sign in sign up
<<
Home , Fasciolidae, Foodborne illness, Genic, Hepatitis A, Heterophyes, Heterophyidae

WHO ESTIMATES OF THE GLOBAL BURDEN OF FOODBORNE

FOODBORNE BURDEN EPIDEMIOLOGY REFERENCE GROUP 2007-2015 WHO ESTIMATES OF THE GLOBAL BURDEN OF FOODBORNE DISEASES

FOODBORNE DISEASE BURDEN EPIDEMIOLOGY REFERENCE GROUP 2007-2015 WHO Library Cataloguing-in-Publication Data WHO estimates of the global burden of foodborne diseases: foodborne disease burden epidemiology reference group 2007-2015. I.World Health Organization. ISBN 978 92 4 156516 5 Subject headings are available from WHO institutional repository

© World Health Organization 2015 All rights reserved. Publications of the World Health Organization are available on the WHO web site (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: bookorders@ who.int). Requests for permission to reproduce or translate WHO publications –whether for sale or for non- commercial distribution– should be addressed to WHO Press through the WHO website (www. who.int/about/licensing/copyright_form/en/index.html). The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. Printed in Switzerland Design by Paprika-Annecy Cover : © Rania Spatha - Photos : © Shutterstock : Chatrawee Wiratgasem, Business, Chepko Danil Vitalevich, Jenny Sturm, giulio napolitano, vanillaechoes, africa924, Max Topchii, charnw V

CONTENTS

ACKNOWLEDGEMENTS VIII

FOREWORD IX

EXECUTIVE SUMMARY X

1 INTRODUCTION 1 1.1 Motivation: the importance of safety ...... 3 1.2 The value of foodborne disease burden estimates ...... 3 1.3 Purpose and audience ...... 4 1.4 Scope ...... 4 1.5 History and structure ...... 4 1.6 Objectives ...... 5 1.7 Other relevant burden of disease estimates ...... 6 1.8 Timeline: FERG Meetings ...... 7 1.8.1 Overview ...... 7 1.8.2 Extracts from reports of major meetings ...... 7 1.8.3 Strategic revisions ...... 10 1.8.4 Key decisions ...... 10 1.8.5 Country-level involvement ...... 11 1.9 Task Force Meetings ...... 12 1.10 Participants ...... 12 1.11 Declarations of Interest ...... 12

2 COMMISSIONED WORK 13 2.1 Enteric Diseases Task Force ...... 15 2.1.1 spp...... 15 2.1.2 Diarrhoeal disease ...... 15 2.1.3 ...... 15 2.1.4 Shiga -producing ...... 15 2.1.5 ...... 16 2.1.6 Invasive non-typhoidal enterica ...... 16 2.1.7 monocytogenes ...... 16 2.2 Parasitic Diseases Task Force ...... 16 2.2.1 ...... 16 2.2.2 Trematodes (includes spp., buski, spp. and spp.) ..16 2.2.3 Echinococcus multilocularis ...... 17 2.2.4 Trichinella spp...... 17 2.2.5 ...... 17 2.3 Chemicals and Task Force ...... 17 2.3.1 Aflatoxins ...... 17 2.3.2 Arsenic ...... 17 2.3.3 Cassava ...... 17 2.3.4 Peanut Allergens ...... 18 2.3.5 Dioxins ...... 18 2.4 Source Attribution Task Force ...... 18 2.5 Computational Task Force ...... 18 2.6 Country Studies Task Force ...... 19 2.7 Other relevant publications ...... 19 VI

3 OVERVIEW OF THE SCIENTIFIC APPROACH 21 3.1 The DALY metric ...... 23 3.2 Overarching methodology decisions by FERG in relation to DALY estimates ...... 23 3.2.1 Hazard-based approach ...... 23 3.2.2 -based approach ...... 23

4 HAZARD-SPECIFIC METHODOLOGY 27 4.1 Hazard Selection ...... 30 4.2 Enteric Hazards ...... 31 4.2.1 Estimating cases, sequelae and for diarrhoeal diseases...... 32 4.2.2 Estimating cases and sequelae of, and deaths due to, extra-intestinal diseases ...... 33 4.3 Parasitic Hazards...... 35 4.4 Chemicals and toxins ...... 38 4.4.1 Cyanide in cassava ...... 38 4.4.2 Peanut allergen ...... 39 4.4.3 Aflatoxin ...... 40 4.4.4 Dioxin ...... 40 4.5 Outcomes and disability weights ...... 42 4.6 Attribution ...... 45 4.6.1 Identification of experts ...... 45 4.6.2 Selection of experts ...... 46 4.6.3 Expert panels ...... 47 4.6.4 Analytical method ...... 47 4.6.5 Seed questions ...... 47 4.6.6 Target questions ...... 49 4.6.7 Data analysis ...... 49 4.7 Computation ...... 51 4.7.1 Disease models and epidemiological data ...... 52 4.7.2 CTF database template ...... 56 4.7.3 Imputation ...... 56 4.7.4 Probabilistic burden assessment ...... 58

5 RESULTS 61 5.1 Attribution ...... 63 5.1.1 Expert performance...... 64 5.1.2 Pathway attribution results ...... 65 5.2 DALY Estimates: Overview ...... 72 5.3 DALY Estimates: Enteric diseases ...... 84 5.4 DALY Estimates: Parasites ...... 86 5.5 DALY Estimates: Chemicals ...... 89

6 DISCUSSION 93 6.1 Attribution ...... 98 6.2 Enteric Diseases ...... 102 6.3 Parasites ...... 104 6.4 Chemicals ...... 110 VII

7 COUNTRY STUDIES 113 7.1 Aim and Objectives of the Task Force and Objectives ...... 115 7.2 Tools and resources to facilitate national burden of foodborne disease studies ...... 115 7.3 Pilot Studies...... 116 7.4 Process ...... 116 7.4.1 Albania ...... 117 7.4.2 ...... 117 7.4.3 Thailand ...... 118 7.4.4 ...... 118 7.5 Findings and Lessons Learned ...... 119 7.5.1 Data gaps ...... 119 7.5.2 Public and private data sources ...... 119 7.5.3 Foodborne versus waterborne disease ...... 120 7.5.4 Situation analysis and knowledge translation ...... 120 7.6 Discussion ...... 120

8 CONCLUSION 123 8.1 Reflections on the WHO Initiative to Estimate the Global Burden of Foodborne Diseases ...... 125

APPENDICES 129 Appendix 1. Formal Description of the Project and Participants ...... 131 Appendix 2. Subregions ...... 143 Appendix 3. Preliminary hazards considered by each task force ...... 144 Appendix 4. Hazard-specific input parameter sources and methods ...... 145 Appendix 5. Disease models ...... 190 Appendix 6. Derivation of Disability Weights ...... 195 Appendix 7: Attribution – Expert Elicitation Results ...... 199 Appendix 8. Data tables for individual hazard classes: enteric, parasitic, chemical ...... 211

REFERENCES 223

GLOSSARY 251

ABBREVIATIONS 252 VIII

ACKNOWLEDGEMENTS

WHO gratefully acknowledges the financial and in-kind contributions from , Denmark, Germany, Ireland, Japan, Republic of Korea, Netherlands, of Great Britain and Northern Ireland, of America and other donors. WHO also acknowledges the invaluable contribution of the members of the Foodborne Disease Burden Epidemiology Reference Group, the resource advisors, and their affiliated organizations who supported this initiative. They are named in Appendix 1. This report is a compilation of the main reports of the task forces of the Reference Group. Numerous persons contributed to or read the manuscript and provided comments, including several units in WHO headquarters and Regional Offices. The text was then edited to introduce conformity in language and style. IX

FOREWORD

Foodborne diseases have been an issue The objective of the initiative was not for all societies since the beginning limited to providing estimates on the of humanity. The types, severity global burden of foodborne diseases and impacts of these illnesses have for a defined list of causative agents of changed through the ages and are microbial, parasitic and chemical origin. still diverse across regions, countries The initiative also aimed at strengthening and communities. the capacity of countries to conduct assessments of the burden of foodborne Yet there are some challenges common disease, and encouraging them to use to all countries. Only a fraction of the burden of foodborne disease estimates people who become sick from food they for cost-effectiveness analyses of have eaten seek medical care. Only a prevention, intervention and control fraction of those cases are recognized measures including implementation of as having been caused by a hazard in standards in an effort to food, treated accordingly, reported to improve national food safety systems. authorities and recorded in official disease statistics. Certain chronic Six taskforces were established under diseases, such as cancer, kidney or FERG, focusing on groups of hazards failure, that result from contaminated or aspects of the methodology. These food appear long after the of taskforces commissioned systematic food and the causal link is never made reviews and other studies to provide for each case. This points to some of the data from which to calculate the the challenges inherent in measuring the burden estimates. burden of foodborne diseases and the This report is an outcome of a decade toll they take on lives and economies. of work by WHO, key partners and Up to now, the global burden of illness a number of dedicated individuals. and deaths caused by foodborne Some additional findings, which cannot disease has never been quantified. In be integrated into this report, will be order to fill this data vacuum, the World published and user-friendly online tools Health Organization (WHO), together made available separately. with its partners, launched in 2006 the This report and related tools should Initiative to Estimate the Global Burden enable governments and other of Foodborne Diseases. After an initial stakeholders to draw public attention to consultation, WHO in 2007 established a this often under-estimated problem and Foodborne Disease Burden Epidemiology mobilize political will and resources to Reference Group (FERG) to lead the combat foodborne diseases. initiative.

Kazuaki Miyagishima Director Department of Food Safety and Zoonoses WHO Estimates of the global burden of foodborne diseases X

EXECUTIVE SUMMARY

Foodborne diseases are an important in 2010; 40% of the foodborne disease cause of morbidity and mortality, and burden was among children under 5 a significant impediment to socio- years of age. Worldwide, 18 (95% UI economic development worldwide, but 12–25) million DALYs were attributed to the full extent and burden of unsafe food, foodborne diarrhoeal disease agents, and especially the burden arising from particularly NTS and enteropathogenic chemical and parasitic contaminants, Escherichia coli (EPEC). Other foodborne has been unknown. Precise information hazards with a substantial contribution on the burden of foodborne diseases to the global burden included Salmonella can adequately inform policy-makers Typhi and Taenia solium. and help to allocate appropriate Foodborne burden estimates are also resources for food safety control and reported for a further 4 bacterial and intervention efforts. 1 chemical hazards, but only for some This report, resulting from the WHO subregions; a global estimate was Initiative to Estimate the Global Burden not feasible. of Foodborne Diseases and prepared by There were considerable differences the WHO Foodborne Disease Burden in the burden of foodborne disease Epidemiology Reference Group (FERG), among subregions delimited on the provides the first estimates of global basis of child and adult mortality. The foodborne disease incidence, mortality, highest burden per population was and disease burden in terms of Disability observed in (AFR) (AFR D Adjusted Life Years (DALYs). For the and AFR E subregions), followed by global estimates, thirty-one foodborne South-East (SEAR) (SEAR B and hazards causing 32 diseases are included, SEAR D) subregions and the Eastern being 11 diarrhoeal disease agents (1 Mediterranean (EMR) D subregion. , 7 , 3 ), 7 invasive Diarrhoeal disease agents were the infectious disease agents (1 virus, 5 leading cause of foodborne disease bacteria, 1 protozoon), 10 helminths and burden in most subregions. NTS was 3 chemicals. an important burden in all subregions, Together, the 31 global hazards caused particularly in Africa. Other main 600 (95% uncertainty interval [UI] diarrhoeal causes of foodborne disease 420–960) million foodborne illnesses burden were EPEC, enterotoxigenic and 420,000 (95% UI 310,000–600,000) E. coli (ETEC) and cholerae in deaths in 2010. The most frequent causes low-income subregions, and of were diarrhoeal spp. in high-income disease agents, particularly norovirus subregions. The burden of aflatoxin and Campylobacter spp. Foodborne was high in the AFR D, Western Pacific diarrhoeal disease agents caused (WPR) B and SEAR D subregions. 230,000 (95% UI 160,000–320,000) In the SEAR subregions there was a deaths, particularly non-typhoidal considerable burden of Salmonella Typhi. Salmonella enterica (NTS, which causes The burden of Opisthorchis spp. was diarrhoeal and invasive disease). Other concentrated in the SEAR B subregion, major causes of foodborne deaths were where the -borne trematodes Salmonella Typhi, Taenia solium, Paragonimus spp. and A virus, and aflatoxin. The global burden were also important. In the Americas of foodborne disease by these 31 hazards (AMR) B and D subregions, Taenia solium was 33 (95% UI 25–46) million DALYs and Toxoplasma gondii contributed national studies ofthefoodborne resources were created to facilitate informed policies.Asuite oftools and burden information insettingevidence- studies, andencouraging theuseof national foodborne diseaseburden This involved capacitybuildingthrough to promote actionsatanationallevel. regional estimates, theInitiative sought In additionto providing globaland gaps andlimitations ofthisstudy. studies are neededto address thedata These estimates are conservative; further of theworld. persons livinginlow-income subregions by children under5years ofage, andby individuals ofallages,butparticularly foodborne diseasesisborneby regional variations. The burden of diseases isconsiderable, withmarked burden. The globalburden offoodborne significantly to thefoodborne disease Executive summary national, regional andinternational levels. estimates into policydevelopment at the food chainby incorporating these improvements infood safety throughout All stakeholders cancontribute to in low-income subregions oftheworld. under 5years ofageandpersonsliving of allages,butparticularlychildren is considerable, andaffects individuals the globalburden offoodborne disease these initialestimates, itisapparent that Despite thedata gapsandlimitations of developed. surveillance andlaboratory capacityis offer aninterim solution,untilimproved regional estimates provided by FERG national studies, andtheglobal the foodborne diseaseburden inthese were themajorhurdle inestimating Japan, Thailand andUganda).Data gaps conducted infour countries (Albania, burden ofdisease, andpilotstudies were XI

EXECUTIVE SUMMARY WHO Estimates of the global burden of foodborne diseases 1

INTRODUCTION1

WHO Estimates of the global burden of foodborne diseases 3

INTRODUCTION

1.1 Motivation: the importance of Foodborne take advantage of food safety weak immune systems. and young children, pregnant women, the elderly Safer food saves lives. With every bite as as those immuno-compromised, one eats, one is potentially exposed are particularly at risk of contracting to illness from either microbiological and dying from common food-related or chemical contamination. Billions of diseases. Malnourished infants and people are at risk and millions fall ill every children are especially exposed to year; many die as a result of consuming foodborne hazards and are at higher risk unsafe food. of developing serious forms of foodborne Concerns about food safety have diarrhoeal diseases; these skyrocketed in more affluent societies. in turn exacerbate malnutrition thus However, the real tragedy of foodborne leading to a vicious circle of debilitation diseases is played out in the developing and mortality. Those who survive may world. Unsafe water used for the suffer from delayed physical and mental cleaning and processing of food; poor development, depriving them of the food-production processes and food- opportunity to reach their full potential handling (including inappropriate use of in society. agricultural chemicals); the absence of Beyond the individual level, foodborne adequate infrastructure; diseases affect economic development, and inadequate or poorly enforced particularly challenging the tourist, regulatory standards–these all contribute agricultural and food (export) industries. to a high risk environment. Moreover, Developing countries’ access to food as a country’s economy develops, the export markets will depend on their agricultural landscape changes. Intensive capacity to meet the international husbandry practices are put in regulatory requirements determined by place to maximize production, resulting the Agreement on the Application of in the increased of pathogens Sanitary and Phytosanitary Measures in flocks and herds. The tropical climate (SPS) of the World Trade Organization of many developing countries favours (WTO). Unsafe exports can lead to the proliferation of pests and naturally significant economic losses. occurring toxins, and the risk of contracting parasitic diseases, including worm infestations. 1.2 The value of foodborne disease burden estimates While exposed to more hazardous environments, people in developing Foodborne diseases (FBD) are an countries often have difficulty coping important cause of morbidity and with foodborne disease. For many living mortality worldwide but the full extent at or below the poverty line, foodborne and cost of unsafe food, and especially illness perpetuates the cycle of poverty. the burden arising from chemical The symptoms of foodborne diseases and parasitic contaminants in food, range from mild and self-limiting (, is still unknown. Detailed data on the and diarrhoea) to debilitating economic costs of foodborne diseases in and life-threatening (such as kidney and developing countries are largely missing. liver failure, brain and neural disorders, Despite the growing international paralysis and potentially cancers), leading awareness of foodborne diseases as to long periods of absenteeism and a significant risk to health and socio- premature . information on theInitiative. it provides acomprehensive source of scientific issues inmore detail.Assuch, itself, aswell asexamining particular background andcontext ontheproject results, thisreport isintended to provide Diseases. Inadditionto collating the Estimate the GlobalBurden ofFoodborne generated by theWHOInitiative to (PLOS), whichcover theestimates from thePublicLibrary ofScience scientific paperspublishedinjournals This report isasupplementto the 1.3 other and policy-makers enable To the Initiative is: Foodborne Diseases.The primarygoalof to Estimate theGlobalBurden of with itspartnerslaunchedtheInitiative and Foodborne Diseases(FOS) together Department ofFood Safety, Zoonoses World Health Organization (WHO) In order to fillthisdata vacuum, the safety control andintervention efforts. allocate appropriate resources for food adequately inform policy-makers and on theburden ofFBDisneededto economic impact.Precise information if connected to majorpublichealthor uninvestigated, andmay onlybevisible often gounrecognized, unreported or the most visiblefoodborne outbreaks particularly inthedeveloping world. Even on foodborne diseasesremain scarce, allocate resources. Epidemiologicaldata makers to setpublichealthprioritiesand diseases, whichwould enablepolicy- the fullextent andcost offoodborne concerns isthelackofaccurate dataon to adequately addressing food safety remains marginalized. Amajorobstacle economic development, food safety Introduction stakeholders to set appropriate, appropriate, set to stakeholders evidence-based priorities in the area of of area the in priorities evidence-based food safety. food Purpose andaudience discussions ofFERG.The meeting consultation thatinformed thetechnical commenced withastakeholder 2007. This multi-disciplinarymeeting who metfor thefirst timeinNovember General appointed theFERGmembers the scientificpress theWHODirector- Following apubliccallfor advisersin Reference Group (FERG)to engagein: Foodborne DiseaseBurden Epidemiology and mandated WHOto establish a for theassessment ofFBDburden, provided thestrategic framework international experts. This consultation consultation attended by over 50 of Foodborne Diseasesataninternational Initiative to Estimate theGlobalBurden In September 2006,FOS launchedthe 1.5 This report covers: 1.4 f f f f f f f f f f f f

future plans. results; and outputs, implications andcontext of hazards-based approach; methods, results, discussion, usinga burden offoodborne disease; overview ofapproach to estimating the project; scientific work commissioned by participants; history oftheproject; burden ofFBDstudies at country level. developing user-friendlytools for diseases that are foodborne, and models to estimate the proportion of developing causeandsource attribution FBD burden where dataare lacking; providing modelsfor theestimation of each ofthemajorFBDs; mortality, morbidityanddisabilityin conducting epidemiologicalreviews for disease estimates; currently existing burden offoodborne assembling, appraising andreporting on History andstructure Scope 4

INTRODUCTION WHO Estimates of the global burden of foodborne diseases 5

saw the establishment of the FERG and external resource and technical Core or Steering Group (coordinating advisers who are invited on an ad hoc and overseeing the burden work) as basis to provide specific expertise. well as several thematic Task Forces (TFs) to advance the work in specific 1.6 Objectives areas, including: The first report from the Initiative, f Enteric Diseases Task Force (EDTF); published in 2008, described the f Parasitic Diseases Task Force (PDTF); following objectives1: and, f Chemicals and Toxins Task f To strengthen the capacity of countries Force (CTTF). in conducting burden of foodborne disease assessments and to increase Subsequently, three additional Task the number of countries who have Forces were established to address the undertaken a burden of foodborne following topics: disease study. f Source Attribution Task Force (SATF) f To provide estimates on the global (established 2008); burden of foodborne diseases f Country Studies Task Force (CSTF) according to age, sex and regions for (established 2009), with a sub-group, a defined list of causative agents of the Knowledge Translation and Policy microbial, parasitic and chemical origin. Group (KTPG) (established 2010); f To increase awareness and f Computational Task Force (CTF) commitment among Member States (established 2012). for the implementation of food As shown in Figure 1, FERG consists of a safety standards. Core (or Steering) Group to coordinate f To encourage countries to use burden and oversee the scientific work, Thematic of foodborne disease estimates for TFs advancing the work in specific areas; cost-effective analyses of prevention, intervention and control measures.

1 http://www.who.int/foodsafety/foodborne_ disease/Summary_Doc.pdf?ua=1 Accessed 9 July 2014 sease estimates published by the following: comprehensive estimates are those number ofresearch groups. The most foodborne have beenpublishedby a considered to beatleast partially Estimates for theburden ofdiseases 1.7 Initiative took two approaches. To meetthesegoalsandobjectives, the Figure 1. Introduction f f

disease studies. to conduct theirown burden of the work ofFERGandenablecountries In-depth country studies to supplement disease estimates. and report onburden offoodborne was established to assemble, appraise Epidemiology Reference Group (FERG) A Foodborne DiseaseBurden organizations withastake and/or burdenofdisease. External expertswhojoin eight WHODepartments the FERGtosupplement Composed ofstafffrom in foodbornedisorders Other relevant burden ofdi- Resource Advisors WHO Secretariat the group’sskills. and UNpartner FERG adhoc Structure oftheinitiative to estimate theglobalburden offoodborne diseases

Core /Steering Group FERG Task Forces the estimates. share dataand promote consistency of communicated withthese groups, to of foodborne diseaseproject, FERG Throughout the course oftheburden 4 3 2 f f f http://www.healthdata.org/gbd Accessed 24 http://www.who.int/topics/global_burden_of_ http://globocan.iarc.fr/Default.aspx Accessed 24

Cancer (GLOBOCAN) International Agency for Research on and Prevalence 2012,publishedby the Estimated Cancer Incidence, Mortality of WHO Mortality andBurden ofDiseaseUnit Evaluation (IHME) the Institute ofHealth,Metricsand (GBD2010) study, undertaken by Global Burden ofDisease2010 disease/en/ Accessed 24 September 2014 September 2014 September 2014 3 (expressed inDALYs). burden offoodbornediseaseestimate generated byothertaskforcestocalculate Utilizing epidemiologicalinformation COMPUTATIONAL TASKFORCE safety interventions. skills tomonitortheprogressoffood analysis andequippingCountrieswiththe disease burdenstudiesandpolicysituation Countries intheconductionoffoodborne Developing userfriendlytoolstoaid COUNTRY STUDIESTASKFORCE responsible foodsource. relevant fractionofdiseaseburdento contamination andaimingtoattributethe burden thatisdirectlyduetofood Seeking toidentifytheproportionofdisease SOURCE ATTRIBUTIONTASKFORCE chemicals andtoxins. Advancing theburdenworkinareaof CHEMICALS ANDTOXINSTASKFORCE to parasites. Specializing infoodbornediseasesrelated PARASITIC DISEASESTASKFORCE viral &bacterialdiseasesinnature. Specializing infoodbornediseasesthatare ENTERIC DISEASESTASKFORCE 2 4

6

INTRODUCTION WHO Estimates of the global burden of foodborne diseases 7

1.8 Timeline: FERG Meetings The objectives of the meeting were:

f to launch an appeal for wider 1.8.1 Overview collaboration, with a detailed plan of f 25–27 September 2006 – Establishment action and time frame; 5 of the initiative, Geneva f to develop a strategic framework for f 26–28 November 2007 – FERG burden of disease estimation that 1, Geneva6 involved all relevant partners; and f 17–21 November 2008 – FERG 2, f to propose elements of a standard Geneva7 (plus Stakeholder Meeting)8 protocol for conducting burden f 26–30 October 2009 – FERG 3, Geneva of illness studies in countries to (plus Stakeholder Meeting) obtain estimates. f 8–12 November 2010 – FERG 4, Geneva9 f 7–10 November 2011 – Strategy Meeting The result of the Consultation was a draft and Commencement of Country strategic framework for the assessment Studies, Durrës, Albania of burden of foodborne diseases, which f 8–12 April 2013 – FERG 5, Geneva10 included: f 23–25 June 2014 – Review Meeting, f the outline of an evidence map for Copenhagen assimilating existing information on the 1.8.2 Extracts from reports of burden of disease [along themes of major meetings – (i) infectious diseases, – (ii) chronic manifestations of 25–27 September 2006– Establishment infectious diseases; and of the initiative, Geneva – (iii) acute and chronic non-infectious WHO’s Department of Food Safety, illness]; and Zoonoses and Foodborne Diseases f a time frame outlining the individual (FOS) launched an initiative to estimate strategic activities in relation to the the global burden of foodborne evidence framework. diseases from all major causes, In order to complete the strategic including chemicals and zoonoses, at an and technical framework, participants international consultation. This was held mandated WHO to establish a in Geneva, Switzerland, from 25 to 27 Foodborne Disease Burden Epidemiology September 2006, and was attended by Reference Group (FERG) and proposed over 50 experts from around the world. the relevant skill mix required for this group. A number of funding agencies were identified that might be approached by WHO to enable the execution of this work. The Consultation concluded with the drafting of a Joint Statement of 5 http://www.who.int/foodsafety/publications/ burden_sept06/en/ Accessed 21 April 2015 Support for the Initiative. 6 http://www.who.int/foodsafety/publications/ A summary document describing the burden_nov07/en/ Accessed 21 April 2015 initiative was published in 200811. 7 http://www.who.int/foodsafety/publications/ferg2/ en/ Accessed 21 April 2015 26–28 November 2007 – FERG 1, Geneva 8 http://www.who.int/foodsafety/publications/ferg- stakeholders/en/ Accessed 21 April 2015 9 http://www.who.int/foodsafety/publications/ferg4/ en/ Accessed 21 April 2015 10 http://www.who.int/foodsafety/publications/ferg5/ 11 http://www.who.int/foodsafety/foodborne_disease/ en/ Accessed 21 April 2015 ferg/en/ accessed 21 April 2015 the next year. technical steps to betaken by FERGover work; and(3) agreed onthelogistic and to commission theindividualburden concrete andvery detailedwork plans should beconducted; (2) developed agents for whichburden assessments provided: (1) prioritylists ofcausative In theirrespective areas, theTFs first year, whichincluded: the progress madeduringtheInitiative’s November 2008(FERG2)highlighted The second formal meetingofFERGin Geneva 2, FERG – 2008 November 17–21 areas, including: (TFs) to advance thework inspecific well asseveral thematicTask Forces overseeing theburden ofthework), as or Steering Group (coordinating and the establishment oftheFERGCore discussions ofFERG.The meetingsaw consultation thatinformed thetechnical meeting commenced withastakeholder November 2007. This multi-disciplinary members, whometfor thefirst timein General ofWHOappointed theFERG in thescientificpress, theDirector- Following apubliccallfor advisers Introduction f f f f f f (plus Stakeholder Meeting) Stakeholder (plus technical recommendations; (SATF) andexecution ofits Source Attribution Task Force establishment ofthe FERG commissioned; including new burden work to be work plansfor allFERGTFs for 2009, the development ofdetailed new as mortality; causes offoodborne diseases,aswell of enteric, parasitic andchemical reviews commissioned intheareas preliminary results often systematic an appraisal ofthemethods,and enteric diseases. chemicals andtoxins; and parasitic diseases; meeting was to enablestakeholders to: and scientificmedia.The purposeofthe consumer groups; industry; andpublic governmental organizations (NGOs); bilateral andmultilateral donors;non- They included:WHOMember States; production, consumption oradvocacy. it through decision-making,research, an interest inensuringfood safety, be various constituencies andsectors with international representatives from the Stakeholder Meetingbrought together The Third Foodborne Diseases Geneva 3, FERG – 2009 October 26–30 presented anddiscussed atFERG4: mortality andburden estimates were of new foodborne disease morbidity, previous FERGmeeting,alarge number Geneva 4, Continuing onthe path taken duringthe FERG – 2010 November 8–12 f f f f f f f f (plus Stakeholder Meeting) Stakeholder (plus

communication andpolicy. in theareas oftechnical reviews, input andrecommendations to WHO meeting, whichprovided valuable A major, multisectoral stakeholder of FERGactivities;and processes andoutputsofthefirst year formal evaluation oftheactivities, Force (CSTF) in2009; of thenew FERGCountry Studies Task agreement ontheterms ofreference trematodiasis; the globalburden offoodborne diseases; the globalburden of diarrhoeal evidence andpolicy. about how to bridgethegapbetween provide direct inputinto discussions cooperation; and open new channelsfor multisectoral its research; Reference Group (FERG)and Disease Burden Epidemiology actively engage withtheFoodborne 8

INTRODUCTION WHO Estimates of the global burden of foodborne diseases 9

f the global burden of cystic favourably in comparison with other echinococcosis; international advisory bodies in which f the global burden of neurocysticercosis; some of the experts had been involved. f the global burden of aflatoxicosis; and FERG experts recognized the complexity f the global burden of cassava cyanide of the Initiative, and some reported that ingestion. at the outset they had doubts about In addition, the FERG experts appraised whether it was achievable. However, the progress made on the systematic they had found that challenges had reviews commissioned for other enterics, been overcome and continued to be parasites and chemicals, and on the addressed, many products were being protocols to be used in the source produced and some had already been attribution expert elicitation process and finalized. The project was being managed in the national FBD burden assessments very energetically, and they expected and policy situation analyses. Each TF successful outcomes in due course. also made recommendations for new There was also a high satisfaction level commissioned work. with the guidance and direction of The various TFs adopted their work plans the FERG and Task Force Chairs. The for 2011 and beyond, which covered the global and regional representation of continuation of the systematic reviews, the FERG membership was valued and the finalization of the priority FERG experts reported that through lists, and the further strengthening of the their involvement, many of them had interfaces between the different TFs. The increased their own capacity. Stakeholder Country Studies Task Force made the involvement was valued by FERG experts final preparations for initiating the pilot and by the stakeholders themselves. country studies in 2011. Four countries Continued expansion of stakeholder were selected for these studies: Albania, constituencies was also suggested by Japan, Thailand and Uganda. both groups. July 2010 Mid-term evaluation A high quality of all outputs was commissioned from an considered very important by FERG external consultant experts, and should be maintained. Most FERG experts were satisfied The overall verdict of this evaluation with the outputs already produced– of the World Health Organization pathogen- and hazard-specific mortality Initiative to Estimate the Global Burden and morbidity reports– although there of foodborne diseases was that it was was acknowledgement that there had making good progress. FERG experts been delays (some of which might not and stakeholders considered it to be have been avoidable), and that there a very important Initiative and were in remained a lot more work to be done. agreement with its goals and objectives. The delays occurred initially and were They recognized that information on mostly considered inevitable. They were the burden of foodborne diseases is dealt with, and FERG experts considered required at country, regional and global that the Initiative was progressing levels in order to prioritize food safety according to plan. Stakeholders were interventions. The leadership and satisfied with the results presented at management of the Initiative by the stakeholder meetings to date, and they WHO Secretariat was highly praised by looked forward to the production of the FERG experts, and described very more results. existing skills.FERGexperts requested team attheSecretariat, with more ofthe and therefore requested anexpanded sustainability andlack ofa‘safety net’, could fail. They were concerned about Initiative would be very vulnerable and there were any changesto personnel,the FERG experts were concerned thatif so far was ascribedto these qualities. motivation, and muchofthesuccess enthusiasm, energy, dedicationand excellent in terms oftechnical expertise, These few key peoplewere considered key personnelintheWHOSecretariat. its success onsuchasmallnumberof the dependence oftheInitiative and a majorthreat to theInitiative, namely FERG experts were concerned about change that could bemade. resources was themost appropriate the Initiative’s andfinancial foodborne diseases.Therefore, increasing estimation oftheglobalburden of State anddonorexpectations for initial there was alsotheneedto fulfilMember to theriskofloss ofmomentum,and a limitation ontimelineextension due and stakeholders stated thatthere was can bereviewed, butFERGexperts to timelinesandresources. Timelines maintained, adjustments must bemade and appropriate. Becausequalitymust be expansion ofthescope was necessary experts were inagreement thatthe expanded scope oftheInitiative. FERG safety– allthesewere partofthe Assembly (WHA)resolution onfood appropriately to the63rd World Health knowledge translation andrespond methodological challenges,integrate plan to collect primarydata, overcome the scope oftheInitiative. The needto was how to dealwiththeexpansion to The mainchallengeto theInitiative considered to bevery effective. of theInitiative were praised and The advocacy efforts ofthe coordinator Introduction that hadbeenmade. Initiative andtheconsiderable investment resources to ensure thesuccess ofthe Initiative through providing thenecessary WHO reiterate theirsupportfor the that highlevel seniormanagementat 1.8.4 Key decisions objectives of: Secretariat convened ameetingwiththe the Initiative was operating, theWHO as thechangedenvironment inwhich of theWHOFERGInitiative, aswell In view oftheincreased complexity 1.8.3 Strategic revisions Meeting Strategy 2011– November 7–10 f f f f f f f and Commencement of Country Studies, Studies, Country of Commencement and Durrës, Albania Durrës,

framework, itsmilestones andtimelines; updating theInitiative’s strategic new Computational Task Force Force: New FERG Computational Task metric indicatives. compatibility withother existing health in order to ensure accuracy, utilityand in Albania,andactionsagreed upon burden were discussed atthemeeting estimation offoodborne disease methodological issues linked to the range ofimportanttechnical and Methodological decisions: mortality estimates by the end of2012. they intended to deliver incidence and pathogens andhazards for which TFs, established ashortened list of TF chairs,inconsultation withtheir Scope oftechnical work: and responsibilities. updating FERGprocesses, roles needs for implementation; and identifying key activitiesandresource burden estimation; priority areas for foodborne disease Initiative, includingtheselectionof redefining thetechnical scope of the Continuing inthisvein, a The thematic A 10

INTRODUCTION WHO Estimates of the global burden of foodborne diseases 11

(CTF) to work on the mathematical the pilot studies and finishing the country modelling to calculate DALYs would tools. Each TF had outlined its priority be established. The TF was currently activities for the coming year and WHO set to be operational by the end of would use these to solicit the funding February 2012. required to complete the FERG project. f Source attribution expert elicitation: The hazard TFs– EDTF, PDTF and CTTF– An expert elicitation would be completed the technical review of the conducted in 2012 to determine what systematic reviews; reviewed and revised proportion of the burden of each the final outcome trees; and made plans hazard was foodborne, and which for completion for each hazard. were the major associated with . A list of hazards SATF finalized the expert elicitation that would be included in the expert protocol for: chemicals and toxins elicitation was established. (inorganic arsenic, lead, cadmium and dioxins); for parasitic diseases 1.8.5 Country-level involvement (, The ‘kick-off’ meeting of the FERG spp., spp., Echinococcus pilot country studies marked a major granulosus, Toxoplasma gondii, milestone for the work of the Initiative Echinococcus multilocularis and Ascaris in fostering national studies of the spp.); and for enteric diseases (diarrhoeal burden of foodborne disease. For the diseases [non-typhoidal Salmonella first time, representatives of the FERG spp., Campylobacter spp., Shiga-toxin pilot countries met to present progress producing, enteropathogenic and on the implementation of a national enterotoxigenic E. coli, norovirus, foodborne disease burden study. They spp., ], typhoid, brucellosis also learnt about study tools that FERG and ). had developed, as well as the future technical support that would be provided The methodology and elicitation by FERG. instrument were agreed with each of the hazard TFs. This expert elicitation would The pilot countries during the kick- be the first time that the methodology off meeting: had been applied at a global level for f drafted pilot country study work plans food safety and would involve disease outlining the way forward; experts from all six WHO regions. The f provided recommendations and input logistics of such an enormous task were to align FERG procedures and tools also mapped out and agreed during for national foodborne disease burden the meeting. estimation and food safety policy CTF (established October 2012) was situation analyses specific to country able to agree on the disease models requirements; and for the majority of the pathogens, as f delivered feedback and agreed on well as meeting individually with each processes to communicate between TF to advance the DALY calculations. participating countries, and between The database was revised, methods the countries and FERG Secretariat. for imputation of missing data were 8–12 April 2013 – FERG 5, Geneva advanced, and disability weights (DWs) were mapped to all outcomes. There was a very clear path towards the end goal of publishing the estimates of CSTF and KTPG agreed the aims, burden of foodborne disease, completing objectives and outline for the joint SATF PDTF CTTF EDTF draft meetingreports are available. For themeetingsmarked*, finalized or teleconferences were heldby eachTF. addition to thesemeetings,numerous Only face-to-face meetingsare listed. In 1.9 with thoseusedby theWHOunit. estimation approaches to beharmonized WHO, Geneva. This enabledtheFERG Health System andInnovation Cluster at and Burden ofDiseaseUnitinthe Mathers, theCoordinator oftheMortality greatly by theattendance by DrColin all theTFs. Progress was enhanced discussions ofburden estimation across This working meetinginvolved detailed Meeting, Strategy 2014– June 23–25 the commissioned work. analysis documentand theoutcome of results, andreviewed thesituational strategy for theglobalandregional FERG development ofthecommunications country study workshop, initiated the Introduction Copenhagen Copenhagen 14 -18July2010–Tunis, Tunisia 20 -22April2010 –Atlanta, USA* Lumpur, Malaysia* 28 -30April2008–Kuala 14 -18July2010–Tunis, Tunisia 7 -9June2009–Rome* 14 -18July2010–Tunis, Tunisia 14 -16July2009–Geneva* 14 -18July2010–Tunis, Tunisia 7 -9June2009–Rome* Task Force Meetings CSTF CTF unbiasedness ofthework. Secretariat to ensure theneutrality and interests were assessed by theWHO matter ofthemeeting. Alldeclared information relevant to thesubject interests, andto provide any additional participants were asked to confirm their Interests. At thestart ofeachmeeting,all standard form for Declaration of completed beforehand theWHO invited to participate inFERGmeetings All experts andresource advisers 1.11 See Appendix1. 1.10 (Uganda), Kampala, Uganda 4 -6March 2012 –Kickoffmeeting Durrës, Albania meeting (Albania,Japan,Thailand), 7 -10November 2011–Kickoff 18 -20March 2010–Atlanta, USA* 10 -12June2009–Rome, Italy* meeting, Antwerp, Belgium February 2014–Dataimputation Disability Weights meeting,Brussels 31 January2014–DALY calculation& The Netherlands meeting. RIVM,Bilthoven, 2 August 2013–Dataimputation FERG 5Geneva, Switzerland April 2013 –Sunday pre-meeting at meeting, Antwerp, Belgium* 2 -4October 2012–Establishment Declarations ofInterest Participants 12

