beef/sheep of which 1% is from FMD-infected

carcasses. 114 Table 14.9 Total animals slaughtered in infected premises in UK 2001

outbreak (to 24 Feb 2002). 114

Table 14.10 Summary of predicted FMD risks from composting of catering

waste from 960,313 FMD-infected "bone-in" sheep carcasses

during an FMD outbreak. 115

Table 14.11 Summary of predicted numbers of FMD cases in cattle and

sheep from composted catering waste allowing for different no

grazing time intervals (assuming soil decay according to Figure

14.5). 115 Table 15.1 Classical swine fever viral titres in tissues of 64 pigs four or five

days after experimental infection. Data from Farez and Morley

(1997). 117

Table 15.2 Classical Swine Fever Virus loadings in an infected pig 118 Table 15.3 CSFV titres in meat samples from four pigs infected with ASF.

Data from McKercher et al. (1978). 120

Table 15.4 Summary of predicted number of CSF infections from

composted catering waste from the illegal importation of

10,000 CSF-infected "bone-in" porcine carcasses – assumes

no time interval period between application of compost and

grazing. 122 Table 15.5 Summary of predicted risks of CSF cases in pigs from

composted catering waste assuming the illegal importation of

10,000 CSF-infected "bone-in" porcine carcasses. 123 Table 15.6 Summary of predicted numbers of CSF cases in pigs from

composted catering waste. Numbers based on 0.52% of UK

pigs housed on land to which compost has been applied –

assumes 1 year ban with 1% of animals spending 7 days

accidentally gaining entry to that land. 124 Table 15.7 Summary of predicted numbers of CSF cases in pigs from

composted catering waste allowing for different no grazing time

intervals (assuming soil decay according to Figure 15.2). 124 Table 16.1 Swine Vesicular Disease Virus loadings in an infected pig

(Data from Burrows at al. (1974) and Mann and Hutchings,

1980). 127 Table 16.2 Summary of predicted SVD risks to pigs from composting of

catering waste from the illegal importation of SVD-infected

"bone-in" porcine carcasses. Risk calculated through oral

challenge. 129 Table 17.1 African Swine Fever Virus loadings in an infected pig. Tissue

HAD50 titres from Farez and Morley (1997). 131 Table 17.2 ASFV titres in meat samples from four pigs infected with ASF.

Data from McKercher et al. (1978). 132 Table 17.3 Summary of predicted number of ASF infections from

composted catering waste from the illegal importation of 1,000

ASF-infected "bone-in" porcine carcasses – assumes no time

interval period between application of compost and grazing. 136 Table 17.4 Summary of predicted risks of ASF infection in pigs from

composted catering waste assuming the illegal importation of

1,000 ASF-infected "bone-in" porcine carcasses. 136 Table 17.5 Summary of predicted numbers of ASF cases in pigs from

composted catering waste assuming the illegal importation of

10,000 ASF-infected "bone-in" porcine carcasses. Numbers

based on 0.52% of UK pigs housed on land to which compost

has been applied – assumes 1 year ban with 1% of animals

spending 7 days accidentally gaining entry to that land. 137

Table 17.6 Summary of predicted numbers of ASF cases in pigs from

composted catering waste allowing for different no grazing time

intervals (assuming soil decay according to Figure 17.3). 138 Table 18.1 Newcastle Disease EID50 loadings in an infected chicken 139 Table 19.1 A quantitative risk assessment for Toxoplasma gondii in

domestic cat faeces in composted MSW. 145

Table 19.2 Summary – Qualitatative risk assessment for Toxoplasma in

composted catering waste. 146 Table 20.1 Percentage of Salmonella spp. positive raw chicken obtained

from local supermarket chains (n = 175) and butchers' shops

(n = 125) during a seven month study in South Wales ( Harris on

et al. 2001). 147

Table 20.2 Breakdown of the percentage of salmonella spp. positive raw

chicken samples by chicken type. Data from Harris on et al.

(2001) for South Wales. Number of samples in parentheses. 147 Table 20.3 Frequency distribution for counts on salmonellas on ground

poultry meat. 148 Table 20.4 Counts of samonellas (MPN cacrass-1) estimated from graphs

in Figure 1 from Dufrenne et al. (2001). 149 Table 20.5 Campylobacter counts (log10) recovered from external and

internal organs of prescald broiler carcasses from a

commercial processing plant. Data from Berrang et al. (2000a). 150 Table 20.6 Breakdown of the percentage of Campylobacter spp. positive

raw chicken samples by chicken type. Data from Harris on et al.

(2001) for South Wales. Numbers of samples in parentheses. 150 Table 20.7 Campylobacter populations recovered from breasts, thighs and

drumsticks of broilers purchased at a retail outlet and skinned

in the laboratory. Data from Berrang et al. (2001). 151 Table 20.8 Campylobacter populations recovered from breast meat, thigh

meat and drumstick meat of broilers purchased at a retail outlet

with and without skin. Data from Berrang et al. (2001). 151 Table 20.9 Summary of E. coli O157:H7 growth on bovine carcass tissue

under aerobic conditions. Data from Berry and Koohmaraie

(2001). 153 Table 20.10 Estimating E. coli O157 loadings in composted catering waste

in UK. 155

Table 20.11 Comparison of predicted E. coli O157 loadings (cfu) in "stored"

manure, conventionally-treated sewage sludge and composted

catering waste in England/Wales. 157

Table 21.1 Summary – Qualitative risk assessment for T. spiralis  in

composted catering waste. 162 Table 22.1 Parameters for dose-response curves in Figure 22.1 166

LIST OF FIGURES

Figure 1.1 Event tree for BSE agent from abattoir to soil through

application of sewage sludge (from Gale & Stanfield, 2001). 33 Figure 4.1 Event tree for transmission of pathogens in composted

catering wastes to root crops. 44 Figure 4.2 Dilution of composted catering waste residues in soil. Based

on assumption that any pathogens in compost leach down to a

depth of 10 cm. 45

Figure 6.1 Laboratory scale thermal treatment of FMDV incubated at

different temperatures with time in pig slurry (solid line) and

Glasgow Eagles medium (dashed line). Data from Turner et al.

(2000). 52 Figure 6.2 Laboratory scale thermal treatment of Classical Swine Fever

Virus incubated at different temperatures with time in pig slurry

(solid line) and Glasgow Eagles medium (dashed line). Data

from Turner et al. (2000). Starting titre for all experiments was

measured at 7.0 log10. 53 Figure 6.3 Laboratory scale thermal treatment of Aujeszky's Disease

Virus (ADV) incubated at different temperatures with time in pig

slurry (solid line) and Glasgow Eagles medium (dashed line).

Data from Turner et al. (2000). 54 Figure 6.4 Temperature measurements for an "in-vessel" system taken on

one day during the hotter stages of processing. Data kindly

provided by Dr Joe Short of London Remade Organics Eco-

site. 58

Figure 6.5 Temperature measurements recorded for an "in-vessel" system

taken at 1.5 days. Data kindly provided by Dr Joe Short of

London Remade Organics Eco-site. 59 Figure 6.6 Effect of number of turns of a windrow on pathogen survival.