INTRODUCTION WHO Estimates of the global burden of foodborne diseases 13

COMMISSIONED WORK 2

WHO Estimates of the global burden of foodborne diseases 15

COMMISSIONED WORK

Each TF commissioned specific pieces f Fischer Walker, C.L., & Black, R.E. 2010. of work to provide scientific evidence Diarrhoea morbidity and mortality on which to base estimates. Most of in older children, adolescents, and these were systematic reviews, either of adults. Epidemiology and , available data on diseases, or reviews 138(9): 1215–1226. Available at http:// of methodology. The majority of journals.cambridge.org/action/ commissioned work resulted in published displayAbstract?fromPage=online&aid= papers, as listed below. Some of these 7849267&fileId=S0950268810000592 publications were part funded by FERG, Accessed 2015-10-16. while others were generated as “in kind” f Pires, S.M., Fischer Walker, C.L., Lanata, contributions by the authors. C.F., Devleesschauwer, B., Hall, A., Kirk, M.D., Duarte, A.S.R., Black, R.E., 2.1 Enteric Diseases Task Force & Angulo, F.J. Aetiology-specific estimates of the global and regional The EDTF commissioned the following incidence and mortality of diarrhoeal systematic reviews: diseases commonly transmitted through food. PLOS ONE, vol 10, iss 12, 2.1.1 Brucella spp. DOI: 10.1371/journal.pone.0142927 f Dean, A.S., Crump, L., Greter, H., 2.1.3 Mycobacterium bovis Hattendorf, J., Schelling, E. & Zinsstag, J. 2012. Clinical manifestations of human f Muller, B., Durr, S., Alonso, S., brucellosis: a systematic review and Hattendorf, J., Laisse, C.J., Parsons, S.D., meta-analysis. PLOS Neglected Tropical van Helden, P.D. & Zinsstag, J. 2013. Diseases, 6(12): Art. e1929. Available Zoonotic Mycobacterium bovis-induced at http://www.plosntds.org/article/ tuberculosis in . Emerging info%3Adoi%2F10.1371%2Fjournal. Infectious Diseases, 19(6): 899–908. pntd.0001929 Accessed 2015-10-16. Available at: http://wwwnc.cdc.gov/eid/ article/19/6/12-0543_article f Dean, A.S., Crump, L., Greter, H., Schelling, E. & Zinsstag, J. 2012. f Durr, S., Muller, B., Alonso, S., Global burden of human brucellosis: Hattendorf, J., Laisse, C.J., van Helden, a systematic review of disease P.D. & Zinsstag, J. 2013. Differences frequency. PLOS Neglected Tropical in primary sites of infection between Diseases, 6(10): Art e1865. Available zoonotic and human tuberculosis: at http://www.plosntds.org/article/ results from a worldwide systematic info%3Adoi%2F10.1371%2Fjournal. review. PLOS Neglected Tropical pntd.0001865 Accessed 2015-10-16. Diseases, 7(8): Art e2399. Available 2.1.2 Diarrhoeal disease at http://www.plosntds.org/article/ info%3Adoi%2F10.1371%2Fjournal. f Fischer Walker, C.L., Sack, D. & Black, pntd.0002399 Accessed 2015-10-16. R.E. 2010. aetiology of diarrhoea in older children, adolescents and 2.1.4 -producing adults: a systematic review. PLOS Escherichia coli Neglected Tropical Diseases, f Majowicz, S.E., Scallan, E., Jones-Bitton, 4(8): Art e768. Available at A., Sargeant, J.M., Stapleton, J., Angulo, http://www.plosntds.org/article/ F.J., Yeung, D.H. & Kirk, M.D. 2014 info%3Adoi%2F10.1371%2Fjournal. Global incidence of human Shiga toxin- pntd.0000768 Accessed 2015-10-16. producing Escherichia coli infections and deaths: a systematic review and Commissioned work 16

knowledge synthesis. Foodborne review and meta-analysis. Lancet Pathogens and Disease, 11(6): 447– 455. Infectious Diseases, 14(11): Available at http://online.liebertpub. 1073– 1082. Available at com/doi/full/10.1089/fpd.2013.1704 http://www.sciencedirect.com/science/ Accessed 2015-10-17. article/pii/S1473309914708709 Accessed 2015-10-17. 2.1.5 Norovirus WORK f Ahmed, S.M., Hall, A.J., Robinson, A.E., 2.2 Parasitic Diseases Task Force Verhoef, L., Premkumar, P., Parashar, PDTF commissioned the following COMMISSIONED U.D., Koopmans, M. & Lopman, B.A. systematic reviews: 2014. Global prevalence of norovirus in cases of : a systematic 2.2.1 Taenia solium review and meta-analysis. Lancet f Carabin, H., Ndimubanzi, P.C., Budke, Infectious Diseases, 14(8): 725– C.M., Nguyen, H., Qian, Y., Cowan, L.D., 730. Available at http://www. Stoner, J.A., Rainwater, E. & Dickey, M. sciencedirect.com/science/article/pii/ 2011. Clinical manifestations associated S1473309914707674 Accessed with neurocysticercosis: a systematic 2015-10-17. review. PLOS Neglected Tropical f Verhoef, L., Hewitt, J., Barclay, L., Diseases, 5(5): Art e1152. Available Ahmed, S.M., , R., Hall, A.J., at http://www.plosntds.org/article/ Lopman, B., Kroneman, A., Vennema, H., info%3Adoi%2F10.1371%2Fjournal. Vinje, J. & Koopmans, M. 2015. Norovirus pntd.0001152 Accessed 2015-10-17. genotype profiles associated with f Ndimubanzi, P.C., Carabin, H., Budke, foodborne transmission, 1999– 2012. C.M., Nguyen, H., Qian, Y.J., Rainwater, Emerging Infectious Diseases, E., Dickey, M., Reynolds, S. & Stoner, 45: 95– 99. Available at: http://wwwnc. J.A. 2010. A systematic review of the cdc.gov/eid/article/21/4/14-1073_article frequency of neurocysticercosis with 2.1.6 Invasive non-typhoidal a focus on people with epilepsy. PLOS Salmonella enterica Neglected Tropical Diseases, 4(11): Art e870. Available at http://www.plosntds. f Ao, T.T., Feasey, N.A., Gordon, M.A., org/article/info:doi/10.1371/journal. Keddy, K.H., Angulo, F.J. & Crump, pntd.0000870 Accessed 2015-10-17. J.A. 2015. Global burden of invasive non-typhoidal Salmonella disease, 2.2.2 Trematodes (includes 2010. Emerging Infectious Diseases, Echinostoma spp., Fasciolopsis 21(6): 941– 949. buski, Heterophyes spp. and f Crump, J.A. & Kirk, M.D. Estimating Metagonimus spp.) the burden of febrile illnesses. PLOS f Furst, T., Keiser, J. & Utzinger, J. 2012. Neglected Tropical Diseases, (in press). Global burden of human food-borne 2.1.7 trematodiasis: a systematic review and meta-analysis. Lancet Infectious f Maertens de Noordhout, C., Diseases, 12(3): 210– 221. Available at Devleesschauwer, B., Angulo, F.J., http://www.thelancet.com/journals/ Verbeke, G., Haagsma, J., Kirk, M., laninf/article/PIIS1473– 3099(11)70294– Havelaar, A. & Speybroeck, N. 2014. The 8/fulltext Accessed 2015-10-17. global burden of : a systematic WHO Estimates of the global burden of foodborne diseases 17

2.2.3 Echinococcus multilocularis /10408444.2011.575766 Accessed f Torgerson, P.R., Keller, K., Magnotta, 2015-10-17. M. & Ragland, N. 2010. The global f Wu, F. 2010. Global Burden of aflatoxin- burden of alveolar echinococcosis. induced disease: Final Report for PLOS Neglected Tropical Diseases, 4: the World Health Organization e722. Available at http://www.plosntds. (WHO) Foodborne Disease Burden org/article/info:doi/10.1371/journal. Epidemiology Reference Group (FERG) pntd.0000722 Accessed 2015-10-17. Chemical Task Force. Department of Environmental and Occupational 2.2.4 Trichinella spp. Health, University of Pittsburgh f Devleesschauwer B, Praet N, Graduate School of Public Health, Speybroeck N, Torgerson P R, Haagsma Pittsburgh, PA, USA. J A, De Smet K, Murrell K D, Pozio E f Liu, Y. & Wu, F. 2010. Global burden and Dorny P (2014) The low global of aflatoxin-induced hepatocellular burden of trichinellosis: evidence and : a risk assessment. implications. International Journal of Environmental Health Perspectives, Parasitology, 45(2-3): 95– 99. Available 118(6): 818– 824. Available at http:// at http://www.sciencedirect.com/ www.ncbi.nlm.nih.gov/pmc/articles/ science/article/pii/S0020751914001374 PMC2898859/ Accessed 2015-10-17. Accessed 2015-10-17. f Liu, Y., Chang, C.C., Marsh, G.M. & Wu, f Murrell, K.D. & Pozio, E. 2011. Worldwide F. 2012. Population attributable risk of occurrence and impact of human aflatoxin-related liver cancer: systematic trichinellosis, 1986– 2009. Emerging review and meta-analysis. European Infectious Diseases, 17(12): 2194– 2202. Journal of Cancer, 48(14): 2125– 2136. Available at http://www.ncbi.nlm. 2.2.5 Toxoplasma gondii nih.gov/pmc/articles/PMC3374897/ f Torgerson, P.R. & Mastroiacovo, P. Accessed 2015-10-17. 2013. The global burden of congenital : a systematic review. 2.3.2 Arsenic Bulletin of the World Health f Oberoi, S., Barchowsky, A. & Wu, F. Organization, 91(7): 501– 508. Available 2014. The global burden of disease at http://www.ncbi.nlm.nih.gov/pmc/ for skin, lung, and bladder cancer articles/PMC3699792/ Accessed caused by arsenic in food. Cancer 2015-10-17. Epidemiology Biomarkers and Prevention, 23(7): 1187– 1194. Abstract 2.3 Chemicals and Toxins available at http://cebp.aacrjournals. Task Force org/content/23/7/1187.abstract Accessed 2015-10-17. CTTF commissioned several systematic reviews and reports. 2.3.3 Cassava cyanide f Cliff, J. 2011. Incidence and prevalence 2.3.1 Aflatoxins estimates of cassava-cyanide induced f Khlangwiset, P., Shephard, G.S. & Wu, F. diseases. Report for the FERG 2011. Aflatoxins and growth impairment: Chemicals and Toxins Task Force. A review. Critical Reviews in , Universidade Eduardo Mondlane, 41(9): 740– 755. Available at http:// . informahealthcare.com/doi/abs/10.3109 Commissioned work 18

2.3.4 Peanut Allergens methods in an age of globalized risks: f Ezendam, J. & van Loveren, H. 2012. lesson from the global burden of Parameters needed to estimate the foodborne disease expert elicitation. global burden of peanut : Risk Analysis DOI: 10.1111/risa.12385 Systematic literature review. European f Aspinall, WP., Cooke, RM., Havelaar, Journal of Food Research and Review, AH., Hoffman, S., Hald, T. 2015. Science-

2(2): 46– 48. Available at http:// based global attribution of foodborne WORK www.rivm.nl/en/Library/Scientific/ diseases: Findings of WHO expert Reports/2012/april/Parameters_ elcitiation. PLOS ONE. (in press). COMMISSIONED needed_to_estimate_the_global_ burden_of_peanut_allergy_Systematic_ 2.5 Computational Task Force literature_review Accessed 2015-10-17. f McDonald, S.A., Devleesschauwer, B., 2.3.5 Dioxins Speybroeck, N., Hens, N., Praet, N., f Zeilmaker, M.J., Devleesschauwer, B., Torgerson, P.R., Havelaar, A.H., Wu, F., Mengelers, M.J.B., Hoekstra, J., Brandon, Tremblay, M., Amene, E.W. & Döpfer, D. E.F.A. & Bokkers, B.G.H. The disease 2015. Data-driven methods for imputing burden of dioxins: A global perspective. national-level incidence in global RIVM Report National Institute for burden of disease studies. Bulletin Public Health and the Environment of the World Health Organization, (RIVM), Netherlands. 93(4): 228– 236 doi: http://dx.doi. org/10.2471/BLT.14.139972 2.4 Source Attribution Task Force The following two papers were not SATF commissioned the commissioned by the Computational Task following papers: Force, but several of the authors were TF members, and the papers are relevant to f Pires, S.M. 2013. Assessing the the Initiative estimates. applicability of currently available methods for attributing foodborne f Devleesschauwer, B., Havelaar, A.H., disease to sources, including food Maertens de Noordhout, C., Haagsma, and food commodities. Foodborne J.A., Praet, N., Dorny, P., Duchateau, Pathogens and Disease, 10(3): 206– 213. L., Torgerson, P.R., Van Oyen, H. & f Pires, S.M., Evers, E.G., van Pelt, W., Speybroeck, N. 2014. Calculating Ayers, T., Scallan, E., Angulo, F.J., disability-adjusted life years to quantify Havelaar, A. & Hald, T. and the Med- burden of disease. International Journal Vet-Net Workpackage 28 team. 2009. of Public Health, 59(3): 565– 569. Attributing the human disease burden f Devleesschauwer, B., Havelaar, A.H., of foodborne infections to specific Maertens de Noordhout, C., Haagsma, sources. Foodborne Pathogens and J.A., Praet, N., Dorny, P., Duchateau, Disease, 6(4): 417– 424. L., Torgerson, P.R., Van Oyen, H. & f Hoffman, S., Aspinall, W., Cooke, R., Speybroeck, N. 2014. DALY calculation Cawthorne, A., Corrigan, T., Havelaar, in practice: a stepwise approach. A., Gibb, H., Torgerson, P., Kirk, M., International Journal of Public Health, Angulo, F, Lake R, Speybroeck N, 59(3): 571– 574. Devleesschauwer B, Hald T. 2015 Perspective: Research synthesis WHO Estimates of the global burden of foodborne diseases 19

2.6 Country Studies Task Force WHO is responsible? PLOS Neglected Tropical Diseases, 1: Art e161. Available Two systematic reviews at http://journals.plos.org/plosntds/ were commissioned: article?id=10.1371/journal.pntd.0000161 f Polinder, S., Haagsma, J.A., Stein, C. & Accessed 2015-10-17. Havelaar, A.H. 2012. Systematic review f Havelaar, A.H., Cawthorne, A., Angulo, of general burden of disease studies F., Bellinger, D., Corrigan, T., Cravioto, A., using disability-adjusted life years. Gibb, H., Hald, T., Ehiri, J., Kirk, M., Lake, Population Health Metrics, 10: Art R., Praet, N., Speybroeck, N., de Silva, N., 21. Available at http://www.ncbi.nlm. Stein, C., Torgerson, P. & Kuchenmüller, nih.gov/pmc/articles/PMC3554436/ T. 2013. WHO Initiative to Estimate the Accessed 2015-10-17. Global Burden of foodborne diseases. f Haagsma, J.A., Polinder, S., Stein, C.E. Lancet, 381(Suppl. 2): S59. & Havelaar, A.H. 2013. Systematic f Hird, S., Stein, C., Kuchenmüller, T. & review of foodborne burden of Green, R. 2009. Meeting report: Second disease studies: quality assessment of annual meeting of the World Health data and methodology. International Organization Initiative to estimate Journal of , the global burden of foodborne 166(1): 34– 47. Available at http:// diseases. International Journal of Food www.sciencedirect.com/science/ Microbiology, 133: 210– 212. article/pii/S0168160513002778 f Kuchenmüller, T., Hird, S., Stein, C., Accessed 2015-10-17. Kramarz, P., Nanda, A. & Havelaar, The results from one of the country A.H. 2009. Estimating the global studies have been published: burden of foodborne diseases – a collaborative effort.Eurosurveillance, Kumagai, Y., Gilmour, S., Ota, E., Momose, Y., 14(18): 1– 4. Available at http://www. Onishi, T., Bilano, V.L.F., Kasuga, eurosurveillance.org/ViewArticle. F., Sekizaki, T. & Shibuya, K. 2015. Estimating aspx?ArticleId=19195 Accessed the burden of foodborne diseases in Japan. 2015-10-17. Bulletin of the World Health Organization, f Lake, R.J., Havelaar, A.H. & 93(8): 540–549. Kuchenmüller, T. 2013. New research The preparation of material to augment on estimating the global burden of the resources developed by the CSTF foodborne disease. pp. 260– 271, in: was commissioned from Sandy Campbell J. Sofos (ed.). Advances in microbial (Knowledge Translation Consultant, food safety. Vol. 1. Woodhead New Mexico, USA). The results of that Publishing, Oxford, UK. work have been included in the Situation f Lake, R.J., Stein, C.E. & Havelaar, analysis, knowledge translation and risk A.H. 2014. Estimating the burden of communication guidance manual, one of the foodborne disease.. pp. 73– 79, in: tools and resources developed by the CSTF. Y. Motarjemi (ed.). Encyclopedia of Food Safety. Vol. 1. Academic Press, 2.7 Other relevant publications Waltham, MA, USA. f Kuchenmulller, T., Abela-Ridder, The following articles were written by B., Corrigan, T. & Tritscher, A. 2013. FERG members and WHO staff: World Health Organization Initiative f Stein, C., Kuchenmuller, T., Hendrickx, to Estimate the Global Burden of S., Pruss-Ustun, A., Wolfson, L., Engels, foodborne diseases. Revue Scientifique D. & Schlundt, J. 2007. The Global et Technique-Office International des Burden of Disease assessments – Epizooties, 32(2): 459– 467. Commissioned work 20 WORK COMMISSIONED COMMISSIONED WHO Estimates of the global burden of foodborne diseases 21

OVERVIEW OF THE SCIENTIFIC APPROACH3

WHO Estimates of the global burden of foodborne diseases 23

OVERVIEW OF THE SCIENTIFIC APPROACH

3.1 The DALY metric incidence, i.e. both current and future health outcomes are included. Future As mentioned in the report from the outcomes include sequelae and mortality initial consultation, the Initiative was resulting from the initial disease within a encouraged to use summary measures defined time period. of public health as the metric for the burden of FBD. The disability adjusted life To define life expectancy for the year (DALY) metric was chosen for the calculation of YLL life expectancy tables following reasons: for the population being studied may be used. Alternatively, life expectancy that f It is an established WHO metric with reflects an ideal of human potential may international application; and be used. f It is consistent with the Global Burden of Disease project. 3.2 Overarching methodology de- The DALY is calculated by adding the number of years of life lost to mortality cisions by FERG in relation to DALY (YLL) and the number of years lived with estimates disability due to morbidity (YLD): 3.2.1 Hazard-based approach DALY = YLL + YLD The burden of disease estimation is The YLL due to a specific disease in a hazard-based because: specified population is calculated by the f it allows a complete estimate of the summation of all fatal cases (n) due to burden of disease due a specific hazard; the health outcomes (l) of that specific f it includes all related sequelae; and disease, each case multiplied by the f measures to address foodborne expected individual life span (e) at the diseases are often hazard specific. age of death. 3.2.2 Incidence-based approach DALYs, and more specifically their YLD component, may be calculated from an YLD is calculated by accumulation over incidence or a prevalence perspective. all health outcomes (l), the product of While incidence-based YLDs are defined the number of cases (n), the duration of as the product of the number of incident the illness (t) and the severity weight (w) cases and the duration and disability of a specific disease. It should be noted weight (DW) of the concerned health that the calculation for YLL implicitly state, prevalence-based YLDs are includes a severity weight factor. The defined as the product of the number of severity weight or disability weight (DW) prevalent cases and the corresponding factors are in the range zero to one, with DW [1,6]. In the incidence-based the severity weight for death being equal approach, all health outcomes, including to one. those in future years, are assigned to the initial event (e.g., exposure to a certain hazard). This approach therefore reflects the future burden of disease resulting DALYs may be calculated using a from current events. In the prevalence- prevalence approach which estimates the based approach, on the other hand, the current burden of disease in a population, health status of a population is assessed considering previous events. However, at a specific point in time, and prevalent the more common approach is to use Overview of the scientific approach 24

diseases are attributed to initial events Regions that happened in the past. This approach Several options were available for therefore reflects the current burden reporting on a regional basis of disease resulting from previous (14 subregions based on child and adult events. For burden of FBD studies, the mortality, as described by Ezzati et al. incidence-based YLD approach was [5]; 21 GBD regions [6]; and 13 GEMS deemed the most appropriate approach, Cluster Diet Regions1). The subregions because (1) this approach is more based on mortality were chosen.2 sensitive to current epidemiological Countries grouped into each of the trends [2]; (2) is more consistent with 14 subregions are listed in Appendix 2. the hazard-based approach, since it has the point of infection (or primary health Reference year effect from exposure) as starting point The reference year for the calculation of for the calculations; and (3) is consistent absolute numbers was 2010. OVERVIEW OVERVIEW with the estimation of YLLs, which by APPROACH

Attribution OF SCIENTIFIC definition follows an incidence-based The choice of a method to attribute approach, as mortality can be seen as a proportion of disease incidence the incidence of death [3]. Nevertheless, to foodborne transmission was a the prevalence- and incidence-based major decision for the project. The approaches yield similar overall results if rationale for choosing a global expert the epidemiology of disabilities and the elicitation process was developed population age-structure are constant after consideration of alternatives, as over time [2]. However, burden estimates described below. for specific age groups will always differ between the prevalence- and incidence- Estimating the burden of FBD is based approaches, because the former complicated because most of the hazards assigns the burden to the age at which causing foodborne disease are not the burden is experienced, while the transmitted solely by food. The relative latter assigns the burden to the age of impact of each route differs depending disease onset [4]. on the epidemiology of the disease causing micro organism (bacteria, virus Using the incidence for the burden or parasite) or chemical hazards. Other estimations is important for diseases factors such as the geographical region, having a long period between exposure season and food consumption patterns and appearance of clinical signs. An also influence the role of different incidence-based approach for the exposures routes [7, 8]. The estimation burden estimations fits better with of the burden of FBD, therefore, requires a hazard-based approach. However, incidence figures are not always 1 http://www.who.int/foodsafety/chem/gems/en/ available. For example, in the case of index1.html Accessed 23 July 2014 peanut [Arachis hypogaea] allergy, only 2 The subregions are defined on the on the basis of prevalence figures are available. When child and adult mortality, as described by Ezzati et al. [5] Stratum A = very low child and adult only prevalence figures are available, mortality; Stratum B = low child mortality and very incidence can be estimated based on the low adult mortality; Stratum C = low child mortality prevalence figures and on the duration of and high adult mortality; Stratum D = high child and adult mortality; and Stratum E = high child mortality the disease. and very high adult mortality. The use of the term ‘subregion’ here and throughout the text does not identify an official grouping of WHO Member States, and the “subregions” are not related to the six official WHO regions. WHO Estimates of the global burden of foodborne diseases 25

a delineation of the major transmission designed as randomized controlled routes, including contaminated food, intervention trials, have been conducted water, soil, air or contact with infected to assess the importance of water, or humans. Previous efforts to particularly for the transmission of quantify the contribution of specific diarrhoeal diseases (reviewed by [27] and sources (including types of foods) and [28]). However, other transmission routes, transmission routes have been gathered such as soil, air and direct contact with under the term ‘source attribution’ or infected humans or animals, are generally ‘human illness attribution’ [9, 10]. The not considered in those studies. Thus, for applicability of available methods for most countries, and at the global level, source attribution of FBD at the global relevant studies and data for quantifying level was recently assessed by Pires [7]. attribution of potential FBD to the major transmission routes do not exist. Source attribution is an important tool for identifying and prioritizing In such situations, structured elicitation effective interventions to prevent and of scientific judgment may be used control FBD [11]. The need for reliable [7, 29]. When data are not available, source attribution estimates has or undertaking primary research is not prompted a growing body of research feasible, a structured elicitation offers focusing on attribution, particularly for a transparent and mathematically infectious agents [7, 10, 12, 13]. However, rigorous way of evaluating and comprehensive attribution studies enumerating uncertainty distributions, based on surveillance data and/or food from the judgments of many individual monitoring and exposure data are still researchers, for quantifying risk models. limited in scope, and to date have been Within food safety, the approach performed for a few hazards only, or in a has been applied to provide national limited number of countries [14– 26]. estimates for the proportion of illnesses attributable to food for specific infectious In addition, existing studies have focused diseases [30– 37], or to inform modelling mainly on identifying specific food of foodborne disease risk assessment sources or animal , whereas models by estimating specific model other potential transmission routes are parameters and their uncertainty often not quantified due to lack of data, [38, 39]. or neglected due to the complexity of attribution models. Many studies, often Overview of the scientific approach 26 OVERVIEW OVERVIEW APPROACH OF SCIENTIFIC WHO Estimates of the global burden of foodborne diseases 27

HAZARD- SPECIFIC METHODOLOGY4

WHO Estimates of the global burden of foodborne diseases 29

HAZARD-SPECIFIC METHODOLOGY

The following material is derived from, Health Organization estimates of the and in some parts repeated verbatim global and regional disease burden of from, text in the suite of papers in which 22 foodborne bacterial, protozoal and the FERG results have been published.1 viral diseases, 2010: A data synthesis These primary outputs are listed below, PLOS , DOI:10.1371/journal. and have been collated in a dedicated pmed.1001921 PLOS collection entitled “The World Torgerson, P.R., Devleesschauwer, B., Health Organization Estimates of the Praet, N., Speybroeck, N., Willingham, Global Burden of Foodborne Diseases”, A.L., Kasuga, F., Rokni, M.B., Zhou, X.-N., which can be accessed at the website: Fèvre, E.M., Sripa, B., Furst, T., Budke, http://collections.plos.org/ferg-2015. We C.M., Carabin, H., Kirk, M.D., Angulo, F.J., acknowledge the PLOS for permission to Havelaar, A. & de Silva, N. 2015. World incorporate this material into this report. Health Organization estimates of the In addition to the series of published global and regional disease burden of 11 papers, the estimates of foodborne foodborne parasitic diseases, 2010: disease burden have been made available a data synthesis. PLOS Medicine, as an on-line tool which will be accessible DOI:10.1371/journal pmed.1001920 via the WHO FERG web page.2 Gibb, H., Devleesschauwer, B., Bellinger, Havelaar, A.H., Kirk, M.D., Torgerson, D., Bolger, P.M., Zeilmaker, M., Barchowsky, P.R., Gibb, H.J., Hald, T., Lake, R.J., A., Oberoi, S., Wu, F., Ezendam, J., Zang, Praet, N., Angulo, F.J., Bellinger, D.C., de J., Carrington, C., Cliff, J., Verger, P., Pitt, Silva, N.R., Gargouri, N., Speybroeck, J., Adegoke, G., Afshari, R., Baines, J., N., Cawthorne, A., Mathers, C., Stein, Bokkers, B., Mengelers, M., van Loveren, C., Devleesschauwer, B. on behalf H., Rainis, H., O’Leary, K. & Liu, Y. 2015. of the World Health Organization World Health Organization estimates of Foodborne Disease Burden Epidemiology the global and regional disease burden Reference Group. 2015. World Health of four foodborne chemicals and toxins, Organization global estimates and 2010: a data synthesis. F1000 Research. regional comparisons of the burden of foodborne disease, 2010. PLOS Medicine, Hald, T., Aspinall, W., Devleesschauwer, DOI: 10.1371/journal.pmed.1001923 B., Cooke, R., Corrigan, T., Havelaar, A., Gibb, H., Torgerson, P., Kirk, M., Angulo, Kirk, M.D., Pires, S.M., Black, R.E., Caipo, F.J., Lake, R., Speybroeck, N. & Hoffmann, M., Crump, J.A., Devleesschauwer, B., S. 2015. World Health Organization Döpfer, D., Fazil, A., Fischer-Walker, estimates of the relative contributions C.L., Hald, T., Hall, A.J., Keddy, K.H., of food to the burden of disease due to Lake, R., Lanata, C.F., Torgerson, P.R., selected foodborne hazards: a structured Havelaar, A.H. & Angulo, F.J. 2015. World expert elicitation. PLOS ONE. in press. 1 All estimates of burden of foodborne disease were reviewed by relevant WHO focal points before Devleesschauwer, B., Haagsma, J.A., submission. In particular, each paper was reviewed Bellinger, D., Cole, D., Döpfer, D., Fazil, by the Mortality and Burden of Disease Unit, Health A., Fèvre, E., Lake, R., Maertens de Statistics and Information Systems Department, Health System and Innovation Cluster (Coordinator: Noordhout, C., McDonald, S.A., Pires, S.M., Dr Colin Mathers) and cleared as required through Speybroeck, N., Thomas, K., Torgerson, Food Safety and Zoonoses Department located in P.R., Wu, F., Havelaar, A.H. & Praet, N. Health Security Cluster at WHO Headquarters. 2015. Methodological framework for 2 http://www.who.int/foodsafety/areas_work/ foodborne-diseases/ferg/en/ accessed 3 World Health Organization estimates of November 2015 the global burden of foodborne disease. Hazard-specific methodology 30

PLOS ONE, vol 10, iss 12 DOI: 10.1371/ f Likely magnitude of foodborne journal.pone.0142498 component of burden of disease. Lake, R., Devleesschauwer, B., Nasinyama, Each of the three papers from the G., Havelaar, A.H., Kuchenmüller, T., hazard-based TFs (EDTF, PDTF and Haagsma, J.A., Jensen, H., Jessani, N., CTTF) includes supplementary material Maertens de Noordhout, C., Angulo, F.J., discussing the sources and methodology Ehiri, J., Molla, L., Agaba, F., Aungkulanon, for the parameters used to estimate: S., Kumagai, Y. & Speybroeck, N. 2015. incidence, clinical outcomes, duration, National studies as a component of the DW, mortality, age and sex distribution. World Health Organization initiative to The full details have been combined estimate the global and regional burden in Appendix 4. The material below of foodborne disease. PLOS ONE, vol 10, explains the rationale for the sources iss 12, DOI: 10.1371/journal.pone.0140319 and methods. Burden of foodborne disease estimates 4.1 Hazard Selection were prepared for the 40 foodborne hazards causing 41 diseases shown in At the first meeting after the Figure 2. establishment of FERG, each hazard- based TF compiled a comprehensive The following methodology section universal list of foodborne hazards that describes the inputs and processes used could be addressed (see Appendix 3). to generate DALY estimates. This material

Pragmatic decisions were then made is broadly structured as follows: METHODOLOGY about specific hazards for further work, HAZARD SPECIFIC f estimation of incidence (or population based on the knowledge of TF members attributable fraction); and applying the following criteria: f health states and disability weights; f Availability of data to estimate f attribution of foodborne incidence; and transmission; and f computation. WHO Estimates of the global burden of foodborne diseases 31

Figure 2. Hazards for which burden of foodborne disease estimates were prepared by FERG, grouped according to TF. Hazards in grey boxes were addressed by individual TFs but were not included in the global overview. Hazards in blue boxes are pending.

EDTF (HAZARDS CAUSING EDTF (HAZARDS CAUSING PDTF CTTF HEALTH EFFECTS OTHER ENTERIC DISEASE) THAN ENTERIC DISEASE) Ascaris spp. Aflatoxin Brucella spp. cereus1 Echinococcus multilocularis Arsenic botulinum3 Campylobacter spp.2 Cadmium Hepatitis A virus Cryptosporidium spp Clonorchis sinensis Cassava cyanide Listeria spp. Clostridium perfringens1 spp. Dioxin Mycobacterium bovis Entamoeba histolytica Salmonella enterica (invasive Enteropathogenic E. coli Intestinal flukes4 Lead infections) non-typhoidal (EPEC) Salmonella enterica Opisthorchis spp. Methyl Enterotoxigenic E. coli (ETEC) Paratyphi A Paragonimus spp. Peanut allergens5 Salmonella enterica Typhi Giardia spp. Taenia solium Norovirus Salmonella enterica Toxoplasma gondii6 (non-invasive infections) non-typhoidal Trichinella spp. Shigella spp. Shiga toxin-producing E. coli (STEC) Staphylococcus aureus1 Vibrio cholerae

Note that and invasive salmonellosis are counted as a single hazard causing two diseases.

Notes: (1) 61 EUR and other subregion A (low mortality) countries only. (2) Includes Guillain-Barré Syndrome cases and deaths. (3) 61 EUR and other subregion A (low mortality) countries only, excluding WPR countries. (4) Includes selected species of the families Echinostomatidae, , Gymnophallidae, , Nanophyetidae, Neodiplostomidae and Plagiorchiidae (depending on data availability). (5) Only the burden for AMR A, EUR A and WPR A was assessed. (6) Separate estimates for congenital and acquired toxoplasmosis.

4.2 Enteric Hazards for global estimation and they were infrequent causes of foodborne disease. The overall aim of the EDTF was to provide estimates of disease incidence During the course of the project, it was and mortality (by age, sex and country or identified that burden estimates for region) for diarrhoeal and other illnesses three parasitic hazards (Giardia spp., Cryptosporidium spp. and Entamoeba due to bacteria and , by all causes histolytica) should be included with and by selected aetiological agents, and the hazards addressed by EDTF, given including sequelae. Despite its name, the that that the primary disease caused by EDTF was not exclusively concerned with these organisms was diarrhoea, and the hazards causing enteric disease. approach taken to estimate the burden of The initial list of hazards considered by these hazards was applicable. EDTF is given in Appendix 3. From this As shown in Figure 2, there were 21 list, entero-aggerative Escherichia coli, hazards causing 22 diseases for which , V. vulnificus final burden estimates were prepared by and Yersinia spp. were excluded on the EDTF. Of these diseases, four are distinct basis that there were insufficient data manifestations of Salmonella enterica Hazard-specific methodology 32

infection: invasive infections due to For this approach, the WHO Child S. enterica Typhi Health Epidemiology Reference Group (S. Typhi); invasive infections due to (CHERG) method was modified to S. serotype Paratyphi A (S. Paratyphi A); estimate diarrhoeal incidence and invasive infections due to non-typhoidal mortality for all age groups [50]. First, S. enterica (iNTS); and diarrhoeal disease the overall incidence of diarrhoea from all due to non-typhoidal S. enterica. causes (i.e. the “envelope” of diarrhoeal incidence) was estimated for 2010 Diarrhoea is a dominant feature for by combining estimates of diarrhoeal 14 of these diseases – ten caused by incidence for children <5 years of age bacteria, three by protozoa, and one and persons 5 years of age [51, 52]. by a virus. One or more extra-intestinal * The overall diarrhoeal mortality (i.e. the manifestations, including bacteraemia, envelope for diarrhoeal deaths) derived hepatitis and , are the by WHO for 2010 was used.3 dominant feature for the other eight An aetiological proportion for each diseases – seven caused by bacteria and disease by region was derived from one caused by a virus. systematic reviews of stool sample 4.2.1 Estimating cases, sequelae and isolation or detection proportions from inpatient, outpatient and community- deaths for diarrhoeal diseases based studies of persons with diarrhoea. For diarrhoeal diseases caused by Following the CHERG standard approach, Campylobacter spp., Cryptosporidium developed because there is limited spp., Entamoeba histolytica, ETEC, METHODOLOGY information on pathogens among HAZARD SPECIFIC EPEC, Giardia spp., norovirus, non- people who have died, it was assumed typhoidal Salmonella spp. and Shigella that the distribution of pathogens spp., because national estimates of observed among inpatients hospitalized foodborne diseases were only available with severe diarrhoea represented the from a limited number of countries, two pathogen prevalence among diarrhoeal approaches were used depending on the deaths [50]. To derive aetiological level of development of the country. The proportions for children <5 years of age, approaches have been described by a it was assumed that the distribution of key accompanying publication [40]. pathogens in outpatient and community The first approach, based on national studies represented the pathogen estimates of the incidence of foodborne prevalence among diarrhoeal episodes diseases, was applied to the 61 countries for those who did not die. The same in low-mortality (EUR and other assumption was made for persons subregion A) countries [41– 49]. For *5 years of age but due to sparseness of countries with national estimates of data, inpatient studies were also included. incidence and mortality, these data For some pathogens it was assumed were used. The median and associated that different aetiological agents, such uncertainty intervals for diarrhoeal as Shigella spp., NTS and Campylobacter diseases for the subregion were used spp. had similar clinical profiles. to estimate incidence and mortality of Initial estimates for the 61 countries in diarrhoeal diseases for other countries low-mortality (EUR and other subregion within these subregions without national A) countries had been prepared data [40]. using this second “CHERG approach”. The second approach was applied to However, it was recognized that for the remaining 133 countries worldwide. 3 http://www.who.int/gho/en/ Accessed 6 June 2014 WHO Estimates of the global burden of foodborne diseases 33

these developed countries this would caused by these pathogens in these overestimate incidence of diseases, in countries. The median CFR from national comparison with published estimates studies was 0.003% for C. perfringens from the national studies available. and 0.0025% for S. aureus; there were no Estimates for were based on B. cereus deaths. the incidence among populations at It was considered that 31% of Guillain- risk for cholera in and non- Barré Syndrome (GBS) cases globally endemic countries [53]. The case fatality were associated with antecedent ratio (CFR) for cholera was 1% in WPR Campylobacter infection and that the subregion B; 1% in SEAR B (except CFR for GBS was 4.1% [55, 56]. 1.5% in Bangladesh); 1.3% in EMR B; 3% Assignment of aetiology of diarrhoeal in SEAR D; 3.2% in EMR D; and 3.8% diseases when using the CHERG in AFR [53]. For all other countries, it approach for middle- and high mortality was assumed cholera occurred only countries was refined by adding in an among international travellers and did not result in deaths. In this instance, the aetiological proportion for pathogens not median incidence from non-endemic associated with foodborne transmission countries with available data for cholera (, , ) was applied. and for unspecified diarrhoeal agents (pathogens that are possibly foodborne Shiga-toxin producing E. coli (STEC) but with insufficient data for estimation, infection incidence and mortality were and unknown agents not yet discovered). based on a systematic review [54]. Sequelae, more common with O157 In our study, norovirus resulted in the infections, were haemolytic uraemic largest number of cases of foodborne syndrome (HUS) and end-stage renal diseases and overall burden, highlighting disease (ESRD). Based on review, it was the global importance of this agent. estimated 0.8% of O157 infections and However, the disease model we used 0.03% of infections caused by other in the 135 middle- and high-mortality result in HUS, and 3% of HUS counties included only norovirus cases result in ESRD. It was further infections that resulted in a diarrhoeal estimated that the CFR for HUS was illness. If we also included estimates 3.7%; for ESRD the CFR was 20% in the for norovirus infections that resulted in 35 subregion A countries; and 100% in vomiting without diarrhoea, there would other countries. be an estimated additional 163 million norovirus cases in these countries [57]. For the incidence and mortality of foodborne intoxications caused by 4.2.2 Estimating cases and , sequelae of, and deaths due to, and , data from extra-intestinal diseases national studies conducted in low- For diseases caused by hepatitis A virus, mortality countries were used. The Brucella spp., Listeria monocytogenes, median incidence from national studies Mycobacterium bovis, iNTS, S. Paratyphi was applied to the 61 countries in EUR A and S. Typhi, a variety of approaches and other subregion A countries. The were used, depending on availability burden due to these three foodborne of data. intoxications in high- and middle- mortality countries was not estimated Institute of Health Metrics and Evaluation due to the absence of data on diseases (IHME) Global Burden of Disease 2010 Hazard-specific methodology 34

(GDB2010) data were used to estimate region except those free from M. bovis the burden of disease for typhoid, in cattle were assumed to have the same paratyphoid and hepatitis A [58]. proportion of human TB infections due IHME provided country-specific, age- to M. bovis. To account for internationally standardized prevalence data for typhoid acquired infections, all countries free and paratyphoid . These data were of M. bovis in cattle were assigned the converted to incidence by dividing by lowest observed proportion of human duration, and partitioned into typhoid TB infections due to M. bovis (0.3%). and paratyphoid assuming a 1.0 to 0.23 To derive estimates of human M. bovis ratio [59]. Country-specific hepatitis A incidence, WHO country-specific human mortality data, stratified by age and sex, TB incidences were multiplied by the were converted to incidence assuming a estimate of the proportion of human CFR of 0.2%. TB infections that were due to M. bovis Rates of iNTS are highly correlated with [65]. To estimate mortality associated HIV prevalence and malaria risk [60]. To with M. bovis that accounted for HIV estimate iNTS incidence globally, age- co-morbidity estimates were used of specific estimates of incidence from a mortality due to human TB in HIV- systematic review [60] were used to negative persons (WHO data). Mortality construct a random effect log linear data were adjusted by assuming that the model using covariates of country- CFR for M. bovis was 20% lower than specific HIV and malaria deaths, and human TB, as M. bovis infections are

more likely to be extrapulmonary [66]. METHODOLOGY

the log of Gross Domestic Product. As HAZARD SPECIFIC data were sparse, incidence for all ages To estimate the incidence and mortality was predicted, which was converted to for brucellosis a systematic review was age-specific incidence based on age updated and included additional data profiles for iNTS cases in low and high on 32 countries that were considered incidence settings [60]. From this, iNTS Brucella-free in livestock (free of B. incidence was predicted among persons abortus in cattle and B. melitensis not infected with HIV [61, 62]. To estimate in sheep and ) [67]. Incidence deaths, it was assumed that the CFR data were imputed to countries for iNTS in non-HIV infected individuals without estimates using a Bayesian was a uniform distribution with a range log-normal random effects model, 5– 20% in sub-region B-E countries except for countries that were Brucella- and range 3.9– 6.6% in sub-region A free in livestock [68]. To account for countries [63]. internationally acquired infections, all Estimates for M. bovis infections were countries that were Brucella-free in based on a systematic review where livestock were assigned the median the proportion of human tuberculosis incidence of human brucellosis reported (TB) infections due to M. bovis ranged from these countries. The CFR for from 0.3% in AMR to 2.8% in AFR [64]. brucellosis was 0.5%, and 40% of cases Fifty-one countries were identified that resulted in chronic infections, and 10% of were free from M. bovis in cattle, based cases in males resulted in orchitis [69]. on European Union certification and The incidence and mortality for listeriosis the World Animal Health Information were estimated using a systematic System (WAHIS) of World Organization review that is described elsewhere [70]. for Animal Health.4 All countries in a In accordance with standard burden of 4 OIE – www.oie.int/wahis disease practice, stillbirths were excluded WHO Estimates of the global burden of foodborne diseases 35

in our baseline burden estimates. The the other intestinal protozoa, as citation CFR was 14.9% for perinatal cases and frequency had remained constant over 25.9% for other cases. the same period. Incidence and mortality data for Foodborne trematodes of high priority were only available from countries were Fasciola spp., Clonorchis spp., in and . The Opisthorchis spp., Paragonimus spp. estimation was limited to the 55 countries and intestinal trematodes such as in EUR and AMR subregion A, which was Fasciolopsis buski, Heterophyes spp. and based on the median incidence derived Metagonimus spp. Three cestode species from countries with national estimates were considered important: Echinococcus of botulism. It was estimated that 35% of granulosus, E. multilocularis and Taenia botulism cases were severe and that the solium. The cestode CFR of severe botulism was 15%. was considered likely to have a very low burden for human health because of the lack of serious sequelae resulting 4.3 Parasitic Hazards from intestinal taeniosis, and hence was At the first formal meeting of FERG, excluded from the priority list. Foodborne the PDTF initially reviewed all parasitic Chagas disease was also considered for diseases that could be potentially possible inclusion at the second FERG transmitted by food (Appendix 3) with meeting, but resources were not available 14 parasitic diseases selected as high to commission work on the foodborne priority. The selection criteria of these transmission of a primarily -borne 14 diseases was based on: proportion of disease. Finally the species foodborne transmission; severity of illness believed to have high impact were and/or sequelae; frequency of illness Anisakidae, Ascaris spp. and Trichinella and/or sequelae causes; global relevance; spp. Disease caused by the Anisakidae particular regional relevance; propensity was later considered to be an uncommon to cause outbreaks; and availability foodborne disease and was subsequently of existing evidence to derive burden removed from the priority list. estimates (see meeting reports from The incidence of each of the parasitic FERG 1 and 2). diseases was estimated where possible. Three intestinal protozoa genera – For cysticercosis, the burden was Cryptosporidium, Entamoeba and Giardia estimated from a proportion of the – were considered a priority, as they prevalent epilepsy cases, i.e. the number were likely to result in a high disease of actual cases of disease, as further burden, and the frequency of citations detailed below. Those incident cases with for these parasites had been markedly sequelae (or diseased individuals) were increasing between 1990 and 2008. For assigned years of life lost (YLLs) if fatal, methodological reasons, the burden of or years lived with disability (YLDs) with the three priority intestinal protozoa that a DW that depended on the severity of cause diarrhoeal disease was estimated the disease. For some diseases, such by the EDTF as described above. as toxoplasmosis, many of the incident cases do not have sequelae (i.e. they are Toxoplasma gondii was also considered sub clinical). Such cases were given a DW to be of high priority because of the of zero. potential serious sequelae. Cyclospora was also initially considered, but a Systematic reviews were undertaken to decision was made to target resources to estimate the incidence, sequelae and Hazard-specific methodology 36

mortality due to these diseases [71– 77]. between age of T and T+1 can be used to Where possible, public health records estimate incidence. describing numbers of cases presenting Incidence estimates and clinical sequelae, for treatment were reviewed. These data for diseases caused by foodborne were only available for some diseases trematodes, were mainly based on the in some countries. In others surveillance results of two review articles [77, 78]. data were used (for example laboratory Incidence rates for countries without data on sero-conversion rates in reported national prevalence were the population). imputed, but only where there were For congenital toxoplasmosis (CT) reports of at least one autochthonous a systematic search of nine major human infection, by using a hierarchical databases for published and unpublished random-effects model and incidence sources was conducted, alongside direct information from other countries as input contact with the authors of source data [79]. In highly endemic zones, adult materials. Searches were country specific. subjects either maintain the parasites To be included, studies had to report acquired when young or can be newly on the incidence of CT, on positivity infected as the consequence of inhabiting to Toxoplasma-specific IgM in infants a zone of high infection risk. This and pregnant women (including sero- suggests that, in those areas, the majority conversion results) or on positivity to of infected adults should be chronically Toxoplasma-specific IgG in the general infected. However, acute lesions by population. Various modelling techniques METHODOLOGY

repetitive infections are frequently HAZARD SPECIFIC were used, depending on the country- superimposed on chronic disease [80]. specific data available, to estimate the Therefore, it is reasonable to assume CT incidence and burden in each country. that such overlapping series of repeat Reports of children born with CT, IgM infections result in life-long sequelae. serology of infants and pregnant women, Thus the incidence of trematode infection and age-stratified sero-prevalence in was estimated from the numbers of new women and the general population, cases in each age cohort. combined with fertility rates of specific age groups, were used to directly To estimate the incidence of alveolar estimate the incidence of CT, or the data echinococcosis (AE), due to infection was used to input into models that were with the larval stage of Echinococcus able to generate CT incidences from IgM- multilocularis, literature searches were sero-positive rates in children or pregnant undertaken in any relevant databases women, or from the IgG-sero conversion that could be accessed. These data rates in women, combined with age- sources were synthesized to obtain specific fertility rates. These data were estimates of the incidences of AE in then synthesized into an estimate of the countries where E. multilocularis was global incidence of CT and of the global known to be endemic. Further details of burden of CT in disability-adjusted life the strategy to obtain the data, together years (DALYs). Further details of the with the methodology to estimate methodology, inclusion criteria, PRISMA incidences from the data, are described statement and the modelling techniques in the report from FERG 2. For cystic used are given in [76]. Data on sero- echinococcosis (CE), due to infection prevalence were also used to estimate with the larval stage of E. granulosus, the the incidence of acquired toxoplasmosis. results of a systematic review [73] and Thus changes in sero-prevalence other databases were used. WHO Estimates of the global burden of foodborne diseases 37

T. solium neurocysticercosis (NCC) is outbreak reports were analysed. Searches known to cause epilepsy and other of six international databases yielded neurological sequelae [73]. A meta- 494 reports, of which 261 were selected analysis revealed that brain lesions due to for data extraction after applying strict NCC are present in approximately 29.0% relevance and reliability criteria. From (95% UI 22.9%– 35.5%) of people with 1986 to 2009, there were 65 818 cases epilepsy in populations living in reported from 41 countries, with T. solium endemic areas in settings with 42 deaths. The apparent annual poor and pig management incidence of and mortality caused by practices, and where pork is consumed trichinellosis was calculated by dividing [74]. Consequently, the incidence, the average number of cases and deaths prevalence, mortality and burden of in this 24-year period by the 1997 mid- disease due to epilepsy (including both year population. Due to the important idiopathic and secondary) used in the variability in reporting of the disease, Global Burden of Disease Study 2010 the apparent incidence and mortality (GBD2010) [58, 81– 83] were used rates per billion persons per year were to estimate the burden of epilepsy- adjusted to account for under-reporting associated NCC. of the cases due to under-ascertainment, Once the population at risk was known, medical misclassification, and/or absence 29% of the burden of epilepsy from of effective surveillance systems. The GBD2010 was applied to that population data analysis focused on incidence, age to estimate the burden of epilepsy and sex of patients, major clinical aspects attributable to NCC. Although NCC including sequelae, and meat sources can show many other neurological of infection. Full details of the search and psychiatric symptoms [81], in the criteria, data sources and analysis are absence of available consistent data on described in [71]. The global burden of these other sequelae, only the burden of trichinellosis was subsequently estimated, NCC-associated epilepsy was estimated which is described elsewhere [84], where in this study. full details of the methodology are given. In the case of NCC, prevalence-based Of the 12 PDTF hazards (including YLDs were used. However, in the congenital and acquired Toxoplasma absence of evidence of strong temporal gondii as separate entities), two hazards trends in incidence, this is a reasonable did not need imputation. For epilepsy approximation for incidence-based YLDs. due to Taenia solium, we applied the GBD2010 burden envelopes [81]. For Data on the global prevalence of human trichinellosis, the regional estimates ascariosis, stratified by age, gender and generated by Devleesschauwer et al. country, were provided by the Institute [84] were applied. For the 10 remaining for Health Metrics and Evaluation hazards, the total number of countries (IHME). Based on these data and using with missing data ranged from 5 to 90 the life expectancy of the parasite (out of 194 countries included). Among (approximately 1 year), the equivalent the 194 countries included, the number of incident cases were estimated from the hazards for which no data were available prevalence data. The sequelae proposed ranged from 0 to 6 (out of 10 hazards). in GBD2010 [82], were used in this study. For the five most populous countries in To assess the global incidence and the world, the number of hazards with no clinical effects of human trichinellosis, data were 0 (China), 6 (India), 3 (United Hazard-specific methodology 38

States of America), 2 () and insects. Appropriate processing before 3 (Brazil). consumption can reduce cyanogenic glucoside content of cassava. High 4.4 Chemicals and toxins dietary cyanide exposure occurs when high-cyanogenic cassava and insufficient At its first meeting, the CTTF identified processing combine, usually in a context groups of chemicals and toxins that of food shortage. Cyanide in cassava is are of highest priority in estimating associated with acute cyanide the burden of foodborne disease and several diseases, including konzo (Appendix 3). The hazards were [86]. Worldwide reports exist of acute ranked on: (1) the severity of potential poisoning from cyanide in cassava [86], health effects; (2) the prevalence of but the data are inadequate to make exposure; and (3) the availability of burden estimates. The data are sufficient, data to make burden estimates. After however, to make burden estimates of considerable discussion, the final list konzo. Konzo is an irreversible spastic of chemicals and toxins for which the paraparesis of sudden onset, associated CTTF believed that burdens could be with the consumption of bitter cassava estimated were aflatoxin5, cyanide in [87, 88] and a low protein intake [89]. cassava, peanut allergen, dioxin and It is a disease of extreme poverty. dioxin-like compounds6, methylmercury, Konzo mostly occurs in , but lead, arsenic and cadmium. Only the sporadic cases are also reported. The results for aflatoxin, cyanide in cassava, case definition includes the following METHODOLOGY peanut allergen, and dioxin are presented criteria: (1) a visible, symmetrically spastic HAZARD SPECIFIC here. The results for the metals will be abnormality of gait while walking and/ provided in a subsequent publication. or running; (2) a history of abrupt onset For each of the four chemicals, a (less than one week), followed by a non- systematic literature review was progressive course in a formerly healthy conducted. It was concluded that burden person; and (3) bilaterally exaggerated estimates could be developed for: knee and/or ankle jerks without signs of (1) cyanide in cassava, and associated disease in the spine [89, 90]. konzo syndrome; (2) peanut (Arachis Because konzo mostly affects remote hypogaea) allergy; (3) aflatoxin and rural areas where health infrastructure is (HCC); poor or non-existent, many cases remain (4) dioxin and hypothyroidy; and undiagnosed or unreported, so the true (5) dioxin and decrease in sperm count. burden of disease remains unknown. No cases have been reported from urban 4.4.1 Cyanide in cassava areas. A total of 2376 konzo cases have Cassava is an important staple for over been reported in 5 countries in Africa 800 million people in approximately (Cameroon, Central African Republic, 80 countries, mostly in sub-Saharan Democratic Republic of Congo (DRC), Africa but also in Asia, the Pacific, Mozambique, and United Republic of and [85]. Cassava ) [86], corresponding to 149 tubers contain a varying quantity of cases per year for 122 million people. cyanogenic glucosides, which protect Dividing the average annual number the root against attack by animals and of cases for each country by the 5 The term, "aflatoxin," refers to all aflatoxins. corresponding country population 6 The term, “dioxin,” refers to dioxins and dioxin- produces an observed incidence ranging like PCBs. from 0.043 to 0.179 per 100 000. The WHO Estimates of the global burden of foodborne diseases 39

degree of underestimation is difficult onset, with a most likely value of three to determine as konzo occurs in rural years after onset, following Banea et al. areas, often under conditions of war, and [94] and Tylleskar et al. [96]. the disease is not notifiable. The only There is no DW specifically for konzo. reported calculation of underestimation WHO defined three severity levels for was that of Tylleskar [91] in the DRC in konzo: (1) Mild = able to walk without 1994, when he estimated that at least support; (2) Moderate = uses one or two twice as many cases may have occurred sticks or crutches to walk; and (3) Severe as those reported. The underestimation in = not able to walk [89]. The GBD2010 the DRC is likely to be much greater more DWs for mild, moderate, and severe recently, due to war and displacement. motor impairment are 0.012, 0.076 and It was therefore decided to account for 0.377, respectively [82]. The distribution the uncertainty in the under reporting of konzo severity among 753 patients by applying an expansion factor ranging from nine different studies were mild uniformly from 1 to 10 to the observed (63%), moderate (27%) and severe (10%) cases. The mean annual incidence rate [91, 93, 94, 96– 101]. This distribution and was therefore estimated as 0.9/100 000 the DWs described above were used (0.04 to 1.8/100 000). This estimate of to assign a disability weight of 0.065 the burden of konzo is restricted to the to konzo. 5 African countries described above, and Angola. The decision to include Angola is 4.4.2 Peanut allergen based on a report to the World Congress Prevalence data on peanut [Arachis on Neurology suggesting that cases have hypogaea] allergy were used to make occurred in that country [92]. Although estimates of incidence, since allergy cassava consumption occurs in tropical occurs early in life (<5 years) and is areas throughout the world, the term believed to be lifelong [102– 106]. All konzo has only been used to describe peanut allergy cases are assumed to be cases in Africa. The incidence of konzo in the result of eating peanuts or peanut other countries in Africa and other parts products. In western countries, the of the world is assumed to be zero. prevalence of clinical peanut allergy in The age of onset and gender distribution children is 0 to 1.8% of the population of these cases was assumed to be that [102], corresponding to incidence rates observed by Tylleskar [90]. The konzo of 0 to 22.6 per 100 000. Limited data case-fatality ratio is approximately 21% exist on the mortality rate of peanut- based on four studies [90, 93– 95]. The induced anaphylaxis, but the majority of studies found similar rates, ranging age and gender distribution of fatal from 0 to 0.006 deaths per 100 000 cases was assumed to be that of Tshala- person-years [102]. Incidence was Katumbay [93]. estimated only for the A level (high The onset of paraparesis in konzo is income) subregions; too few data exist abrupt, usually within minutes or hours, to make estimates for other subregions with occasional progression during [102]. Several studies have reported that the first days of the illness. After that 63– 66% of cases are male [102], but time, the paraparesis is non-progressive given the uncertainty in this number, and permanent. As a result, duration is the gender distribution was assumed defined as lifelong for non-fatal cases. to be equal for the burden of disease For fatal cases, it was assumed that calculations. No DW exists for peanut death occurred one to seven years after allergy. Mullins et al. [103] reported that Hazard-specific methodology 40

52% of cases referred to a specialist derived [111] and global populations, the allergy medical practice in population attributable fractions (PAFs) suffered from mild symptoms (skin and by country were estimated, and applied subcutaneous tissue involvement only), to HCC incidence and mortality based on 42% from moderate symptoms (features information from WHO [112, 113]. suggestive of respiratory, cardiovascular A Bayesian log-normal random effects or gastro intestinal involvement), and model [79] was used to extrapolate 6% from severe symptoms (cyanosis, available PAFs to countries without hypotension, confusion, collapse, loss of data. Age-specific incidence estimates consciousness, incontinence). were derived from a study in China The DW for peanut allergy was assigned comparing age-specific incidence of as a weighted average accounting for HCC in Qidong, a city in China with high this severity distribution. GBD2010 aflatoxin exposure, and Beijing, a city DWs [82] for the health states defined with low aflatoxin exposure [114]. The in the category “Asthma: controlled” YLD and YLL envelopes for HCC available (DW=0.009) are considered applicable from WHO were multiplied by the for mild and moderate cases (94%), and proportion of the burden due to aflatoxin. “Generic uncomplicated disease: anxiety Thus no DW was directly involved in about the diagnosis” (DW=0.054) for the calculation. severe cases (6%), because anxiety is known to affect Quality of Life in 4.4.4 Dioxin food allergic patients [107], leading to Dioxins are mainly by-products of METHODOLOGY a severity-weighted DW of 0.012 for industrial processes, but can also HAZARD SPECIFIC clinically relevant peanut allergy. Unlike result from natural phenomena, such other childhood , such as cow’s as volcanic eruptions and forest fires. and egg allergy, peanut allergy rarely More than 90% of human exposure is resolves [108, 109]. through food, mainly meat and dairy products, fish and shellfish [115]. Due 4.4.3 Aflatoxin to the bio-accumulating and lipophilic Aflatoxins are secondary metabolites characteristics of dioxins, daily dietary of the fungi Aspergillus flavus and A. exposure leads to accumulation of these parasiticus, and less frequently other compounds in human body fat. In adults Aspergillus species such as A. nomius this accumulation is thought to reach [110]. These species are prevalent in a constant level (i.e. a steady state). food crops – particularly maize, peanuts Consequently, the dioxin body burden, (groundnuts), oilseeds and tree nuts rather than the daily exposure, is taken as – in tropical and subtropical regions the metric for chronic risk worldwide [110]. It is believed that all and the assessment of dioxins [116– 121]. aflatoxin exposure results from food In this context the dioxin concentration consumption. A multiplicative model in breast milk fat directly reflects the concentration in body fat [121– 124]. was assumed for the effects of aflatoxin exposure and virus (HBV) Many national authorities have infection on hepatocellular carcinoma. programmes in place to monitor dioxin Aflatoxin exposure by country is that in the food supply and breast milk described by Liu and Wu [110]. To [124– 126]. Dioxin-induced pre-natal and account for differences in background post-natal hypothyroidy and pre-natally rates between the study population from induced reduced sperm production have which the cancer potency factor was been found to be the most sensitive WHO Estimates of the global burden of foodborne diseases 41

non-cancer toxic endpoints for dioxins. the reduction in cauda epididymis Estimates for dioxin-induced pre-natal sperm count in male offspring [136]. The and post-natal hypothyroidy and reduced resulting dose response data were used fertility due to disturbed sperm formation to calculate a BMD lower confidence were based on an exposure assessment, limit (BMDL) and upper confidence toxicity assessment and the comparison limit (BMDU) dioxin body burden for of both assessments [127, 128]. The various levels of reduction in sperm exposure assessment is based on breast count. A WHO reference cut-off value milk concentrations of dioxin from for impaired fertility of 20 ×106 sperm 50 countries [129]. The toxicity cells/mL was used to link toxicity (sperm assessment utilizes the benchmark count reduction) to a disease status dose (BMD) approach [130– 132] in (impaired fertility) (i.e. the calculation of which the dose response of post- the probability of a male being born with natal total thyroxine (TT; decrease of dioxin-impaired fertility) [137]. TT4 in adult ), pre-natal A BMD analysis of a National Toxicology stimulating hormone (TSH; increase Program (NTP) two-year feeding study in in TSH in neonatal blood), and sperm rats was used to make estimates of dioxin- production (reduced concentration of induced thyroid toxicity. The NTP study sperm cells) is analysed. The toxicity and administered 2,3,7,8 tetrachlorodibenzo- exposure assessments are compared p-dioxin (TCDD) [138] and 2,3,4,7,8– to derive the transgression of a dioxin- pentachlorodibenzofuran [139] for periods induced decrease in TT4, decrease of 14, 31 and 53 weeks. The concentrations in sperm count and increase in were converted to Toxic Equivalent TSH across a physiological threshold Quotients [140] to enable a combined indicating a disease status (i.e. incidence analysis of both congeners. BMDL and of hypothyroidy or impaired fertility). BMDU body burdens for reduction in TT4 Additional details of these assessments were calculated for each of the exposure may be found in Zeilmaker et al. [133]. periods. A distribution of TT4 in human The BMD analysis was performed on blood has been reported by Aoki et al. studies that served as the starting point [135]. The 5th percentile of this distribution for the derivation of a Tolerable Weekly (65 nmol/L) was used as the cut-off for Intake (TWI) [117– 120] or Reference Dose overt clinical hypothyroidism in adults. for dioxin (RfD) [121]. The results of the BMD analyses and In a study of a mother-child cohort, the breast milk concentrations for Baccarelli et al. determined the 50 countries were compared, taking relationship between maternal plasma account of possible differences between dioxin concentration and TSH level [134]. experimental animals and humans and A BMD analysis of these data resulted in among individual humans [127, 128]. a population distribution of the maternal This comparison provided country- body burden of dioxin corresponding specific estimates of the incidence of to an increased TSH level of 5 ųU/mL dioxin induced pre-natal and post-natal in offspring, a level not to be exceeded hypothyroidy and impaired fertility. in 3% of newborns in iodine-replete The estimates were extrapolated to populations [135]. other countries for which no breast Following administration of an acute oral milk concentrations were available, dose to pregnant Long Evans rats on by means of Bayesian random effects day 15 of gestation, Gray et al. measured modelling [79]. Hazard-specific methodology 42

4.5 Outcomes and [4]. For dioxin-induced hypothyroidy, disability weights the GBD2013 DW for hypothyroidy was adopted, as this health state was not DALYs incorporate the severity of health included in the GBD2010 DW study [142]. states through the DW, reflecting the corresponding relative reduction in Several FBDs present with unique clinical healthy life on a scale from zero to one. signs, for which no DWs have been Table 1 lists the DWs used for the health derived. Acute trichinellosis, for instance, states associated with each hazard. typically presents with myalgia and facial Further details are given in Appendix 5 oedema, for which no specific DWs are – Structuring of the health states into available [84]. When DWs were missing, disease models for computation, and proxy health states were selected by a Appendix 6 – Sources and derivation medical expert and DW expert in the CTF of DWs. and confirmed by disease experts in the hazard-specific TFs. DWs for several health states have In other instances, DWs were available been derived for the GBD studies and for severity levels that were not explicitly for various national burden of disease considered in the disease models. studies [141]. To ensure comparability, the For diarrhoea, for instance, DWs were CTF adopted the DWs that were used available for mild, moderate and severe for WHO’s Global Health Estimates [4]. diarrhoea, although the disease models These DWs were based on those derived only included diarrhoea as such. In those for the GBD 2010 study [82], but with an

cases, weighted averages were calculated METHODOLOGY alternative value for primary infertility (i.e. based on published reviews of severity HAZARD SPECIFIC 0.056 instead of 0.011). The latter revision distributions, avoiding an over- or under- was motiviated by an analysis showing estimation of YLDs that would occur if that the GBD 2010 weights undervalued only one DW would have been selected. the health states associated with fertility WHO Estimates of the global burden of foodborne diseases 43

Table 1. FERG hazards, causally related health states and corresponding disability weights (DWs). Details on the derivation of the DWs are provided in Appendix 4.

HAZARD HEALTH STATE DW DIARRHOEAL DISEASE AGENTS Norovirus Diarrhoeal disease 0.074 Campylobacter spp. Diarrhoeal disease 0.101 Guillain-Barré syndrome 0.445 Enteropathogenic E. coli Diarrhoeal disease 0.074 Enterotoxigenic E. coli Diarrhoeal disease 0.074 Shiga toxin-producing E. coli Diarrhoeal disease 0.091 Haemolytic uraemic syndrome 0.210 End-stage renal disease 0.573 Non-typhoidal S. enterica Diarrhoeal disease 0.101 Invasive salmonellosis 0.210 Shigella spp. Diarrhoeal disease 0.101 Vibrio cholerae Diarrhoeal disease 0.194 Cryptosporidium spp. Diarrhoeal disease 0.074 Entamoeba histolytica Diarrhoeal disease 0.074 Giardia spp. Diarrhoeal disease 0.074 INVASIVE INFECTIOUS DISEASE AGENTS Hepatitis A virus Hepatitis 0.108 Brucella spp. Acute brucellosis 0.108 Chronic brucellosis 0.079 Orchitis 0.097 Listeria monocytogenes, perinatal 0.210 Central nervous system infection 0.426 Neurological sequelae 0.292 Listeria monocytogenes, acquired Sepsis 0.210 Central nervous system infection 0.426 Neurological sequelae 0.292 Mycobacterium bovis Tuberculosis 0.331 Salmonella Paratyphi Paratyphoid fever 0.210 Liver abscesses and cysts 0.254 Salmonella Typhi 0.210 Liver abscesses and cysts 0.254 Toxoplasma gondii, congenital Intracranial calcification 0.010 Hydrocephalus 0.360 Chorioretinitis, early in life 0.033 Chorioretinitis, later in life 0.033 CNS abnormalities 0.360 Toxoplasma gondii, acquired Chorioretinitis, mild 0.004 Chorioretinitis, moderate 0.033 Chorioretinitis, severe 0.191 Acute illness 0.053 Post-acute illness 0.254 ENTERIC INTOXICATIONS Bacillus cereus (1) Acute intoxication 0.061 (1) Moderate/mild botulism 0.198 Severe botulism 0.445 Hazard-specific methodology 44

HAZARD HEALTH STATE DW Clostridium perfringens (1) Acute intoxication 0.061 Staphylococcus aureus (1) Acute intoxication 0.061 CESTODES Echinococcus granulosus, cases seeking Pulmonary cystic echinococcosis 0.192 treatment Hepatic cystic echinococcosis 0.123 CNS cystic echinococcosis 0.221 Echinococcus granulosus, cases not Pulmonary cystic echinococcosis 0.015 seeking treatment Hepatic cystic echinococcosis 0.012 CNS cystic echinococcosis 0.054 Echinococcus multilocularis Alveolar echinococcosis 0.123 Taenia solium Epilepsy: treated, seizure free 0.072 Epilepsy: treated, with recent seizures 0.319 Epilepsy: severe 0.657 Epilepsy: untreated 0.420 Ascaris spp. Ascariasis infestation 0.030 Mild abdominopelvic problems due to 0.012 ascariasis Severe wasting due to ascariasis 0.127

Trichinella spp. Acute clinical trichinellosis 0.637 METHODOLOGY HAZARD SPECIFIC Trematodes Abdominopelvic problems due to heavy Clonorchis sinensis 0.123 clonorchiosis Abdominopelvic problems due to heavy Fasciola spp. 0.123 Abdominopelvic problems due to heavy Intestinal flukes (2) 0.123 intestinal fluke infections Abdominopelvic problems due to heavy Opisthorchis spp. 0.123 opisthorchiosis Central nervous system problems due to Paragonimus spp. 0.420 heavy paragonimosis Pulmonary problems due to heavy 0.132 paragonimosis ORGANIC POLLUTANTS Dioxin Infertility 0.056 Hypothyroidy due to pre-natal exposure 0.019 Hypothyroidy due to post-natal exposure 0.019 TOXINS AND ALLERGENS Hepatocellular carcinoma: diagnosis and Aflatoxin 0.294 primary therapy Hepatocellular carcinoma: metastatic 0.484 Hepatocellular carcinoma: terminal phase 0.508 with medication Hepatocellular carcinoma: terminal phase 0.519 without medication Cyanide in cassava Konzo 0.065 Peanut allergens (1) Living with peanut-induced allergy 0.012

Notes: (1) Excluded from global burden assessments. (2) Includes Echinostoma spp., Fasciolopsis buski, Heterophyes spp., Metagonimus spp. and other foodborne intestinal trematode species. WHO Estimates of the global burden of foodborne diseases 45

4.6 Attribution through food contamination, resulting in cysticercosis. In the complete absence of Overall, the study was designed to pork consumption, there would be no provide estimates of the proportion of T. solium taeniosis and hence no illness acquired through different major cysticercosis. routes of exposure. Major exposure routes considered were: food, environmental The regions selected for this study were (water, soil, air), human-to-human based on mortality. Six general regions: transmission, direct animal contact, and a Africa (AFR), the Americas (AMR), the variety of potential lead exposure sources. Eastern Mediterranean (EMR), Europe Exposure route attribution estimates were (EUR), South-East Asia (SEAR) and the developed for 19 individual hazards for Western Pacific (WP) were then divided each of the fourteen subregions (Table 2). into subregions on the basis of child and Three hazard-based TFs within FERG adult mortality, where Stratum A = very low (EDTF, PDTF and CTTF) identified, from child and adult mortality; Stratum B = low their prioritized lists of hazards, those to be child mortality and very low adult mortality; included in the expert elicitation. Stratum C = low child mortality and high adult mortality; Stratum D = high child and Certain hazards were considered 100% adult mortality; and Stratum E = high child foodborne, i.e. Listeria monocytogenes, mortality and very high adult Mycobacterium bovis, all foodborne mortality [5, 144]. trematodes, Taenia solium, Trichinella spp., cyanide in cassava and peanut 4.6.1 Identification of experts allergens. For aflatoxin, inorganic arsenic, An iterative peer nomination process cadmium, dioxin and methyl mercury, based on a social network sampling CTTF determined that adequate data on technique called “snowball sampling” was foodborne exposure existed to allow use of used to identify a pool of potential expert a risk assessment approach for estimating participants for this study. The first points the foodborne disease burden, thus of contact were identified through FERG negating the need for attribution. members and other networks (e.g. Global The remaining hazards were included in the Foodborne Infections Network – GFN; structured expert elicitation (Table 1). Global Environment Monitoring System -borne trematodes and Trichinella – GEMS; International Network of Food spp. were assumed to be 100% foodborne, Safety Authorities – INFOSAN; Joint FAO/ based on the nature of their life cycle. In WHO Expert meeting on Microbial Risk addition, Fasciola spp. were assumed to Assessment – JEMRA; Joint FAO/WHO be 100% foodborne, although there may Expert Committee on Food Additives – be small opportunities for waterborne JECFA; European Food Safety Authority transmission [77, 143]. Taenia solium – EFSA scientific panels; and WHO regional cysticercosis was assumed to be 100% food safety advisors). These persons were foodborne, but indirectly. In other words, asked to use their professional networks the T. solium life cycle cannot persist and recognized expertise in relevant areas without foodborne transmission of the to nominate additional experts. Since the parasite between pigs and humans. purpose of this process was to identify Humans become infected by the adult an adequately large pool of appropriate stage of T. solium by eating pork, resulting experts, rather than to identify the entire in intestinal taeniosis. However individuals expert network, the process of referral who have T. solium taeniosis infect continued until an adequate size pool was themselves or others by eggs excreted in identified to fill panels of typically their faeces, which are then ingested, often 8 to 12 experts per panel. Hazard-specific methodology 46

Table 2. Foodborne hazards, and structure of the expert panels.

NO. OF HAZARD GROUPS HAZARDS PANEL STRUCTUREa PANELS

DIARRHEAL DISEASE

Bacteria Campylobacter spp., enteropathogenic Escherischia coli (EPEC), enterotoxigenic E. coli (ETEC), Shiga-toxin Sub regional 7 producing E. coli (STEC), non-typhoidal Salmonella spp., Shigella spp., Vibrio cholerae

Virus Norovirus Sub regional 1

Intestinal protozoa Cryptosporidium spp., Entamoeba histolytica, Giardia spp. Global 3

OTHER INFECTIOUS DISEASE

Bacteria Brucella spp., Global 2 Salmonella Typhi Sub regional

Virus Hepatitis A virus Global 1

Protozoa Toxoplasma gondii Global 1

Helminths Ascaris spp., Echinococcus granulosus, Echinococcus Global 3 multilocularis

CHEMICALS

Lead Global 1

Total 19 a Experts on a global panel were asked to provide estimates for all 14 sub regions, whereas experts on a sub-regional panel could METHODOLOGY choose a set of sub regions depending on their expertise. HAZARD SPECIFIC

4.6.2 Selection of experts TFs and the SATF reviewed the expert’s In collaboration with the FERG hazard- information and CVs, and a final selection based TFs, SATF defined a set of criteria was made. FERG TF chairs and members for inclusion of experts. These criteria of the SATF were not eligible for the considered each expert’s background study. DOIs were evaluated by WHO. (education, and current and past Given the broad nature of the attribution positions), years of experience within task, care was taken to include a suitably the field, and geographic coverage of wide range of scientific backgrounds and expert within the panel. The WHO invited professional experience, and to ensure nominated experts to participate, and adequate geographical representation. the experts were asked to complete a Frequently, expert elicitations use declaration of interests (DOI), and an publication record as the measure of expert sheet providing information on recognized expertise [145]. However, for their research/working area, highest this study, restricting expert selection education, current position, geographical to choices based solely on publication experience, and years of experience. records would have eliminated important The experts were asked also to indicate groups of experts, in particular public- for which panel(s) they believed health field workers and food-safety themselves to be best suited for. professionals in developing countries. The experts were not offered any compensation for their participation. The chairs of the three hazard-based WHO Estimates of the global burden of foodborne diseases 47

4.6.3 Expert panels (narrow) are the distributions provided). The panels for Brucella spp., hepatitis Experts are thus scored with regard to A virus, parasitic diseases including statistical accuracy (calibration score) intestinal protozoa, and lead were and informativeness (information structured as global panels, meaning that score). The statistical accuracy is all experts in those panels were asked measured as the p-value at which one to provide estimates for all fourteen would falsely reject the hypothesis that subregions. the expert’s probability assessments The panels for the eight bacterial and were statistically accurate, and viral pathogens and Salmonella Typhi informativeness is measured as Shannon were structured as subregional panels. relative information with respect to a user-supplied background measure. The experts on subregional panels were Informativeness scores are not absolute, free to decide the subregions for which but relative to a set of experts assessing they provided their judgments. Experts the same variables. The calculated participating in panels addressing more calibration and information scores are than one hazard could also choose used to aggregate experts' judgments to provide estimates only for those on target variables. The same measures hazards for which they felt they had can be applied to any combination of adequate expertise. the experts’ assessments to implement criteria for aggregating the assessments. 4.6.4 Analytical method The study used Cooke’s [145– 147] The Cooke Classical Model provides “Classical Model” for expert elicitation. a rigorous, quantitative means for This approach uses “calibration” or estimating model parameters and their “seed” questions to develop performance uncertainties and is the only elicitation weights used in aggregating experts’ procedure that has objective empirical judgments. The paradigmatic seed control on expert scoring. Moreover, it question is one for which the true value allows formal optimization of aggregated is not known at the time the experts uncertainty distributions in terms of answer the question, but will be known statistical accuracy and informativeness or is expected to become known post [146]. The expert judgment processing software EXCALIBUR (http://www. hoc. So the experts are not expected to lighttwist.net/wp/excalibur) also allows know these values, but should be able direct comparison of the results that to capture a majority of them reliably by would be obtained from unweighted defining suitable credible intervals. aggregation of expert judgments versus Analysis of the experts’ performance those produced by weighted linear on the seed variables has two main pooling (or other combination schemes). purposes: 1), to evaluate the expert’s statistical accuracy when assigning 4.6.5 Seed questions values to probability outcomes against It is not always possible to develop seed the seed values (i.e. how reliably the questions that are in the paradigmatic expert’s credible interval responses form of asking about a future event or capture the true values of the seed measurement that has not been made, variables, statistically), and 2), to evaluate but could be made, in principle. The the expert’s informativeness when essential feature of a viable seed question providing uncertainty distributions over is that the expert is not expected to know the seed variables (i.e. how concentrated the exact value but, if they are a subject- Hazard-specific methodology 48

matter expert, should be able to define a were designed to focus on questions that narrow uncertainty range that captures are substantively related to foodborne the value. Therefore, an alternative is to illness source attribution. Further, to ask about selected data or values in the account for the wide range of scientific topic domain, about which the expert backgrounds and experiences, seed will not have perfect knowledge, nor questions covered a range of substantive access to realization values at the time topics relevant to source attribution. Five they are answering the seed questions, main categories of seed questions were but for which the values are known to the identified for the panels on biological analyst. Such “retrospective” questions hazards (diarrhoeal pathogens and are frequently used in expert elicitations parasites): (1) dietary patterns and food applying the Cooke Classical Model (see supply; (2) under 5 years mortality e.g. [31, 148]). rate; (3) access to improved water and sanitation; (4) disease surveillance; and In the present case, the seed questions (5) systematic reviews related to these formulated were a mixture of and other scientific topics relevant to retrospective and prospective seed source attribution. For the panel on lead, variables. It is possible that expert questions were categorized as: (1) mean uncertainty judgments vary by subject blood levels; (2) dietary exposure; matter domain. In this study, the and (3) dietary patterns and food possibility of such biases relevant to supply. Examples of seed questions are foodborne illness source attribution was presented in Table 3. of concern. Therefore, the seed questions METHODOLOGY HAZARD SPECIFIC

Table 3. Examples of calibration seed questions

TOPIC HAZARD QUESTION Dietary patterns All microbial hazards Among all subregions in 2010, what was the proportion of regional and food supply vegetable supply (tonne) that was imported rather than produced domestically in the subregion with the highest such percentage? Under 5 mortality Brucella spp., Echinococcus Based on WHO estimates, think of the country in the African Region that rate spp., intestinal protozoa, had the largest percentage point decrease from 2000 to 2010 in all-cause diarrhoeal pathogens <5 mortality due to diarrhoea. What was that percentage point decrease? Disease surveillance Ascaris spp., Echinococcus spp., intestinal protozoa, What will be the rate per 100 000 population of laboratory-confirmed hepatitis A virus, diarrhoeal human cases of in 2012 in all EU member states as pathogens (developed reported in EFSA’s annual report? subregions only) Systematic review All microbial hazards Fewtrell et al. (2005) conducted a systematic review and meta-analysis to compare the evidence of relative effectiveness of improvements in , sanitation facilities and practices in less developed countries in reducing diarrhoeal illness. The meta-analysis of 5 studies was used to estimate the relative risk of diarrhoeal illness with and without multiple interventions. What was the estimated relative risk? Mean blood level Lead What was the geometric mean blood lead concentration for all participants ages 1 year and older in the 2007– 2008 U.S. NHANES survey? Please express your answer as positive micrograms per deciliter (ųg/dL). WHO Estimates of the global burden of foodborne diseases 49

All experts comprising each panel were the questionnaires did provide experts asked the same set of seed questions, with an option to indicate additional and several sets of seed questions were routes of transmission if they disagreed used across panels. This allows some with the TF’s evaluation. consistency checks to be performed Experts were asked to complete a set between panels on performance and of tables for each assigned hazard and scoring outcomes. The number of subregion. Experts were provided with questions varied from nine to twelve in the tables at the end of the telephone total, depending on the panel. Experts interview during which the seed were asked to provide a central judgment questions were asked, and the facilitator in terms of the median value, and a went through several target questions 90% credible interval for each question. Seed questions were administered with the experts to ensure that they by facilitators through one-to-one understood the task. For the target items telephone interviews. The experts were (but not the seed questions), the experts not presented with the seed questions were free to consult any information before the interview and they were asked sources they felt appropriate in the four- to provide estimates based on their week period they were given to return experience, knowledge and judgment, the target item spreadsheets. without referring to other sources As with the seed questions, the experts of information. were asked to provide their 5th, 50th and 95th percentile values for each 4.6.6 Target questions question. Technically, the median values Target questions are the substantive of a joint distribution do not need to add questions of interest. In this study, for all up to 100%, but because we included identified hazards, we enquired about the a category “other”, we asked about a percentage of all human disease cases joint distribution that logically spanned caused by exposure through each of a all possible exposure routes. Therefore, number of indicated exposure routes. the experts’ median values for source The point of exposure was chosen as the attribution percentages for a hazard point of attribution, i.e. the experts were should sum to a value close to 100%. asked to distribute the disease burden In individual cases, where these sums on the sources that was the direct cause were found to differ significantly (i.e. of human exposure. So, for example, outside 100% ± 10%), the experts someone with a norovirus infection might concerned were asked to review their be exposed by eating food contaminated responses. with the virus, although the food may have been contaminated by wastewater 4.6.7 Data analysis at an earlier stage. Weights for individual experts were Exposure routes varied between hazards, computed using the Classical Model as indicated in Table 4. In order to formulation, using the EXCALIBUR reduce the time and effort burden of software, by multiplying their calibration the elicitation on expert panelists, the and informativeness scores, with hazard-based TFs decided which hazard the products then jointly normalized exposure routes were relevant for present to sum to unity over all experts in purposes. So, for example, human-to- the group. An expert was positively human transmission was excluded as an weighted only if his/her p-value was exposure route for Brucella spp. However, above a certain threshold, chosen to Hazard-specific methodology 50

optimize the combined score across all The quantiles corresponding to these seed items. See [146, 149] for further random deviates were then obtained details on the computation of expert via linear interpolation. To ensure that performance weights. the random attributional proportions summed to 100%, a “normalization” step The normalized experts’ performance was applied per iteration, in which each weights were used to construct the joint random value was divided by the sum probability distribution complying with of random values for each exposure their assessments for individual target pathway. The resulting 10 000 normalized questions (i.e. attributable proportion random attributional proportions were of illness per pathway, subregion and then summarized by their median and a hazard). In a final step, 10 000 random 95% uncertainty interval defined by the values from the marginal cumulative 2.5th and 97.5th percentiles. These final distributions were simulated using manipulations were performed in R 3.1.1 probability integral transform [34]. [150] using functions available in the First, independent vectors of 10 000 ‘FERG’ package [151]. random deviates from a Uniform (0,100) distribution, per exposure category within a hazard-subregion combination, were generated. METHODOLOGY HAZARD SPECIFIC WHO Estimates of the global burden of foodborne diseases 51

Table 4. Exposure routes included in the expert elicitation, per hazard

HAZARD AIR SOIL TOYS FOOD PAINT OTHER WATER AND WILD) GLASSWARE COOKWARE, POTTERY OR POTTERY COOKWARE, HUMAN TO HUMAN CONTACT HUMAN TO ANIMAL CONTACT (DOMESTIC (DOMESTIC ANIMAL CONTACT

DIARRHOEAL DISEASE

Campylobacter spp. xxxxxnaN/AN/AN/Ax

Non-typhoid Salmonella spp.xxxxxN/AN/AN/AN/Ax

Shiga toxin-producing E. coli xxxxxN/AN/AN/AN/Ax

Brucella spp. x x N/A x x N/A N/A N/A N/A x

Shigella spp. x N/A x x x N/A N/A N/A N/A x

Enteropathogenic E. coli xxxxN/AN/AN/AN/AN/Ax

Enterotoxigenic E. coli xxxxN/AN/AN/AN/AN/Ax

Cryptosporidium spp.xxxxN/AN/AN/AN/AN/Ax

Giardia spp. xxxxN/AN/AN/AN/AN/Ax

Tyhoid Salmonella spp. x N/A x x N/A N/A N/A N/A N/A x

Vibrio cholerae x N/A x x N/A N/A N/A N/A N/A x

Entamoeba histolytica x N/A x x N/A N/A N/A N/A N/A x

Norovirus x N/A x x N/A N/A N/A N/A N/A x

Hepatitis A virus x N/A x x N/A N/A N/A N/A N/A x

PARASITIC DISEASE

Toxoplasma gondii x x N/A x x N/A N/A N/A N/A x

Echinococcus granulosus x x N/A x x x N/A N/A N/A x

Echinococcus multilocularis x x N/A x x x N/A N/A N/A x

Ascaris spp. xxxxxN/AN/AN/AN/Ax

CHEMICALS

Lead xN/AN/Axxxxxxx

Notes: N/A = not applicable, meaning that these exposure routes were considered not possible or extremely unlikely by the respective FERG TF.