Model assumes 80% of pathogens are in a high temperature

zone where proportion () of pathogens surviving. 61 Figure 8.1 Composting as set out in Table 7.1 and Table 7.3 achieves a

4.62-log reduction of infectivity in catering waste. 70 Figure 8.2 Composting - Failure or omission of the windrow 2nd stage for

the "meat" fraction reduces the net destruction of composting

to 2.73-logs 71

Figure 8.3 Composting - Stock-piling of the "non-meat" fraction is of little

point if there is no windrow second stage for the "meat"

fraction. 72 Figure 8.4 Biogas treatment as set out in Table 7.2 and Table 7.4

achieves a 4.42-log reduction of infectivity in catering waste. 73 Figure 8.5 Biogas process - Omission of the 18-day storage stage for the

"meat" fraction results in a 3.7-log overall reduction by biogas. 74 Figure 8.6 Biogas – overall effect of adding storage to the "black-bag"

fraction is negligible. 74 Figure 9.1 Prototype event tree for comparison of exposures to animals

through landfill disposal and composting with a 2 month

grazing 78 Figure 10.1 The effect of "by-pass" 81

Figure 10.2 Applying 0.01% of the raw meat in catering waste directly to

land with no grazing ban would "cancel out" the benefits of the

composting processes. This is analogous to the illegal action of

feeding catering waste directly to pigs. 81 Figure 11.1 Decay of FMDV in soil with time according to Equation 8.

Concentrations based on a single FMDV-infected pig carcass

entering the food chain. 85

Figure 12.1 Event tree for partitioning of central nervous system material

from cattle (under 30 months of age) slaughtered at UK

abattoirs into catering waste. 90

Figure 12.2 Event tree for partitioning of central nervous system material

from cattle slaughtered at foreign abattoirs and imported

carcasses 91 Figure 12.3 Event tree for contamination of processed meat with brain

material in foreign abattoirs 92

Figure 12.4 Event tree for transmission of BSE infectivity (bovine oral ID50

units) to root crops through application of composted catering

waste residues to agricultural land. 93 Figure 14.1 Rate of inactivation of FMDV in beef stored at 4C. Data from

Table 3 of Henderson and Brooksby (1948). 103 Figure 14.2 Survival of FMDV in tissues of infected cattle during storage at

1 to 4C. Data from Cottral (1969). 104 Figure 14.3 Dose-response curve for ingestion of FMDV by pigs. The oral

ID50 is about 106.0 TCID50. 108 Figure 14.4 Decay of FMDV in cattle slurry at 4C and 17C. Data from

Haas et al. (1995). 109

Figure 14.5 Decay of FMDV in soil with time according to Equation 8.

Concentrations based on a single FMDV-infected pig carcass

entering the food chain. 111 Figure 15.1 Decay of Classical Swine Fever in pig slurry at 4C and 17C.

Data from Haas et al. (1995). 120

Figure 15.2 Decay of CSFV in soil with time according to Equation 9.

Concentrations based on a 10,000 CSFV-infected pig

carcasses entering the food chain per year. 121 Figure 16.1 Negative exponential dose response curves fitted to data for

oral (r = 00000011) and skin (r = 0.00035) challenge of SVDV

to pigs. Data from Mann and Hutchings (1980). 128

Figure 17.1 Negative exponential dose-response curve (r = 0.000035) for

ingestion of ASFV by pigs. The porcine oral ID50 is about 104.3

HAd50. Data from McVicar (1984). 133 Figure 17.2 Decay of African Swine Fever Virus in pig slurry at 4C and

17C. Data from Haas et al. (1995). 134

Figure 17.3 Decay of ASFV in soil with time according to Equation 10.

Concentrations based on a 1,000 ASFV-infected pig carcasses

entering the food chain per year. 135

Figure 19.1 Negative exponential dose-response curve (r = 0.00085) for

risk of dead birth in sheep from oral exposure to T. gondii. Data

from Buxton (1998) who cited an experiment in which less than

18% of births were live after an oral dose of 2,000 oocysts.

This is plotted as the point. 144 Figure 20.1 Estimation of the arithmetic mean campylobacter loading

(MPN/carcass) for fresh chickens in The Netherlands. See text

for details. Data from Dufrenne et al. (2001). 152

Figure 20.2 Dose-response curve for E. coli O157:H7. "Most-likely" Beta-

Poisson model fitted using = 0.221; = 8,722.46 (from

Powell et al. 2000). Also shown is Beta-Poisson model ( =

0.49; N50 = 1,130) proposed by Crockett et al. (1996) and

supported by data of Strachan et al. (2001). 155

Figure 22.1 Effect of limited intestinal flora on dose-response of spores of

Clostridium botulinum type A in mice. Negative exponential

dose response curves fitted to data of Wells et al. (1982) using

parameters in Table 22.1. 165 Figure 22.2 Predicted growth of C. botulin spores in food at 19C, pH 6.0

(Food MicroModel). 167

Glossary

"Biogas plant" means a plant in which biological degradation of products of animal origin is undertaken under anaerobic conditions for the production and collection of biogas.

"Catering waste" means all waste food originating in restaurants, catering facilities and kitchens, including central kitchens and household kitchens. In terms of waste types, catering waste would includes B – K as set out below. B would be classified as catering, because it could also contain food.

"Composting plant" means  a  plant  in  which  biological  degradation  of  products  of animal origin is undertaken under aerobic conditions.

"Enclosed" in this report means vermin and bird-proof (as from the point of pathogens) and does not refer to gas exchanges (odours).

Greenwaste includes A and sometimes B and C in addition.

"Meat fraction" – waste fraction for composting which contains meat derived from two sources, namely:-

1. waste stream containing the meat which has been separated at source by the waste producer from the residual stream; and
2. residual black bag waste which has not been separated at source and will include meat as well as other waste materials.

"Non-meat fraction" – waste fraction for composting which should be free of most of the meat because waste producer has been instructed to exclude meat by source separation.

"Source separation" – the actions of the waste producer to keep certain parts of their waste (which is required for composting) separate from the residual waste stream.

"Turning" – the process of redistributing the material comprising a windrow – to be defined by a standard operating procedure (SOP).

"Maturation" – the mesophilic phase of  composting  which follows  the thermophilic phase characterised by establishment of natural fungi.

"Windrow" – trapezoidal heaps.

EID50 – egg ID50 (for Newcastle disease)

ID50 – Dose which when given to each and every member of a population infections half of the members of that population.

IDU, infectious dose unit (a hypothetical term) TCID50 – Tissue culture ID50 – measure of virus titre Pfu, plaque-forming unit (measure of virus titre) BSE, bovne spongiform encephalopathy;

TSE, transmissible spongiform encephalopathy; ASF(V), African Swine Fever (Virus)

CSF(V), Classical Swine Fever (Virus)

FMD(V), Foot and Mouth Disease (Virus)

SVD(V), Swine Vesicular Disease (Virus)

MGRT; minimum guaranteed retention time (biogas) HRT; hydraulic retention time (biogas)

DRG, dorsal route ganglia

HSE, Health and Safety Executive

"Tds", tonnes dry solids

VTEC, Vero-toxigenic Escherichia coli

CNS, Central Nervous System

OTMS, Over Thirty Month Scheme (cattle)

SRM, Specified Risk Material,

MHS, Meat Hygiene Service,

EXECUTIVE SUMMARY

A proportion of meat from domestic households and catering establishments will be discarded uncooked with catering waste. A risk assessment is developed here to determine the risks to cattle, sheep, pigs and chickens of infection from pathogens potentially present in that meat after the catering waste has been composted and applied to land. In addition, the risks to humans utilising compost and consuming crops grown in those fields to which compost has been applied are considered. The risk assessment has focused on TSE agents, the exotic pig viruses,  E. coli O157, campylobacters, salmonellas, Newcastle disease and parasites. The overall conclusion is that it is acceptable to apply composted catering waste to land provided: -

1. All steps are taken to eliminate any by-pass of the composting/biogas process, including ensuring that:-

• Raw catering waste material is not keep on livestock farms;

• Birds and small mammals do not gain access to the raw material;

• Raw material is delivered to a housed reception;

2. A two-barrier composting system is used for the "meat" fraction.
3. For each composting barrier, the catering waste reaches a temperature of 60C for two days during composting, with the composting process being continued for at least 14 days;
4. The first treatment barrier be it "in-vessel" or windrow is housed or enclosed;
5. Windrows are turned at least three times;
6. "Dirty" end is kept separate from the "clean" end; i.e. different tools and equipment are used to handle the final product and the raw material;
7. Biogas is performed at 57C; MGRT = 5 h; HRT = 19 days
8. The maximum particle size for composting is <40 cm diameter. This includes large joints of meat, e.g. discarded after freezer failures. For biogas, a maximum 5 cm (diameter) particle size is required;
9. Animals are not allowed to graze on land to which composted catering waste has been applied for a period of 2 months.

The risk assessment approach

A risk assessment is developed with the Source – Pathway – Receptor approach. Environmental risk assessment is not concerned with the complete elimination of a pathogen by any one barrier, but relies on a multiple barriers approach. The barriers include exclusion of meat at source, the composting/biogas processes, stock- pilling/storage, decay in the soil, dilution in the soil, in addition to the fact that only a small proportion (model assumes 1%) of meat is discarded in the kitchen uncooked.

Source Terms

For BSE, scrapie and E. coli O157, the risks are calculated on the basis of the number of bovines and sheep slaughtered in the UK. For BSE, account is also taken of imported bovine carcasses and processed meat. For the purpose of BSE and scrapie it is assumed that 90% of dorsal route ganglia (DRG) and 100% of spinal cord in lamb chops, respectively, present in the food chain are discarded to catering waste. For E. coli O157 is it assumed that 0.01% (w/w) of meat sold in shops is cow/sheep faeces. It is interesting to note that the predicted E. coli O157 concentration is 1.4 cfu g-1 beef, which is in good agreement with the median concentration of 1.5 cfu g-1 of beef measured in burgers in an outbreak in Canada.

For campylobacter and salmonella, the risks are calculated on the basis of published data on bacterial loadings per chicken carcass and the fact that 615 million broilers are slaughtered annually in the UK.

For exotic agents, the Source Term is based on "What if?" scenarios. Thus for Trichinellae, CSF and FMD, the model addresses the question of "What if 10,000- infected pig carcasses entered the food chain each year?". Although unrealistically worst-case for illegally imported meat, such scenarios could occur during outbreaks of FMD and CSF. For SVD and ASF, the risks are based on 1,000-infected pig carcasses entering the food chain annually.

For Clostridium botulinum, the starting point is that 4.18% of bacon contains spores.

For Toxoplasma gondii, the Source Term is based on an estimate of faeces from 15,000 infected domestic cats being disposed of to "black-bag" waste through cat litter.

Composting process parameters

It is proposed here that the composting process (windrow or "in-vessel") should achieve a temperature of 60C for two days. The literature data reviewed here demonstrate that at 55C considerable destruction of FMDV, CSFV and ASFV occurs within 10 – 15 minutes. Furthermore aeration processes at 50C give complete inactivation over 48 h. However, for SVDV in one slurry experiment a temperatures of 56C was required for 2 hours to achieve a 3-log reduction. The "60C for 2 days" is chosen because it takes 40 h for a sphere of diameter 40 cm to reach a temperature of 56C at the centre when the surrounding temperature is 60C. This is analogous to a large leg of pork with the bone in. The bone marrow contains high CSFV loadings in a CSF-infected pig carcass and

would take 40 h to reach 56C if the outside temperature is maintained at 60C.

It is shown here that a temperature of 70C for 1 hr according to the EU Regulations is only acceptable if the particle size is <6 cm in diameter. Indeed the EU Regulations specify <12 mm. It is the 2 day time period of composting proposed here which eliminates the requirement for a 12 mm minimum particle size at 60C. This is important because such a small particle sizes would tend to impair the aerobic processes of composting.

Biogas processing parameters

For biogas, the particle size must be less than 5 cm (diameter). The thermophilic biogas system must achieve a digestion temperature of 57C with MGRT of 5 h and HRT of

>19 d. When operated correctly removals of 4 – 6 logs have been demonstrated at operational scale. Mesophilic digestion (36 – 38C) is not appropriate for treatment of catering waste containing meat.

By-pass of the composting process (or biogas process)

Of critical importance is the degree of by-pass of the processes. Indeed, the net pathogen destructions achievable by any process are ultimately limited by the degree of by-pass. By-pass forms the basis of estimating the net pathogen destruction by composting/biogas in this risk assessment. There is some quantitative information on by-pass for thermophilic biogas systems albeit under conditions of operational failure. By-pass of the 60C/2 day conditions for windrows can be modelled mathematically according to the number of turns. It is calculated here that at least three turns of a windrow are required to ensure <0.2% of the raw material remains in the "cold" part. For "in-vessel" composting, the requirement is that >99.8% of the raw material achieves 60C for two days. (A 99.8% destruction, or 0.2% survival, is equivalent to a 2.7-log reduction as set out in Table 1). The biogas plant must be designed and operated to ensure <0.02% by-pass of untreated material. This is to give a net 3.7-log (5,000-fold) destruction of pathogens by biogas (Table 2) and can be achieved at operational scale.

Stock-piling of compost/Storage of biogas product

The storage period recommended is 18 days. This is to allow for a 1-log decay of CSFV. During this process for composts, there will be some heat generation and greater destructions than just 1-log would be anticipated in practice.

A credit system for modelling the barriers

The credit system is based on log removals. Thus a 1-log removal is a ten-fold destruction, a 2-log removal represents 100-fold destruction, 2.7 logs is 500-fold, 3.7- logs is 5,000-fold.

Source separation

Source separation is the actions of the waste producer to keep certain parts of their waste (which is required for composting) separate from the residual waste stream.

The non-meat fraction is the waste fraction for composting which should be free of most of the meat because waste producer has been instructed to exclude meat by source separation.

The Meat fraction is the waste fraction for composting which contains meat derived from two sources, namely:-

1. waste stream containing the meat which has been separated at source by the waste producer from the residual stream; and
2. residual black bag waste which has not been separated at source and will include meat as well as other waste materials.

Composting of "non-meat" fraction

The "non-meat" fraction could contain meat "by accident" due to inefficient Source Separation. The risk assessment is based on a credit system such that a 4.7-log (i.e. a 50,000-fold) reduction occurs through Meat Exclusion at Source, Composting/Biogas and Stock-piling. Meat Exclusion at Source is assumed to be 90% efficient, i.e. source- separate  "non-meat"  waste  contains  10%  of  the  total  uncooked  meat  waste.  The barriers are set out in Table 1 for composting and Table 2 for biogas.

Table 1 A credit system for the barriers for composting "non-meat" waste.

*Windrow or "in-vessel"

Table 2 A credit system for the barriers for biogas treatment of "non-meat" waste.

Composting the "meat" fraction

Waste containing meat must be composted by a two barrier process. First an "In-vessel" process in which <0.2% fails to reach 60C for 2 days, and secondly a windrow. The windrow need not be housed as it is a secondary barrier. The barriers are set out in Table 3.

Table 3 A credit system for the barriers for composting the "meat" fraction.

Biogas treatment of the "meat" fraction

A two barrier process comprising biogas digestion and a storage is set out in Table 4. A storage stage of the finished product of 18 d is required to allow for a 1-log decay of CSFV.

Table 4 A credit system for the barriers for biogas treatment of the "meat" fraction.