4.7 Computation enterica), or a risk factor (e.g. an exposure that inceases the likelihood of illness such Different approaches can be taken for as unsafe water) [152]. Since FERG is calculating DALYs, depending on whether concerned with the burden of FBDs, which the interest lies in quantifying the burden are caused by a wide range of hazards of a health outcome (such as diarrhoea), (bacteria, viruses, protozoa, parasites, a hazard (e.g. a that may chemicals and toxins), a natural choice is cause illness in humans such as Salmonella the hazard-based approach. This approach Hazard-specific methodology 52

defines the burden of a specific foodborne models, and not merely as as that resulting from the health disease models. While biological disease states, i.e. acute and chronic manifestations models merely reflect the natural history of disease, including death, that are causally of disease, computational disease models related to the hazard transmitted through also reflect the input parameters needed food, and which may become manifest to calculate incidence and mortality of at different time scales. This approach each of the relevant health states. As thus allows for a comprehensive estimate such, computational disease models are of the burden of disease due to a certain a combination of disease biology and hazard, including sequelae, which may have data availability. a higher burden than acute illness alone Computational disease models may be [153– 155]. represented as directed acyclic graphs, defined by parent and child nodes and 4.7.1 Disease models and directed edges (arrows) defining the epidemiological data relationships between nodes. In the CTF The starting point of the CTF workflow framework, parent nodes were either was the outline of disease models incidence, mortality, YLD or YLL rates, for each of the included hazards (as while child nodes were multiplicative chosen by the hazard-based TFs), and elements, such as proportions or ratios the epidemiological data inputs that (reflecting, e.g. the probability of parameterized these disease models. developing a specific symptom following To obtain this information, systematic

infection, or the proportion of illnesses METHODOLOGY reviews were commissioned and attributable to the concerned hazard). HAZARD SPECIFIC managed by three hazard-based TFs, i.e. A specific disease model “language” was the Enteric Diseases Task Force (EDTF), developed to denote the relationship the Parasitic Diseases Task Force (PDTF), and contribution of the different nodes. and the Chemicals and Toxins Task Rectangles defined parent nodes, and Force (CTTF). rounded rectangles defined child nodes. The course of disease is characterized Grey nodes did not contribute directly by various health states (e.g. acute to the DALYs, green nodes contributed or chronic phases; short-term or YLDs, and red nodes contributed YLLs. long-term sequelae), possibly having Nodes that contributed to the incidence different severity levels. A disease of the index disease were identified by model, also referred to as an outcome a thick border. Appendix 5 gives the tree, is a schematic representation of disease models for all 36 FERG hazards. the various health states associated In general, three main approaches can be with the concerned hazard, and the distinguished for estimating the burden possible transitions between these due to a specific hazard in food, i.e., states. A disease model for each categorical attribution, counterfactual hazard was defined by the members analysis, and risk assessment. Table and commissioned experts of each 5 gives an overview of the modelling hazard-based TF, considering relevant strategy applied for each included health outcomes identified in the hazard. As the choice of the modelling respective reviews. strategy was mainly driven by the type In the context of the CTF, disease models of data available, no sensitivity analyses were defined as computational disease could be performed to triangulate different modelling approaches. WHO Estimates of the global burden of foodborne diseases 53

Table 5. Modelling strategies for the hazards included in the WHO global burden of foodborne disease estimates

BURDEN DISEASE MODEL FOODBORNE HAZARD ATTRIBUTION IMPUTATION APPROACH ATTRIBUTION APPROACH DIARRHOEAL DISEASE AGENTS Categorical Modified CHERG Expert elicitation Norovirus Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Campylobacter spp. Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Enteropathogenic E. coli Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Enterotoxigenic E. coli Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Shiga toxin-producing E. coli Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Non-typhoidal S. enterica Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Shigella spp. Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Vibrio cholerae Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Cryptosporidium spp. Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Entamoeba histolytica Attribution attribution approach [40] [156] Categorical Modified CHERG Expert elicitation Giardia spp. Attribution attribution approach [40] [156] INVASIVE INFECTIOUS DISEASE AGENTS Categorical Expert elicitation Hepatitis A virus Direct: GBD2010 [3] N/A (1) attribution [156] Categorical Random effects Expert elicitation Brucella spp. Transition attribution [79, 151] [156] Categorical Random effects Listeria monocytogenes, perinatal Transition 100% attribution [79, 151] Categorical Random effects Listeria monocytogenes, acquired Transition 100% attribution [79, 151] Categorical Mycobacterium bovis Attribution N/A (1) 100% attribution Categorical Expert elicitation Salmonella Paratyphi Direct: GBD2010 [3] N/A (1) attribution [156] Categorical Expert elicitation Salmonella Typhi Direct: GBD2010 [3] N/A (1) attribution [156] Categorical Random effects Expert elicitation Toxoplasma gondii, congenital Transition attribution [79, 151] [156] Categorical Random effects Expert elicitation Toxoplasma gondii, acquired Transition attribution [79, 151] [156] ENTERIC INTOXICATIONS Categorical Bacillus cereus (2) Direct Uniform 100% attribution Categorical Clostridium botulinum (2) Direct Uniform 100% attribution Categorical Clostridium perfringens (2) Direct Uniform 100% attribution Hazard-specific methodology 54

BURDEN DISEASE MODEL FOODBORNE HAZARD ATTRIBUTION IMPUTATION APPROACH ATTRIBUTION APPROACH Categorical Staphylococcus aureus (2) Direct Uniform 100% attribution CESTODES Echinococcus granulosus, cases Categorical Random effects Expert elicitation Transition seeking treatment attribution [79, 151] [156] Echinococcus granulosus, cases Categorical Random effects Expert elicitation Transition not seeking treatment attribution [79, 151] [156] Categorical Random effects Expert elicitation Echinococcus multilocularis Transition attribution [79, 151] [156] Categorical Taenia solium Attribution N/A (1) 100% attribution NEMATODES Categorical Expert elicitation Ascaris spp. Direct: GBD2010 [3] N/A (1) attribution [156] Categorical Trichinella spp. Direct N/A (1) 100% attribution TREMATODES Categorical Random effects Clonorchis sinensis Direct 100% attribution [79, 151] Categorical Random effects Fasciola spp. Direct 100% attribution [79, 151]

Categorical Random effects METHODOLOGY

Intestinal flukes (3) Direct 100% HAZARD SPECIFIC attribution [79, 151] Categorical Random effects Opisthorchis spp. Direct 100% attribution [79, 151] Categorical Random effects Paragonimus spp. Direct 100% attribution [79, 151] ORGANIC POLLUTANTS Random effects Dioxin Risk assessment Direct 100% [79, 151] TOXINS AND ALLERGENS Counterfactual Random effects Aflatoxin Attribution 100% analysis [79, 151] Categorical Cyanide in cassava Direct Uniform 100% attribution Categorical Peanut allergens (2) Direct Uniform 100% attribution

Notes: (1) N/A = not applicable as no imputation had to be performed because data were used that had already been imputed. (2) Excluded from global burden assessments. (3) Includes Echinostoma spp., Fasciolopsis buski, Heterophyes spp., Metagonimus spp. and other foodborne intestinal trematode species. WHO Estimates of the global burden of foodborne diseases 55

Categorical attribution can be used cancer case was caused by aflatoxin. In when a foodborne hazard results in an this situation, the standard approach for outcome (death or a specific syndrome) calculating the burden of environmental that is identifiable as caused by the exposures is to use a counterfactual hazard in individual cases [157]. Following analysis in which the current disease the typology of Devleesschauwer et outcomes with current exposure are al. [152], the burden due to a specific compared with the disease outcomes hazard can then be calculated using under an alternative exposure (a min. an attributional model (in which the risk exposure which could be zero, or incidence of the symptom is multiplied some accepted background level) [158]. with the attributable proportion for a This allows calculation of a population given hazard) or a transitional model attributable fraction (PAF) that can be (in which the incidence of the hazard applied to the all-cause burden estimates is multiplied with the probability for the relevant disease outcome (the of developing a given symptom). so-called burden envelope), leading Categorical attribution was applicable to a special case of the attributional for all viral, bacterial and parasitic model [152]. In the context of FERG, hazards, and for cyanide in cassava and counterfactual analysis was used to peanut allergens, and was therefore estimate the burden of aflatoxin-related the standard method used by FERG. hepatocellular carcinoma. Appendix 5 shows the computational In addition to categorical attribution disease model for Mycobacterium and counterfactual analysis, which can bovis, which is characteristic for the be considered top-down approaches, attributional models. In this model, FBD burden can also be estimated by a the overall incidence and mortality risk assessment approach, which can be of tuberculosis is multiplied with the considered a bottom-up approach. In this proportion attributable to M. bovis, approach, the incidences of the specific resulting in the incidence and mortality health states (e.g. impaired male fertility of M. bovis tuberculosis. Appendix 5 due to pre natal dioxin exposure) are shows the computational disease model estimated by combining exposure and for Echinococcus granulosus, which is dose-response data. The dose-response characteristic for the transitional models. model may for instance define the In this model, the overall incidence of probability of illness at a given exposure infection by this parasite was multiplied level, which can then be translated into with child nodes reflecting the probability an estimate of the number of incident of developing the concerned health or prevalent cases expected to occur states, resulting in the incidences of the in the exposed population [158, 191]. As specific health states. this approach does not involve burden When the hazard elevates the risk of a attribution, it does not necessarily ensure disease or disability outcome that occurs consistency with existing health statistics. in the population from other causes as However, risk assessment may be a valid well, causal attribution can only be made alternative when no burden envelopes statistically, and not on an individual exist or when it can be demonstrated basis. This is the case for many chemicals, that the estimated excess risk is additive including aflatoxin and dioxin. Aflatoxin to the background risk. In the context for instance may increase the risk of of FERG, risk assessment was used to hepatocellular carcinoma, but it is not estimate the burden of dioxin-related possible to specify that a specific liver hypothyroidy and impaired fertility. Hazard-specific methodology 56

4.7.2 CTF database template data such as incidence or mortality rates A database template was developed in [159]. These models estimate parameters Excel™ 7 to collect in a standardized way based on data of neighboring regions the data resulting from the systematic or other time periods. The external reviews. The structure of the database data used must thus be representative was based on the disease models, with of the selected population, region and one sheet per node. Three generic sheets time. The CTF developed, tested and were defined: (1) a “RATE” sheet, for rates evaluated several possible approaches by country; (2) a “PROB-local” sheet, to impute missing incidence data at the for proportions or ratios by country; country level [79]. This exercise identified and (3) a “PROB-global” sheet, for a several pitfalls in the use of explanatory single proportion or ratio that applied to covariates, such as the potential for all countries. overfitting and the arbitrariness in the selection of covariates. Therefore, Each sheet consisted of four tables for and further motivated by a strive for entering: (1) the rate or proportion/ratio parsimony and transparency, we decided data; (2) the age distribution; (3) the to use a log-Normal random effects sex distribution; and (4) if applicable, model as the default model for imputing the duration. Using a drop-down menu, missing country-level incidence data. different formats could be selected for We used the subregions as defined in entering the input parameters, including Appendix 2 as the random effect or a mean and 95% confidence interval; cluster variable. This model assumes minimum, most likely and maximum METHODOLOGY that the log-transformed incidence rate HAZARD SPECIFIC percentiles; the shape and rate of a in country belonging to subregion Gamma distribution (for rates); and the arises from a Normal distribution with shape parameters of a Beta distribution subregion specific mean and a within- (for proportions). Gamma and Beta region (= between-country) variance distributions were chosen because their domains correspond to that of . Each subregion specific mean rates and proportions, respectively, is in turn assumed to arise from a and because their parameters have an Normal distribution with mean and a intuitive interpretation (i.e., number of between-region variance : cases and sample size, respectively, number of positives and number of negatives). Likewise, different levels of stratification could be selected for the duration parameters (i.e. none, by age After fitting this hierarchical random only, by sex only, by age and sex). Age effects model to the available data, distribution, sex distribution and duration incidence values for countries with were allowed to vary by country, by no data were imputed based on defining different “groups” and assigning the resulting posterior predictive countries to “groups”. distributions. In other words, we 4.7.3 Imputation represented missing incidence data by log-Normal distributions based on the Extrapolation or imputation models may fitted mean and variance parameters. For be needed when literature searches countries in a subregion where none of cannot provide essential epidemiological the countries had data, the log-incidence was imputed as multiple random draws 7 Microsoft Corp., Redmond, Washington, USA WHO Estimates of the global burden of foodborne diseases 57

from a Normal distribution with mean while for other hazards we directly equal to the fitted global intercept drew on GBD 2010 estimates (Table and variance equal to the sum of the 5). For the remaining hazards, either fitted between-region variance and epidemiological data were used that did not need (further) correction, or the the fitted within-region variance (thus underreporting factor was included in the imputing the log-incidence as that of disease model (which was the case for a “random” country within a “random” Trichinella spp. and cyanide in cassava). subregion, with the uncertainty interval describing the variability between and For aflatoxin, the same random effects within subregions): model was used to extrapolate PAFs, but now using logit-transformed instead of log-transformed values. For countries in a subregion where at The model was implemented in a least one of the other countries had Bayesian framework, using independent data, the log-incidence was imputed as Normal(0, 1e5) priors for and ; multiple random draws from a Normal a Uniform(0, 10) prior for ; and a distribution with mean equal to the Folded-t(1) prior for , as suggested fitted region-specific intercept and by Gelman [160]. Sensitivity analyses variance equal to the fitted within-region using Gamma priors for the variance variance (thus imputing the log- parameters did not yield meaningful incidence as that of a “random” country differences. The model was run in JAGS within the concerned subregion, with [161] through the ‘rjags’ package in R the uncertainty interval describing the [162]. After a burn-in of 5000 iterations, variability within subregions): another 5000 iterations were retained for inference. Two chains were run, and convergence was ascertained When countries were considered free through density and trace plots, from exposure through the food chain, and the multivariate potential scale they were excluded from the imputation reduction factor (or Brooks-Gelman- model and thus did not contribute to Rubin diagnostic). the subregional estimates. This was the A crucial assumption made by this case for Brucella spp., as discussed in imputation model is that missing data [168], and foodborne trematodes and were considered “missing at random” Echinococcus spp., as discussed in [261]. (MAR), meaning that missingness was For countries with available incidence independent of the unobserved data, data, no imputation was performed. The given the observed data [163, 164]. In our incidence data used in the probabilistic case, this assumption implied that, within burden assessments were thus a each subregion, countries with data combination of actual data and imputed provided unbiased information on those estimates. No additional step had to be without data, and that, across subregions, included to correct incidence data for subregions with data provided unbiased potential underreporting, as this was information on those without data. already captured by the previous steps However, for five hazards (Bacillus cereus, of the framework. Indeed, for the hazards Clostridium perfringens, Clostridium that used an attributional model, disease botulinum, Staphylococcus aureus and envelopes were used that had already peanut allergens), only data from high- been corrected for underreporting, Hazard-specific methodology 58

income subregions, i.e. subregions A or B, Nations World Population Prospects could be retrieved. In those instances, the [166]. All estimates were generated per assumption of MAR was clearly violated, country, and subsequently aggregated and it was decided not to extrapolate per subregion, per region, and globally those data to the rest of the world. As a (Appendix 2). result, those hazards were excluded from The duration component of the YLDs is the global burden of disease estimates. defined as the average observed duration Table 5 shows which imputation strategy until remission or death. For calculating was used for each of the included YLDs when duration was lifelong, we hazards. For the four intoxications, therefore used the country-specific peanut allergens and cyanide in cassava, life expectancy (LE) [166] as duration. the default random effects model was The time component of the YLLs, on not used because of the limited number the other hand, is defined as the ideal of data points. Instead, the burden for residual life expectancy a person would each concerned country was imputed have if the world would be free from as draws from a Uniform distribution disease and provide maximal access defined by the lowest and highest to health care. We used the highest globally observed incidence or mortality projected LE for 2050 as normative LE rates. To ensure consistency with results for calculating YLLs [4]. This LE table has of the Child Health Epidemiology a LE at birth of 92, higher than that of the Reference Group (CHERG), alternative LE tables used in the GBD studies, which

imputation approaches were applied for were based on current death rates [1, 6]. METHODOLOGY estimating aetiological fractions for the Since even for the lowest observed death HAZARD SPECIFIC eleven diarrhoeal agents [40, 50, 168]. rates there are a proportion of deaths For seven other hazards, no imputation which are preventable or avertable, the had to be performed because data were highest projected LE for the year 2050 used that had already been imputed. was deemed to better represent the This was the case for hepatitis A virus, maximum life span of an individual in Salmonella Typhi, Salmonella Paratyphi, good health, while acknowledging that Ascaris spp. and Taenia solium, for it may still not represent the ultimate which GBD2010 data were used, and for achievable human life span [166]. Mycobacterium bovis and Trichinella spp., In line with current global burden of for which other published data were used disease assessments, no age weighting [84, 165]. or time discounting was applied [4, 6]. HIV infected invasive salmonellosis cases 4.7.4 Probabilistic burden assessment and deaths, and HIV infected M. bovis For each hazard, incidence, mortality, deaths, were excluded from the burden YLD, YLL and DALY rates were calculated estimates. No further corrections were for 11 age groups (<1; 1-4; 5-14; 15-24; made for possible co-morbidities. 25-34; 35-44; 45-54; 55-64; 65-74; 75-84; *85) and both sexes. When Parameter uncertainty was taken into necessary, age and sex specific rates account by performing the burden were obtained by multiplying the overall assessments in a probabilistic framework. rates with outcome specific age and sex Ten thousand Monte Carlo (parametric distributions. The reference year for the bootstrap) simulations of the input calculation of absolute numbers was parameters were generated to calculate 2010, with population estimates obtained 10,000 estimates of incidence, mortality, from the 2012 revision of the United YLD, YLL and DALY rates. These 10,000 WHO Estimates of the global burden of foodborne diseases 59

estimates were then summarized by The structured expert elicitation using their median and a 95% uncertainty Cooke’s Classical Method conducted to interval defined as the 2.5th and 97.5th attribute burden to different exposure percentile of the distribution of estimates. routes, providing hazards-specific Special care was taken to deal with estimates for each exposure route correlated uncertainties, for instance per subregion [156]. This process when the disease model included “global” yielded a probabilistic estimate of the probabilities (e.g., when it was assumed proportion foodborne, in the form of an that the probability of developing a empirical cumulative density function certain health state following infection from which random samples could was the same for each country). In such be drawn. Foodborne cases, deaths, cases, a vector of random probabilities YLDs, YLLs and DALYs were then was simulated only once and applied obtained by multiplying the vectors of to the different countries, instead of random values for these parameters incorrectly simulating a new, independent with a vector of random values for the vector of random probabilities for proportion foodborne. As before, the each country. perfect correlation of uncertainty was dealt with by simulating only one vector of random foodborne proportions per subregion, and by applying this vector to all parameters of all countries within the concerned subregion. Hazard-specific methodology 60 METHODOLOGY HAZARD SPECIFIC WHO Estimates of the global burden of foodborne diseases 61

RESULTS 5

WHO Estimates of the global burden of foodborne diseases 63

RESULTS

In this section, the results of the global interest (2). Of the 100 experts enrolled, expert elicitation study are reported first. 78 completed interviews with facilitators An overview of global and regional DALY and 73 returned their spreadsheets with estimates according to hazard follows. their responses to the target questions Subsequent sections report more specific and seed questions. The single main hazard-based estimates, and include reason for not completing the interview estimates for some hazards for which and returning the spreadsheet was time global estimates could not be derived constraints. All responses were reviewed and only regional estimates are reported (e.g. checked for missing estimates, that (peanut allergen; toxin-producing species sums across pathways were close to of bacteria). 100%, and that the 5th percentile < 50th percentile < 95th percentiles), and some 5.1 Attribution experts were contacted for clarification of the responses they had provided. One A total of 299 potential experts were expert was dropped after not responding asked by email of their interest in to requests for clarification. This resulted participating in the study. Of these154 in the responses of 72 experts being replied positively and they were included in the final dataset. Table 6 requested to forward their CV, a filled-in shows the distribution of experts across expert sheet and a signed declaration of panels, and Figure 3 shows distribution of interest. A total of 103 did that. the experts by their geographical areas Of these, 3 were not included due to lack of expertise. of experience (1) or possible conflicts of

Table 6. The number of experts enrolled, interviewed and finally included in the elicitation across panels

EXPERTS EXPERTS RETURNED HAZARD GROUPS ENROLLED INTERVIEWED ANSWERS

DIARRHEAL DISEASE

Bacterial (incl. S. Typhi) pathogens and norovirus Sub regionala 49 37 37

Intestinal protozoa Global 12 9 9

OTHER INFECTIOUS DISEASE

Brucella spp. Global 10 8 7

Hepatitis A virus Global 9 7 7

Toxoplasma gondii Global 11 10 9

Ascaris spp. Global 8 6 7

Echinococcus spp. Global 7 6 6

CHEMICALS

Lead Global 10 9 6

Totalb 100 78 72 a Due to the sub regional structure of these panels, the number of experts varied between 10 and 15 depending on the hazard and sub region. b As several experts served on more panels, the number of experts per panel does not add up to the total number of individual experts included. Results 64

Figure 3. Geographical distribution of experts according to working experience (>3 years) per subregion. Several experts had experience in more than one subregion.

3

18 23 6

5 5 14 10 6

11 5 5 14

5

AFR D EMR B SEAR B AFR E EMR D SEAR D 0 1000 2000 3000 km AMR A EUR A WPR A AMR B EUR B WPR B

AMR D EUR C Not applicable © WHO 2015. All rights reserved

Notes: The subregions are defined on the on the basis of child and adult mortality, as described by Ezzati et al. [5]. Stratum A = very low child and adult mortality; Stratum B = low child mortality and very low adult mortality; Stratum C = low child mortality and high adult mortality; Stratum D = high child and adult mortality; and Stratum E = high child mortality and very high adult mortality. The use of the term ‘subregion’ here and throughout the text does not identify an official grouping of WHO Member States, and the “subregions” are not related to the six official WHO regions.

5.1.1 Expert performance produced by applying the performance

In this study, there were 115 distinct weights. In contrast, the informativeness RESULTS panels (i.e. panels that differed in properties of the equal weights solutions membership or seed questions) and, were much lower than those provided overall, performance weight and equal by performance weights solutions weight combinations showed acceptable (Figure 4). This “trade-off” between statistical accuracy. Only in the case accuracy and informativeness when of the panel considering lead was the applying equal weights or performance p-value of the performance-based weights is often seen, because the combination small enough to cast doubt least accurate experts are typically the on the usual criterion for statistical most informative, and their narrow 90% accuracy, with p = 0.045 (i.e. less than confidence bands often have little or no the 0.05 criterion). With a set of 115 overlap. Moreover, the combined score panels, at least one score this low would using performance-based weights was be expected even if the performance- above that of the equal weights pooling based combination was always in 62% of the cases. It was therefore statistically accurate. decided to use the performance weights combinations for constructing the joint Results obtained by applying equal probability distributions for the pathway weights pooling and performance attribution estimates, as long as the weights pooling were compared. The statistical accuracy was acceptable. It equal weights solutions tended to have should also be noted that the weight higher statistical accuracy than those WHO Estimates of the global burden of foodborne diseases 65

attributed to an expert – comprising the large number of experts were assessed normalized product of their two scores using very similar variables, thereby – is dominated by the accuracy term, allowing their informativeness scores to so that high informativeness cannot be compared. An in-depth analysis of buy down poor accuracy. A unique the experts’ performance has also been feature of the present study is that a published [181].

Figure 4. Statistical accuracy versus informativeness of the experts included, when using equal weight (blue) or performance weight (red) combinations, respectively.

1,8

1,6

1,4

1,2

1,0

0,8

Informativeness 0,6

0,4

0,2

0,0 1,E-02 1,E-01 1,E+00 Statistical accuracy (p-value)

5.1.2 Pathway attribution results Figure 5 shows the subregional The collective results of the performance- estimates of the foodborne proportion based weighted expert responses are for Campylobacter spp., non-typhoidal shown in Appendix 7 (Table A7.1-3 for Salmonella spp., Shiga-toxin producing diarrhoeal disease, Table A7.4 for non- Escherichia coli (STEC), Brucella spp. and diarrhoeal , and Table Shigella spp. For Salmonella spp. and A7.5 for lead). For most estimates there Brucella spp., there is a clear pattern that is considerable uncertainty, reflecting: the foodborne proportion is considered (1) variations in uncertainty estimations more important in the developed between individual experts; (2) that, subregions (AMR A, EUR A and WPR A) for some hazards, the values provided compared with developing subregions. by experts having high performance Although less distinct, this pattern can weights in the analysis did not accord also be seen for Campylobacter spp. with one another; and (3) that, for some and STEC. For Campylobacter spp., subregions or hazards, the number of Salmonella spp. and STEC, the foodborne contributing experts was small (<7). Thus, transmission route was assessed by the the broad uncertainty intervals are most experts to be the most important route in likely reflecting current shortcomings in all subregions, followed by direct animal hard scientific evidence about the relative contact, human-to-human transmission contribution to human disease from each of and waterborne transmission in varying the transmission pathways. order, but generally with medians below 0.25 (Table A7.1 in Appendix 7). For Brucella spp., direct animal contact was considered Results 66

equally or more important than foodborne histolytica, norovirus, and hepatitis A transmission in developing subregions. virus. Overall, foodborne infections were Human-to-human transmission was not assessed by the experts to be the considered the most important route for more important routes in the majority of Shigella spp. in the majority of subregions. subregions. Exceptions were hepatitis A The proportion of foodborne Shigella infections, where foodborne and human- spp. infections ranged from 0.07 (95% UI to-human transmission were evaluated 0.00– 0.46) in EUR A to 0.36 (95% UI 0.01– equally important in most subregions, 0.70) in WPR A (Table A7.1 in Appendix 7). and S. Typhi, where foodborne and Overall, foodborne transmission was waterborne infections were assessed assessed to be more important in equally important in SEAR and WPR South-East Asian and Western Pacific regions (Table A7.3 in Appendix 7). Human- subregions than in other parts of the world. to-human transmission was identified as Transmission through soil or other routes the main exposure route for norovirus and was recognized by the experts to be of E. histolytica in most subregions, whereas minor importance for these five pathogens. waterborne transmission was estimated to be the main transmission route for Figure 6 shows the subregional estimates V. cholerae infections (Table A7.3 in of the proportion foodborne for entero- Appendix 7). pathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), Cryptosporidium spp. Figure 8 shows the subregional and Giardia spp. The estimates for EPEC estimates of the proportion foodborne are seen to follow the same pattern as for Toxoplasma gondii, Echinococcus described above, with the foodborne route multilocularis, Echinococcus granulosus being assessed to be more important and Ascaris spp. The foodborne route was in developed subregions. In developing assessed by the experts to be the most subregions in the African, American and important transmission route for T. gondii Eastern Mediterranean regions (AFR, AMR and Ascaris spp. in most subregions, and EMR), water was identified as the most but there was a clear tendency for soil RESULTS important transmission route. For ETEC, to increase in relative importance in less the estimated foodborne proportions were developed subregions (subregions D and quite similar for all subregions with medians E) (Table A7.4 in Appendix 7). Specifically ranging from 0.33 to 0.43 (Table A7.2 in for Ascaris spp., the foodborne route was Appendix 7), but the foodborne route was assessed to be particularly important in only assessed by experts to be the more developed subregions (A subregions). important route in European subregions. There was only little geographical variation For Cryptosporidium spp. and Giardia spp., between the median estimates for each the foodborne proportions were also quite of the transmission pathways for the two similar across subregions, but generally Echinococcus species. For E. granulosus, considered less important, with medians animal contact was clearly believed to be below 0.20 (Table A7.2 in Appendix the most important route, with medians 7). Human-to-human and waterborne just over 0.50. For E. multilocularis, the transmission were the more important foodborne route was considered most routes for these infections in all subregions. important, with medians ranging from 0.43 in EMR B to 0.58 in AFR D and E, Figure 7 shows the subregional estimates AMR B and D, and SEAR B and D, but of the proportion foodborne for Salmonella the estimates had very large uncertainty Typhi, Vibrio cholerae, Entamoeba intervals (Table A7.4 in Appendix 7). WHO Estimates of the global burden of foodborne diseases 67

Figure 5. Subregional estimates of the proportion of foodborne illnesses caused by Campylobacter spp., non-typhoidal Salmonella spp., Shiga-toxin producing Escherichia coli (STEC), Brucella spp. and Shigella spp. Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles.

Campylobacter spp. non-typhoidal S. enterica AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

STEC Brucella spp. AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Shigella spp. AFR D

AFR E

AMR A

AMR B

AMR D

EUR A

EUR B

EUR C

EMR B

EMR D

SEAR B

SEAR D

WPR A

WPR B

0.00 0.25 0.50 0.75 1.00 Proportion Results 68

Figure 6. Subregional estimates of the proportion of foodborne illnesses caused by enteropathogenic E. (EPEC), enterotoxigenic E. coli (ETEC), Cryptosporidium spp. and Giardia spp.

EPEC ETEC AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Cryptosporidium spp. Giardia spp. AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B RESULTS EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Notes: Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles. WHO Estimates of the global burden of foodborne diseases 69

Figure 7. Subregional estimates of the proportion of foodborne illnesses caused by typhoidal Salmonella, Vibrio cholerae, Entamoeba histolytica, norovirus, and hepatitis A virus.

Salmonella Typhi Vibrio cholerae AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Entamoeba histolytica Norovirus AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Hepatitis A virus AFR D

AFR E

AMR A

AMR B

AMR D

EUR A

EUR B

EUR C

EMR B

EMR D

SEAR B

SEAR D

WPR A

WPR B

0.00 0.25 0.50 0.75 1.00 Proportion

Notes: Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles. HAV = Hepatitis A virus. Results 70

Figure 8. Subregional estimates of the proportion of foodborne illnesses caused by Toxoplasma gondii, Echinococcus multilocularis, Echinococcus granulosus and Ascaris spp.

Toxoplasma gondii Ascaris spp. AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D

SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Echinococcus granulosus Echinococcus multilocularis AFR D AFR D

AFR E AFR E

AMR A AMR A

AMR B AMR B

AMR D AMR D

EUR A EUR A

EUR B EUR B

EUR C EUR C

EMR B EMR B

EMR D EMR D RESULTS SEAR B SEAR B

SEAR D SEAR D

WPR A WPR A

WPR B WPR B

0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Proportion Proportion

Notes: Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles. WHO Estimates of the global burden of foodborne diseases 71

Figure 9 shows the subregional estimates in two subregions in Europe. of the foodborne proportion for lead Air was assessed to be the main exposure exposure. Water, food and air exposure route in seven of the 14 subregions were the main transmission routes and water in four regions. Soil, paint, indicated by the experts with some cookware/pottery/glassware and toys subregional differences (Table A7.5 in were in comparison found to be of only Appendix 7). The foodborne route was minor importance in most subregions assessed to be the most important only (Figure 10).

Figure 9. Subregional estimates of the proportion of disease caused by foodborne exposure to lead.

Lead AFR D

AFR E

AMR A

AMR B

AMR D

EUR A

EUR B

EUR C

EMR B

EMR D

SEAR B

SEAR D

WPR A

WPR B

0.00 0.25 0.50 0.75 1.00 Proportion

Notes: Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles.

Figure 10. Subregional estimates (medians) of the proportion of disease caused by exposure to lead through eight different exposure routes.

exposure AFR D

AFR E food

AMR A water

AMR B soil

AMR D air

EMR B paint

EMR D cookware

EUR A toys

EUR B other EUR C

SEAR B

SEAR D

WPR A

WPR B

0.00 0.25 0.50 0.75 1.00

Proportion Results 72

5.2 DALY Estimates: Overview The global burden of FBD caused by the 31 hazards (including sequelae) Global disease burden in 2010 was 33 million DALYs (Table Of the approximately 600 million cases 7). Eighteen million DALYs, or 54%, of illness caused by the 31 foodborne of the total burden was attributed to hazards in 2010 (see Table 7), infectious diarrheal disease agents, particularly agents that cause diarrhoeal diseases to non-typhoidal S. enterica, which accounted for the vast majority (550 was responsible for 4.0 million DALYs million), in particular norovirus (120 (Figure 11). Six diarrhoeal disease agents million cases) and Campylobacter (norovirus, Campylobacter spp., EPEC, spp. (96 million cases). Among other ETEC, Vibrio cholerae and Shigella hazards, hepatitis A virus, the helminth spp.) each caused a foodborne burden Ascaris spp. and the typhoid bacterium of 1–3 million DALYs. Other foodborne Salmonella Typhi were frequent causes of hazards that contributed substantially to foodborne illness, causing 14, 12 and 7.6 the global burden included Salmonella million cases, respectively. Typhi (3.7 million DALYs), T. solium (2.8 million DALYs), hepatitis A virus Foodborne diarrhoeal disease agents (1.4 million DALYs) and Paragonimus also caused 230,000 of the 420,000 spp. (1.0 million DALYs). By contrast, deaths due to foodborne hazards (Table the global burden of trichinellosis, due 7). Of these, non-typhoidal S. enterica to the widespread nematode parasite accounted for 59,000, enteropathogenic Trichinella, was estimated at only 550 E. coli (EPEC) for 37,000, norovirus DALYs. For full details of results including for 35,000, and enterotoxigenic E. coli foodborne illnesses, deaths, DALYs, YLLs (ETEC) for 26,000 deaths. Of the 59,000 and YLDs for all 31 hazards in this study, global deaths due to non-typhoidal see the Supplementary Information S. enterica, 32,000 were in the two for the overview publication [167]. The African subregions, and included 22,000

Supplementary Information also includes RESULTS deaths due to invasive disease by this the results for total illnesses, deaths, bacterium. The major non-diarrhoeal DALYs, YLLs and YLDs by all exposure causes of foodborne deaths were due to pathways, for all hazards that were Salmonella Typhi (52,000), the helminth included in the source attribution expert Taenia solium (28,000) and hepatitis A elicitation in Appendix 7. virus (28,000) and aflatoxin with 20,000 (95% UI 8,000-51,000) deaths. WHO Estimates of the global burden of foodborne diseases 73

Table 7. Median global number of foodborne illnesses, deaths, Years Lived with Disability (YLDs), Years of Life Lost (YLLs) and Disability Adjusted Life Years (DALYs), with 95% uncertainty intervals, 2010.