Putting these barriers together

The composting processes outlined in Table 1 and Table 3 are put together in a single event tree in Figure 1 to model the overall reduction of infectivity in meat in catering waste by composting. In total 2.36 x 10-5 infectious dose units (IDU) from each IDU in the catering waste remain in the compost. This is equivalent to a 4.62-log reduction overall.

0.998

Meat in "Meat" In-vessel

catering 0.9 fraction0.002 0.998

waste Windrow Compost

1 IDU SSoepuarcraetion 0.002 3.6 x 10 -6 IDU

0.1

0.998

"Non-meat" Composting

fraction 0.002 0.9

Stockpiling (18 d)

0.1 2.0 x 10 -5 IDU Total = 2.36 x 10-5 IDU

4.62-log reduction overall

Figure 1 Composting  as  set  out  in  Tables  1  and  3  achieves  a  4.62-log reduction of infectivity in catering waste.

Similarly for biogas, combining the processes set out in Tables 2 and 4 into an event tree (Figure 2) shows that overall the net reduction of infectivity in catering waste is 4.42-logs.

0.9998

"Meat" Biogas

cMateearti ning 0.9 fraction0.0002 0.9

waste Storage (18 d) Final product

1 IDU SSoepuarcraetion 0.1 1.8 x 10 -5 IDU

0.1

0.9998

"Non-meat" Biogas

fraction 0.0002

2.0 x 10 -5 IDU

Total = 3.8 x 10-5 IDU 4.42-log reduction overall

Figure 2 Biogas treatment as set out in Tables 2 and 4 achieves a 4.42-log reduction of infectivity in catering waste.

Modelling the 2 month no grazing ban

For FMDV, CSFV and ASFV the model allows for a 5-log decay with time as specified by available data. This is shown for FMDV in Figure 3. The risk to grazing cattle (ingesting 0.41 kg soil cow-1 d-1) is calculated from the cumulative exposure between days 61 and 426.

1.00E-08

61 d 426 d 1.00E-09 BAN

Exposure to

1.00E-10

grazing animals

1.00E-11

1.00E-12 1.00E-13 1.00E-14 1.00E-15

0 100 200 300 400

Day

Figure 3 Modelling annual exposure of grazing animals and the effect of the 2 month no grazing ban

Summary of the predicted risks

The results of the risk assessment are presented in Table 5. It should be noted that the annual numbers of animal infections are predicted on the assumption that 0.52% of the England/Wales herd/flock are housed or graze on land to which compost has been applied.

Table 5 Summary  of  risks for  each  pathogen from  composting  of  "non- meat" fraction – assumes 4.7-log credits by Source Separation and composting according to Table 1. To model for composting of all meat (as set out in Figure 1) multiply risks by a factor of 1.2. To model for biogas treatment of all meat (as set out in Figure 2) multiply risks by a factor of 1.9.

Could composted catering waste spread endemic infections to uninfected animal herds?

The question of whether the application of composted catering waste to agricultural land could spread the infection to uninfected herds and flocks is not tackled directly. Instead the risks relative to those from spreading of sewage sludge and manures are compared. Predicted concentrations of salmonellas in soil from application of

compost  are  at  least  1,000-fold  lower  than  those  predicted  for  application  of conventionally-treated sewage sludge. This puts the risks of  spread  of  infection to uninfected herds into perspective. For E. coli O157 and salmonellas, quantitative risk assessment is complicated by the uncertainty in the potential for regrowth, both in the raw catering waste and in the composted product. In the case of campylobacters, no

growth will occur on the meat or in the compost. This, together with the barriers, namely composting,  decay  and  dilution  in  the  soil,  would  mean  that  the  application  of composted  catering  waste  to  land  would  be  a  minor  pathway  for  campylobacter compared to other routes.

Animal Health - Classical Swine Fever Virus – why both a 2 month ban and multiple composting barriers are important.

The model assumes that 10,000 CSF-infected pig carcasses enter the food chain in a year.  Although  unrealistically  worst-case  for  illegally  imported  material,  it  is  not inconceivable that such challenges could occur during an outbreak of CSF in the UK. An acceptable risk would be <1 new CSF case every 50 years, i.e. <0.02 pigs infected year-1.

Composting System

Table 6 demonstrates the importance of both the second barrier (i.e. windrow for "meat" fraction) and the 2 month no gazing ban for composting systems.

Table 6 Predicted risks of CSF to pigs in UK from application of compost- treated meat to land.

*assumes 0.52% of UK pigs are housed on land to which compost applied Biogas system

The predicted risks to pigs from CSFV are set out in Table 7. In the case of biogas, both the 2 month no grazing ban and the additional 18 d storage for "meat" fraction are required.

Table 7 Predicted risks of CSF to pigs in UK from application of biogas- treated meat to land.

*assumes 0.52% of UK pigs are housed on land to which compost applied

The extra 18 d storage stage only reduces the risk by 5.4-fold. However, it is a back-up barrier and reduces the frequency of a CSF case from once every 23 years to once every 125 years.

Public health – salmonellas, E. coli O157, and campylobacters – the case for two barrier composting/biogas

Table 8 compares the predicted risks to humans from ingestion of a gram of compost producted  from  feedstock  containing  meat.  In  the  case  of  campylobacters  and salmonellas, two stage composting and biogas with storage is important to bring the risks down to the order of 10-6 person-1 year-1, which is an accepatable risk according to the HSE.

Table 8  Predicted  individual  risks  (infections  gram-1)  to  humans ingesting 1 g of compost produced from material containing meat. Note it is assumed that 18 d stock-piling (Figure 1) and 18 d storage (Figure 2) give a 1-log reduction of these pathogens.

*Risks calculated using two different dose-response models

Comparison with landfill routes

0.9999

"Meat" Retained in

Meat in landfills

catering 0.9 fraction0.0001 0.9

waste Eaten

Source 0.1 Landfill Route 1 IDU Separation 0.9 x 10 -5 IDU

0.1

0.998

"Non-meat" Composting

fraction 0.002 0.9

Stockpiling

0.99994 Compost with

0.1 Dilution

Decay 2 month grazing ban 0.00006 1.2 x 10 -9 IDU

Figure 4. Prototype event tree for comparison of exposures to animals through landfill disposal and composting with a 2 month grazing ban. Note; 1) model assumes source separation, such that 90% of the meat goes to landfill; 2) no data are  currently  available  for  how  much  meat  is  removed  from  landfills  by scavenging gulls.

The composting route for catering waste  potentially presents lower risks to grazing animals than disposal through land-fill. A prototype model is set out in Figure 4. This is because composting offers extra control points. First, the raw catering waste will be delivered to enclosed receptions where birds and animals cannot gain access. Second, a no-grazing period can be enforced after application of the compost to land.

The significance of the ultimate "by-pass"

The illegal action of feeding raw catering waste directly to pigs would by-pass the 9-log process  (set  out  in  Figure  4)  comprised  of  the  multiple  barriers  of  source separation/composting and dilution/decay in the soil (from the  2 month  no grazing period). If just 0.01% of the meat in catering waste in the UK were applied direct to land (with no 2 month grazing ban), the risks of an FMD or CSF outbreak would increase by some 100,000-fold. This represents feeding catering waste direct to pigs and would in effect remove the benefits of applying the composting and biogas processes set out in Figures 1and 2, respectively. The net reduction in risks across the UK by implementing a multiple barriers composting process with a grazing ban would no longer be 9-logs but just 4-logs. The "take home" message is that if just a small proportion of catering waste was  illegally  fed  directly  to  pigs,  then  there  would  be  no  point  in  the  rest  of  UK composting the waste – simply tilling all catering waste produced in the UK into the land and applying the 2 month grazing ban would not present a higher risk.