FOODBORNE FOODBORNE FOODBORNE FOODBORNE FOODBORNE HAZARD ILLNESSES DEATHS YLDS YLLS DALYS 600 652 361 5 580 028 27 201 701 32 841 428 418 608 TOTAL (417 646 804– (4 780 374– (19 655 451– (24 809 085– (305 128–598 419) 962 834 044) 8 195 314) 38 922 210) 46 274 735)

548 595 679 839 463 16 821 418 17 659 226 Diarrhoeal disease 230 111 (369 976 912– (644 924– (11 700 916– (12 458 675– agents (160 039–322 359) 888 528 014) 1 123 907) 23 579 652) 24 516 338)

124 803 946 2 403 107 2 496 078 34 929 91 357 Viruses (70 311 254– (1 102 397– (1 175 658– (15 916–79 620) (51 047–174 130) 251 352 877) 5 387 672) 5 511 092)

124 803 946 2 403 107 2 496 078 34 929 91 357 Norovirus (70 311 254– (1 102 397– (1 175 658– (15 916–79 620) (51 047–174 130) 251 352 877) 5 387 672) 5 511 092)

349 405 380 13 795 606 14 490 808 187 285 685 212 Bacteria (223 127 469– (9 688 221– (10 303 551– (131 742–254 037) (521 848–921 335) 590 002 559) 18 893 580) 19 681 271)

95 613 970 1 689 291 2 141 926 21 374 442 075 Campylobacter spp. (51 731 379– (1 141 055– (1 535 985– (14 604–32 584) (322 192–587 072) 177 239 714) 2 652 483) 3 137 980)

23 797 284 2 908 551 2 938 407 Enteropathogenic E. 37 077 22 977 (10 750 919– (1 574 520– (1 587 757– coli– EPEC (19 957–61 262) (9 662–66 211) 62 931 604) 4 833 325) 4 865 590)

86 502 735 2 011 635 2 084 229 Enterotoxigenic E. coli– 26 170 70 567 (49 136 952– (1 132 331– (1 190 704– ETEC (14 887–43 523) (40 134–119 017) 151 776 173) 3 407 273) 3 494 201)

1 176 854 Shiga toxin-producing 128 3 486 9 454 12 953 (754 108– E. coli– STEC (55–374) (1 741–6 996) (4 140–27 208) (5 951–33 664) 2 523 007)

78 707 591 3 976 386 4 067 929 Non-typhoidal S. 59 153 78 306 (31 843 647– (2 410 953– (2 486 092– enterica (36 341–89 045) (35 961–185 179) 211 154 682) 6 180 921) 6 271 290)

51 014 050 1 181 231 1 237 103 15 156 51 613 Shigella spp. (20 405 214– (519 372– (554 204– (6 839–30 072) (21 184–114 267) 118 927 631) 2 445 834) 2 520 126)

763 451 1 719 381 1 722 312 24 649 2 721 Vibrio cholerae (310 910– (718 642– (720 029– (10 304–50 042) (1 019–6 020) 1 567 682) 3 487 195) 3 491 997)

67 182 645 492 354 5 558 57 536 432 316 Protozoa (35 794 977– (239 400– (2 593–11 958) (30 526–102 608) (195 372–960 910) 120 556 797) 1 034 790)

8 584 805 3 759 8 155 287 690 296 156 Cryptosporidium spp. (3 897 252– (1 520–9 115) (3 598–17 355) (114 012–711 990) (119 456–724 660) 18 531 196)

28 023 571 1 470 20 851 115 740 138 863 Entamoeba histolytica (10 261 254– (453–5 554) (7 431–53 080) (32 070–476 144) (47 339–503 775) 68 567 590)

28 236 123 0 26 270 0 26 270 Giardia spp. (12 945 655– (0–0) (11 462–53 577) (0–0) (11 462–53 577) 56 996 454) Results 74

FOODBORNE FOODBORNE FOODBORNE FOODBORNE FOODBORNE HAZARD ILLNESSES DEATHS YLDS YLLS DALYS

35 770 163 1 098 675 6 960 656 8 065 581 Invasive infectious 117 223 (18 604 754– (729 530– (3 128 316– (3 983 949– disease agents (54 789–243 482) 70 045 873) 1 796 607) 14 882 637) 16 557 714)

13 709 836 1 258 812 1 353 767 27 731 85 885 Viruses (3 630 847– (325 409– (383 684– (7 169–77 320) (22 118–250 641) 38 524 946) 3 509 844) 3 672 726)

13 709 836 1 258 812 1 353 767 27 731 85 885 Hepatitis A virus (3 630 847– (325 409– (383 684– (7 169–77 320) (22 118–250 641) 38 524 946) 3 509 844) 3 672 726)

10 342 042 5 472 374 5 697 913 85 269 225 792 Bacteria (3 506 116– (2 283 968– (2 394 245– (37 573–196 544) (108 092–604 162) 27 627 480) 12 803 285) 13 384 811)

393 239 110 971 1 957 13 324 124 884 Brucella spp. (143 815– (37 470– (661–45 545) (4 095–315 952) (43 153–2 910 416) 9 099 394) 2 583 081)

14 169 3 175 2 255 116 109 118 340 Listeria monocytogenes (6 112–91 175) (1 339–20 428) (843–14 981) (48 693–740 357) (49 634–754 680)

121 268 10 545 50 733 556 998 607 775 Mycobacterium bovis (99 852–150 239) (7 894–14 472) (38 441–68 052) (417 711–761 851) (458 364–826 115)

1 741 120 829 136 855 730 12 069 26 987 Salmonella Paratyphi A (536 650– (259 990– (268 879– (3 784–29 521) (7 610–72 811) 4 310 983) 2 028 112) 2 100 120) 7 570 087 3 604 940 3 720 565 52 472 117 334 Salmonella Typhi (2 333 263– (1 130 390– (1 169 040– (16 454–128 350) (33 086–316 571) 18 743 406) 8 817 876) 9 130 956) 10 280 089 829 071 684 763 326 62 899 Protozoa (7 403 516– (561 297– (333–1 300) (511 314–1 175 619) (30 575–119 512) 14 904 324) 1 264 567) 10 280 089 829 071 684 763 326 62 899 Toxoplasma gondii (7 403 516– (561 297– RESULTS (333–1 300) (511 314–1 175 619) (30 575–119 512) 14 904 324) 1 264 567) 12 928 944 3 367 987 2 428 929 5 810 589 45 226 Helminths (8 957 617– (2 840 638– (1 869 610– (4 864 518– (34 143–59 035) 24 008 256) 4 358 741) 3 173 545) 7 367 619)

430 864 1 220 578 1 932 154 3 158 826 36 500 Cestodes (334 389– (941 084– (1 387 290– (2 411 585– (25 652–50 063) 774 703) 1 576 600) 2 664 120) 4 122 032)

Echinococcus 43 076 482 12 121 27 626 39 950 granulosus (25 881–371 177) (150–3 974) (5 515–99 213) (8 577–227 715) (16 996–322 953)

Echinococcus 8 375 7 771 8 749 303 039 312 461 multilocularis (656–17 005) (243–15 896) (856–22 576) (8 102–622 954) (9 083–640 716)

1 192 236 1 586 288 2 788 426 370 710 28 114 Taenia solium (916 049– (1 170 461– (2 137 613– (282 937–478 123) (21 059–36 915) 1 522 267) 2 177 848) 3 606 582) 12 285 286 1 012 518 451 80 021 605 738 Nematodes (8 292 732– (388–2 783) (351 732–1 211 907) (30 652–220 274) (411 113–1 301 619) 22 984 630) 12 280 767 1 008 518 096 79 800 605 278 Ascaris spp. (8 287 414– (384–2 781) (351 418–1 211 691) (30 426–220 154) (410 668–1 301 114) 22 980 491)

4 472 4 342 210 550 Trichinella spp. (2 977–5 997) (2–5) (149–646) (116–306) (285–934) WHO Estimates of the global burden of foodborne diseases 75

FOODBORNE FOODBORNE FOODBORNE FOODBORNE FOODBORNE HAZARD ILLNESSES DEATHS YLDS YLLS DALYS 1 616 785 2 024 592 218 569 7 533 403 884 Trematodes (1 257 657– (1 652 243– (167 886–281 872) (6 383–8 845) (342 815–473 423) 2 062 782) 2 483 514) 302 160 31 620 5 770 219 637 522 863 Clonorchis sinensis (247 586– (21 515–45 059) (4 728–6 988) (149 514–312 718) (431 520–635 232) 366 036)

10 635 0 90 041 0 90 041 Fasciola spp. (6 888–24 100) (0–0) (58 050–209 097) (0–0) (58 050–209 097)

18 924 0 155 165 0 155 165 Intestinal flukes* (14 498–24 200) (0–0) (118 920–198 147) (0–0) (118 920–198 147)

16 315 1 498 102 705 85 364 188 346 Opisthorchis spp. (11 273–22 860) (1 230–1 813) (70 849–143 938) (70 123–103 317) (151 906–235 431)

1 033 097 1 048 937 139 238 250 15 535 Paragonimus spp. (730 118– (743 700– (95 610–195 078) (160–371) (9 971–23 035) 1 423 031) 1 438 588) 217 632 247 920 650 157 908 356 19 712 Chemicals and toxins (172 024– (196 490– (283 769– (506 112– (8 171–51 664) 1 140 463) 1 410 260) 1 617 168) 2 714 588) 632 901 636 869 21 757 19 455 3 945 Aflatoxin (265 578– (267 142– (8 967–56 776) (7 954–51 324) (1 551–10 667) 1 606 493) 1 617 081)

1 066 227 2 521 15 694 18 203 Cassava cyanide (105–3 016) (22–669) (249–7 142) (1 514–46 304) (1 769–53 170)

193 447 240 056 240 056 0 0 Dioxin (155 963– (192 608– (192 608– (0–0) (0–0) 1 085 675) 1 399 562) 1 399 562) Notes: * Includes selected species of the families Echinostomatidae, Fasciolidae, Gymnophallidae, Heterophyidae, Nanophyetidae, Neodiplostomidae and Plagiorchiidae (depending on data availability). Results 76

Figure 11. Ranking of foodborne hazards, based on Disability-Adjusted Life Years at the global level, with 95% uncertainty intervals, 2010.

10,000,000

1,000,000

100,000

10,000

1,000 Foodborne Disability-Adjusted Life Years

100

spp. spp. spp. spp. spp. spp. spp. spp. spp. spp. spp. STEC ETEC EPECTyphi Dioxins Aflatoxin Norovirus S. enterica Paratyphi A Giardia Ascaris Shigella Fasciola Brucella Taenia solium Trichinella Intestinal flukes Clonorchis Vibrio cholerae Cassava cyanide Hepatitis A Virus Opisthorchis Paragonimus Salmonella Toxoplasma gondii Campylobacter Cryptosporidium Mycobacterium bovis Entamoeba histolytica Listeria monocytogenes Salmonella Echinococcus granulosus non-typhoidal Echinococcus multilocularis

Notes: White dots indicate the median burden, black boxes the inter-quartile range (50% UI), black lines the 5 and 95 percentiles (90% UI) and grey lines the 2.5 and 97.5 percentiles (95% UI). Note that the y-axis is on a logarithmic scale. Abbreviations: EPEC = Enteropathogenic Escherichia coli; ETEC = Enterotoxigenic E. coli; STEC = Shiga toxin-producing E. coli. RESULTS Regional differences Australia, New Zealand and Japan), which The studies found considerable regional were all in the range of 40-50 DALYs per differences in the burden of FBD (Table 100,000 population. Other subregions 8 and Figure 12). The highest burden per (AMR B and AMR D, EMR B and WPR 100,000 population was observed in the B) had intermediate burdens, all in the two African subregions: 1,300 DALYs per range of 140-360 DALYs per 100,000 100,000 population in AFR D and 1,200 population (see Table 2 for a full list of DALYs per 100,000 population in AFR the countries in each subregion). E. In the South-East Asian subregions, The contribution of individual hazards SEAR B and SEAR D, the burden to the burden of FBD differed markedly was 690 and 710 DALYs per 100,000 between subregions (Figure 12). In population, respectively, and in the both African subregions, nearly 70% Eastern Mediterranean subregion, EMR of the burden was due to diarrhoeal D, 570 DALYs per 100,000 population. disease agents, particularly to non- The lowest burden was observed in the typhoidal S. enterica (including invasive North American subregion AMR A (35 salmonellosis), EPEC, and ETEC; DALYs per 100,000 population), followed additionally, V. cholerae caused an by the three European subregions EUR important burden of diarrhoeal disease A, EUR B and EUR C, and the Western in the AFR E subregion, and T. solium Pacific subregion WPR A (which includes caused a high burden in both African WHO Estimates of the global burden of foodborne diseases 77

subregions (see Table 8 for the detailed burden. In the AMR B region, important data for all hazards and all subregions). causes of FBD burden were T. solium In the SEAR D and SEAR B subregions, (25 DALYs per 100,000 population) diarrhoeal disease agents contributed and T. gondii (20 DALYs per 100,000 approximately 50% of the total disease population). In the AMR D region, the burden, mainly caused by a range of burden of T. solium was particularly high hazards including EPEC, norovirus, at 69 DALYs per 100,000 population; non-typhoidal S. enterica, ETEC and the trematodes Paragonimus spp. Campylobacter spp. In both of these and Fasciola spp. contributing 53 and subregions, there was also a considerable 46 DALYs per 100,000 population, burden of Salmonella Typhi (180 DALYs respectively to the overall disease per 100,000 population in SEAR B and burden. 110 DALYs per 100,000 population in The burden due to chemical hazards was SEAR D). The burden of disease due to also highly localized. Aflatoxin caused the fluke Opisthorchis spp. was almost the highest burden in AFR D, WPR B exclusively concentrated in SEAR B and SEAR B, whereas dioxins caused the (40 DALYs per 100,000 population). In highest burden in SEAR D, EMR B and EMR D, diarrhoeal disease agents were D and EUR A. The burden of cassava responsible for approximately 70% of the cyanide was limited to the AFR regions, total burden of FBD, with Campylobacter and was similar to that of aflatoxin in spp. the leading cause in the region, AFR D. followed by EPEC, non-typhoidal S. enterica, Shigella spp. and ETEC. Other In the three European subregions, important hazards in this region were diarrhoeal disease agents contributed Salmonella Typhi, aflatoxin and hepatitis to 49-68% of the total burden of FBD, A virus. with non-typhoidal S. enterica and Campylobacter spp. being the most In the WPR B subregion, diarrhoeal important hazards. Other important disease agents accounted for hazards included T. gondii in all European approximately 14% of the FBD burden, subregions, Brucella spp. in the EUR B with Campylobacter spp. the leading and Mycobacterium bovis in the EUR C cause. In this region, the seafoodborne subregions. In the WPR A region, 65% trematodes Paragonimus spp. and of the burden was caused by diarrhoeal Clonorchis sinensis were important disease agents, with T. gondii and contributors to the FBD burden. In hepatitis A virus also contributing. Finally, the AMR B and AMR D subregions, in the AMR A region, diarrhoeal disease the contribution of diarrhoeal disease agents contributed approximately 67% agents to the total burden was smaller of the total burden, with non-typhoidal than in other subregions (approximately S. enterica and Campylobacter spp. 40% and 20%, respectively), with the most important hazards; T. gondii Campylobacter spp., norovirus and and L. monocytogenes were also non-typhoidal S. enterica causing most relatively important. Results 78

Table 8. Median rates of foodborne Disability Adjusted Life Years (DALYs) per 100 000 population, by subregion, with 95% uncertainty intervals, 2010.

AFR AFR AMR AMR AMR EMR EMR EUR EUR EUR SEAR SEAR WPR WPR HAZARD D E A B D B D A B C B D A B 1 276 1 179 140 315 362 571 52 685 711 36 293 35 41 49 TOTAL (459– (726– (97– (243– (205– (325– (33– (360– (343– (23– (219– (23–49) (29–64) (33–77) 2263) 1764) 274) 575) 582) 954) 136) 1291) 1335) 170) 406) Diarrhoeal 889 824 277 393 330 380 23 60 72 28 25 24 23 41 disease (196– (447– (153– (217– (154– (159– (13–33) (36–94) (40–117) (17–39) (14–37) (13–35) (14–32) (21–65) agents 1 731) 1 326) 460) 644) 576) 717) 28 33 3 69 3 75 76 3 12 13 4 3 55 4 Viruses (0.5– (0.4– (0.08– (0.8– (0.09– (6–222) (0–225) (0.6–8) (0–38) (0–43) (0–11) (0.2–9) (0–224) (0–19) 79) 90) 9) 263) 7) 28 33 3 69 3 75 76 3 12 13 4 3 55 4 Norovirus (0.5– (0.4– (0.08– (0.8– (0.09– (6–222) (0–225) (0.6–8) (0–38) (0–43) (0–11) (0.2–9) (0–224) (0–19) 79) 90) 9) 263) 7) 787 712 237 347 247 285 19 45 54 24 21 20 19 34 Bacteria (186– (393– (124– (190– (107– (119– (10–28) (26–68) (28–87) (14–32) (11–32) (10–29) (11–27) (17–54) 1 482) 1 160) 403) 576) 429) 506) 75 97 Campyloba- 71 70 9 15 15 10 8 8 37 33 9 10 (40– (54– cter spp. (35–119) (33–117) (5–14) (8–23) (8–26) (6–14) (4–13) (4–12) (14–87) (2–84) (5–14) (4–17) 109) 143) Enteropatho- 0.006 9 0.006 0.006 136 138 7 46 60 0.004 0.004 64 65 5 genic E. coli– (0.001– (0.9– (0.001– (0.001– (11–329) (6–327) (0.4–21) (9–114) (8–151) (0–0.01)(0–0.01) (3–144) (1–162) (0.3–13) EPEC 0.02) 26) 0.02) 0.02) Enterotoxi- 107 105 0.003 7 9 29 37 0.004 0.004 0.004 42 42 0.003 4 genic E. coli– (26– (17– (0– (1–19) (2–25) (6–78) (5–102) (0–0.01)(0–0.01)(0–0.01) (3–106) (2–111) (0–0.01) (0.3–11) ETEC 245) 240) 0.009) Shiga toxin- 0.009 0.08 0.2 0.2 0.2 0.07 0.1 0.2 0.01 0.4 0.5 0.6 0.2 0.4 producing (0.002– (0.02– (0.05– (0.09– (0.1– (0.02– (0.03– (0.007– (0.002– (0.09–1) (0.2–1) (0.2–1) (0.02–1) (0.1–1) E. coli 0.03) 0.2) 0.3) 0.4) 0.5) 0.2) 0.4) 1) 0.04) Non- 338 193 59 9 11 14 50 67 12 12 11 58 9 9 typhoidal (94– (44– (22– (4–16) (2–20) (4–26) (17–82) (26–112) (7–18) (6–21) (5–19) (0–162) (5–14) (4–16) S. enterica 612) 336) 154) RESULTS 25 25 37 37 0.2 2 2 27 37 0.09 0.1 0.2 0.2 4 Shigella spp. (0.6– (0.7– (0–156) (0–148) (0–1) (0–8) (0–9) (0–109) (0–145) (0–0.9) (0–1) (0–1) (0–1) (0.1–11) 84) 90) 0.2 28 36 0.1 Vibrio 70 143 0 0 0 0 0 0 2 0 (0.009– (0.9– (0.2– (0.005– cholerae (2–197) (4–383) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0.2–3) (0–0) 0.5) 96) 133) 0.4) 6 20 21 0.2 2 3 7 0.2 0.2 0.2 10 10 0.2 1 Protozoa (0.9– (0–74) (5–66) (0.01–1) (0.4–6) (0.5–8) (1–33) (0–0.8) (0–0.9) (0–0.9) (2–35) (2–39) (0–0.9) (0.2–4) 27) Crypto- 0.2 0.7 12 12 1 3 3 0.1 0.1 0.1 6 6 0.1 0.3 sporidium (0.01– (0.08– (0–44) (0–45) (0.1–5) (0–21) (0–24) (0–0.8) (0–0.8) (0–0.8) (0–28) (0–32) (0–0.8) (0–3) spp. 0.9) 3) Entamoeba 5 5 0 0.4 0.4 2 2 0 0 0 2 2 0 0.3 histolytica (0–41) (0–41) (0–0) (0–2) (0–3) (0–12) (0–17) (0–0) (0–0) (0–0) (0–17) (0–18) (0–0) (0–1) 0.6 0.7 0.7 0.03 0.4 0.5 0.4 0.03 0.03 0.03 0.1 0.1 0.03 0.3 Giardia spp. (0.005– (0–3) (0–3) (0–0.1) (0–2) (0.01–3) (0–2) (0–0.1) (0–0.1) (0–0.1) (0–0.9) (0–1) (0–0.1) (0–1) 2) WHO Estimates of the global burden of foodborne diseases 79

AFR AFR AMR AMR AMR EMR EMR EUR EUR EUR SEAR SEAR WPR WPR HAZARD D E A B D B D A B C B D A B Invasive 146 147 73 137 244 infectious 10 38 49 10 19 16 272 10 65 (46– (55– (32– (38– (38– disease (6–14) (16–76) (19–144) (7–15) (9–61) (10–29) (71–721) (5–132) (19–145) 342) 343) 148) 334) 623) agents 0.5 0.8 1 27 18 1 2 2 32 1 1 5 58 5 Viruses (0.07– (0.03– (0.07– (4–77) (3–55) (0.1–4) (0.2–7) (0.2–5) (2–102) (0.2–3) (0.3–4) (0.6–15) (6–182) (0.3–17) 2) 2) 3) 0.5 0.8 1 Hepatitis A 27 18 1 2 2 32 1 1 5 58 5 (0.07– (0.03– (0.07– virus (4–77) (3–55) (0.1–4) (0.2–7) (0.2–5) (2–102) (0.2–3) (0.3–4) (0.6–15) (6–182) (0.3–17) 2) 2) 3) 93 104 82 251 165 4 16 19 50 3 8 5 3 50 Bacteria (31– (40– (22– (59– (27– (2–7) (4–47) (5–65) (16–121) (3–5) (3–39) (3–10) (1–126) (12–124) 259) 277) 241) 696) 490) 2 0.3 0.07 2 4 0.8 0.8 0.7 0.6 0.6 1 23 0.3 4 Brucella spp. (0.2– (0.007– (0.02– (0.2– (0.6– (0.07– (0.004– (0.003– (0.02– (0.09– (0.3–16) (3–83) (0.07–1) (0.7–35) 53) 18) 0.6) 38) 68) 6) 112) 92) 125) 9) Listeria 0.3 1 1 3 2 1 1 1 3 0.6 1 1 1 1 mono- (0.2– (0–21) (0–21) (2–5) (0.2–17) (0–21) (0–21) (0–21) (2–4) (0.3–2) (0–21) (0–21) (0.7–2) (1–4) cytogenes 0.8) Myco- 0.03 0.4 0.08 0.1 25 34 2 1 13 0.6 3 11 14 3 bacterium (0.01– (0.2– (0.06– (0.08– (15–39) (21–48) (0.8–4) (0.5–3) (6–25) (0.5–1) (2–5) (4–27) (6–27) (1–5) bovis 0.06) 0.8) 0.1) 0.2) 2 0.01 26 Salmonella 11 12 0.1 2 3 10 0.02 0.3 42 0.1 8 (0.006– (0– (0.6– Paratyphi A (0–39) (0–43) (0–0.4) (0–6) (0–12) (0–36) (0–0.1) (0–2) (7–120) (0–0.5) (1–22) 7) 0.06) 80) 8 184 Salmonella 47 52 0.4 7 14 45 0.09 2 0.04 113 0.6 36 (0.03– (32– Typhi (0–169) (0–187) (0–2) (0–27) (0–51) (0–158) (0–0.6) (0–9) (0–0.3) (3–347) (0–2) (6–95) 29) 522) 21 20 5 20 27 20 18 6 10 10 13 9 5 9 Protozoa (8–41) (9–37) (2–8) (9–33) (10–84) (10–35) (9–31) (3–9) (5–23) (5–18) (6–22) (2–19) (3–8) (4–14) Toxoplasma 21 20 5 20 27 20 18 6 10 10 13 9 5 9 gondii (8–41) (9–37) (2–8) (9–33) (10–84) (10–35) (9–31) (3–9) (5–23) (5–18) (6–22) (2–19) (3–8) (4–14) 186 184 36 185 162 1 5 21 0.4 6 6 52 60 2 Helminths (125– (141– (27– (149– (131– (0.9–4) (2–15) (12–40) (0.2–1) (3–27) (4–15) (42–64) (45–80) (1–3) 308) 240) 134) 229) 202) 172 178 0.4 0.2 0.03 25 71 1 0.7 4 4 3 46 45 Cestodes (112– (136– (0.3– (0.05– (0.007– (19–34) (53–95) (0.2–10) (0.1–19) (2–25) (2–12) (2–5) (34–61) (25–65) 289) 235) 0.6) 0.5) 0.8) 0.4 0.01 0.3 0.1 0.02 0.3 Echino 0.8 2 0.9 0.6 2 0.5 0.001 0.8 (0.06– (0.002– (0.02– (0.02– (0.001– (0.08– granulosus (0.2–16) (0.4–8) (0.2–10) (0.1–19) (0.5–6) (0.09–1) (0–0.1) (0.2–3) 21) 0.03) 5) 0.4) 0.8) 0.9) 0.03 0.005 0.03 0.007 0.008 Echino coccus 0 0 0 0 0 2 2 0 18 (0.005– (0– (0.008– (0– (0.001– multilocularis (0–0) (0–0) (0–0) (0–0) (0–0) (0.5–21) (0.5–11) (0–0) (0–37) 0.06) 0.05) 0.06) 0.04) 0.02) 170 176 0.4 25 69 0 0 0 0 0.9 3 45 0 27 Taenia solium (110– (134– (0.3– (19–32) (51–91) (0–0) (0–0) (0–0) (0–0) (0.6–2) (2–5) (33–60) (0–0) (20–35) 283) 229) 0.6) 0.6 0.04 0.004 13 5 11 12 3 13 1 1 8 13 11 Nematodes (0.3– (0.02– (0.001– (2–28) (1–11) (3–106) (3–24) (1–7) (4–20) (0.3–2) (0.4–2) (2–15) (4–26) (3–22) 0.9) 0.07) 0.007) 0.6 13 5 11 12 3 13 0 1 1 8 13 0 11 Ascaris spp. (0.3– (2–28) (1–11) (3–106) (3–24) (1–7) (4–20) (0–0) (0.3–2) (0.3–2) (2–15) (4–26) (0–0) (3–22) 0.9) 0.001 0.001 0.009 0.009 0.009 0.04 0.04 0.04 0 0 0.004 0.004 Trichinella 0 0 (0– (0– (0.005– (0.005– (0.005– (0.02– (0.02– (0.02– (0– (0– (0.001– (0.001– spp. (0–0) (0–0) 0.002) 0.002) 0.01) 0.01) 0.01) 0.07) 0.07) 0.07) 0.001) 0.001) 0.007) 0.007) Results 80

AFR AFR AMR AMR AMR EMR EMR EUR EUR EUR SEAR SEAR WPR WPR HAZARD D E A B D B D A B C B D A B 0.06 0.02 0.2 0.1 0.3 0.2 0.2 101 7 1 40 0.7 2 106 Trema todes (0.02– (0.008– (0.04– (0.04– (0.2– (0.05– (0.05– (74–135) (4–10) (0.8–1) (32–50) (0.2–2) (1–2) (85–131) 0.2) 0.07) 3) 0.5) 0.5) 0.6) 0.6) 0.04 0.01 0.04 0.05 Clonorchis 0 0 0 0 0 0 0 0 0 31 (0.03– (0.003– (0.01– (0.01– sinensis (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (26–38) 0.04) 0.04) 0.2) 0.2) 0.02 0.01 0.04 0.04 0.2 0.07 0.06 0.04 0.02 0.05 0.07 46 7 0.9 Fasciola spp. (0.008– (0.005– (0.001– (0.02– (0.1– (0.02– (0.02– (0.01– (0.008– (0.02– (0.01– (27–75) (4–10) (0.1–8) 0.07) 0.04) 2) 0.1) 0.3) 0.2) 0.2) 0.1) 0.05) 0.1) 0.4) 0.01 0.1 0.06 0.06 0.08 0.03 0.05 0.09 0.2 0.1 Intestinal 0 0 1 9 (0.005– (0.04– (0.02– (0.02– (0.03– (0.009– (0.02– (0.03– (0.1– (0.03– flukes* (0–0) (0–0) (0.9–2) (7–11) 0.04) 0.5) 0.2) 0.2) 0.2) 0.09) 0.2) 0.2) 0.5) 0.4) 0.07 0.05 Opisth orchis 0 0 0 0 0 0 0 0.9 40 0.4 0 3 (0.02– (0.01– spp. (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0.6–1) (32–50) (0.1–1) (0–0) (2–4) 0.3) 0.3) 0.03 0.008 0.04 0.04 0.02 0.03 0.05 0.06 0.05 Paragon imus 53 0 0 0 60 (0.008– (0.002– (0.004– (0.01– (0.008– (0.01– (0.008– (0.02– (0.02– spp. (38–73) (0–0) (0–0) (0–0) (43–83) 0.08) 0.02) 0.6) 0.1) 0.07) 0.1) 0.5) 0.2) 0.2) 2 0.9 0.3 Chemicals 30 7 0.4 3 0.8 9 2 2 20 18 18 (0.09– (0.4– (0.06– and toxins (8–85) (3–21) (0.2–3) (0.7–16) (0.3–14) (4–66) (1–22) (2–9) (4–75) (13–52) (3–71) 159) 25) 13) 0.04 2 0.3 0.2 28 3 3 0.7 5 0.6 0.5 18 4 17 Aflatoxin (0.006– (0.07– (0.1– (0.04– (7–78) (1–8) (0.6–9) (0.2–3) (1–17) (0.3–1) (0.2–2) (3–52) (0.6–15) (3–69) 0.2) 137) 0.7) 0.8) Cassava 1 3 0 0 0 0 0 0 0 0 0 0 0 0 cyanide (0.1–3) (0.3–9) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) 0.2 0.2 0.1 0.2 0.09 0.3 0.2 0.1 0.06 0.3 3 2 2 14 Dioxin (0.05– (0.09– (0.03– (0.01– (0.004– (0.09– (0.005– (0.02– (0.006– (0.1–3) (2–56) (1–22) (1–8) (12–40) 6) 9) 11) 23) 11) 24) 45) 12) 5)

Notes: * Includes selected species of the families Echinostomatidae, Fasciolidae, Gymnophallidae, Heterophyidae, Nanophyetidae, Neodiplostomidae and Plagiorchiidae (depending on data availability). RESULTS Figure 12. The global burden of foodborne disease (DALYS per 100 000 population) by hazard groups and by subregion, 2010.

1200

800

400 Foodborne DALYs per 100,000 inhabitants

0 AFR D AFR E AMR A AMR B AMR D EMR B EMR D EUR A EUR B EUR C SEAR B SEAR D WPR A WPR B

Diarrheal disease agents Helminths Invasive infectious disease agents Chemicals and toxins WHO Estimates of the global burden of foodborne diseases 81

The relative contribution of mortality and V. cholerae, Listeria monocytogenes, (measured as YLL) and morbidity Salmonella Typhi and Salmonella (measured as YLD) to the total burden of Paratyphi, Echinococcus multilocularis disease varied widely between hazards and aflatoxin). At the other extreme, (Figure 13). For 18 foodborne hazards, more than 75% of the total burden due more than 75% of the total burden was to morbidity (blue columns in Figure 2) due to premature mortality (red columns were accounted for by seven foodborne in Figure 13). These mainly include hazards, of which four (Giardia spp., hazards leading to diseases with known Fasciola spp., intestinal flukes, and dioxin) high case-fatality ratios (non-typhoidal were not assumed to cause fatal illnesses. S. enterica, EPEC, ETEC, Shigella spp.

Figure 13. Relative contribution of Years of Life Lost due to premature mortality (YLL) and Years Lived with Disability (YLD) to the global burden of 31 hazards in food, 2010.

1.00

0.75

0.50

Proportion 0.25

0.00

spp. spp. spp. spp. spp. spp. spp. spp. spp. spp. spp. EPEC Typhi ETEC STEC Dioxins Aflatoxin S. enterica Norovirus Paratyphi A Shigella Ascaris Giardia Brucella Taenia solium Fasciola Vibrio cholerae Clonorchis Trichinella Intestinal flukes Salmonella Hepatitis A Virus Cassava cyanide Opisthorchis Paragonimus Toxoplasma gondii Campylobacter Mycobacterium bovis Cryptosporidium Entamoeba histolytica Listeria monocytogenes Salmonella non-typhoidal Echinococcus granulosus Echinococcus multilocularis

YLL YLD Results 82

Figure 14. Age-distribution of disability adjusted life years for 31 hazards contributing to the global burden of foodborne disease, 2010.

1.00

0.75

0.50 Proportion

0.25

0.00

spp. spp. spp. spp. spp. spp. spp. spp. spp. spp. EPEC ETEC STEC Typhi Dioxins Aflatoxin S. entericaNorovirus Paratyphi A FasciolaGiardia Shigella Ascaris Taenia solium Brucella Intestinal flukes Vibrio cholerae Clonorchis Trichinella Hepatitis A VirusSalmonella Cassava cyanide Paragonimus Opisthorchis spp. Toxoplasma gondii Campylobacter Mycobacterium bovis Entamoeba histolyticaCryptosporidium SalmonellaListeria monocytogenes non-typhoidal Echinococcus granulosus Echinococcus multilocularis

5+ <5

The FERG studies show that children incidence), including uncertainty under five years old bear 40% of the intervals. On the basis of this plot, foodborne disease burden (including, hazards were divided by two criteria for some hazards, the life-long burden with arbitrary cut-offs as indicated RESULTS of sequelae). More than 75% of the by grey-shaded areas in the Figure. V. burden of four hazards (Fasciola spp., cholerae, T. solium and Paragonimus Giardia spp., dioxins, and intestinal spp. were in the H/H category. All other flukes) occurred among children under diarrheal disease agents were in the H/L five (Figure 14). Prenatal infections category, except STEC, E. histolytica accounted for 21% of the burden of and Giardia spp. (L/L). The L/L category L. monocytogenes and for 32% of further included Trichinella spp. The L/H the burden of Toxoplasma gondii. By category contained agents that are of contrast, more than 75% of the burden of relatively low global impact but have 11 hazards occurred among people over a high impact on affected individuals. five years old. These included different parasites, particularly E. multilocularis, the invasive Figure 15 presents a scatterplot of the bacteria Brucella spp., L. monocytogenes burden at individual level (DALYs per and M. bovis. In subregions where the case, a measure for disease severity) burden is higher than the global average, and the burden at population level these agents are of specific relevance to (foodborne DALYs per 100,000 policy makers. population, also accounting for disease WHO Estimates of the global burden of foodborne diseases 83

Figure 15. Scatterplot of the global burden of foodborne disease per 100 000 population and per incident case.

100

Emult AAfla

CassC CClonlon OpisO 10 LLmonomono PParagarag TsoTsoll Fasc Flukesu MMbovbov

VcholV

DDioxiox Egran 1 SSParaPara STypSTyphh

Bruc

Trich EPEEPECC HAVV 0.1 Toxoxo

DALYs per foodborne case AAscsc NTNTSS Cryp ETEETECC CampCamp Shig NoV STEC 0.01

Ehist

0.001 Giar

0.01 0.1 110 100

Foodborne DALYs per 100,000 inhabitants

Diarrheal disease agents Helminths

Invasive infectious disease agents Chemicals and toxins

Abbreviations: NoV = Norovirus; Camp = Campylobacter spp.; EPEC = Enteropathogenic Escherichia coli; ETEC = Enterotoxigenic E. coli; STEC = Shiga toxin-producing E. coli; NTS = non-typhoidal Salmonella enterica; Shig = Shigella spp.; Vchol; Vibrio cholerae; Ehist = Entamoeba histolytica; Cryp = Cryptosporidium spp.; Giar = Giardia spp.; HAV = Hepatitis A virus; Bruc = Brucella spp.; Lmono = Listeria monocytogenes; Mbov = Mycobacterium bovis; SPara = Salmonella Paratyphi A; STyph = Salmonella Typhi; Toxo = Toxoplasma gondii; Egran = Echinococcus granulosus; Emult = E. multilocularis; Tsol = Taenia solium; Asc = Ascaris spp.; Trich = Trichinella spp.; Clon = Clonorchis sinensis; Fasc = Fasciola spp.; Flukes = Intestinal fl ukes; Opis = Opisthorchis spp.; Parag = Paragonimus spp.; Diox = Dioxin; Afl a = Afl atoxin. Results 84

5.3 DALY Estimates: proportion of foodborne infections Enteric diseases caused by EPEC, Cryptosporidium spp. and Campylobacter spp. occurred among It was estimated that the 22 diseases in children <5 years of age (Table A8.3 in the enteric disease study caused Appendix 8) Among the 11 diarrhoeal 2.0 billion (95% UI 1.5–3.0 billion) diseases, the rate ratio of foodborne illnesses in 2010, 39% of which (95% UI cases occurring among children <5 years 26–53%) were in children <5 years of age. of age compared with those >5 years of Among the 1.9 billion cases of diarrhoeal age was 6.44 (95% UI 3.15–12.46). diseases, norovirus was responsible for 684 million illnesses– the largest number It was estimated that the 22 diseases in of cases for any pathogen (Table A8.1 in the enteric diseases study caused 1.09 Appendix 8). The pathogens resulting million (95% UI 0.89–1.37 million) deaths in the next largest number of cases in 2010, of which 34% (95% UI 29–38%) were ETEC, Shigella spp., Giardia spp., in children <5 years of age. Among Campylobacter spp. and non-typhoidal the diarrhoeal diseases, norovirus was Salmonella spp. Campylobacter spp. responsible for the most deaths. Other cases included almost 32 000 GBS diarrhoeal pathogens responsible for cases. There were also 2.48 million large numbers of deaths were EPEC, V. STEC cases, which included 3610 with cholerae and Shigella spp. The 37 600 HUS and 253 with ESRD. Among the deaths attributed to Campylobacter spp. extra-intestinal diseases, the pathogens included 1310 deaths from GBS. Among resulting in the most infections were the extra-intestinal enteric diseases, the hepatitis A virus, S. Typhi and S. pathogens resulting in the most deaths Paratyphi A. Brucella spp. resulted in 0.83 were S. Typhi, hepatitis A virus, iNTS and million illnesses, which included almost S. Paratyphi A. 333 000 chronic infections and 83 300 Overall, the 22 diseases in the enteric episodes of orchitis. L. monocytogenes

diseases study resulted in 351 000 RESULTS resulted in 14 200 illnesses which (95% UI 240 000–524 000) deaths due included 7830 cases of septicaemia, to contaminated food in 2010; with 33% 3920 cases of meningitis, and 666 cases (95% UI 27–40%) in children <5 years of with neurological sequelae. age. The enteric pathogens resulting in Overall, 29% (95% UI 23–36%) of all the most foodborne deaths were S. Typhi, 22 diseases were estimated to be EPEC, norovirus, iNTS, non-typhoidal transmitted by contaminated food, Salmonella spp. and hepatitis A. The equating to 582 million (95% UI mortality rates of foodborne diseases 400–922 million) foodborne cases in were consistently highest in the African 2010; of which 38% (95% UI 24–53%) in subregions, followed by the South children <5 years of age. The pathogens Eastern Asian subregions (Table A8.2 resulting in the most foodborne cases in Appendix 8) Among the 11 diarrhoeal were norovirus, Campylobacter spp., diseases due to contaminated food, the ETEC, non-typhoidal Salmonella spp., rate ratio of deaths in children <5 years and Shigella spp. A high proportion of age compared with those >5 years of of foodborne infections caused by V. age was 8.37 (95% UI 5.90–11.4). For all cholerae, S. Typhi, and S. Paratyphi 22 diseases, the rate ratio of deaths in A occurred in the African region children <5 years of age compared with (Table A8.2 in Appendix 8). A high those >5 years of age was 4.85 (95% UI 3.54–6.59). WHO Estimates of the global burden of foodborne diseases 85

It was estimated that the 22 diseases in age. Figure 17 shows the relative burden the enteric diseases study caused 78.7 of foodborne enteric infections, if iNTS million (95% UI 65.0–97.7 million) DALYS and non-typhoidal Salmonella spp. were in 2010, of which 43% (95% UI 38–48%) grouped together. The pathogen resulting in children <5 years of age. The in the most foodborne DALYs was non- pathogens resulting in the most typhoidal Salmonella spp., if iNTS were DALYs were norovirus, S. Typhi, EPEC, included (4.07 million). Other pathogens V. cholerae, ETEC, and hepatitis A resulting in substantial foodborne DALYs (Table A8.1 in Appendix 8). included: S. Typhi, EPEC, norovirus and Campylobacter spp. The rates of DALYs The rates of foodborne DALYs per for foodborne diseases were highest in 100 000 population are shown by hazard the African subregions. Overall, the 22 and by subregion in Table A8.2 diseases transmitted by contaminated in Appendix 8. food resulted in 10.8 million (95% UI It was estimated that the 22 diseases 7.59–15.3 million) DALYs in children <5 in the enteric diseases study resulted years of age, compared with 14.3 million in 25.2 million (95% UI 17.5–37.0 million) (95% UI 9.42–22.5 million) DALYs in those DALYs due to contaminated food; 43% >5 years of age. (95% UI 36–50%) in children <5 years of

Figure 16. The global burden of foodborne disease by subregion (DALYS per 100 000 population) caused by enteric hazards, 2010.

1000

750

500

250 Foodborne DALYs per 100,000 inhabitants

0 AFR D AFR E AMR A AMR B AMR D EMR B EMR D EUR A EUR B EUR C SEAR B SEAR D WPR A WPR B

Norovirus non-typhoidal S. enterica Giardia spp. Salmonella Paratyphi A Campylobacter spp. Shigella spp. Hepatitis A Virus Salmonella Typhi EPEC Vibrio cholerae Brucella spp. ETEC Cryptosporidium spp. Listeria monocytogenes STEC Entamoeba histolytica Mycobacterium bovis Results 86

Figure 17. Disability Adjusted Life Years for each pathogen acquired from contaminated food ranked from lowest to highest with 95% Uncertainty Intervals, 2010.

10,000,000

1,000,000

100,000 Global foodborne DALYs

10,000

spp. spp. spp. spp. spp. STEC ETEC EPEC Typhi

Norovirus Giardia Hepatitis A Brucella Shigella Vibrio cholerae , non-typhoidal Salmonella Campylobacter MycobacteriumSalmonella bovis Paratyphi EntamoebaCryptosporidium histolytica Listeria monocytogenes Salmonella

Note figure is on a logarithmic scale. The figure shows the median (white dot); Inter-Quartile Range = 50% UI = 25%/75% percentiles (thick black line); 90% UI = 5%/95% percentiles (thin black line); 95% UI = 2.5%/97.5% percentiles (thin grey line). Note, four foodborne intoxications due to Clostridium botulinum, Cl. perfringens, S. aureus, and Bacillus cereus due to a lack of data for global estimation. In addition, data for non-typhoidal Salmonella enterica infections and invasive non-typhoidal S. enterica have been combined. RESULTS

5.4 DALY Estimates: Parasites caused by cysticercosis were 2.79 million (95% UI 2.14–3.61 million) DALYs. The parasitic diseases with the largest Foodborne trematodiasis resulted in total number of symptomatic incident 2.02 million (95% UI 1.65–2.48 million) cases and symptomatic incident cases DALYs. Toxoplasmosis had a burden attributable to contaminated food in (congenital and acquired combined) of 2010 are acquired toxoplasmosis and 1.68 million (95% UI 1.24–2.45 million) ascariosis. The incidence in 2010 of DALYs, with ascariosis also resulting in each parasitic disease per 100 000 1.32 million (95% UI 1.18–2.70 million) population by region are given in DALYs. Echinococcosis (alveolar and (Table A8.4 in Appendix 8). Also of note cystic combined), had a burden of were the relatively few cases of human approximately 871 000 DALYs (CE trichinellosis, with a global estimate of 184 000, 95% (UI 88 100–1.59 million] just 4400 cases and 4 deaths in 2010. DALYs; AE 688 000, 95% [UI 409 000–1.1 The number of DALYs associated with million] DALYs). This gives a 2010 global each parasite and the proportion of burden of these 11 parasitic diseases of DALYs that were foodborne in 2010 8.78 million (95% UI 7.62–12.5 million) are given in Table A8.5 in Appendix 8. DALYs, of which 6.64 million (95% UI In 2010 the burdens estimated to be 5.61–8.41 million) DALYs were estimated WHO Estimates of the global burden of foodborne diseases 87

to be foodborne. Contaminated food may DALYs per 100 000, respectively, whereas be responsible for 48% (95% UI 38%– the lowest was found in the European 56%) of incident cases and approximately subregions, with 11 (95% UI 8–24) DALYs 76% (95% UI 65%–81%) of DALYs (Table per 100 000 (Appendix 8 Table A8.4). 1). Stillbirths were excluded, although in However, the relative importance of the the case of congenital toxoplasmosis, different parasitic infections varied across if counted as deaths as an alternative regions and this is clearly illustrated in scenario, this would result in 4470 (95% Figure 18A. For example, the burden of UI 969–12 400) additional deaths and opisthorchiosis is largely confined to hence an addition of approximately SEAR subregion D, whilst cysticercosis is 411 000 (95% UI 89 100–1.14 million) rarely seen in either EMR or EUR regions. YLLs. Of these, approximately 2180 The absolute and relative foodborne (95% UI 470–6090) deaths and 200 000 burdens of these parasitic diseases, (95% UI 43 200–560 000) YLLs would including the three enteric protozoa, be foodborne. are illustrated in Figure 18B. The relative The largest global incidence rate of proportion of the burden of each of the DALYs was found in the Western Pacific foodborne parasitic diseases contributed and African subregions, with 156 (95% by YLLs and YLDs is illustrated in UI 127–193) and 208 (95% UI 159–283) Figure 19.