Conclusions

1. The risks from prion diseases (BSE and scrapie) both to animals and humans are remote, even allowing for no destruction by composting/biogas/stock-piling and no decay in the soil.
2. It is concluded that the risks posed to animal health from the application of composted catering waste to land are acceptably low, providing the two barrier composting and biogas processes as set out in Figures 1 and 2, respectively, are used, and a 2 month no grazing ban is enforced.
3. The no grazing ban (2 month) and 2nd barriers in the composting/biogas processes are essential for CSF, and are also important during FMD outbreaks.
4. The 2nd barrier is important in reducing the risks from campylobacter and salmonellas to acceptable levels if compost is ingested.
5. Clostridium spores will not be destroyed by composting or biogas. The risk assessment cannot eliminate Clostridium botulium spores in compost as a potential risk to infants (under 6 months of age). However, the spores levels predicted in compost are no higher than reported for some soils.
6. The composting approach outlined here provides more control points than landfilling. Composting could potentially present lower risks to animal health than the current practice of disposal of catering waste to landfill.
7. The risks to human health from consumption of crops grown on land to which compost has been applied are very low.

Notes on the risk assessment

The models described in Figures 1 and 2 allow for composting of all of the uncooked meat (i.e. both the "meat" and "non-meat" fractions), which is discarded each year to catering waste in the UK. The risks predicted in Table 5 are based on the assumption that the Source-Separated "non-meat" fraction of  all catering waste in the UK is composted (either by "in-vessel" or a windrow) and that compost is then applied randomly to land across the UK. The model assumes that any pathogens surviving the composting process are contained in the 500,000 tonnes of compost currently produced annually in the UK. Clearly, if composting expands within the UK, then the total tonnage of compost produced will increase greatly, thus lowering the pathogen concentration (per tonne) in the compost. In this sense the predicted risks to individual humans ingesting compost as presented in Table 5 are too pessimistic and would decrease in magnitude. In contrast, the estimates of the numbers of infected animals will remain constant because the greater tonnage of compost will be spread over a greater surface area of land, potentially exposing a greater number of animals (albeit to lower individual risks).

1.  Introduction

The Animal By-Products Order 1999, as amended, requires that all catering wastes that could contain or have been in contact with meat or other products of animal origin be disposed of so that livestock and wild birds cannot gain access to them. Since swill- feeding is banned, this material is currently being disposed of primarily to landfill or incineration. Land-filling is not a sustainable option in the long term. There is, therefore, increasing pressure to find more sustainable disposal routes. One option is the use of composting and Biogas plants to treat the materials, with land-spreading to dispose of the compost or residues.

The Government strongly supports the composting of waste. It is expected to be a vital tool in helping the UK meet stringent targets for reducing landfill of biodegradable municipal waste (BMW) under the Landfill Directive. By 2020 the amount of BMW landfilled must be reduced to 35% of that produced in 1995. Composting is also expected to be a vital component of meeting Waste Strategy 2000 targets for recycling and composting. The Government has set targets to recycle or compost at least 25% of household waste by 2005, rising to at least 33% by 2015. New EC rules, coming into force in mid-2002, will permit the use of composting and Biogas plants to dispose of low- risk (category 3) animal by-products, manure, gut contents and catering wastes containing meat. The Animal By-Products Regulation will require that the material is reduced to a particle size of no more than 12 mm and then treated, in a closed unit, so that all the material in the unit reaches a temperature of 70C for 60 minutes. Providing the material complies with certain microbiological standards, the compost or residues may then be spread to non-pasture land. Manure and gut contents may also be treated in this way and spread to any type of land. Category 2 animal by-products (e.g. fallen stock) could only be used in a composting or Biogas plant if they had first been rendered to 133C at 3 bar pressure for 20 min. Category 1 material (e.g. Specified Risk Material) could not be used at all.

1. Objectives

The overall objective of this project is to determine the risks to animal, public and plant health from the land application of various categories of animal by-products and catering wastes containing meat.

More specifically, the objectives are:-

1. To compare the risks from the following three options

• Maintain the current ban on the use of composting and biogas to dispose of animal by-products and catering waste containing meat;

• Adopt the new EC rules;

• Adopt specific UK standards;

2. To determine any minimum standards that might be needed to reduce those risks to an acceptable level.

2. Project appreciation

The objective "adopt specific UK standards" requires the alternative disposal strategies to be examined with respect to their ability to prevent the spread of a range of pathogens identified as being of concern to human, animal and public health. Given the wide range of pathogens of concern and the number of treatment options to be considered, the only feasible approach is to carry out a series of microbiological risk assessments.

The approach taken for environmental risk assessment is the Source, Pathway, Receptor approach. The Source Term defines the reservoirs of infectivity, and includes the frequency of infections and the loadings of pathogen in the catering wastes and other animal by-products studied. The Receptor Term defines the animal, human and plant categories that may be exposed. Central to the Receptor Term are dose-response data (i.e. how susceptible to infection is each species), and information on the nature of the disease. In this respect, information on the transmissibility through secondary spread is of major importance. For example, horizontal transmission of BSE (from cow- to-cow) occurs at such low frequency that it has not yet been detected by epidemiology. In contrast, foot and mouth disease rapidly spreads both within and between herds.

The Pathway Term defines the routes by which the various receptors might be exposed to pathogens in the Source Term. Some pathways will be well understood through epidemiological studies from outbreaks and sporadic cases. Other pathways will not have been identified, either because they are rare events or because epidemiological studies have not yet been able to separate them out.

Central to the Pathway Term are the barriers, which protect or attenuate exposure. There are two types of barrier. The pathway barriers include rendering, composting, decay on the soil. Dilution in the soil is also important if the composted waste is sub- surface injected. The biomedical barriers control how infectious the microbiological agent is to the receptors. Examples of biomedical barriers include the species barriers in BSE, acquired protective immunity in C. parvum infection and the natural gut microbiota in Salmonella infection (see Gale, 2001).

The pathways and pathway barriers will be presented in the form of event trees. An example of an event tree is shown in Figure 1.1. This defines the pathways and barriers for transmission of BSE agent from abattoirs to agricultural land through the application of sewage sludge. The fractions define the proportion of BSE infectivity through the pathway and must add up to 1.0 for the arrows coming from each node.

Rendering

Brain/spinal Plant

cord of OTMS 0.99 Sewage

carcass at  Effluent

abattoir 0.01 0.0 Destroyed

Sewage by Sludge Treatment

1.0 0.0

SeRwaawge ? Decay in soil

Sludge 1.0 0.0 and leaching

Treated

Sewage

Sludge 1.0

Soil particle Infectivity 0.9982

remaining

on top soil Dilution

0.0018

Sludge particle

Figure 1.1 Event tree for BSE agent from abattoir to soil through application of

sewage sludge (from Gale & Stanfield, 2001).

3. Variation and uncertainty

There will undoubtedly be natural variation in the Source, Pathway and Receptor Terms. Natural variation in concentrations of pathogens is accommodated in risk assessment either by using Monte Carlo simulations or more simply, by using the arithmetic mean exposure (Haas, 1996; Gale and Stanfield, 2000; Gale and Stanfield, 2001). Indeed, the risk assessment for BSE in sewage sludge was carried out simply by estimating the arithmetic mean concentration for BSE in sewage sludge in England and Wales.

4. General Approach

The approach for environmental risk assessment is based on the Source – Pathway – Receptor approach.

The risk assessment involves gathering information and modelling seven areas. These are:-

1. What is the prevalence of pathogen occurrence in cattle, sheep, pork and poultry food products in the UK. This is relatively easy to assess for diseases such as scrapie and BSE, but is very difficult to address for the exotic pathogens such as FMDV and CSFV. Indeed, levels will fluctuate depending on the incidence in illegally imported meats and whether there is an outbreak in the UK;

2. What is the pathogen loading in the different tissues of infected animals;
3. What happens to individual animal tissues at abattoirs, for sheep, cattle, pigs and poultry;
4. What happens to these different tissues in catering outlets (as catering waste) and domestic kitchens (as MSW);
5. What happens to pathogens during the various composting/biogas processes;
6. How much decay of pathogens is there in the soil; and
7. Dose-response

2. Source Term

"Catering waste" means all waste food originating in restaurants, catering facilities and kitchens, including central kitchens and household kitchens.

Household kitchens will tend to discard kitchen waste to the dustbin for collection by the council. The meat component of municipal solid waste (MSW) is therefore considered.

1. Household consumption of meat

Meat consumption from a survey of 14,584 households by DEFRA is presented in Table

2.1. The average consumption of uncooked meat across  households  was  1,208 g person-1 week-1.

The total annual consumption of meats by household in the UK is therefore 1,208 x 52 x 60 x 106 g = 3.77 million tonnes year-1.

Table 2.1 Household consumption of meats by households in UK (Source

DEFRA)

2. Meat discarded to the waste bin in household surveys

Some preliminary results of the bin-survey carried out in the Wales are presented in Table 2.2 for the summer months and Table 2.3 for the winter months.

Table 2.2 Household waste survey (summer). 250 households sampled.

Table 2.3 Household waste survey (winter). 156 households sampled.

1. Composition of the putrescible material household biowaste

Limited data on the composition of waste produced weekly by 23 households out of a total of around 1600 households participating in a biowaste source-separation scheme have been reported (Barr et al. University of Leeds).

The mean amount of biowaste waste produced by these households was 3.5 kg per household per week. All (99.9 %) of the biowaste was putrescible material although 23.4% of the black bin waste (total 196.1 Kg per week) collected from the same households also contained putrescible material but was observed to be composed mainly of garden waste. The biowaste was composed of fresh food (81%) and cooked food (19%) but no data was given on the type of food that comprised these two categories.

A study was carried out by Greenfinch Limited on behalf of WRc-NSF Ltd in support of this study. The contents of 8 bags (10% of total) collected from a scheme for treating source separated kitchen waste from individual households in Burford, Shropshire. The average kitchen bag weight was 3.4 kg per household per week. Uncooked fruit and vegetables comprised the greatest majority (60% by weight) of the putrescible material. Cooked meat, including bone, accounted for 12% of the total weight whilst the uncooked meat accounted for an ever smaller proportion (1%) of the total weight.

2. Summary

The three studies give remarkably good agree with between 3 and 4 kg household-1 week-1 (Table 2.4).

Table 2.4 Summary of total kitchen waste discarded weekly by households

from three studies

Assuming there are 20 million households in the UK, and on the basis that 1% of the 4.0 kg of kitchen waste discarded weekly is uncooked meat, the total uncooked meat discarded annually would be 4.0 kg x 0.01 x 20 x 106 x 52 x 10-3 = 41,600 tonnes year-1.

The total annual consumption of meat in the UK is calculated as 3.77 million tonnes (Section 2.1). Therefore, 1.1% of meat is discarded uncooked to the bin.

It is assumed for the purpose of the risk assessment that 1% of meat purchased by households and catering establishments is discarded to the bin uncooked.

A survey undertaken on this contract by WRc-NSF of the amount of uncooked meat discarded to bins by local catering establishments is presented in the Appendix together with a similar survey undertaken for domestic households. The results for the domestic kitchen waste are presented in Table 2.5. No account is taken of perception bias.

Table 2.5 Summary of results of survey by WRc-NSF of 39 domestic kitchens.

Note, percentages are only estimates and were not weighed.

The survey suggests that the average percentage discarded is 3.4%. This is 3.4-fold higher than the 1% used in the risk assessment. However, a value of 1% is used in the

risk assessment because it appears for the survey that catering establishments discard very little uncooked meat in catering waste. This is because most of the meat is purchased pre-butchered.

3. Catering Outlets

Catering  waste  includes  hotels,  restaurants,  pubs,  popular  catering,  leisure,  staff catering, health care, and education canteens.

1. Meat use in catering outlets in UK

The quantities of beef, lamb, pork, bacon and ham used in catering outlets in the UK are summarised in Table 2.6.

Table 2.6 Breakdown of meat distribution to catering outlets in UK (The

Foodservice market meat monitor, 2001). Tonnes purchased per week by sector.

2. Amount of uncooked meat discarded by catering outlets

As part of this project WRc-NSF undertook a telephone survey to find out exactly how much  uncooked  meat  is  discarded  by  commercial  catering  operations.  The questionnaire is presented in Appendix 1.

4. Other pathogen sources in household domestic waste In addition to kitchen and food wastes, domestic waste will include:-

• Pet food;

• Dog and cat faeces (cat litter);

• Dead pets (hamsters, mice, cats?);

• Nappies.

Microbial contamination of municipal solid waste (MSW) is mainly of faecal origin (Deportes et al. 1998). For example, 1 – 4% of the dry weight of MSW consists of soiled disposable diapers.

5. Poultry

In England and Wales in 2000, some 615 million broilers were slaughtered (DEFRA). This excludes "Dead on Arrivals" and includes registered slaughter houses only. The average live weight was 2.26 kg. This is therefore 615 x 106 x 0.00226 = 1,389,900 tonnes.

In 2000, household consumption was 214 and 39 g person-1 week-1 of uncooked and cooked poultry (Food Standards Agency). The total consumption of poultry is therefore

253 x 52 x 60 x 106 g = 789,360 tonnes.

Over 700 million chickens are sold per year in the UK (cited in Harris on et al. 2001). These figures appear to be in relatively good agreement.

3. Animal Tissues Composition

1. Cattle

In the UK in 2000, some 2.28 million head of prime cattle were slaughtered for the human food chain (data from Meat and Livestock Commission (MLC)). This included 4,700 live cattle imports, and accounted for 708,000 tonnes of beef products entering the human food chain.

In addition 307,000 tonnes of meat was imported. This comprised 202,000 tonnes of carcasses and 105,000 tonnes of processed meat.

Table 3.1 Cattle by-products (data from MLC). Organs discarded as SRM in

italics.

2. Sheep

According to MLC data, 19.14 million heads of sheep were marketed in the UK in 2000. This included 169,000 head imported. DEFRA statistics (www.defra.gov.uk) confirm 15.96  million  "other  sheep  and  lambs"  and  2.42  million  "ewes  and  rams"  were slaughtered for meat in the United Kingdom in 2000.

In addition 123,000 tonnes of sheep meat was imported into the UK in 2000. The masses of sheep-by-products (per carcass) are summarised in Table 3.2.

Table 3.2 Sheep by-products – weights in lambs. Organs discarded as SRM in italics. For sheep tissues multiply by a factor of 1.6. Data from MLC.

KKCF, Kidney Knob and Channel Fat

3. Pigs

The total number of pigs in the UK June census is 6.5 million in 2000 (MLC pig year book 2001).

DEFRA statistics (www.defra.gov.uk) confirm 12.4 million "clean pigs" and 0.32 million "sows and boars" were slaughtered for meat in the United Kingdom in 2000.