Figure 18A. The relative contribution to the DALY incidence by each agent for each of the subregions. This includes enteric protozoa to complete the picture on foodborne parasitic diseases. However the detail is reported in the accompanying manuscript on foodborne enteric pathogens [168].

200

150

100

50

Foodborne DALYs per 100,000 inhabitants 0 AFR D AFR E AMR A AMR B AMR D EMR B EMR D EUR A EUR B EUR C SEAR B SEAR D WPR A WPR B

Toxoplasma gondii, congenital Taenia solium Fasciola spp. Entamoeba histolytica Toxoplasma gondii, acquired Ascaris spp. Intestinal flukes Cryptosporidium spp. Echinococcus granulosus Trichinella spp. Opisthorchis spp. Giardia spp. Echinococcus multilocularis Clonorchis spp. Paragonimus spp. Results 88

Figure 18B. Disability Adjusted Life Years for each parasite acquired from contaminated food ranked from lowest to highest with 95% Uncertainty Intervals, 2010. This includes enteric protozoa to complete the picture on foodborne parasitic diseases. However the detail is reported in the accompanying manuscript on foodborne enteric pathogens [168].

10,000,000

1,000,000

100,000

10,000

1,000 Foodborne Disability-Adjusted Life Years

100

Giardiosis Fasciolosis Ascariosis Trichinellosis Clonorchiosis Cysticercosis Entamoebosis Paragonimosis Opisthorchiosis

Cystic echinococcosis Intestinal trematodosis Alveolar echinococcosisAcquired toxoplasmosis Congenital toxoplasmosis

Figure 19. The relative proportion of the burden of each of the foodborne parasitic diseases RESULTS contributed by YLLs and YLDs

1.00

0.75

0.50

Proportion 0.25

0.00

Giardiosis Ascariosis Fasciolosis ClonorchiosisCysticercosis Trichinellosis Entamoebosis Paragonimosis Opisthorchiosis Cryptosporidiosis

Cystic echinococcosis Intestinal trematodosis Alveolar echinococcosis Acquired toxoplasmosis Congenital toxoplasmosis

YLL YLD WHO Estimates of the global burden of foodborne diseases 89

5.5 DALY Estimates: Chemicals disease per 100 000 population for each of the regions is reported in Table A8.7 The analyses presented here show that in Appendix 8. Peanut allergy is not four selected chemicals already have reported in Table A8.7 in Appendix 8 a substantial impact on the foodborne because burden was estimated only burden of disease, particularly in low- for AMR A (United States of America, and middle-income countries. Just Canada and Cuba), EUR A (primarily these four agents are estimated to be countries in Western Europe), and WPR associated with 339 000 illnesses (95% A (Australia, Brunei Darussalam, Japan UI 186 000–1 239 000); 20 000 deaths and New Zealand) subregions. Burden (95% UI 8 000–52 000); and 1 012 000 estimates for cyanide in cassava are DALYs (95% UI 562 000–2 822 000) in provided only for the African region 2010. These should be considered the (AFR) and assumed to be zero for “tip of the iceberg” in terms of foodborne other regions. chemicals and their impact on the global burden of disease. For peanut allergens, Figure 20 provides the DALYs per we were unable to estimate a burden 100 000 inhabitants by global region. for low- and middle-income countries The regions with the highest burden due to data gaps. We also had to use an per 100 000 inhabitants are the approximate disability weight, as there subregions in SEAR, WPR and AFR. are data only on quality of life of patients The American Region (AMR), Eastern with [102] and no specific Mediterranean Region (EMR), and data are available for peanut allergy. European Region (EUR) have the lowest DALYs per 100 000. Aflatoxin is the The estimated number of incident largest contributor to the burden in cases, deaths and DALYs for each of AFR and WPR. Dioxin makes the largest the diseases associated with chemicals contribution in SEAR. Figure 21 contrasts is given in Table A8.6 in Appendix 8. the proportion of DALYs due to YLL The chemical associated with the most and YLD for each of the four chemicals. number of illnesses is dioxin; however, Virtually all of the DALYs for aflatoxin no deaths have been reported from the and most of the DALYs for cyanide in presence of dioxin in the food supply. The cassava are due to YLL, whereas most chemical associated with the greatest of the DALYs for peanut allergen and all number of DALYs is aflatoxin. The DALY of the DALYs for dioxin are due to YLD. estimates for aflatoxin and dioxin have Figure 22 shows the uncertainty around the least uncertainty; more uncertainty is the DALY estimates for each of the four associated with the DALY estimates for chemicals. The chemical with the least peanut allergen and cyanide in cassava. uncertainty and the greatest number of The annual incidence, mortality, and DALYs is aflatoxin. DALY rate of each chemical-associated Results 90

Figure 20. The relative contribution to the DALY incidence by each of four chemicals for each of the WHO Regions.

30

20

10 Foodborne DALYs per 100,000 inhabitants

0 AFR D AFR E AMR A AMR B AMR D EMR B EMR D EUR A EUR B EUR C SEAR B SEAR D WPR A WPR B

Aflatoxin Cassava cyanide Dioxin Peanut

Notes: Peanut allergy burden was estimated only for the AMR A (United States of America, Canada and Cuba); EUR A (primarily countries in western Europe); and WPR A (Australia, Brunei Darussalam, Japan and New Zealand) subregions. Burden estimates for cyanide in cassava are provided only for the African region (AFR), and assumed to be zero for other regions. RESULTS Figure 21. The relative contributions from YLLs and YLDs for each of four chemicals.

1.00

0.75

0.50

Proportion 0.25

0.00

Peanut Dioxin Aflatoxin

Cassava cyanide

YLL YLD WHO Estimates of the global burden of foodborne diseases 91

Figure 22. Disability Adjusted Life Years for each of four chemicals from contaminated food ranked, from lowest to highest, with 95% uncertainty intervals

10,000,000

1,000,000

100,000

10,000

1,000 Foodborne Disability-Adjusted Life Years

100

Peanut Dioxin Aflatoxin

Cassava cyanide

Notes: The dot in the middle of each box represents the median, the box the 50% uncertainty interval, the dark bar the 95% uncertainty interval, and the light bar the 95% uncertainty interval. Results 92 RESULTS WHO Estimates of the global burden of foodborne diseases 93

DISCUSSION 6

WHO Estimates of the global burden of foodborne diseases 95

DISCUSSION

This study estimates that 31 foodborne in the FERG estimates whereas in the hazards resulted in 33 million DALYs GBD 2010 study they were recorded in in 2010, which shows the considerable different disease categories. Furthermore, public health impact of contaminated the GBD 2010 study used a different life food. Importantly, children <5 years of table than FERG and more extensive age experienced 40% of the foodborne mathematical modeling to account for disease burden, despite representing only data gaps, which smoothed the data 9% of the global population. The study considerably, resulting in narrower provides a substantial expansion of the uncertainty intervals than in our study. available data on the public health impact The GBD 2010 and FERG studies used of FBDs. the same set of disability weights, but the FERG included some updates Several high-income countries have as recommended by WHO. Neither published national estimates of FBD. study applied time discounting or age- Estimates of food-related illnesses and weighting in their baseline estimates. deaths in the USA were estimated in the late 1990s [169] and updated to cover The GBD 2010 study, which looked at the period 2000–2008 [170, 171]. Similar all sources of disease, found that the studies are available from Australia [173], key hazards and risk factors for disease Canada [175], [174] and the UK burden were dietary risk factors (254 [172]. Some countries have extended this million DALYs), unimproved water and work to estimate DALYs, including Greece sanitation (211 million DALYs), HIV/AIDS [177], the Netherlands [178], New Zealand (82 million DALYs), malaria (82 million [176] and the USA [179]. While the range DALYs), air (76 million DALYs) of hazards covered in these studies and tuberculosis (49 million DALYs). differed from those in the FERG studies, Recently published findings from WHO1 the focus was on enteric diseases and a for 2012 were: HIV/AIDS (92million limited number of invasive and parasitic DALYs); malaria (55 million DALYs) and diseases. The FERG data, by contrast, tuberculosis (44 million DALYs). Hence, cover numerous countries across the the burden of FBD (33 million DALYs) globe and provide a more complete is of a similar order of magnitude as picture of FBD. the ‘big three’ infectious diseases (HIV/ Comparisons of the FERG estimate of AIDS, malaria and tuberculosis) and the burden of FBD with other estimates, air pollution, but clearly lower than such as those of the Institute for Health the burden of dietary risk factors or Metrics and Evaluation’s GBD 2010 study unimproved water and sanitation. [81], must be made with care because The FERG estimate of 29,000 deaths due of differences in the methodology and to foodborne transmission of invasive data used. For example, the GBD 2010 non-typhoidal S. enterica only included study used prevalence-based DALYS, infections in non-HIV infected individuals. whereas our study used incidence-based Ao et al. [180] estimated there were DALYs. As a consequence, the impact of approximately 680,000 deaths due to sequelae such as Guillain-Barré syndrome invasive non-typhoidal salmonellosis in (due to Campylobacter spp.), hemolytic 2010. Of these, approximately 350,000 uremic syndrome (due to Shiga toxin- would be due to foodborne transmission, producing E. coli) and invasive disease assuming 52% of all non-typhoidal (due to non-typhoidal S. enterica) were 1 http://apps.who.int/gho/data/node.main. attributed to the diarrheal disease agents DALYNUMWORLD?lang=en; accessed July 22, 2014 Discussion 96

salmonellosis cases is transmitted by relevance [182]. In particular, we did not food [156]. Even though this high number include burden estimates for several of deaths among HIV infected people chemicals (arsenic, cadmium, lead and is not included in the FERG estimates methylmercury), because methods for of the burden of FBD, they would be estimation of the fraction of illnesses preventable by food safety interventions. attributed to foodborne exposure to This study is subject to several limitations, these chemicals are not readily available. notably due to uncertainties in the data Even for the hazards we have studied, limitations on burden estimates and it was not always possible to include attribution estimates. For most hazards all relevant disease outcomes in our (25 of the 31 studied), the 95% DALY estimates of burden. For example, we did uncertainty interval (UI) ranged from not include functional bowel disorders as one-fourth to four times the median. The potential outcomes for enteric infections uncertainty was markedly greater for E. [183]. Inclusion of these outcomes would multilocularis (because of uncertainty in likely considerably increase the burden of the attribution estimates), E. granulosus enteric infections [184}. and L. monocytogenes (because of Aflatoxin burden was estimated using uncertainties in the imputation results). a counterfactual approach, estimating In low-income countries, where the population attributable fractions from burden is highest, data availability was exposure assessment estimates and generally most problematic. Furthermore, cancer potency factors, and applying in these countries, the proportions of these to WHO estimates for incidence diseases transmitted by food, water and mortality by hepatocellular and the environment are difficult to carcinoma. Risk assessment, as used to disentangle, as contaminated water may assess the burden of dioxins [133] has also result in contamination of foods. Due been proposed as an alternative basis to these limitations, it was not possible for estimating this particular burden, to present reliable estimates at country and would result in considerably higher level, and elected to present results at estimates of the burden of aflatoxin [110]. subregion level. Moreover, both aflatoxin and dioxin can For some hazards (e.g. M. bovis and cause other adverse health effects than E. multilocularis, aflatoxin and dioxin), the ones considered (e.g. diarrhoeal incident illness is related to past diseases [185] and aflatoxin as causes exposures due to long incubation of malnutrition and stunting, dioxin and DISCUSSION times of disease. For such hazards, the immune effects or cancer), for which estimated burden reflects exposure data were not available to allow disease dating back to the average incubation burden estimates. period of the disease rather than current A further limitation of this study is that exposure. For some hazards (e.g. DALYs do not quantify the full societal dioxins), the impact on the child depends impact of FBD. The economic burden on the lifelong exposure of the mother. (cost-of-illness, losses in the agricultural The FERG estimates of the FBD and food sectors and trade impacts) burden are probably conservative, is also an important factor to consider i.e. underestimates rather than in national and international decision- overestimates. Limited resources and making. Also, the process of food data obliged us to focus on only a subset production can cause human diseases of more than 100 hazards of potential by mechanisms other than direct WHO Estimates of the global burden of foodborne diseases 97

transmission of pathogens through regional estimates, the consideration of food. For example, animal husbandry local hazards that may not have been is an important source of zoonotic addressed at a global level, and the disease agents that spread from pigs, translation of burden estimates into food , cattle, etc., by direct contact or safety policy. The estimates developed by through the environment, and may also this WHO initiative will be invaluable for affect livestock health. It is increasingly countries where local data gaps prevent necessary to consider holistically all the development of a full picture of FBD. aspects of food-related disease in a One Health Framework [186]. The considerable difference in the burden of foodborne disease between low- and Despite its data gaps and assumptions, high-income regions suggests that a this study presents the first ever major proportion of the current burden estimates of the global burden of FBD is avoidable. The WHO is working with and should serve as an important governments and partners, including resource to focus activities that will food producers, caterers and consumers, reduce this burden. A sustainable, to reduce food contamination throughout multi-sectoral response is needed the food chain, and particularly at from governments and international the point of consumption, to levels at organizations to reduce the visible which the exposure to pathogens and and ‘hidden’ burden; this includes contaminants does not pose significant enforcement of food safety standards risks for human health. There is, therefore, and effective surveillance networks an urgent need to develop cost-effective at country, regional and global levels. food hygiene interventions that can be This will require a concerted effort by implemented in resource-poor settings. all stakeholders in the food chain, from This research and development should be primary production to consumers. The informed by estimates of the burden of diversity of foodborne hazards suggests specific food vehicles, taking all hazards the need for a multi-faceted strategy, with priorities tailored to each region. into account. While national studies may further General principles for strengthening food refine these priorities and are highly safety systems have been suggested by recommended, the current findings the WHO; they include integrating food could already be a basis for developing safety into nutrition and food security strategies at the global, regional and policies and programs, and fostering national levels. closer collaboration between the various The diversity of foodborne hazards and sectors involved (, human regional differences in their importance health, animal health, trade, tourism, etc.). suggest the need for consideration of The WHO recommends governments put these estimates at the national or even in place risk-based food control systems subnational level. As one of its aims, and implement international food the FERG has fostered national studies safety standards as established by the of the burden of FBD, and pilot studies Commission. Food have been conducted in Albania, Japan, handlers and consumers should handle Thailand and Uganda. The tools and and prepare food safely, practicing the protocols developed by the FERG to WHO’s ‘Five Keys to Safer Food’ and support such national studies emphasize grow and vegetables using the the collation of local data to validate its WHO’s ‘Five Keys to Growing Safer Fruits Discussion 98

and Vegetables’ to decrease microbial We were unable to identify contamination2. epidemiological studies in the literature that delineate and quantify the FBDs are closely linked to poverty in importance of each transmission pathway developing countries but they are also as investigated in this study. This makes it a global public health issue because difficult to formally validate the findings growing international trade increases of the expert elicitation. Still, a discussion the risk of contamination in transported of summary findings in the context of foods; also, migration and travel can other scientific knowledge may be of expose populations to new hazards. value. Achievement of the internationally agreed Millennium Development The hazards can be grouped into several Goals and the proposed Sustainable categories with respect to their major Development Goals, including the pathways. For Campylobacter, non- overarching goals of poverty reduction, typhoidal Salmonella, STEC, T. gondii, and achieving food security and ensuring E. multilocularis, the foodborne route was healthy lives, will depend in part on considered the most important route in successful reduction of the burden all subregions. These pathogens are all of FBD. zoonotic and known to have one or more animal reservoirs. The zoonotic nature of these organisms is also reflected in 6.1 Attribution experts’ judgments by the identification In the attribution study, the results are of direct contact with animals as an presented of the first world-wide study important transmission route as well. on the contribution of contaminated Other pathogens with animal reservoirs food and other exposure routes to include E. granulosus and Brucella spp., human disease caused by 18 major and here direct contact with animals was microbiological hazards and a chemical considered equally or more important hazard. The study highlights the than food as routes of transmission. importance of the foodborne route As described in the results section for of transmission for these hazards several pathogens, there was a clear and– when combined with estimates pattern that the experts considered the of incidence, severity, duration and foodborne route less important in low- mortality– allows estimation of the and middle-income subregions, where DISCUSSION burden of foodborne disease. Attempting other routes (animal contact, water to estimate foodborne transmission at and soil) were believed to contribute the subregional level is an ambitious relatively more in comparison with high- goal. However, this was vital given income subregions. This is consistent the geographically localized nature of with data showing lower levels of access exposure to many pathogens. The results to improved water and sanitation in are significant due to the global nature less developed regions compared of the estimation, the number of experts with high-income countries. This participating, and the rigorous approach ranking of subregions across different taken to assessing and including expert pathogens provides some confidence performance in the final estimates. in the results, as the estimates were 2 http://www.who.int/campaigns/world-health- done independently and partly by day/2015/en/ different experts. WHO Estimates of the global burden of foodborne diseases 99

An expert elicitation was used to primarily a human (e.g. hepatitis estimate source attribution parameters A virus, S. Typhi and V. cholerae). Some because of not only the lack of globally of the differences observed may occur consistent data on which to base such because we are comparing national estimates, but also a general lack of estimates with subregional estimates, data and research on source attribution where the latter could be interpreted as a in most of the world. The generally weighted average across all countries in wide uncertainty bounds provided by the subregion. the expert elicitation in this study are For hepatitis A, there is a strong presumed to reflect both uncertainty disagreement between the national and variation, where uncertainty arises studies and this study, where the due to the sparseness of hard evidence proportion foodborne is estimated to data for, or the presence of conflicting be less, but at the same level in the evidence for, the contribution of different four national studies that investigated transmission pathways, and variation this pathogen. This difference cannot reflects the experts’ beliefs concerning be readily explained, but it should be variations between countries within any noted that there was also disagreement given subregion. A study operating with between the experts in this study, smaller regions or at country level might where three of the six experts serving have reduced the uncertainty due to on the hepatitis A virus panel provided variation. estimates in line with those published There exist a few recent national studies for the national studies, whereas the that estimate the proportion of illnesses remaining three experts provided attributable to the foodborne route considerably higher estimates. The for specific infectious diseases [33–35, disagreement between the experts is also 37, 187, 188]. Table A7.6 in Appendix 7 reflected in the uncertainty bounds for provides the main results from these the estimates, which are quite wide and studies. Four of the six studies used contain the estimates from the national some kind of formal expert elicitation, studies. For S. Typhi and V. cholerae, where enrolled experts were asked to the estimates from the national studies provide a central estimate and their are higher than those found in this uncertainty bounds around this. The study. One explanation could be that estimates published by Gkoga et al. [187] the national studies [35, 187, 188] were only attributing domestically acquired and Scallan et al. [42] were derived by cases. In the study by Scallan et al. [42], the authors using a synthesis of data the proportion foodborne for S. Typhi from different public health surveillance was estimated based on data from 17 systems and the literature. As all six domestic outbreaks reported in a 19– studies were conducted in developed year period, where 13 outbreaks were countries, we compare the results confirmed foodborne and 4 outbreaks only with the results from the relevant were of unknown origin. The same study subregions (i.e. EUR A, AMR A and WPR included also only data from domestically A) in this study. acquired cases of V. cholerae, but as For the zoonotic pathogens, particularly around 70% of all cases were estimated non-typhoidal Salmonella spp., the to be travel related, including all cases estimates agree more closely and the could change the proportion foodborne uncertainty ranges tend to be relatively significantly. Infections with S. Typhi narrower than for pathogens with and V. cholerae are typically linked with Discussion 100

contaminated water sources and poor food or water depending on the point- hygiene in developing regions [27], so of-attribution. It is clear that there exists these transmission routes are likely to no single “right” way of delineating the be relatively more important among foodborne route from other transmission cases that have been travelling to these routes; however, it is critical that point of regions. This is probably what is reflected attribution be clearly defined. Definition in the attribution results in this study. of point-of-attribution should depend on the objective and focus of the specific The operational definition of different study carried out, and could involve many transmission routes, in particular the factors, including the foodborne hazard food and waterborne routes, will in question, the food production systems affect attribution estimates. Hazards and routines in place, the geographical transmitted by multiple routes can occurrence of the hazard, sanitation and “change” source or vehicle during the hygiene in the region, and consumption transmission from primary source to patterns. If point of attribution is clearly humans, meaning that the burden of defined, then additional modelling or illness caused by a particular hazard further research can be used to adjust attributed to a specific transmission route attribution to exposure at other points of may vary, depending on the point-of- interest in the transmission chain. If the attribution chosen [7, 21]. The choice of point of attribution is not unambiguously point-of-attribution seems particularly defined, then not only are the results of critical for delineating foodborne and the study unclear, but it will be difficult to . This is because use them to model other relationships in water in itself is ingested just as a food, the transmission chain. is used for irrigation of food plants, for washing and cleaning of food during FERG agreed that the point of human preparation and constitutes an essential exposure was the most simple and ingredient in many food products. In understandable point-of-attribution to addition and particularly related to be used across all hazards for delineating zoonotic diseases, the water source the major transmission routes (Figure 23). itself is often contaminated by an animal FERG recognizes that for other purposes, reservoir, including food-producing e.g. for risk management, other points animals. Another situation relates to of attribution may be more appropriate the consumption of water-based foods (e.g. primary production, processing and such as shellfish harvested from areas retail, or preparation). Yet, the FERG’s

where the water is contaminated with definition of point of attribution for DISCUSSION pathogenic organisms, such as Vibrio major transmission routes directly links spp. or enteric viruses. The burden attribution to disease incidence, as it is related to the consumption of foods that the end of the transmission chain. Further have had contact with contaminated modelling can then be used to work water at some stage of the production backward from exposure to identify the may, therefore, be attributed to either important points of contamination. WHO Estimates of the global burden of foodborne diseases 101

Figure 23. Major transmission routes of human foodborne diseases indicate two points of attribution: the reservoir level and the exposure level.

Food and Environment Food animals Pets and wildlife Reservoir level Feed plants

Processing Processing

Preparation Consumption

Foodborne

Water, soil, Direct animal Exposure level Human contact etc.

Human- human

In this study, we identified potential with the usual markers of professional experts through peer-nomination, for qualifications, such as publication records the purpose of enumerating parameter [190]. Structured expert judgment uncertainties through structured expert studies are, therefore, not as sensitive to judgment. It is important to recognize selection bias and low response rates as that the goal of a structured expert other types of surveys. In our study, we judgment is not to characterize the approached 299 potential experts and properties of the group of experts in ended up with a final pool of 72 (24%), some sense, but to obtain uncertainty which is actually a fair response rate quantifications of target variable compared with population surveys. The which are statistically accurate and response rate of those that committed informative. The degree to which this to participate in the study by forwarding goal is realized is assessed objectively by their CV, DOI, etc. was 70%. The motives referencing elicited target uncertainties for some people declining to participate to the experts’ performances in judging were not specifically asked for, but most similar, factual realizations of calibration of those giving a reason indicated lack variables in the subject matter field. of time. A few declined because they did This empirical validation of expert not perceive themselves experts in the (and combined experts) uncertainty field. The reason given by the majority quantifications is what distinguishes for not finally submitting responses to formalized structured expert judgment the target question, even though they from surveys or statistical sampling had gone through the interview, was also [189]. Assessment of the uncertainty time constraint. Although some selection judgment capabilities of experts using bias cannot be ruled out absolutely, calibration variables has also been we do not believe that it has had major demonstrated to be a better predictor impact on the study results due to of expert performance compared the formalized basis of the elicitation Discussion 102

process and objective judgment- availability of estimates of the envelope pooling methodology. of diarrhoeal deaths, along with recent advances in diarrhoeal disease diagnosis, 6.2 Enteric Diseases such as widespread application of polymerase chain reaction (PCR) for This study, for the first time, estimates norovirus detection [192, 193]. the substantial burden of foodborne In our study, norovirus resulted in the bacterial, viral and protozoal diseases in humans, particularly among largest number of cases of foodborne children. Although children <5 years diseases and overall burden, highlighting of age represent only 9% of the global the global importance of this agent [192]. population, 43% of the disease burden However, the disease model we used from contaminated food occurred in in the 135 middle- and high-mortality this group. Foodborne illnesses from counties included only norovirus diarrhoeal and invasive non-typhoidal infections that resulted in a diarrhoeal Salmonella spp. resulted in the largest illness. If we also included estimates for disease burden, reflecting the ubiquitous norovirus infections in these countries nature of Salmonella, the severe nature of that resulted in vomiting without illness, and the fact that young children diarrhoea, there would be an estimated are commonly infected [54]. Large additional 163 million norovirus cases human disease burdens are also imposed [194]. We also found, similar to what by foodborne infections due to norovirus has been reported in national studies, and typhoid. It is important to recognize that in the countries where we applied that diseases with a lower immediate the modified CHERG approach, the burden may still warrant intervention. In aetiological cause of almost half of particular, certain foodborne diseases diarrhoeal cases and deaths remained may represent a larger problem in unknown. This was probably due, in large some regions. For example, the most part, to pathogens that are possibly substantial burden due to foodborne foodborne but with insufficient data for cholera occurs in African and Asian estimation, or unknown agents not yet regions. Similarly, the burden of discovered. In this study, we focused brucellosis and M. bovis infections were our attention on the burden due to highest in the Middle Eastern and African pathogens that were known to be regions. transmitted by contaminated food.

To develop these comprehensive “When we examined the human health DISCUSSION estimates of the disease burden of impact of different pathogens, various foodborne diseases, we adopted an serotypes of Salmonella resulted in innovative approach to incorporate the greatest foodborne burden. If the highest quality data available for we consider the combined burden each foodborne disease [40]. Due to attributable to S. enterica from all their quality, we gave highest priority invasive (including iNTS, Salmonella to studies with national estimates of Typhi and Salmonella Paratyphi A) foodborne diseases. Since studies with and diarrheal infections, there were national estimates were only available in an estimated 21.2 million (95% UI 12.8– a few countries, we adapted the CHERG 35.8 million) DALYs from all transmission approach for estimating the disease sources and 8.76 million (95% UI 5.01– burden of diarrhoeal diseases [50, 51]. 15.6 million) DALYs from contaminated This approach was facilitated by the food. This highlights the significant WHO Estimates of the global burden of foodborne diseases 103

public health importance of Salmonella typhoid fever and incorporated these infections and the urgency of control, data into our estimates. The CFR for particularly for invasive infections in low- each of the diseases included in our and middle-income settings where most estimate are comparable to those of the mortality occurs [51–53].” reported in national studies. There is a continuing need for high quality studies Twelve of the diseases included in assessing the foodborne disease burden our study were also included in the at the national level. Our methodology GBD2010 study [6, 58, 81]. For three for estimating the disease burden diseases (typhoid, paratyphoid and attributable to foodborne transmission hepatitis A) we used GBD2010 data could be used in future studies. to derive estimates of incidence. For the other nine diarrhoeal diseases, we In comparing the overall burden of our elected to conduct our own analysis or findings, the diseases we included in used updated data from commissioned our study resulted in 79 million DALYs in systematic reviews to derive estimates. 2010. This represents approximately 3% Our study builds upon the GBD2010 of the 2.49 billion DALYs reported in the study by providing estimates of the GBD2010 study [81]. GBD2010 estimated proportion of each disease acquired that approximately 25% of DALYs globally from food; we also provide, in addition were due to deaths and disability in to estimates of deaths, estimates of children <5 years of age, while we the number of illnesses for each of the estimated that 43% of the DALYs in our diseases [6]. Before we applied our study were among children <5 years estimate of foodborne proportions to of age. each pathogen, our estimates of the There are obvious policy implications of disease burden for a few pathogens, in our findings. Countries and international terms of the estimated number of deaths agencies must prioritize food safety to and DALYs, were relatively similar for minimize foodborne illness, particularly diseases common to GBD2010 and FERG. among young children. The highest However, there were important burden of disease due to contaminated differences in other estimates. The food was in the African region, largely GBD2010 estimates of deaths due to E. due to iNTS in children. In contrast to histolytica, Cryptosporidium spp. and waterborne disease, the interventions Campylobacter spp. were 10 times, 4 for foodborne diseases are less clear times and 3 times greater, respectively, and have a much weaker evidence than the FERG estimates [58]. The FERG base. For example, there are virtually no estimate for DALYs for non-typhoidal randomized studies examining the impact Salmonella spp., combining diarrhoeal of interventions on reducing foodborne and invasive infections, was 4 times disease. Our results should stimulate greater than GBD2010 [81]. The FERG research into prevention of foodborne estimates are relatively similar to previous illnesses and better understanding of the global estimates of cholera, typhoid fever, epidemiology of these infections. salmonellosis and shigellosis [53, 54, 59, A limitation of our estimates of the 198]. GBD2010 estimates included ‘cysts consequences for human health of and liver abscesses’ as a of foodborne diseases is that, due to data typhoid fever, which has been questioned gaps, particularly in middle- and high- [199]. However, we understood this mortality countries, we included only a categorization to be a proxy for serious few sequelae in our estimates. We did not Discussion 104

include data on post-infectious sequelae, variable proportions attributed to food, such as and irritable depending on the nature of the experts bowel syndrome following foodborne included in elicitation studies [201, 202]. infections. Studies of the burden of Without specific studies attributing enteric pathogens in low-mortality sources of infection, it is difficult to countries highlight that excluding these obtain accurate estimates for foodborne sequelae under-represents the true transmission, but this finding regarding burden of disease, but reliable data were the need for more attribution studies not available from middle- and high- is an important outcome of our study mortality countries [200]. Where we did [203]. For example, the FERG expert include sequelae, there were insufficient elicitation study estimated that 18% of data to account for age-specific effects. norovirus was foodborne, compared with We also excluded stillbirths; this exclusion 14% estimated from a recent study based only affected disease burden estimates on outbreak genotyping [204]. for L. monocytogenes. The major limitation of our study was A recent review estimated that listeriosis, the lack of reliable data from many which we assume is all foodborne, causes regions of the world. In particular, we 273 stillbirths globally annually [70]. had the least data for some pathogens We were unable to distinguish between for the most populous regions of the the effects on health from the condition world [50]. We tried to use the best under study and that of co-morbid data available and attempted to make conditions, which is common to many reasoned assumptions wherever possible studies of human health. For example, [205]. For some agents, such as toxin- salmonellosis and M. bovis infections mediated illnesses, we elected to limit our may occur as HIV co-infections. If estimates to countries where diseases FERG included deaths among HIV- were endemic or where there was infected persons, there would have been sufficient data. Further data on burden of substantial additional deaths due to enteric diseases from low- and middle- invasive non-typhoidal Salmonella spp. income settings, particularly high quality deaths, some of which could presumably epidemiological data, are needed to be averted by improvements in food improve our understanding of foodborne safety. For some other pathogens in our diseases [40, 205]. estimates, such as L. monocytogenes and Cryptosporidium spp., due to a 6.3 Parasites lack of reliable data we were unable to DISCUSSION In this study, we estimate for the first account for the excess mortality due to time the disease burden imposed by HIV infection. foodborne parasites. The results highlight Another important limitation of our the significant burden in low- and attempt to quantify the disease burden middle-income countries, where cycles due to foodborne disease is the inherent of parasitic infection are highly specific difficulty in estimating the proportion of to food sources. In addition to those illness acquired from food [7]. We relied detailed here, a further 357 million cases, on a structured expert elicitation study. 33 900 deaths and 2.94 million DALYs We were unable to estimate differences are due to enteric protozoa, of which 67.2 in mode of transmission by age, despite million cases, 5560 deaths and 492 000 this potentially being important. Expert DALYs are attributable to foodborne elicitation studies can result in highly transmission (see [168] and Table 1), WHO Estimates of the global burden of foodborne diseases 105

completing the picture for the foodborne in the present study. The former study parasitic diseases, given data available. [206] also undertook age weighting and discounting, which we decided not to We used the best evidence available incorporate into this study. In addition combined with the natural history of different life tables were used. Our use the disease to obtain estimates of the of DWs was guided by GBD2010 and incidence, mortality and sequelae of the results of a systematic review of each parasitic disease. Several of the the clinical manifestations of CE [75]. diseases were included in GBD2010 [81]. However, a median estimate in excess In a number of cases our estimates for of 188 000 cases of CE per year, with the global burden of disease differ quite the possibility of up to 1.77 million new substantially from those of GBD2010. cases, indicates a substantial burden. The estimate for echinococcosis (which With a low case-fatality rate the burden combined AE and CE in one estimate) in terms of DALYs is highly dependent on in GBD2010 is 144 000 DALYs [81]. the DW and duration of illness. Neither of This is less than a fifth of our combined these is defined with certainty. The lack median estimate of 871 000 DALYs. This of defined DWs specific for the differing discrepancy probably reflects different sequelae of CE must be seen as a major methodologies between the two studies. data gap. GBD2010 relied heavily on vital records for mortality attributed to these diseases, When arriving at the estimates for AE, whereas we used an approach based it was assumed that in excess of 90% on the natural history of the disease. of cases outside of Europe would be Our choice of approach was strongly fatal. This assumption was supported by influenced by the chronic nature of these survival analyses, confirming that in the diseases, and that often only prevalence absence of aggressive treatment of this data were available. In addition, these disease, including chemotherapy, most diseases often have their highest impact cases die [207, 208]. Our results suggest in low income countries, where vital it is possible that the global burden of AE records are likely to be poor and hospital may be somewhat higher than that of CE, treatment unavailable. Our estimates for which may at first sight seem surprising the global burden of CE would arguably as there are many more cases of CE be more consistent with an earlier globally and the parasite has a more estimate [206], if there had been no cosmopolitan distribution. Hence, we substantial methodological differences. have a median estimate of CE incidence The earlier report suggested a median that is ten times higher than the median estimate of 285 000 DALYs assuming no estimate of AE incidence. The high case under-reporting, rising to 1 million DALYs fatality ratio of AE, results in the loss where under-reporting was assumed. The of 37 DALYs per case compared with earlier report also used DWs ranging from 0.98 DALYs for each case of CE. Thus 0.2 to 0.809, depending on the severity the global burden of AE was driven by of the disease. In the present study we the large number of YLLs. For CE it was used a maximum DW of 0.221, and that driven by the YLDs. was only applied to the relatively small Our estimates for cysticercosis were number of neurological cases. higher than that of GBD2010, as a result Echinococcosis of the abdominal organs of assigning a substantial proportion – the most common presentation – had a of epilepsy burden to cysticercosis, DW of 0.123 for treatment-seeking cases based on the results of a systematic Discussion 106

review [74]. Furthermore, a subsequent and where human food habits suggest systematic review has largely confirmed transmission to humans to be likely. We our findings in terms of the fraction tried to correct for this lack of data by of epilepsy attributable to NCC [209]. imputing incidence rates for all countries However, our results are not inconsistent with at least one autochthonous human with GBD2010 [81] because we have infection reported in the reviewed allocated some of the burden from literature. Nevertheless, and in line with epilepsy to a specific aetiological agent. the original study [78], very conservative Nevertheless, the present estimate in this estimates from the imputation were report may still underestimate the burden accepted in an attempt to avoid inflating of cysticercosis, as there are other the burden estimates for human important clinical symptoms associated foodborne trematode infections based on with neurocysticercosis, such as chronic unclear evidence. headache, hydrocephalus, stroke and Some diseases, such as toxoplasmosis, depressive disorders [73]. Better were not estimated in GBD2010 and estimates of the role that cysticercosis will inevitably have been included plays in stroke and depressive in other syndromes. For example, disorders globally could considerably congenital defects in GBD2010 will have increase its burden estimates, since incorporated the DALYs for congenital these conditions are ranked third and toxoplasmosis that we have estimated in eleventh, respectively, in the GBD2010 the present study. [81] estimates. It is also unclear how GBD2010 arrived at their estimates It can be argued that congenital for cysticercosis. If, for example, it was toxoplasmosis is a vertically transmitted assumed that cysticercosis-related disease rather than foodborne. However epilepsy can only be attributed in public health measures are largely individuals who are serologically undertaken to prevent maternal (i.e. positive for cysticercosis, this would horizontal) infection, which will, as a lead to substantive underestimates. A consequence, reduce the risk of foetal large proportion of cases of epilepsy infection. There is relatively little evidence attributed to cysticercosis, as shown that treatment to prevent vertical by imaging studies, are nevertheless transmission (such as antiprotozoal sero-negative. For example Montano treatment of acutely infected pregnant et al. [210] describes 15 cases of women) is effective in reducing disease epilepsy aetiologically confirmed as burden [211].Thus it was considered a DISCUSSION neurocysticercosis, but only 7 of these horizontally transmitted infection to the were sero-positive. mother, although the burden of disease is suffered mostly by the foetus, following Likewise, the estimates for the burden of subsequent vertical transmission. foodborne trematode infections may also Accordingly the proportion of foodborne represent under estimates. Our estimates disease suffered by the foetus is the were based on the results of an earlier proportion of the horizontal transmission study, which used estimation methods to susceptible women that occurs that were conservative [78]. Often, through food. population-level information on human foodborne trematode infections were With the exception of NCC, we have used completely lacking from areas where an incidence approach to estimating the parasites are endemic, as indicated the YLDs. This is where the YLD part of by substantial rates of animal infections the DALY was estimated from number WHO Estimates of the global burden of foodborne diseases 107

of incident cases per year multiplied by methods to deal with data deficiencies. the DW and duration. This is in contrast Until these are published we will only be to the GBD2010 approach, which used able to hypothesize the reasons for some a prevalence approach to YLDs, where of the differences in the estimates. YLDs were estimated by number of The limitations in this study are similar prevalent cases multiplied by the DW. to others in this series. There were often For acute disease in generally stable substantial data gaps that had to be epidemiological situations (i.e. no filled by imputation and suffer from considerable shifts in the epidemiological the uncertainties that surround such key indicators of prevalence, incidence, models. Excluding stillbirths is consistent duration, severity, remission and with the approach used to estimate the mortality) and settings with more or burden due to enteric pathogens [168]. less stable population size, the approach Congenital toxoplasmosis is the only makes little difference [2]. But for chronic pathogen investigated that could result diseases in populations that are rapidly in a substantial incidence of stillbirths. increasing, the prevalence approach may However, an estimate for the burden underestimate the numbers of YLDs. of congenital toxoplasmosis, which Parasitic diseases are often chronic includes stillbirths as equivalent to and are often of highest incidence in neonatal deaths, has been reported as low income countries with increasing 1.2 million DALYs per annum [76]. FERG populations. Many parasitic diseases has assumed that acquired toxoplasmosis have durations of many years, or in usually results in a relatively mild acute the case of congenital toxoplasmosis, illness, with some cases suffering the sequelae are usually lifelong. Thus, for a few months [212]. Although as we adopted the GBD2010 data for fatal cases have been recorded [213], epilepsy to estimate the burden of NCC, these were assumed to be uncommon the YLDs will be prevalence-based. and hence zero YLLs were estimated. Nearly all of the burden of NCC is in We have also assumed that although low income countries, which usually acquired chorio retinitis occurs following have increasing populations. Therefore toxoplasmosis, it only occurs in a small the cohort at the time of infection, with proportion of cases (see Appendix 4). the burden attributed in an incidence- This results in approximately 1.15 million based approach, will be larger than DALYs in 2010 from an estimated 20.7 earlier cohorts that are still affected by million people having clinical disease NCC but are reported in the prevalence- following exposure to the pathogen for based approach. Accepting this the first time. However, there is increasing limitation means that the estimates for evidence that acquired toxoplasmosis epilepsy attributed to NCC will result in may result in a number of neurological a further under-estimate of the burden or psychiatric diseases, such as of cysticercosis. schizophrenia and epilepsy. In GBD2010 We have summarized the differences these diseases resulted in 15.0 million between the estimates for GBD2010 and and 17.4 million DALYs, respectively. the FERG estimates for these pathogens, From two meta-analyses [214, 215] and including the enteric protozoa in Table 9. a large cross-sectional study conducted In addition, an issue that appears in China [216], it is possible to estimate common to many hazards is that that the population-attributable GBD2010 [9] has not published many of fraction of schizophrenia associated the search strategies used, or modelling with seropositivity to toxoplasmosis is Discussion 108

approximately 9%, which on a crude detect and remove the parasite from the level could account for approximately 1.3 food chain [217]. Likewise, other cestode million additional DALYs. zoonoses, where the adult tapeworm is located in the (e.g. There were also some notable omissions spp.) with few clinical from our study. Taenia saginata, signs, were also excluded. In contrast, which causes human taeniosis and trichinellosis was considered to be an is transmitted solely from beef, was important foodborne pathogen with not considered because the parasite potentially serious disease. However, produces very mild, unapparent clinical this study has suggested that the global disease in affected humans, which would burden of trichinellosis is small. This is result in a DW of close to zero and hence discussed elsewhere [84]. For reasons a very low burden of human disease. of resource limitations, we were not able However, it is accepted that this parasite to consider foodborne Chagas disease, generates substantial economic damage although it was suggested as a possible because of meat inspection and trade priority pathogen during the second regulations required in many countries to FERG meeting.