Table 3.3 Pig by-products (Data from MLC)

Farez and Morley (1997) describe the various tissues in which pig viruses replicate. Pork and pork products are comprised principally of skeletal muscle, bone and fat. Bone marrow, blood within the capillaries of skeletal muscles and lymph nodes (prepectoral, presternal, precrural, superficial inguinal, politeal, iliac, lumbar and renal) amount to a very small fraction of the swine carcase. With respect to pork portions, lymph nodes and bone marrow may not be present as a result of trimming, deboning on the cut. Many of the lymph nodes are removed through carcass trimming due to their fat-embodied location on the carcass. The blood, respiratory, GI and reproductive tracts, the head, the respective lymph nodes of these parts and the tonsils are not pork tissues.

The tissues in which the various pig viruses replicate are an important consideration in the risk assessment.

4. Chickens

Giblets

The giblets are defined as the heart, liver and gizzard of a poultry carcass.

A recent survey by the Food Standards Agency showed that 268 of 4,881 (5%) of chickens sampled on retail sale contained giblets.

4. Modelling The Pathways

Destroyed

during

composting

Catering

Waste Composting

Decay No leaching

Residue Decay

in soil Dilution in soil

0.9933

Soil after Probability of collision

X? months with a residue particle 0.0067 Remains in

Residue 0.98 ground 0.02

Proportion of Residue residue transferred on crops

on collision

Figure 4.1 Event tree for transmission of pathogens in composted catering

wastes to root crops.

An event tree outline the main barriers for pathogens in composted catering waste applies to soil is shown in Figure 4.1.

1. Growth of pathogenic bacteria in catering waste

This only applies to the bacterial pathogens (see Section 20.2) and includes spores of Clostridium  botulinum  after germination.  Viruses,  protozoa  and  TSE  agents  cannot replicate outside the host. The main problem in modelling growth of bacterial pathogens in catering waste is allowing for the competition effect of indigenous bacteria. Berry and Koohmaraie (2001) studied the influence of various levels of endogenous beef bacteria on the growth and survival of E. coli O157:H7 on bovine carcass surface tissue. This is considered in greater detail in Section 20.2.

2. Dilution in the soil

There will undoubtedly by dilution of any pathogens present in the compost once it has been applied to the soil. The question for quantitative risk assessment is estimating the dilution factor. This is relatively straight-forward for application of sewage sludge to land (Gale and Stanfield, 2001) because the sludge is tilled into the soil to a depth of 0.25 m. Compost, however, may not necessarily be tilled in, particularly for grasslands. In terms of risk assessment, dilution can be modelled as the probability of a grazing animal ingesting a soil particle, compared to the much smaller probability of ingesting a sludge or  compost  particle.  For  the  purpose  of  risk  assessment,  it  is  assumed  that  any pathogens present in compost are "diluted" down to a depth of 0.1 m (i.e. 10 cm). This could occur through leaching of the pathogens during the 2 month no grazing ban, or through mechanical disruption of soil (e.g. ploughing before sowing of vegetable crops).

The dilution factor assuming 10 tonnes ha-1 of composted residue is tilled in to a depth of 10 cm of soil is 150-fold (Figure 4.2). Thus, the probability of a crop colliding with a compost particle is 0.0067 (Figure 4.1).

10 tonnes residue (dry weight) per ha 100 m

100 m

0.10 m

Volume = 100 x 100 x 0.10 = 1,000 m3

Density soil (dry weight) = 1.5 g/cc Mass of soil = 1,500 tonnes

Dilution = 1,500 / 10 = 150-fold (w/w)

Figure 4.2 Dilution of composted catering waste residues in soil. Based on

assumption that any pathogens in compost leach down to a depth of 10 cm.

3. Decay of pathogens on land

There is a lot of information on the decay of endemic pathogens on soil to which slurries and sludges have been applied. This is now summarised.

4.3.1  Decay of pathogens in sludge-treated soil. Salmonellas

The die-off of salmonellas in sludge-treated soil is affected by several factors such as moisture, temperature and sunlight. Andrews et al. (1983) reported a T90 (time for 1-log decay) for the winter period of 17 d. In the summer the T90 was 3.7 d. Over a period of seven weeks between April and June 1976, Watson (1980) demonstrated a 2.2-log reduction in salmonellas in soil to which treated-sewage sludge had been applied. Watkins and Sleath (1981) demonstrated a 2-log reduction for salmonellas over five weeks in soil to which raw sewage sludge had been sprayed. The experiment was conducted in mid-winter. The reduction over 8 weeks may well have been greater than 2-logs since undetectable readings were recorded at 6-8 weeks. The approach adopted here in the risk assessment is conservative in not extrapolating the decay rates to the 12 month interval specified by the Safe Sludge Matrix between application of conventionally-treated sewage sludge and harvesting of root crops. The model thus allows for only 2-log decay of salmonellas (Table 4.1).

Campylobacters

Unpublished studies of campylobacter survival following land spreading of dairy farm yard manure on a clay loam grassland soil demonstrated a 2.4-log destruction after 16 d (ADAS, pers. comm.). Decays of 2-log and 3-log were reported for beef farm yard manure and beef slurry after 16 d. These are comparable with a 0.74-log decline in campylobacter counts after 5 d in matured unaerated slurry sprayed to land (Stanley et al. 1998). The risk assessment assumes a 2-log decay of campylobacters on the soil (Table 4.1).

E. coli O157

Maule (1995) compared survival of cultured E. coli O157:H7 in soil cores, cattle slurry, cattle faeces and river water at laboratory scale. While die-off was particularly rapid in slurry, E. coli O157:H7 survived best in the soil cores. Counts in the soil increased slightly between day 0 and day 7. Regression analysis of the data of Maule (1995) showed a 1.05-log decay between day 14 and day 63. A 1-log decay is used in the risk assessment model (Table 4.1). An increase of about 1-log (10-fold) in faecal coliform counts was observed by Stanley et al. (1998) in slurry after application to topsoil after one day. Thereafter the counts declined. Field studies of slurry containing E. coli applied to land (Fenlon et al. 2000; Ogden et al. 2001) suggested T90 of 16 days. Strachan et al. (2001) allow for a T90 of 16 days in a model for an environmental outbreak (New Deer, Scotland, May – June 2000) of E. coli O157 infection.

Protozoa

Whitmore and Robertson (1995) studied the survival of C. parvum oocysts in sludge- amended soil mesocosms (2% w/w) At 10C, the viability decreased from 91% to 60- 70% over 30 d. Olson et al. (1999) demonstrated that decay of Cryptosporidium oocysts in soil samples was strongly influenced by temperature. Thus, a 1-log decay of oocysts required 12 weeks at 4C, but only seven weeks for 25C. At soil temperatures of -4C,

however, only 50% of the oocysts had decayed at ten weeks. Viability was determined by dye exclusion. The risk assessment assumes a 1-log decay for Cryptosporidium oocysts in the soil (Table 4.1).

Giardia cysts do not survive as well as Cryptosporidium oocysts in the soil environment. Olson et al. (1999) demonstrated a greater than 1-log loss of viability in cysts at one week in unautoclaved soil at -4C. At 25C, a l-log loss of cyst viability occurred at two weeks. At 4C, a 1-log loss of viability occurred at about six weeks. Hu et al. (1996) demonstrated destruction of Giardia cysts within 12 weeks following soil amendment of anaerobically digested sewage sludge. In the second trial of Hu et al. (1996), cysts were initially present at a concentration of approximately 600 g-1. The population had declined to less than 100 cysts g-1 after one week, and no cysts were detectable after 12 weeks. It would therefore be reasonable to apply a 600-fold (2.78-log) reduction on the basis of the data of Hu et al. (1996). For the purpose of risk assessment, a 2-log destruction in the soil is allowed for over a 12 week period (Table 4.1). This is consistent with an extrapolation of the 4ºC data of Olson et al. (1999) from six weeks to a 12 week period.

Viruses

The retention and persistence of enteroviruses in soil is influenced by a number of factors including virus type, type and texture of the soil, and temperature (Straub et al. 1992; Sobsey et al. 1995). The survival of viruses is enhanced by a combination of low soil temperature and sufficient moisture. Tierney et al. (1977) spiked raw sewage sludge with poliovirus and monitored the decay after application to soil plots. In the winter months a 3-log decay was recorded in about 80 d, while during the summer months a 3- log decay occurred in about 7 d. Straub et al. (1992) applied anaerobically digested sludge, spiked with poliovirus to desert soil in laboratory studies. At 15ºC, a 1-log decay occurred in the region of 10 – 16 d depending on the soil type. The risk assessment (Table 4.1) uses a 3-log reduction based on the experiments of Tierney et al. (1977) which were conducted in Ohio (USA) and are therefore more representative with respect to the climatic conditions in the UK than the desert conditions studied by Straub et al. (1992).

Table 4.1 Summary of parameters for decay of pathogens in sewage sludge

after application to soil.

4. Receptor Term

Information supplied by growers suggested that root crops contain 2% (w/w) soil at point of harvest (Gale and Stanfield, 2001). Thus when a tonne of root crops collides with a tonne of compost residue in the event tree (Figure 4.1), 0.02 tonnes of compost will transfer to the root crops.

Cattle ingest 0.41 kg soil cow-1 day-1 on average (EUSES 1997). It is assumed here that sheep and lambs ingest half this amount. It is assumed that pigs ingest the same amount of soil as cattle.

5. Exposure to grazing cattle, pigs and sheep across England and Wales.

The Composting Development Group (1998) reported an estimated 18 million ha of agricultural land in the UK, of which 11 million ha are land where the addition of compost would have little benefit in increasing prevailing high levels of organic matter (e.g. grassland, natural grazing or woodland). They estimated that compost could be used on about 2.6 million hectares of land. A typical application rate would be 20-25 t ha-1. To avoid excessive applications of phosphate, compost is unlikely to be applied more than one year in four in the longer term.

According to "The State of Composting 1999" (Stater et al. 2001), a total of 462,768 tonnes of composted material was produced in the UK during 1999 (see Section 5.2) from some 833,044 tonnes of feedstock material (see Section 5.1). For the purpose of risk assessment it is assumed that 500,000 tonnes of composted material containing catering waste is applied to 50,000 ha of agricultural land in England and Wales. This represents an arithmetic mean application rate of 10 tds ha-1 year-1. The total tillage and grassland in England and Wales is 9.5 million ha (Table 4.2). Thus, composted catering waste is applied across 0.52% of total tillage and grass land in England and Wales. Assuming compost is applied at a rate of 25 tds ha-1, then 500,000 tonnes could only be spread over 20,000 ha of agricultural land. On this basis, 0.21% of the total tillage and grassland in England and Wales would be used.

Table 4.2 Total tillage and grass land in England and Wales (Anon 1997)

The total number of cattle, sheep and pigs in England and Wales according to the 1996 survey (Anon 1997) are presented in Table 4.3.

Assumption

Assuming cattle, pigs and sheep graze randomly over the total tillage and grassland in England and Wales, then 0.52% of the animals would be exposed to soil to which composted catering waste had been applied. The total numbers of animals exposed are presented in Table 4.3.

Table 4.3 Total number of cattle, pigs and sheep in England and Wales (Anon

1997).

This  is  a  worst  case  assumption  in  that  it  assumes  cattle,  pigs  and  sheep  graze randomly on land across the UK to which composted catering waste may have been applied.

Furthermore, a proportion of pigs may be housed on intensive systems inside and therefore not exposed to land to which compost has been applied.

5. Amount Of Material Composted In The Uk

1. Total feedstock material

The overall picture for composting in the UK is one of continued expansion. One of the key challenges facing the industry over the next few years is the continued expansion required to provide an alternative to landfill for biodegradable waste to contribute to national statutory recycling and composting targets and EU Landfill Directive obligations.

In 1999, Slater et al. (2001) report there were a total of 90 operators running 197 sites and processing approximately 833,044 tonnes of material within the UK. Approximately 74% of this material came from municipal sources and 26% came from non-municipal sources. Of the 618,517 tonnes of municipal waste composted, 72% was green waste from bring sites, 17% was green waste from Local Authority parks and gardens and only 7.5% was collected from the kerbside. Municipal waste composted was comprised of 80% (493,520 tonnes) household waste and 20% (124,997 tonnes) non-household waste.

2. Types of composted product

Slater et al. (2001) report that 462,768 tonnes of composted material was produced in the UK during 1999, accounting for 55% of the total feedstock material. The quantities and proportions of composted material are presented in Table 5.1.

Table 5.1 Quantity and proportion of composted material

6. Particle Size, Temperature And Time For Pathogen Destruction By Composting And Biogas

1. Objectives

The objectives of this section are to:-

1. Review data for heat inactivation data of pathogens for different time/temperature regimes;
2. Consider the effect of particle size in relation to surrounding temperature;
3. On the basis of the most resistant pathogens, define a minimum temperature;
4. Identify a time/temperature combination to be achieved by composting/biogas;
5. Consider the degree of by-pass of different composting/biogas processes; and
6. Formulate credit system of log-removals for use in the risk assessment.

BSE and scrapie prions and  C. botulinum spores are excluded because for the purposes of risk assessment they are considered not to be affected by the temperatures achievable by composting and biogas processes.

2. Introduction

The on going discussions have also been noted concerning the need to achieve a balance between pathogen kill, through the use of high temperatures, and obtaining a stabilised compost.

The use of high temperatures can apparently lead to a loss of the soil conditioning properties of compost and the use of lower process temperatures has been advocated by some. It is proposed that where lower temperatures are used the indigenous microflora of the material being composted will be preserved and that competition for nutrients created by this natural microflora will cause pathogen numbers to decline. Microbiologically this appears to be a sound argument in that composting processes will be an alien and therefore inimical environment for pathogens. The indigenous microbes will be better suited to this environment and will flourish at the expense of any pathogens present through competition for nutrients and predation. Furthermore, the indigenous microbes are important in preventing salmonella regrowth in compost (Sidhu et al. (2001)).

3. Heat inactivation of exotic viruses

Turner  et al. (2000) report data on decontamination of pig slurry containing exotic viruses of pigs (namely FMDV, Aujeszky's disease virus and CSF) by heat inactivation. The work demonstrated the suitability of thermal treatment in ensuring the safety of pig slurry following a disease outbreak.

1. Foot and Mouth Disease Virus

The inactivations of FMDV over a 10-min time period at three temperatures, namely 55C, 60C and 65C, are shown in Figure 6.1. The net destructions, calculated on pooling both the slurry and medium data, are presented in Table 6.1. The data show that even at the lower temperatures of 55C and 60C considerable inactivation occurs over  the  10  minute time  period  of the  experiment.  Thus  even  at  55C  a  1.92-log inactivation was measured over just 10 minutes. It should be noted that during the composting process, meat in catering waste will be exposed for much longer time periods than 10 minutes.

Table 6.1 Net destruction of FMDV over 10 minutes. Data from Turner et al.

(2000). Data averaged for medium and pig slurry counts.