Table 9. Comparisons of the total burden of parasitic diseases (foodborne and non-foodborne) with 95% uncertainty intervals, estimated by FERG and by GBD2010 [9]

HYPOTHESIZED REASONS FOR DIFFERENCES PARASITE GBD FERG BETWEEN GBD2010 AND FERG ESTIMATES 8 372000 2 159 331 Cryptosporidium Differences in DALYS estimated by GBD2010 and FERG are (6 473 000– (1 392 438– spp. largely due to differences in how aetiological-specific deaths were 10 401 000) 3 686 925) estimated. FERG estimated aetiology specific deaths using the methodology adopted by CHERG*[35]. GBD2010 used a modelling- 2 237 000 515 904 Entamoeba spp. based approach to estimate aetiology-specific deaths, but there is (1 728 000– (222 446– () no description of the GBD2010 model available to review. GBD2010 2 832 000) 1 552 466) has not published the studies included, their search strategy, nor 171 100 modelling methods; until these are published it is not possible to Giardia spp. Not estimated (115 777–257 315) completely compare GBD2010 and FERG estimates. 1 684 414 Toxoplasma Assumed to be included in congenital diseases and non-specific Not estimated (1 236 005– gondii communicable diseases in GBD2010. 2 452 060) 183 573 Echinococcus GBD2010 used vital records, which are often missing in low (88 082– resource countries. FERG used a natural history approach based granulosus 152 000 1 590 846) on surveillance data. GBD2010 used prevalence-based YLDs, which (60 000– will underestimate burden for a chronic disease like echinococcosis. 359 000) 687 823 Echinococcus Methods for imputation of missing data were different. GBD2010 has

(409 190– DISCUSSION multilocularis 1 106 320) not published their modelling methods for missing data. GBD2010 used vital records relying on a diagnosis of cysticercosis. 514 000 2 788 426 FERG assigned a substantial proportion of the epilepsy envelope to Taenia solium (398 000– (2 137 613– cysticercosis in resource-poor, pork-consuming communities, based 650 000) 3 606 582) on evidence from a systematic review and meta-analysis. GBD2010 has not published their modelling methods for missing data. 1 315 000 1 317 535 Only subtle differences as FERG and GBD2010 used the same source Ascaris spp. (713 000– (1 182 187– data, but FERG estimated incidence-based YLDs whereas GBD2010 2 349 000) 2 700 572) used prevalence-based. 550 Trichinella spp Not estimated (285–934) 1 875 000 2 024 592 Only subtle differences as FERG and GBD2010 used the same source Foodborne (708 000– (1 652 243– data, but FERG estimated incidence-based YLDs whereas GBD2010 Trematodes 4 837 000) 2 483 514) used prevalence-based.

Notes: *Child Health Epidemiology Reference Group of the WHO/UNICEF. WHO Estimates of the global burden of foodborne diseases 109

However, particularly recently, the The expert elicitation for routes of assumption that Chagas disease is transmission estimated that a median primarily a vector-borne disease is being of approximately 15% (95% UI 7–27%) questioned [218]. For example, 70% of of Giardia infections were transmitted cases of acute Chagas disease recorded via contaminated food. This is was in Brazil between 2000 and 2010 were higher than we expected for this enteric associated with food consumption protozoan. For example, Scallan et [219]. As GBD2010 made an estimate al. [188] suggested that 7% of Giardia of the burden of Chagas disease of infections acquired in the United States 546 000 DALYs [81] there could be a of America were of foodborne origin. significant additional burden through However, in contrast, a recent 40-year foodborne transmission if these data summary of outbreaks of giardiosis are representative. Indeed, foodborne reported to the United States Centers for Chagas disease may turn out to have Disease Control and Prevention identified a higher burden than the foodborne that 16% of 242 outbreaks were true burden of some of the pathogens FERG results of foodborne transmission [222]. considered, such as Trichinella and Both these studies suggested that the Giardia spp. proportion of foodborne giardiosis is within the 95% uncertainty limits of our FERG was also unable to estimate the study. Furthermore, a recent report by burden of foodborne cyclosporosis. the Food and Agriculture Organization This has caused outbreaks in the United of the United Nations (FAO) and WHO States of America, such as the multi- presented a multi-criteria ranking of 24 state outbreak of 631 cases in 2013 [220]. [groups of] foodborne parasites, and However, the total numbers of cases concluded that Giardiosis was the 11th over the medium to long term appears most important foodborne parasite [223, to be quite small, with a median annual 224], with fresh produce likely to be the incidence of 0.03 cases per 100 000 vehicle of transmission. This indicates [221]. Thus any contribution to the that it is accepted that this parasite has burden of disease by this pathogen is a foodborne transmission route and puts likely to be small. our estimates in this context. A further important limitation was We used epilepsy and ascaris prevalence relying on expert elicitation for the data from GBD2010 to inform our proportion of disease that is foodborne. estimates of cysticercosis and foodborne This was an important issue with those ascariosis, respectively. Therefore the parasitic diseases such as ascariosis, accuracy of our estimates will be limited toxoplasmosis and echinococcosis, to the accuracy of the GBD2010 data that can have several pathways of from which it was derived. transmission. Expert elicitation studies can result in a highly variable proportions Toxoplasma gondii is globally attributed to food. However, as data distributed, with a high proportion of on source attribution for a number the world population estimated to be of parasites were not available, the seropositive. Ascaris spp. is the most structured elicitation undertaken offered frequently encountered human helminth, a transparent way of evaluating and although the burden is confined to enumerating this uncertainty, and thus low- and middle-income countries. represents the best available source of However, a number of diseases had information [156, 168]. very high burdens limited to distinct Discussion 110

geographical populations. Most of the The report by FAO/WHO presented a global burden of AE is in China, and ranking of foodborne parasites, based mainly on the Tibetan plateau [72]. In on multi-criteria analysis [224]. In our this highland region there are specific study, we present data on the foodborne factors that promote transmission disease burden for 13 parasites included between wildlife, dogs and humans in the FAO/WHO report. Comparing the that are not present in other endemic results of the ranking from the FAO/WHO areas. This results in large numbers of model with the results of the present human cases in certain communities study, the parasites selected by FERG [225]. Such unique epidemiological had the highest rank orders in the FAO/ conditions are not present elsewhere, WHO report (i.e. ranking from #1to #14), even where the parasite is endemic. only Trypanosoma cruzi at rank #11 and Taenia solium transmission can only be at rank #13 were maintained where pork is consumed, not assessed by FERG. Taenia solium pigs are left roaming, and where there was ranked #1 by both approaches and is poor sanitation. Thus it is largely Toxoplasma gondii #3 by FERG and #4 absent from upper-income countries by FAO/WHO. There were, however, also and from communities where pork is remarkable differences in the ranking of not consumed, such as countries in the individual parasites. Paragonimus spp. the Middle East. Sporadic cases are was ranked #2 by FERG, but only #14 in occasionally reported and these are often the FAO/WHO report, and E. granulosus #12 by FERG, but #2 by FAO/WHO. The linked to the employment of immigrants disease burden of E. multilocularis was who originate from endemic countries considerably higher than the burden of E. and hence transmit the infection through granulosus (310 000 vs. 40 000 DALYs), poor hygienic practices [226]. Foodborne but nevertheless was ranked lower at trematodes also have a limited #3 by FAO/WHO. The disease burden distribution, but they cause a high burden of intestinal flukes was #9 by FERG. of disease in the at-risk populations such This was higher than the #22 ranking of as in South-East Asia. Trematodes have heterophyidae by FAO/WHO. FAO/WHO complex life cycles that include various used 9 criteria for ranking, of which 6 species of molluscs. This limits their were health-related criteria and 3 non- distribution to specific regions where health criteria. This weighting of the suitable life-cycle hosts are endemic, different criteria may be responsible for which may be adapted to specific the FAO/WHO report having a different DISCUSSION climatic and hydrological conditions ranking order for the various parasites. [227]. The human disease is further For example, E. granulosus has a global limited to populations that are likely to distribution and a relatively important consume the raw fish or undercooked measure in the FAO/WHO ranking. In aquatic vegetables that are the sources contrast, E. multilocularis is only found in of transmission. Consequently, although the northern hemisphere. we are reporting the global burden of these parasitic diseases, this is often borne almost completely by relatively 6.4 Chemicals small populations in limited geographical The assessment of the burden of disease areas. Therefore, in such communities, from chemicals in food is a challenge these diseases have a major impact on on several levels. There are thousands the health of the population. of chemicals in production and many WHO Estimates of the global burden of foodborne diseases 111

naturally occurring toxins. How many of the disease that is attributable to the of these chemicals and toxins make it exposure. A probabilistic version of this into the food supply is unknown. The method, which is applied in chemical risk health effects of chemicals may not be assessment, was used for dioxin [127, 128]. observed for years following exposure The two approaches would result in (e.g. aflatoxin and liver cancer; lead and the same outcome if perfect data were ). Longitudinal available, and if it can be assumed that studies of these effects are expensive and the risk of exposure to a chemical is time-consuming. Sufficient information is additive to the background risk from available, however, to make estimates of other causes. In reality, the available data the burden for arsenic, cadmium, methyl for both approaches are limited and mercury and lead, and possibly for other there is insufficient information to decide chemicals and toxins (e.g. fish toxins, conclusively whether risks are additive, aristolochic acid). Other chemicals (e.g. multiplicative or otherwise. This may result Persistent Organic Pollutants) may not in considerable discrepancies between require elaborate epidemiological studies results from these methods. In this study, because the burden can be derived from bio-monitoring data in combination with FERG chose a “top-down” approach for relevant toxicity data. Estimates of the aflatoxin because the cancer potency burden for these chemicals will provide a factor derived by JECFA [111] was based much more comprehensive understanding on a multiplicative model, and there is of the impact that chemicals in the food evidence for a high background rate in the supply have on the burden of disease. study population underlying this estimate and the global population (see Appendix As the relevant disease endpoints due 4). Using the population-attributable to foodborne chemicals may arise from fraction approach, it was estimated different causes, various approaches there were approximately 22 000 (95% are possible for estimating incidence UI 9 000–57 000) cases of aflatoxin- and mortality. A “top-down” approach related HCC in 2010. A dose-response uses an existing estimate of morbidity approach [110] estimated that, annually, or mortality of the disease endpoint by 25 200–155 000 cases of HCC might all causes (the “envelope”) as a starting be attributable to aflatoxin exposure. point. A population-attributable fraction Even though the uncertainty intervals is then calculated for the hazard under overlap, the differences between these consideration, and applied to the envelope two approaches are considerable. There to estimate the hazard-specific incidence. is evidence for a high background rate This method, which is the standard in in the study population underlying this Global Burden of Disease estimations, was estimate and the global population (see used for aflatoxin. Appendix 4), which may result in over- A “bottom-up” or dose-response approach estimation of mortality by the dose- uses dose-response and exposure response approach. In contrast, the information. The approach begins with global liver cancer envelope may be selection of the appropriate dose-response underestimated, particularly in Africa relationship between the chemical and [228, 229], leading to under estimation the particular disease. This dose-response of the aflatoxin-attributable incidence. relationship is then combined with the Hence, there is considerable data and distribution of exposure within a population model uncertainty in our estimates, which to derive an estimate of the incidence should be addressed by further studies. Discussion 112

DISCUSSION WHO Estimates of the global burden of foodborne diseases 113

COUNTRY STUDIES 7

WHO Estimates of the global burden of foodborne diseases 115

COUNTRY STUDIES

7.1 Aim and Objectives of To specifically address the second the Task Force objective above, a subgroup, the Knowledge Translation and Policy Group The WHO initiative to estimate the (KTPG), was established in 2010. global and regional burden of foodborne diseases has four stated objectives, two of which involve actions at a 7.2 Tools and resources to facilitate national level: national burden of foodborne di- sease studies f To strengthen the capacity of countries in conducting burden of foodborne The initial activity by the CSTF was to disease assessments, and to increase develop of a series of tools and resources the number of countries that have to facilitate national burden of foodborne undertaken a burden of foodborne disease studies. These were intended to disease study. promote a methodology that would be f To encourage countries to use burden consistent with the global and regional of foodborne disease estimates for burden estimates being developed by cost-effective analyses of prevention, FERG, in particular estimating burden intervention and control measures. using the disability-adjusted life- The Country Studies Task Force (CSTF) year (DALY) metric, and strengthen was established in 2009 to advise WHO capacity to develop science-based on the initiation, conduct and completion policies. These tools and resources were of national burden of foodborne developed by members of the CSTF, diseases studies. as well as commissioned scientists, to be made available on a WHO website The objectives of the Task Force were to dedicated to the burden of foodborne advise WHO on: disease initiative. The tools and f the development of burden of resources included: foodborne disease pilot protocols that f reviews of existing burden -of disease can be used by countries to estimate studies and protocols [230, 231]; their national burden of foodborne f a manual on how to conduct a national disease from enteric pathogens, burden of foodborne disease study parasites and chemicals and toxins; (adapted from the WHO manual f the development or commissioning of on national burden -of disease all relevant training materials needed estimation [232]); to assist countries to build capacity f a hazard selection tool, including and undertake a national burden of a listing of priority hazards being foodborne disease study; addressed by the WHO initiative at the f oversight of the initiation, conduct global and regional levels, and guidance and completion of an agreed number for identification of hazards that may of national foodborne disease be locally important; pilot studies; f guidance on data collection, describing f the evaluation of the protocols after the information needed to estimate the pilot studies are completed, and on foodborne burden-of disease, and making necessary revisions; and potential sources of data, such as f oversight of the initiation, conduct and surveillance systems, demographic completion of 18 national burden of databases, etc. This tool also suggests foodborne disease studies, 3 in each contextual information that helps to WHO region. assess data quality; and a Country studies 116

f FERG Situation Analysis/Knowledge 7.4 Process Translation/Risk Communication Manual Each pilot country was asked to (SA/KT/RC Manual).1 The development assemble a team to conduct their study. of this resource benefited from previous The members of these teams included burden of food- and waterborne representatives from government and disease studies in the Caribbean, under academic institutions. Early in the the auspices of the Pan American process, KTPG recommended that each Health Organization (PAHO) [233]. study team conduct a situation analysis A WHO Global Foodborne Infections according to the guidelines in the SA/ Network capacity building workshop KT/RC Manual, to describe the regulatory in July 2012 resulted in the creation of and economic status of food safety in 13 issue briefs, with context-specific the country; identify actors, policies and target audiences, and immediately practices; and generally provide context implementable recommendations. The for the scientific data. This analysis would template for these issue briefs was also identify stakeholders who should be included in the guidance manual aware of the study, could contribute data (Dr Enrique Perez, PAHO, pers. comm.). and information, and might ultimately use the results of the study to guide 7.3 Pilot Studies decision-making. In 2010, WHO invited countries to The initial step in each study was to express interest in conducting national identify hazards in the food supply that burden of foodborne disease studies as a were relevant to the pilot country. Lists pilot process. Countries which expressed of hazards and associated diseases interest were sent an overview of a that were considered of global and national burden of foodborne disease regional importance by FERG were study from the FERG perspective, and provided; each country was able to add a request for information relevant to hazards considered important from the conduct of the study. Following their perspective. Available information an assessment process undertaken by was then collated on the incidence of the Department of Food Safety and diseases associated with the hazards, Zoonoses (FOS) at WHO headquarters, as well as data on the prevalence of the four countries were selected for pilot hazards in the food supply. These data studies: Albania, Japan, Thailand and were summarized, and then attribution Uganda. A commencement meeting of disease burden to foodborne for the Albanian, Japanese and Thai transmission was considered, as studies was held in November 2011, and data allowed. for the Ugandan study in March 2012. KTPG sought to promote knowledge The studies were supported by ongoing translation and risk communication communication between the countries throughout the development and

and CSTF. STUDIES

implementation of the study. Tools for COUNTRY 1 A situation analysis report or resource is designed these processes, as described in the to collect and summarize the contextual SA/KT/RC Manual, are intended to information concerning food safety in the country undertaking the national foodborne burden of involve stakeholders from the outset so disease study, including policies and practices, as to promote ownership, share results capacities, key agencies and actors in the food with stakeholders, and promote efforts safety system, and to document factors that will affect the development of policies and to use the information for developing their implementation. evidence-based policies. WHO Estimates of the global burden of foodborne diseases 117

Here we provide a brief overview of key and Environment, is responsible for data and food safety systems associated general hygiene and sanitation across with each pilot study. As enteric disease all businesses, including food-related is a common outcome of exposure to businesses. The Ministry of Agriculture, microbial foodborne hazards, there is a Food and Consumer Protection focus on data related to enteric disease. Food Safety Directorate includes the National Food Authority (NFA), which 7.4.1 Albania is responsible for official control, risk Human health surveillance of foodborne assessment, and communication. Official diseases in Albania is led by the Public control involves the inspection of food Health Institute within the Ministry of production hygiene, and certification Health, which collates data supplied by of hazard analysis regional departments of public health. (HACCP)-based systems. An early warning surveillance system Data on the prevalence of hazards in operates across all of Albania (similar the food supply are limited. Official to the system that operates in Serbia monitoring programmes for shellfish and Macedonia [234]), and the case (algal toxins and Escherichia coli) have definitions are the same as for syndromic been in place since 2005 to support surveillance under the International exports to the European Union. Health Regulations. Key indicators of foodborne disease are the annual rates 7.4.2 Japan of reported gastro intestinal illness The major objectives of the Japanese (approximately 56 000 cases per year, country study were to assess the disease approximately 2 000 cases per 100 000 burden from major foodborne diseases population) and cases reported as food in Japan and to analyse the policies poisoning (approximately 2800 cases on foodborne disease using the FERG per year, approximately 100 cases per framework. The study has now been 100 000 population). Food poisoning published [238]. cases are reported on the basis of assessment by physicians from primary As a pilot study, three major foodborne health care, as well as hospitalized diseases caused by Campylobacter spp., non-typhoid Salmonella spp., and cases. aetiology for cases in these entero haemorrhagic E. coli (EHEC) general disease categories is rarely were prioritized, based on food investigated. Surveillance for parasitic or poisoning statistics in 2011 and an expert viral infections is not routine, apart from consultation. First, the annual incidence infection with Entamoeba histolytica. was estimated from reported surveillance Cross-sectional studies of faecal samples data, adjusted for probabilities of for viral and parasitic infections have case confirmation and physician visits. been carried out (e.g. [235, 236]). The estimated annual incidence was Access to health care is limited, significantly higher than that reported in particularly in rural areas. A lack of the routine surveillance data, suggesting awareness of entitlements, and informal a marked under estimation of the payment systems, mean that 20–30% magnitude of foodborne diseases. of people cannot access primary health A series of systematic reviews of care [237]. disabling sequelae from the three Another section of the Ministry of priority diseases was conducted. Health, the Department of Health Subsequently, the estimated incidence Country studies 118

was adjusted for food-attributable and interoperability of all three health proportions, which were estimated by insurance databases contributed to data an expert elicitation process, similar reliability and ensured entitlement to the to that carried out in the Netherlands health services covered [239]. [33]. Together with the cause-of-death Extrapolations from these data sources data from vital registration, the disease allowed an estimate of the incidence burden in terms of DALYs was estimated. of acute diarrhoea in the community In 2011, foodborne disease caused of 10–35 million illnesses in 2009 by Campylobacter spp., non-typhoid (for the National Notifiable Disease Salmonella spp. and EHEC led to an Surveillance System, acute diarrhoea estimated 6099, 3145 and 463 DALYs in Japan, respectively. The burden from is defined as at least 3 loose stools disabling sequelae was consistently within 24 hours or any abnormal stools higher than that due to gastroenteritis [e.g. watery, with mucous, or bloody]). among the three major foodborne Information on aetiology is limited, but diseases. Data gaps in estimating the incidence of salmonellosis, cholera, foodborne disease burden in Japan, in shigellosis and E. coli infection were particular population-based data on estimated from diagnoses in the National incidence, were also identified. Hospital Record. Building on the FERG framework, the In addition, the prevalence of liver fluke policy situation analysis provided an infection (, a locally overview of the food safety policies and important foodborne hazard transmitted systems in Japan. As a Japan-specific via fish) and the incidence of rotavirus issue, a rigorous policy situation analysis infection have been estimated. of the management of risks associated Food safety regulatory activity in with possible radio active substances in Thailand is led by the Bureau of Food and food, due to the nuclear power plant Water Sanitation, Department of Health, accident in Fukushima after the Great Ministry of Public Health. The popularity East Japan Earthquake in 2011, was of street food has led to the development also completed. of a sanitation standard for vendors.

7.4.3 Thailand 7.4.4 Uganda The Thai country study focused on The Ugandan country study established the incidence of diarrhoeal disease, teams to separately address enteric, using data from the National Notifiable parasitic and chemical hazards, and Disease Surveillance System maintained source attribution [240]. A detailed by the Bureau of Epidemiology of situation analysis was prepared, which the Thai Ministry of Public Health. described the context for food safety These data were supplemented by in terms of legislation, regulatory information from National Hospital authorities, the food supply, production STUDIES

Records (both in-patient and out- COUNTRY and consumption. The Ugandan country patient), the National Health and Welfare study was undertaken in conjunction with Survey, and community-based studies a project by FAO on the use of multi- of young children. In this study, the criteria decision analysis for food safety hospital data accessed all three health in Uganda. insurance systems, including the universal coverage, social security, and civil servant Data were collated from surveillance benefit health insurance. The sharing sources (particularly the Health WHO Estimates of the global burden of foodborne diseases 119

Management Information System in parts of Uganda, no reports of acute administered by the Ministry of Health, , konzo or tropical and the Central Public Health Laboratory) ataxic neuropathy were found. on acute diarrhoea (1.9 million reported It was important that both waterborne outpatient cases in 2012, approximately and foodborne transmission of diseases 5700 cases per 100 000; case definition: were included in the Ugandan study, as three or more watery stools in 24 hours food safety was not considered to be but not lasting for more than 14 days), independent from water safety. It was cholera, dysentery, brucellosis, hepatitis difficult to generate DALY estimates E and typhoid fever. Parasitic infections from the available data, particularly due are reported as worm infections or to the shortage of community-level intestinal worm infections. Although incidence data. such infections are very common (approximately 1.8 million outpatient infections reported annually), aetiological 7.5 Findings and Lessons Learned data are few. 7.5.1 Data gaps The reliability of these data has improved steadily with increased access to A lack of data prevented DALY healthcare since 2000. Uganda has calculations in several of the pilot studies. undergone a number of reforms that The data gaps included: have influenced health service delivery. f information to assign aetiology Among the major reforms, conducted in for important syndromes such as the early 1990s, was the decentralized acute and governance of districts, with attendant parasitic infections; devolution of powers to allocate f data on the incidence of diseases resources and deliver services, including caused by some hazards, particularly health care. Physical access to health chemical hazards; and facilities for the population living within f limited outbreak and other data 5 km of a health facility increased from on which to base attribution for 49% in 2001 to 72% in 2004 [241]. foodborne transmission. Other sources of data included the 7.5.2 Public and private data sources Ministry of Agriculture, Animal Industry In some countries, private hospitals and Fisheries; the Ministry of Trade, provide a significant proportion of the Industry and Cooperatives; the Ministry of available healthcare, and may not have Water and Environment; the Ministry of the same reporting requirements as Local Government and Local Authorities; public hospitals [242]. Engagement and research and academic institutions. with private hospitals and other Of the chemical hazards, the most facilities to provide a complete picture data were available for aflatoxins, of the incidence of diseases caused with information on the prevalence of by foodborne hazards may need to contamination for relevant foods being be specifically addressed. Data from available. The incidence of hepatocellular primary producers and the carcinoma, an important health outcome concerning foodborne hazards can be of aflatoxin exposure, is also available. gathered, but economic implications, Acute poisoning due to methanol in illicit particularly for trade, mean that such alcoholic beverages is often reported. data should be carefully handled and Despite cassava consumption being high with discretion. Country studies 120

7.5.3 Foodborne versus Barriers to knowledge translation include: waterborne disease f Limitations resulting from lack of The separation of food and water data and information. Incomplete as exposure vehicles for attribution information, with associated caveats and uncertainty, may prevent clear purposes is often useful as different conclusions being drawn for policy. regulatory agencies may have f Differing time pressures. Research may responsibility for each source. However, take months or years to complete, at a community level, the differentiation whereas policy-makers usually between food and water may not need to produce decisions in much shorter timeframes. be sensible in terms of how risks are f The weighting of evidence may differ. managed. These issues should be Scientists are likely to value data specifically considered in a national and analysis most highly, whereas burden study. policy-makers may be also influenced by personal experience, anecdotal 7.5.4 Situation analysis and information, political and economic knowledge translation considerations, and other factors. Social scientists, stakeholders and Knowledge translation can be facilitated by: decision-makers need to be included in the study team from the earliest f Strong personal relationships between stages in order to effectively support researchers and policy-makers. Face to face meetings and direct conversations knowledge translation and the can promote trust and credibility, and development of science-based policies. support formal written reports. Their involvement includes developing f Presenting the results of research so a situation analysis (for an example see that they address risk management [243]), and early and continuous efforts questions. Such questions are best formulated and delivered by policy- to recognize and incorporate knowledge makers at the commencement of translation and risk communication to the research, but researchers should audiences identified in the situation always expect to address questions analysis. Differences in experience and of effectiveness, cost, and high perspectives can make collaboration risk groups. between the social scientists and 7.6 Discussion epidemiological/food safety technical participants challenging. The pilot studies of national burdens- of-foodborne disease, initiated by

Knowledge translation and risk WHO, have promoted the importance STUDIES COUNTRY communication are usually specialist of such studies amongst the activities, and require on-going participating countries and disseminated internationally accepted methodology commitment and resources [244]. In for such estimates. Few DALY estimates order to promote uptake of research could be calculated, but this was not results, identified barriers and facilitators unexpected, due to data gaps. The first are described in the SA/KT/RC Manual. attempt at conducting such studies has WHO Estimates of the global burden of foodborne diseases 121

identified challenges in both process and a situation analysis (such as the existing information, including the recognition national food control system). In addition, that data collection and analysis, the WHO initiative sought to foster the development of situation analysis, and knowledge translation of burden of on-going knowledge translation and risk disease data into policy through on- communication, require commitment of going cross-agency communication. time and financial resources. Such activities are best undertaken by The WHO initiative has provided burden people from within a country. of foodborne disease estimates from National burden of foodborne disease global and regional perspectives. These studies, particularly in developing estimates provide context and can fill countries, now have an opportunity to many of the data gaps for individual fill data gaps, and assign aetiology and countries undertaking foodborne attribution to the incidence of foodborne burden-of disease studies. In particular, diseases, using the data from the WHO the provision of aetiology estimates initiative to augment local data. Such for syndromic surveillance data, and local data can also be used as a cross attribution estimates for foodborne check to validate national estimates disease, will be particularly difficult derived from regional estimates. This for studies in developing countries to should allow the generation of at least address individually. preliminary burden estimates to inform The Global Burden of Disease 2010 Study national policy. The effective delivery of (GBD2010), undertaken by IHME, Seattle, this information can be guided by the USA, covers a broad range of disease considerations and tools provided in the and injuries, and has published country- SA/KT/RC Manual. In the longer term, specific estimates for these on its website burden of foodborne disease information [245]. Foodborne diseases are a subset should be a fundamental component of of these estimates, although estimates a systematic approach to food safety, are typically not stratified by transmission such as the risk management framework route. National foodborne disease studies advocated by Codex [246]. Such an as promoted by WHO and FERG include approach can enhance both public health consideration of the national context in and trade. Country studies 122 STUDIES COUNTRY WHO Estimates of the global burden of foodborne diseases 123

CONCLUSION8

WHO Estimates of the global burden of foodborne diseases 125

CONCLUSION

This report presents the first global army of more than a hundred scientists, and regional estimates of the burden specialized in their own fields, it turned of foodborne diseases. The large out to be possible to present the first disease burden from food highlights the ever estimates of the global burden of importance of food safety, particularly in foodborne disease. The process took Africa, South-East Asia and other more eight years and an uncounted number greatly affected regions. Our results of hours. All involved donated their time indicate that some hazards, such as and experience to WHO, finding own non-typhoidal S. enterica, are important sources of funding in addition to the causes of FBD in all regions of the limited means available and invested world, while others – such as certain liberal amounts of personal time. In parasitic helminths and aflatoxin – are particular the Core Group (Task Force of highly focal nature resulting in high chairs and senior advisers) spent their local burden. time in numerous teleconferences at Despite the data gaps and limitations of sometimes highly inconvenient hours, these initial estimates, it is apparent that in particular for the colleagues from the global burden of FBD is considerable, Australia and New Zealand. Initially and affects individuals of all ages, but annual meetings were organized, creating particularly children <5 years of age and momentum and commitment. The global persons living in low-income regions financial crisis inevitably hit FERG, and of the world. By incorporating these much more reliance was placed on estimates into policy development at teleconferences and other means of both national and international levels, remote communication, slowing down all stakeholders can contribute to the process and limiting the involvement improvements in safety throughout the to the Core Group mainly. Nevertheless, food chain. These results will also help to all FERG members and resource advisers direct future research activities. continued to believe in and support the Initiative. 8.1 Reflections on the WHO Initia- The global burden of foodborne disease tive to Estimate the Global Burden was estimated in several distinct steps, building on established methods for of Foodborne Diseases estimating burden, as expressed in When the WHO Foodborne Disease Disability Adjusted Life Years (DALYs). Burden Epidemiology Reference Group First, incidence of food-related diseases, (FERG) first met in September 2007, including some chronic sequelae and they were convinced of the necessity to mortality, were estimated for 31 hazards present estimates of the global burden of that were considered to contribute foodborne disease, but did not yet know significantly to the burden, and for if, and how, it could be done. They were which sufficient data were available. The aware of national studies on the burden hazards included 18 enteric pathogens, 10 of foodborne diseases, but recognized parasitic diseases and 3 toxic chemicals. that attempting a global estimate was For 5 hazards, the data were insufficient a daunting task. The sheer complexity to present global estimates, and data of the problem was challenging: food were presented for high-income regions consumption across the globe is highly only. Next, information was generated diverse and the range of potential on duration and severity of the incident contaminants in the food supply is cases of disease to produce estimates astounding. Yet, with the help of an of Years Lived with Disability (YLD) Conclusion 126

and on the number of Years of Life Lost needs for burden of illness estimates are (YLL) due to premature mortality. Many high, and crucial information was often foodborne hazards are not exclusively lacking, particularly for some of the transmitted by food, and a separate world’s most populous countries, such effort was set up for the attribution of as China, India and . FERG used exposure to different sources, including statistical models and expert input to food, the environment and direct contact estimate some missing data. In particular, between humans or with animals. As Bayesian regression modelling has many data are lacking for attribution, been used to estimate missing disease it was decided to apply structured incidence data. expert elicitation to provide a consistent Due to the limitations in data availability, set of estimates. The global expert FERG decided to present its estimates elicitation study involved 73 experts and on a regional level, even though all 11 elicitors, and was one of the largest, calculations were made on a national if not the largest study, of this kind ever level. The regional estimates are undertaken. Combining all streams of considered more robust as they build data resulted in estimates of the global on data from several countries in most burden of foodborne disease. regions. Yet, the regional estimates Unlike previously completed national do not reflect the diversity of risks burden of illness studies, FERG decided between countries in a region, or even to also include chemical hazards. The within a country. Maps are therefore inclusion of chemical hazards was not presented as it was considered particularly challenging, and it was that these would not adequately reflect only through determined efforts by regional heterogeneity. the Chemicals and Toxins Task Force The results of the FERG project are (CTTF) that several chemical hazards presented in several formats. A PLOS could be included. Whereas WHO collection entitled “The World Health committees such as the Joint FAO/WHO Organization Estimates of the Global Expert Committee on Food Additives Burden of Foodborne Diseases”, which (JECFA) and Joint FAO/WHO Meeting can be accessed at a dedicated website.1 on Residues (JMPR) typically The website presents the key results in use a risk assessment approach, a a series of seven peer-reviewed papers, counterfactual attribution approach and also provides access to a large is commonly applied in global burden and growing number of reviews and estimates of cancer, cardiovascular and description of methods that have been other diseases. Deciding which of these published in different peer-reviewed approaches was most appropriate for journals. This large body of evidence FERG was a difficult, and as yet not fully reflects the considerable support given to resolved, process. As a result, burden FERG by the global scientific community. estimates for several important chemical These papers are also accessible through contaminants (methylmercury, lead, a dedicated WHO website.2 WHO has arsenic and cadmium) are expected to be also produced this report, documenting presented at a later stage. the results and the process of estimating Even though all efforts were made to 1 http://collections.plos.org/ferg-2015 accessed 2 include the best available science in the December 2015 2 http://www.who.int/foodsafety/areas_work/ estimates, FERG is fully aware of the foodborne-diseases/ferg/en/ accessed 3 limitations of the current work. Data November 2015 CONCLUSION WHO Estimates of the global burden of foodborne diseases 127

the global burden of foodborne disease (as documented for aflatoxin, see and an interactive website allowing Section 6.4). stakeholders to explore the results from Countries who want to build their different perspectives. national food safety strategies are Even though the currently presented advised to combine the global estimates burden of foodborne disease is with national data. It is our experience substantial, it was not feasible to that a vast amount of additional data document the full burden, which is likely exist but has not yet been mined because to be considerably higher. Not all relevant it is not available in easily accessible contaminants could be included, and for databases but rather in paper form. those that were included, not all relevant Building on such data may provide endpoints could be taken into account. sources of validation for any estimates FERG selected a shortlist of hazards at derived from FERG numbers. As a next the onset, reducing a list of more than step, further development of national 100 contaminants to 40. Exclusions laboratory-based surveillance programs, were based on initial judgments about should be a priority. the importance of the global or regional A crucial element of the initiative, often burden, but also on data availability. Of taking a back seat during the huge the 40 contaminants selected, analyses effort in generating global and regional have not been completed for lead, burden estimates, was therefore the methyl mercury, arsenic and cadmium promotion of foodborne burden of for inclusion in this report. Of potentially disease studies and capacity building in relevant endpoints, only Guillain-Barré individual countries. FERG was only able syndrome and haemolytic uraemic to make limited progress towards this syndrome and invasive salmonellosis objective, in the form of pilot studies in were included as outcomes for diarrhoeal four countries. Since some of these pilot diseases, but not irritable bowel studies encountered significant resource syndrome or other functional bowel barriers and data shortages, it is hoped disorders that are increasingly linked to that one legacy of the initiative would diarrhoeal disease in developed countries be to help overcome these through local and are associated with a substantial use of regional estimates. Individual burden. FERG estimates do not include countries can evaluate and apply the the effects of foodborne diseases FERG regional burden estimates to on malnutrition and development in generate national DALY-based burden low- and middle-income countries, data for foodborne illness prioritization. and invasive salmonellosis in HIV co- Such a process should include local data morbid cases was also excluded, even for validation where available, and be though a major proportion of these undertaken by local scientists with an infections may be foodborne. No awareness of the food safety context in stillbirths were included for listeriosis their country. FERG has also sought to and toxoplasmosis, but many would promote knowledge translation of burden be preventable by appropriate food of disease estimates into food safety safety interventions. The counterfactual policy at a national level. approach for chemicals produces lower estimates than risk assessment approach Conclusion 128 CONCLUSION WHO Estimates of the global burden of foodborne diseases 129

APPENDICES

WHO Estimates of the global burden of foodborne diseases 131

APPENDIX 1. Formal Description of the Project and Participants

A1.1 Terms of Reference for A1.2 Foodborne Disease Burden WHO’s Foodborne Disease Epidemiology Reference Group Burden Epidemiology Reference The Foodborne Disease Burden Group (FERG) Epidemiology Reference Group The Foodborne Disease Burden (FERG) is composed of internationally Epidemiology Reference Group renowned experts in a broad range of (FERG) will act as an advisory body to disciplines relevant to global foodborne WHO on matters of global foodborne disease epidemiology. Members were diseases epidemiology. appointed by the WHO Director- General, Dr Margaret Chan, following a Functions: transparent selection process. The FERG shall have the The expert group is charged to: following functions: f assemble, appraise and report on f To review epidemiological data on the current, the projected, as well foodborne disease burden. as the averted burden of foodborne f To identify technical gaps and priorities disease estimates; for research activities. f conduct epidemiological reviews for f To make recommendations to WHO mortality, morbidity and disability in on the establishment of FERG TFs and each of the major foodborne diseases; other means through which scientific f provide models for the estimation of and technical matters are addressed. FBD burden where data are lacking; Composition: f develop cause attribution models to estimate the proportion of diseases that f FERG members shall serve in their are foodborne; and, most importantly, personal capacities to represent the f use the FERG models to develop user- broad range of disciplines relevant to friendly tools for burden of foodborne global foodborne disease epidemiology. disease studies at country level. f Members of the FERG, including the Chair, shall be selected by the Director- To estimate the global human health General. burden (expressed in Disability-Adjusted f Members of the FERG, including the Life Years– DALYs), FERG will initially Chair, shall be appointed to serve for a focus on microbial, parasitic, zoonotic period of one year, and shall be eligible and chemical contamination of food with for re-appointment. an emphasis on: Operation: f diseases whose incidence and severity The FERG shall usually meet at least is thought to be high; and twice a year. WHO shall provide any f pathogens and chemicals that are most necessary scientific, technical and other likely to contaminate food, and that support for the FERG, including for the have a high degree of preventability. preparation of meeting reports. FOS shall provide secretarial support. Appendices 132

A1.3 Participants Formerly National Institute for Public Health and FERG Members the Environment f Formally appointed by the WHO Bilthoven Director-General, following a Netherlands selection procedure. Vice-chair f Allocated to Core Group and TFs. 2007 - 2010 f Have full participation rights in all Nilanthi DE SILVA technical discussions. Dean and Professor of Parasitology Resource advisers Faculty of Medicine, University f Not formally appointed by the of Kelaniya Director General. Ragama, Sri Lanka f Allocated to TFs on an ad hoc basis (as required). 2011 - 2015 f Have full participation rights in Alejandro CRAVIOTO technical discussions. Global Evaluative Sciences Seattle, WA. WHO Secretariat and other USA UN Organizations Members f Have full participation rights in technical discussions. Gabriel O Adegoke f Allocated to TFs on an ad hoc basis. Department of Animal Science, National University of Lesotho Observers Lesotho f Nominated by FERG members (one per member). Reza Afhshari f No ‘formal’ right of intervention Head, Development of Research and in plenary. Education Development f Participation in TFs, as appropriate. Mashhad University of Medical Sciences Stakeholders Iman Rez Hospital (Islamic Rep.) f Invited by WHO to designated sessions. f Formal right of intervention in Frederick J. ANGULO designated sessions. Associate Director for Science f No participation in technical discussions Division of Global Health Protection to avoid conflicts of interest. Center for Global Health, CDC, Atlanta GA USA A1.4 Members of the Foodborne Disease Burden Epidemiology Janis BAINES Reference Group (FERG) (past Manager and present) Food composition, Evaluation and Modelling Section Chair Food Standards Australia New Zealand Arie HAVELAAR Canberra BC ACT Emerging Pathogens Institute Australia University of Florida Gainesville, FL USA APPENDICES WHO Estimates of the global burden of foodborne diseases 133

Kalpana BALAKRISHNAN Department of Clinical Medicine, School Professor and Head, Department of of Environmental Health Engineering, University of Wisconsin– Madison, Director, WHO Collaborating Center for WI, USA Occupational Health Sri Ramachandra University John EHIRI India Professor and Director Division of Health Promotion Sciences in David C. BELLINGER the Mel and Enid Zuckerman College of Professor of Neurology, Harvard Medical Public Health School, and University of Arizona, Tucson, AZ Professor in the Department of USA Environmental Health, Harvard School of Public Health Aamir FAZIL Neuroepidemiology Unit, Risk Assessment Specialist and Children’s Hospital Environmental Engineer Boston, MA Public Health Agency of Canada, Guelph, USA N1G 5B2, Canada

Wan Mansor BIN HAMZAH Catterina FERRECCIO Disease Control Division Profesora Titular Ministry of Health Departamento de Salud Publica, Facultad Federal Government Administrative de Medicina Pontificia Universidad Complex, Putrajaya, Malaysia Catolica de Chile, Chile

Robert BLACK Eric FÈVRE Edgar Berman Professor and Chairman, Chair of Veterinary Infectious Diseases; Department of International Health International Livestock Research Institute The John Hopkins University Bloomberg PO Box 30709–00100, Nairobi, School of Public Health Institute of Infection and Global Health Baltimore, MD University of Liverpool, Leahurst Campus USA Neston, UK

Wanpen CHAICUMPA Neyla GARGOURI Professor Emeritus Director, Medical Affairs Dept. Parasitology Hikma Pharmaceuticals Faculty of Medicine Siriraj, Amman, Mahidol University Bangkok, Thailand Herman J. GIBB Gibb Epidemiology Consulting LLC Brecht DEVLEESSCHAUWER Arlington, VA USA Global Food Safety and Zoonoses Emerging Pathogens Institute and Tine HALD Department of Animal Sciences Head of Epidemiology and Risk Modelling University of Florida, FL, USA Division of Epidemiology and Microbial Genomics Dörte DÖPFER National Food Institute, Technical Farm Animal Production Medicine Group University of Denmark, Søborg, Denmark Appendices 134

Gillian HALL Xiumei LIU Senior Lecturer Technical Consultant National Centre for Epidemiology and China National Center for Food Safety Population Health Risk Assessment College of Medicine / Health Sciences Ministry of Health Australian National University Beijing, People’s Republic Of China Canberra, ACT, Australia Ben MANYINDO Fumiko KASUGA Deputy Executive Director Director Uganda National Bureau of Standards National Institute of Health Sciences Uganda Ministry of Health, Labour and Welfare , Japan George NASINYAMA Associate Professor of Epidemiology and Karen Helena KEDDY Food Safety, Senior Consultant, and Head of the Department of Centre for Enteric Diseases, and Preventive Medicine, and Director National Institute for in charge of Research, Innovations and Communicable Diseases Knowledge Transfer Sandringham, South Africa Makerere University, Kampala, Uganda

Martyn KIRK, Pierre ONGOLO-ZOGO Associate Professor Head, Centre for Development of Best Convener, Master of Philosophy in Practices in Health Applied Epidemiology (MAE) Yaoundé Central Hospital National Centre for Epidemiology and Yaoundé, Cameroon Population Health The Australian National University, ACT John PITT Australia Honorary Research Fellow Commonwealth Scientific and Industrial Robin LAKE Research Organization (CSIRO) Institute of Environmental Science and Australia, North Ryde, Research (ESR) Ltd NSW Australia Christchurch, New Zealand Nicolas PRAET Claudio F. LANATA Institute of of Antwerp Senior Researcher and Professor Antwerpen, Belgium Instituto de Investigación Nutricional Lima Mohammad Bagher ROKNI Professor, Department of Medical Haichao LEI Parasitology and Mycology Deputy Director-General School of Public Health and Institute of Beijing Municipal Health Bureau Public Health Research China Teheran University of Medical Sciences Tehran, Islamic Republic Of Iran APPENDICES WHO Estimates of the global burden of foodborne diseases 135

Niko SPEYBROECK Philippe VERGER (later WHO staff) Institute of Health and Society (IRSS) Head, Research Unit Faculty of Public Health (FSP) National Institute for Université catholique de Louvain Agricultural Research Brussels, Belgium Paris, France

Banchob SRIPA Arve Lee WILLINGHAM (later WHO staff) Tropical Disease Research Laboratory Head, WHO Collaborating Centre for Department of Pathology Parasitic Zoonoses Faculty of Medicine, Khon Faculty of Life Sciences, University Kaen University of Copenhagen, Khon Kaen, Thailand Frederiksberg, DENMARK

Paul TORGERSON Xiao-Nong ZHOU Section of Epidemiology Professor and Deputy Director Vetsuisse Faculty, National Institute of Parasitic Diseases Zurich, Switzerland Chinese Center for Disease Control Shanghai, People’s Republic Of China Rolaf VAN LEEUWEN Professor in Food Toxicology Center for Substances and Integrated Risk Assessment National Institute for Public Health and the Environment (RIVM) Bilthoven The Netherlands

Resource/Technical Advisers and Commissioned Scientists Deena ALASFOOR Laura BOEHM Director of Nutrition Research Assistant Ministry of Health, Oman Department of Mathematics, Statistics and Computer Science Aden ASEFA St. Olaf’s College Center for Food Safety and Northfield, MN 55057, USA Applied Nutrition US FDA, USA Philip Michael BOLGER, Senior Managing Scientist Willie ASPINALL Center for Chemical Regulation and Food Cabot Professor in Natural Hazards and Safety Exponent Risk Science Annapolis, MD 21401, USA School of Earth Sciences University of Bristol, UK Eric BROWN Chief Molecular Methods and Subtyping Branch, Division of Microbiology US FDA, USA Appendices 136

Robert L. BUCHANAN, Thomas FUERST Director and Professor Swiss Tropical and Public Health Institute Center for Food Safety and Switzerland Security Systems University of Maryland, MD 20742, USA Elissavet GKOGKA Wageningen University Christine BUDKE The Netherlands Assistant Professor of Epidemiology Department of Veterinary Integrative David GOLDMAN Biosciences, and College of Veterinary Assistant Administrator Medicine and Biomedical Sciences Office of Public Health Science Texas A&M University, USA. Food Safety and Inspection Service, USA

Sandy CAMPBELL Juanita HAAGSMA Knowledge Translation Specialist Erasmus Medical Centre Taos, NM, USA Department of Public Health 3000 CA Rotterdam, The Netherlands Alessandro CASSINI European Centre for Disease Prevention Aron HALL and Control (ECDC) Epidemiologist, Division of Viral Diseases 171 83 Stockholm, National Center for Immunization and Respiratory Diseases Julie CLIFF CDC, MS A-34, Atlanta, Georgia Health Alliance International 30333, USA Eduardo Mondlane University, Mozambique Olga HENAO Epidemiologist Dana COLE National Center for Zoonotic, Centers for Disease Control and Vectorborne and Enteric Diseases Prevention (CDC) CDC, Mailstop D-63, Atlanta, Georgia Atlanta, Georgia 30333, USA 30333, USA

Roger COOKE Brianna HIRST Chauncey Starr Chair for Risk Analysis Research Assistant Resources for the Future, USA Department of Mathematics, Statistics and Computer Science Amélie CREPET St. Olaf’s College, Northfield, MN French Agency for Food Safety, and 55057, USA Environmental & Occupational Health Safety (ANSES) Sandra HOFFMANN France Economic Research Service USDA, Washington, DC, USA John A CRUMP McKinlay Professor of Global Health and Co-Director Centre for International Health Dunedin School of Medicine University of Otago, Dunedin 9054, New Zealand APPENDICES WHO Estimates of the global burden of foodborne diseases 137

Helen H. JENSEN Charline MAERTENS DE NOORDHOUT Department of Economics Institute of Health and Society (IRSS)- Food and Nutrition Policy Division Faculty of Public Health (FSP) Center for Agriculture and Rural Université catholique de Louvain Development (CARD) Clos Chapelle-aux-champs, 1200 Brussels, Iowa State University, Ames, Iowa 50011– Belgium 1070, USA Shannon MAJOWICZ Nasreen JESSANI Assistant Professor Department of International Health School of Public Health and Johns Hopkins Bloomberg School of Health Systems Public Health University of Waterloo Baltimore, MD 21205, USA 200 University Avenue West Waterloo, Ontario N2L 3G1, Canada Tim JONES Deputy State Epidemiologist Scott MCDONALD Communicable and Environmental National Institute for Public Health and Disease Services the Environment Tennessee Department of Health, USA Centre for Infectious Disease Control PO Box 1, 3720 BA Bilthoven, Ina KELLY The Netherlands Senior Medical Officer Department of Public Health, Ireland Sara MONTEIRO PIRES Division of Epidemiology and Microbial Genomics National Institute for Public Health and Technical University of Denmark the Environment, RIVM Department of Microbiology and Risk Antonie van Leeuwenhoeklaan 9 Assessment 3721 MA Bilthoven, The Netherlands National Food Institute Technical University of Denmark Bocar KOUYATE DK-2860 Søborg , Denmark Health Adviser Ministry of Health, Burkina Faso Gerald MOY Private Consultant (and former WHO Julie LEGLER staff member) Director, Statistics Program Switzerland Department of Mathematics, Statistics and Computer Science Darwin MURRELL St. Olaf’s College, Northfield, MN Honorary Professor 55057, USA Department of Veterinary Disease Biology Myron LEVINE Faculty of Life Sciences Grollman Distinguished Professor University of Copenhagen, Denmark and Director University of Maryland School of Medicine Center for Development , USA Appendices 138

Arun NANDA Kate THOMAS European Centre for Disease Prevention Epidemiologist and Control (ECDC) Enteric Surveillance and Population Surveillance Unit Studies Division 17183 Stockholm, Sweden Centre for Food-borne, Environmental and Zoonotic Infectious Diseases Shilpi OBEROI, Public Health Agency of Canada University of Pittsburgh Guelph Ontario, N1H 8J1, Canada Graduate School of Public Health Department of Environmental & Juerg UTZINGER Occupational Health Bridgeside Point Swiss Tropical and Public Health Institute Pittsburgh, PA 15219, USA Switzerland

Sarah O’BRIEN Sommer WILD Department of Epidemiology and Research Assistant Population Health Department of Mathematics, Statistics Institute of Infection and Global Health and Computer Science University of Liverpool St. Olaf’s College, Northfield, MN Liverpool, L69 7BE, UK 55057, USA

Edoardo POZIO Felicia WU Head of the European Union Reference Department of Environmental and Laboratory for Parasites Occupational Health Istituto Superiore di Sanità, Italy Graduate School of Public Health University of Pittsburgh, USA Elaine SCALLAN Colorado School of Public Health Marco ZEILMAKER University of Colorado National Institute for Public Aurora, Colorado 80045, USA Health (RIVM) Bilthoven, The Netherlands Dana SCHNEIDER Health Scientist Yu ZANG Division of Public Health Systems and Toxicology Team, Division of Workforce Development Petition Review Center for Global Health Office of Food Additive Safety Centers for Disease Control and FDA CFSAN Prevention, USA College Park, MD 20740, USA

Kurt STRAIF Jakob ZINSSTAG International Agency for Research Assistant Professor on Cancer Department of Epidemiology and 69372 Lyon Cedex 08, France Public Health Swiss Tropical and Public Health Institute, Switzerland APPENDICES WHO Estimates of the global burden of foodborne diseases 139

WHO Secretariat and other UN Organizations

Initiative Leader for WHO Secretariat Anthony BURTON 2006– 2010 Expanded Programme on Claudia STEIN Immunization Plus Director WHO, Geneva, Switzerland Division of Information, Evidence, Research and Innovation Tim CORRIGAN World Health Organization Department of Food Safety Regional Office for Europe and Zoonoses Copenhagen Ø, Denmark WHO, Geneva, Switzerland

2010 – 2011 Chrystelle DAFFARA Danilo LO-FO-WONG Department of Food Safety Department of Food Safety and Zoonoses and Zoonoses WHO, Geneva, Switzerland WHO, Geneva, Switzerland Leonardo DE KNEGT 2011 – 2013 Division of Epidemiology and Tanja KUCHENMUELLER Microbial Genomics Department of Evidence and Intelligence Department of Microbiology and for Policy-making Risk Assessment WHO Regional Office for Europe National Food Institute Technical Copenhagen Ø, Denmark University of Denmark Soborg, Denmark 2013 – 2015 Amy CAWTHORNE Mohamed ELMI Department of Food Safety Environmental Health Risk and Zoonoses & Regional Adviser for Food and WHO, Geneva, Switzerland Chemical Safety Regional Centre for Environmental 2015 – 2015 Health Action Natsumi Regional Office for the Eastern Department of Food Safety Mediterranean and Zoonoses WHO, Amman, Jordan WHO, Geneva, Switzerland Keiji FUKUDA Assistant Director-General Other UN and WHO Staff: WHO, Geneva, Switzerland Awa AIDARA-KANE Department of Food Safety David HEYMANN and Zoonoses Former Assistant Director-General WHO, Geneva, Switzerland WHO, Geneva, Switzerland

Peter Karim BEN EMBAREK Lisa INDAR Department of Food Safety Foodborne Diseases and Zoonoses Caribbean Epidemiology Centre (CAREC) WHO, Geneva, Switzerland Appendices 140

Mary KENNY Jørgen Schlundt Food Safety and Quality Unit Former Director FAO, Rome, Italy Department of Food Safety and Zoonoses Hilde KRUSE WHO, Geneva, Switzerland WHO Regional Office for Europe Copenhagen Ø, Denmark Annette PRUESS-ÜSTÜN Occupational and Environmental Health Doris MA FAT WHO, Geneva, Switzerland Health Statistics and Evidence WHO, Geneva, Switzerland Agneta SUNDEN-BYLEHN UNEP Chemicals Colin MATHERS Châtelaine, Switzerland Country Health Information WHO, Geneva, Switzerland Andrea SWINTEK Department of Food Safety Yuki MINATO and Zoonoses Department of Food Safety WHO, Geneva, Switzerland and Zoonoses WHO, Geneva, Switzerland Angelika TRITSCHER Department of Food Safety Kazuaki MIYAGISHIMA and Zoonoses Director WHO, Geneva, Switzerland Department of Food Safety and Zoonoses Philippe VERGER WHO, Geneva, Switzerland Department of Food Safety and Zoonoses Linda MOLONEY WHO, Geneva, Switzerland Department of Food Safety and Zoonoses Steven WIERSMA WHO, Geneva, Switzerland Expanded Programme on Immunization Plus Desiree M. NARVAEZ WHO, Geneva, Switzerland Mercury and Other Metals Programme UNEP Chemicals DTIE, Switzerland Maged YOUNES Former Director Enrique PÉREZ GUTIÉRREZ Department of Food Safety Alert and Response and and Zoonoses Waterborne Diseases Unit WHO, Geneva, Switzerland Department of Communicable Diseases and Health Analysis PAHO/WHO, Washington DC, USA APPENDICES WHO Estimates of the global burden of foodborne diseases 141

Observers Ermias Woldemariam AMENE Patricia GRIFFIN School of Veterinary Medicine CDC University of Wisconsin-Madison, Atlanta WI, USA Georgia, USA

Valerie DAVIDSON Erika OTA School of Engineering Department of Global Health Policy University of Guelph Graduate School of Medicine Canada The University of Tokyo Tokyo, Japan Anou DREYFUS Section of Epidemiology Todd REED Vetsuisse Faculty Acting Director, Data Analysis and Zurich, Switzerland Integration Group Office of Data Integration and Food Protection Food Safety Inspection Service, USA

Task Force Members: EDTF: PDTF: Co-chairs: Martyn Kirk, Fred Angulo Chair: Neyla Gargouri (2006 - 2010), Nilanthi da Silva (2010 - 2012) EDTF and FERG members: George Paul Torgerson (2012 - 2015) Nasinyama, Fred Angulo, Aamir Fazil, Arie Havelaar, Dörte Döpfer, Black, PDTF and FERG Members: Nicolas Tine Hald, Lake, Claudio Lanata, Praet, Niko Speybroeck, Fumiko Kasuga, Alejandro Cravioto, Karen Keddy, Claudia Mohammad B Rokni, Xiao-Nong Zhou, Lanata, Xu-Mi Liu, George Nasinyama, Eric M. Fèvre, Banchob Sripa Paul Torgerson Commissioned scientists/resource Commissioned scientists/resource advisers/consultants: Arve Lee advisers/consultants: Marisa Caipo, Willingham, Thomas Fürst, Christine M Brecht Devleesschauwer, Christa Fischer- Budke, Hélène Carabin Walker, Sara Pires, Shannon Majowicz, CTTF: Aron Hall, John Crump, Tony Ao, Charline Maertens de Noordhout, Anna Dean, Chair: Herman Gibb Borna Muller, Jakob Zinsstag CTTF and FERG members: Gabriel Adegoke, David Bellinger, Reza Afshari, Michael Bolger, John Pitt, Janis Baines, Yan Liu, Arie H. Havelaar, David C. Bellinger Commissioned scientists/resource advisers/consultants: Kurt Straif, Felicia Wu, Janine Ezendam, Julie Cliff, Marco Zeilmaker, Philippe Verger, Bas Bokkers, Henk van Loveren, Marcel Mengelers Appendices 142

SATF: CSTF: Chair: Tine Hald Chair: Niko Speybroeck (2009 - 2011), Rob Lake (2012 - 2015) SATF and FERG members: Arie Havelaar, Fred Angulo, Fumiko Kasuga, Nilanthi CSTF and FERG members: Gabriel de Silva, Muhammad Rokni, David Adegoke, Fred Angulo, David Bellinger, Bellinger, Robert Black, Herman Gibb, Dr Alejandro Cravioto, Dörte Döpfer, Nicolas Wan Mansor Bin Hamzah, Dörte Döpfer, Praet, Herman Gibb, Arie Havelaar, Rob Lake Fumiko Kasuga, Bocar Kouyate, George Nasinyama, Robert Buchanan, Catterina Commissioned scientists/resource Ferriccio, Bocar Kouyate, Myron Levine, advisers/consultants: Sara Pires, Roger Sarah O’Brien, Nilanthi de Silva, Paul Cooke, Sandra Hoffmann, Willie Aspinall Torgerson CTF: Consultants: Brecht Devleesschauwer, Chair: Nicolas Praet Charline Maertens de Noordhout, CTF and FERG members: Brecht Juanita Haagsma Devleesschauwer, Paul Torgerson, Aamir KTPG: Fazil, David Bellinger, Arie Havelaar, Rob Chair: Pierre Ongolo-Zogo (2010 - 2011), Lake, Dörte Döpfer, Niko Speybroeck, John Ehiri (2012 - 2015) Eric Fèvre Members: Deena Alasfoor, Commissioned scientists/resource Nasreen Jessani, Helen Jensen, advisers/consultants: Dana Cole, Sara Tanja Kuchenmüller, Haichao Lei, Pires, Juanita Haagsma, Kate Thomas, Bocar Kouyate Scott McDonald, Felicia Wu Consultant: Sandy Campbell APPENDICES WHO Estimates of the global burden of foodborne diseases 143

APPENDIX 2. Subregions

SUBREGION(1) [5] WHO MEMBER STATES AFR D Algeria; Angola; Benin; Burkina Faso; Cameroon; Cabo Verde; Chad; Comoros; Equatorial Guinea; Gabon; Gambia; Ghana; Guinea; Guinea-Bissau; Liberia; Madagascar; Mali; Mauritania; Mauritius; Niger; Nigeria; Sao Tome and Principe; Senegal; Seychelles; Sierra Leone; Togo. AFR E Botswana; ; Central African Republic; Congo; Côte d’Ivoire; Democratic Republic of the Congo; Eritrea; Ethiopia; Kenya; Lesotho; ; Mozambique; Namibia; ; South Africa; Swaziland; Uganda; United Republic of Tanzania; ; Zimbabwe. AMR A Canada; Cuba; United States of America. AMR B Antigua and Barbuda; ; Bahamas; Barbados; Belize; Brazil; Chile; Colombia; Costa Rica; Dominica; Dominican Republic; El Salvador; Grenada; Guyana; Honduras; Jamaica; Mexico; Panama; Paraguay; Saint Kitts and Nevis; Saint Lucia; Saint Vincent and the Grenadines; Suriname; Trinidad and Tobago; Uruguay; (Bolivarian Republic of). AMR D (Plurinational State of); Ecuador; Guatemala; Haiti; ; Peru. EMR B Bahrain; Iran (Islamic Republic of); Jordan; Kuwait; Lebanon; Libya; Oman; Qatar; Saudi Arabia; Syrian Arab Republic; Tunisia; United Arab Emirates. EMR D Afghanistan; Djibouti; Egypt; Iraq; Morocco; Pakistan; Somalia; South Sudan(2); Sudan; Yemen. EUR A Andorra; Austria; Belgium; Croatia; Cyprus; Czech Republic; Denmark; Finland; France; Germany; Greece; Iceland; Ireland; ; Italy; Luxembourg; Malta; Monaco; Netherlands; Norway; Portugal; San Marino; Slovenia; Spain; Sweden; Switzerland; United Kingdom. EUR B Albania; Armenia; ; Bosnia and Herzegovina; Bulgaria; Georgia; ; Montenegro; Poland; Romania; Serbia; Slovakia; Tajikistan; The Former Yugoslav Republic of Macedonia; ; ; Uzbekistan. EUR C Belarus; Estonia; Hungary; ; Latvia; Lithuania; Republic of Moldova; Russian Federation; Ukraine. SEAR B Indonesia; Sri Lanka; Thailand. SEAR D Bangladesh; Bhutan; Democratic People’s Republic of Korea; India; Maldives; Myanmar; Nepal; Timor- Leste. WPR A Australia; Brunei Darussalam; Japan; New Zealand; Singapore. WPR B Cambodia; China; Cook Islands; Fiji; Kiribati; Lao People’s Democratic Republic; Malaysia; Marshall Islands; Micronesia (Federated States of); ; Nauru; Niue; Palau; Papua New Guinea; Philippines; Republic of Korea; Samoa; Solomon Islands; Tonga; Tuvalu; Vanuatu; Viet Nam.

Notes: (1) The subregions are defined on the basis of child and adult mortality as described by Ezzati et al. [5]. Stratum A = very low child and adult mortality; Stratum B = low child mortality and very low adult mortality; Stratum C = low child mortality and high adult mortality; Stratum D = high child and adult mortality; and Stratum E = high child mortality and very high adult mortality. The use of the term ‘subregion’ here and throughout the text does not identify an official grouping of WHO Member States, and the “subregions” are not related to the six official WHO regions, which are AFR = African Region; AMR = Region of the Americas; EMR = Eastern Mediterranean Region; EUR = European Region; SEAR = South-East Asia Region; WPR = Western Pacific Region.

(2) South Sudan was re-assigned to the WHO African Region in May 2013. As this study relates to time periods prior to this date, estimates for South Sudan were included in the WHO Eastern Mediterranean Region. Appendices 144

APPENDIX 3. Preliminary hazards considered by each task force

At the FERG 1 meeting (26–28 November Dientamoeba fragilis 2007), each of the hazard-based TFs Diphyllobothrium latum considered a comprehensive list of potential Echinococcus spp. foodborne hazards for the development of Echinostoma spp. burden estimates. During the course of the Entamoeba histolytica project these lists had to be reduced, largely Fasciola spp. for practical reasons concerning the ability Fasciolopsis buski to generate burden estimates. For reference, hominis the complete list is given here. Giardia spp. EDTF spinigerum Adenovirus Aeromonas spp. Astrovirus Isospora belli Bacterial toxins (B. cereus) Linguatula serata Bacterial toxins (C. perfringens) Bacterial toxins (S. aureus) Nanophytes salmincola Brucella spp. Campylobacter spp. Opisthorchis viverrini Clostridium botulinum Paragonimus spp. Enteroaggerative E. coli (EAggEC) Sarcocystis hominis Entero-pathogenic E. coli (EPEC) Taenia saginata Entero-toxigenic E. coli (ETEC) Taenia solium Toxocara spp. Helicobacter pylori Toxoplasma gondii Hepatitis A virus Trichinella spp. virus spp. Leptospira spp. Listeria monocytogenes CTTF Mycobacterium bovis Elemental contaminants (e.g. lead, Non cholera mercury, cadmium, manganese, arsenic) Norovirus (e.g. aflatoxins, , , trichothocenes) Rotavirus Food additives (e.g. sulphites, nitrites/ Salmonella (non-typhoidal) spp. nitrates, ) Salmonella (typhoid) spp. (e.g. organophosphates, Shiga-toxin producing E. coli (STEC) carbamates, DDT, pyrethrins) Shigella spp. Vibrio cholerae 01/0139 Organic industrial pollutants (e.g. Yersinia spp. persistent organic pollutants) PDTF Veterinary drugs/residues (e.g. , hormones–but not residues) Angiostrongylus cantonensis Seafood toxins (e.g. , Angiostrongylus costaricensis ciguatera, shellfish toxins, DSPs, simplex PSPs, histamines) Ascaris spp. Blastocystis hominis Process contaminants (e.g. acrylamide, Capillaria philippinensis PAHs, choropropanol) Clonorchis sinensis Allergens (e.g. peanuts) Cryptosporidium spp. Natural (e.g. cyanide in Cyclospora spp. cassava, aminoglycosides) Radionuclides and depleted uranium APPENDICES WHO Estimates of the global burden of foodborne diseases 145

APPENDIX 4. Hazard-specific input parameter sources and methods

A4.1 Brucellosis brucellosis incidence estimates for 81 countries. The FERG Computational Incidence Task Force imputation model was There were 32 countries identified as then used to impute an incidence of “free of brucellosis in livestock”, using human brucellosis in all countries with 2006–2012 data reported to the World missing incidence. Organisation for Animal Health (OIE) Clinical Outcomes [248], and a list of European countries recognized by the European Union as The FERG-commissioned systematic “officially brucellosis free” in cattle, sheep review assisted in determining the clinical and goats in 2010 [249]. Using 2001– outcomes for human brucellosis [254]. 2004 OIE data, a previous review [250] These were: acute brucellosis (severe); estimated human brucellosis incidence acute brucellosis (moderate); chronic for 9 of the countries identified as free brucellosis; brucellosis orchitis; and of brucellosis in livestock. The median brucellosis death. For acute brucellosis, human brucellosis incidence from these it was assumed that 50% of cases were 9 countries free of brucellosis in livestock severe, 50% of cases were moderate, was used as the estimated human 40% of brucellosis cases resulted in brucellosis incidence for each of the chronic brucellosis, and 10% of brucellosis 32 countries free of brucellosis in livestock. cases in males resulted in orchitis [254]. A FERG-commissioned systematic review was then used to screen 2385 articles Duration [251] and a literature review for national Acute brucellosis: duration 14 days (min. human brucellosis incidence estimates 7 days–max. 21 days). Chronic brucellosis: [174, 175, 187, 188, 252, 253], to extract duration 6 months (min. 3 months– max. brucellosis national incidence estimates 24 months). Brucellosis orchitis: duration for 17 countries (Argentina, Canada, 6 months (min. 3 months–max. Chad, China, Egypt, France, Greece, Iraq, 24 months) [254]. Iran, Italy, Kyrgyztan, Jordan, Mexico, Oman, Saudi Arabia, Turkey, and the Disability weight United States of America). The human Acute brucellosis (severe): GBD2010 brucellosis incidence estimates in each disability weight of 0.210 (95% UI of these countries were compared with 0.139–0.298) for infectious disease, human brucellosis incidence estimates acute episode, severe. Acute brucellosis in the same country in a previous review, (moderate): GBD2010 disability weight which used 2001–2004 OIE data [250], to of 0.053 (95% UI 0.033–0.081) for estimate a multiplier (mean=5.4, range infectious disease, acute episode, mild. 1.6–15.4) to account for under-reporting. Chronic brucellosis: GBD2010 disability This multiplier was used to estimate weight 0.079 (95%UI 0.053–0.115) national human brucellosis incidence for for musculoskeletal problems, legs, countries with OIE human brucellosis data moderate. Brucellosis orchitis: GBD2010 in the previous review but without national disability weight of 0.097 (95% UI 0.063– human brucellosis incidence estimates 0.0137) for epididymo-orchitis [82]. identified in the current systematic review or literature review. By multipling the Mortality human brucellosis incidence reported to Acute brucellosis and chronic brucellosis OIE by the multiplier, there were 32 such case fatality ratio 0.5% (min. CFR 0.25%– countries. These steps yielded human max. CFR 0.75%) [255, 256]. Appendices 146

Age distri bution Clinical Outcomes Acute brucellosis, chronic brucellosis, Clinical outcomes were M. bovis brucellosis orchitis and brucellosis death tuberculosis and M. bovis death. age distribution: 3% <15 years; 29% 15–24 years; 24% 25–34 years; 16% 35–44 years; Duration 13% 45–54 years; 12% 55–64 years; and M. bovis tuberculosis duration was 3% >65 years [257]. estimated using data in the 2014 WHO Global Tuberculosis Report on incidence Sex distri bution and prevalence of human TB infections Acute brucellosis, chronic brucellosis [165]; these data yielded a duration of and brucellosis deaths sex distribution: 1.5 years in all regions except AFR, where 55% male (95% UI 50%–60% male) [254]. the duration was 1 year. Brucellosis orchitis: 100% male. Disability weight M. bovis tuberculosis: GBD2010 disability A4.2 Mycobacterium weight of 0.331 (95% UI 0.222–0.450) for bovis infections tuberculosis without HIV infection [82].

Incidence Mortality There were 51 countries identified as Deaths from M. bovis were estimated “free of Mycobacterium bovis in cattle” following the same approach for using 2005–2012 data reported to OIE estimating M. bovis cases after [248] and a list of European countries reducing the mortality by 20% due to recognized by the European Union as the recognition from another FERG- “officially free of bovine tuberculosis” commissioned review that M. bovis in 2010 [249]. A FERG-commissioned infections are more likely to result in systematic review screened 1203 articles extrapulmonary infections [259] and that [258] with data from 91 countries, and extrapulmonary infections have a lower estimated the median proportion of case-fatality ratio (CFR) than pulmonary human tuberculosis cases due to M. bovis tuberculosis infections; a 20% reduction at the region level as 2.8% for AFR, 0.4% in mortality was based on a review of the United States of America national for EUR and 0.3% for AMR; the overall surveillance data from 2009–2010, which median proportion from studies in the found that the CFR for extrapulmonary review (1.0%) was used in the three tuberculosis infections was approximately other regions. These proportions were 20% lower than the CFR for pulmonary applied to all countries in each respective tuberculosis infections. Therefore, region except for the 51 countries free of country-level human tuberculosis M. bovis in cattle. The lowest observed mortality rates of tuberculosis among proportion (0.3%) was assigned to the persons not infected with HIV were 51 countries free of M. bovis in cattle. abstracted from the WHO Global Country-level human tuberculosis Tuberculosis Report [165], reduced by incidence was abstracted from the WHO 20%, and then multiplied by population Global Tuberculosis Report [165] and estimates and the proportion of human multiplied by population estimates and tuberculosis cases due to M. bovis to the proportion of human tuberculosis estimate M. bovis deaths. cases due to M. bovis to estimate human M. bovis cases. APPENDICES WHO Estimates of the global burden of foodborne diseases 147

Age distribution survey in 1997 [262]; similarly for typhoid It was assumed that the age distribution abscesses and cysts. We used the of M. bovis cases and M. bovis deaths GBD2010 range of estimates around the was the same as the age distribution of mean estimate of global deaths due to human tuberculosis cases and deaths, typhoid and paratyphoid (190 242 and therefore used the age distribution with UI 23 786–359 075) to derive a from Table 3.2 of the WHO Global range of estimates for typhoid incidence. Tuberculosis Report: 2% <15 years; 60% Clinical Outcomes 15–44 years; 28% 45–64 years; 10% >65 years [165]. Clinical outcomes were typhoid fever, typhoid liver abscesses and cysts, and Sex distribution typhoid death [6]. It was assumed that the sex distribution Duration of M. bovis cases and M. bovis deaths was the same as the sex distribution of Typhoid fever: duration 28 days (min. human tuberculosis cases and deaths, 7 days–max. 42 days). Typhoid liver and therefore used the sex distribution abscesses and cysts: duration 42 days from Table 3.2 of the WHO Global (min. 28 days–max. 56 days). Duration Tuberculosis Report: 65% male [165]. was estimated based on median duration before hospitalization for typhoid fever or typhoid abscesses/cysts of 10 days, A4.3 Typhoid recommended treatment duration for typhoid fever of 10–14 days and for Incidence typhoid abscesses/cysts of 28–112 days, FERG reviewed available burden of and presumed longer duration in patients disease estimates for typhoid fever with typhoid fever or typhoid abscesses/ [6, 260] before selecting the IHME Global cysts who are not hospitalized [263]. Burden of Disease 2010 (GBD2010) estimates because these estimates were Disability weight published in peer-reviewed literature Typhoid fever: GBD2010 disability and were available for all countries. At weight of 0.210 (95% UI 0.139–0.298) the request of FERG, IHME provided for infectious disease, acute episode, GBD2010 data with country-specific, severe. Typhoid liver abscesses and cysts: age-standardized prevalence (per GBD2010 disability weight of 0.254 (95% 100 000 population) of “typhoid and UI 0.170–0.355) for infectious disease, paratyphoid fever”, and “typhoid and post-acute consequences, severe [82]. paratyphoid liver abscesses and cysts” [6]. Assuming a steady disease state, Mortality prevalence of typhoid and paratyphoid GBD2010 country-specific mortality data fever was converted to incidence by for “typhoid and paratyphoid fevers” dividing by duration; similarly for typhoid were obtained by sex and 20 age groups and paratyphoid abscesses and cysts. from the IHME website [58]. Typhoid Typhoid fever incidence was determined mortality was determined using a ratio using a ratio of 1.0 Salmonella serotype of 1.0 Salmonella serotype Typhi cases Typhi cases to 0.23 Salmonella serotype to 0.23 Salmonella serotype Paratyphi A Paratyphi A cases observed in national cases observed in national laboratory- laboratory-based surveillance in the based surveillance in the United States United States of America and in a global of America and in a global survey in 1997 Appendices 148

[262]. The GBD2010 range of estimates of 0.23 Salmonella serotype Paratyphi A around the mean estimate of global cases to 1.0 Salmonella serotype Typhi deaths due to typhoid and paratyphoid cases observed in national laboratory- fevers (190 242 with UI 23 786–359 075) based surveillance in the United States were used to derive a range of estimates of America and in a global survey in 1997 for paratyphoid deaths. [262]; similarly for paratyphoid abscesses and cysts. We used the GBD2010 range Age distribution of estimates around the mean estimate Using data from IHME, the age of global deaths due to typhoid and distribution for typhoid fever, typhoid paratyphoid fevers (190 242 with UI liver abscesses and cysts, and typhoid 23 786–359 075) to derive a range of deaths was 5% <1 year; 16% 1–4 years; estimates for paratyphoid incidence. 22% 5–14 years; 19% 15–24 years; 14% Clinical Outcomes 25–34 years; 9% 35–44 years; 6% 45–54 years; 3% 55–64 years; 3% 65–74 years; Clinical outcomes were paratyphoid fever, 1% 75–84 years; and 1% >85 years [6]. paratyphoid liver abscesses and cysts, and paratyphoid deaths [6]. Sex distri bution Duration Using data from IHME, the sex distribution for cases of typhoid fever, Paratyphoid fever: duration 28 days (min. and typhoid liver abscesses and cysts 7 days–max. 42 days); paratyphoid liver was 56% male, and the sex distribution abscesses and cysts: duration 42 days for typhoid deaths was 58% male [6]. (min. 28 days–max. 56 days). Duration was estimated based on median duration before hospitalization for paratyphoid A4.4 Paratyphoid fever or paratyphoid abscesses and cysts of 10 days, with a recommended Incidence treatment duration for paratyphoid FERG reviewed available burden of fever of 10–14 days and for paratyphoid disease estimates for typhoid and abscesses and cysts of 28–112 days, and paratyphoid fever [6, 260] before presumed longer duration in patients selecting the IHME Global Burden of with paratyphoid fever or paratyphoid Disease 2010 (GBD2010) estimates abscesses and cysts who are not because these estimates were published hospitalized [263]. in peer-reviewed literature and were available for all countries. At the request Disability weight of FERG, IHME provided GBD2010 data Paratyphoid fever: GBD2010 disability with country-specific, age-standardized weight of 0.210 (95% UI 0.139–0.298) for prevalence (per 100 000 population) infectious disease, acute episode, severe. of “typhoid and paratyhoid fever”, and Paratyphoid liver abscesses and cysts: “typhoid and paratyphoid liver abscesses GBD2010 disability weight of 0.254 (95% and cysts” [6]. Assuming a steady UI 0.170–0.355) for infectious disease, disease state, prevalence of typhoid post-acute consequences, severe [82]. and paratyphoid fever was converted to incidence by dividing by duration; Mortality similarly for typhoid and paratyphoid GBD2010 country-specific mortality data abscesses and cysts. Paratyphoid fever for “typhoid and paratyphoid fevers” incidence was determined using a ratio were obtained by sex and 20 age groups APPENDICES WHO Estimates of the global burden of foodborne diseases 149

from the IHME website [58]. Paratyphoid Clinical Outcomes mortality was determined using a ratio Clinical outcomes were acute hepatitis of 0.23 Salmonella serotype Paratyphi A A (severe), acute hepatitis A (mild), and cases to 1.0 Salmonella serotype Typhi hepatitis A death. For acute hepatitis, cases observed in national laboratory- it was assumed that 50% of cases were based surveillance in the United States severe and 50% of cases were mild [264]. of America and in a global survey in 1997 [262]. We used the GBD2010 range of Duration estimates around the mean estimate Acute hepatitis A: duration 21 days (min. of global deaths due to typhoid and 14 days–max. 30 days) [264]. paratyphoid fevers (190 242 with UI 23 786–359 075) to derive a range of Disability weight estimates for paratyphoid deaths. Acute hepatitis A (severe): GBD2010 Age distribution disability weight of 0.210 (95% UI Using data from IHME, the age 0.139–0.298) for infectious disease, acute distribution for paratyphoid fever, episode, severe. Acute hepatitis A (mild): paratyphoid liver abscesses and cysts, GBD2010 disability weight of 0.005 (95% and paratyphoid deaths was 5% <1 year; UI 0.002–0.011) for infectious disease, 16% 1–4 years; 22% 5–14 years; 19% 15–24 acute episode, mild [82]. years; 14% 25–34 years; 9% 35–44 years; 6% 45–54 years; 3% 55–64 years; 3% Mortality 65–74 years; 1% 75–84 years; and 1% >85 GBD2010 country-specific mortality years [6]. data for “hepatitis A” were obtained by sex and 20 age groups from the IHME Sex distribution website [58]. The GBD2010 range of Using data from IHME, the sex estimates around the mean estimate of distribution for cases of paratyphoid global deaths due to hepatitis A (102 850 fever, and paratyphoid liver abscesses with UI 51 157–228 057) were used to and cysts, was 56% male, and the sex derive a range of estimates for hepatitis distribution for paratyphoid deaths was A deaths. 58% male [6]. Age distribution A4.5 Hepatitis A infection Using data from IHME, the age distribution for acute hepatitis A cases Incidence and hepatitis A deaths was 10% <1 year; Assuming a case-fatality ratio of 0.2% 5% 1–4 years; 2% 5–14 years; 3% 15–24 [264], the IHME Global Burden of Disease years; 5% 25–34 years; 11% 35–44 years; 2010 (GBD2010) country-specific data 17% 45–54 years; 20% 55–64 years; 17% on “Hepatitis A”, available on the IHME 65–74 years; 5% 75–84 years; and 5% >85 website by sex and 20 age groups years [58]. [58] were converted to incidence. the GBD2010 range of estimates around Sex distribution the mean estimate of global deaths due Using data from IHME, the sex to hepatitis A (102 850 with UI 51 157– distribution for acute hepatitis A 228 057) to derive a range of estimates cases and hepatitis A deaths was 57% for hepatitis A incidence. male [58]. Appendices 150

A4.6 Shiga toxin-producing moderate diarrhoea, and 80% of STEC Escherichia coli (STEC) infection infections resulted in mild diarrhoea [266]. We assumed that the following Incidence percent of STEC infections were serotype O157: 36% in AMR A, AMR B, EUR and Using a FERG-commissioned systematic WPR A; 10% in AMR D, AFR and SEAR; review that screened 17 178 articles, and and 0% in EMR. We assumed that 0.8% a search for national surveillance data, (min. 0.7%–max. 0.9%) of O157 STEC Shiga toxin-producing Escherichia coli infections and 0.03% (min. 0.01%–max. (STEC) incidence data from 21 countries 0.04%) of non-O157 STEC infections were identified. Using a hierarchical study resulted in HUS, and the 3% (min. 0%– selection process, in the subregions with max. 30%) of HUS cases resulted in prospective cohort studies or multipliers ESRD [265]. studies that estimated national STEC incidence, that STEC incidence was Duration assigned to all countries in the subregion. Acute STEC diarrhoea: duration 7 days In subregions with only STEC-notifiable (min. 5 days–max. 10 days) [266]. STEC disease data, STEC incidence for all haemolytic uraemia syndrome: duration countries in the subregion was estimated 28 days (min. 14 days–max. 42 days). using a multipler of 36 (range 7.4–106.8) STEC end-stage renal disease: results in to account for under-reporting; the lifelong disability in countries in AMR A, STEC incidence from notifiable disease EUR A and WPR A, and death in other data was multiplied by the multiplier to countries [265, 267, 268]. estimate the national STEC incidence. In subregions with no STEC incidence Disability weight data, geographical proximity was used Acute diarrhoea (severe): GBD2010 to extrapolate the STEC incidence to all disability weight of 0.281 (95% UI countries in the subregion [265]. These 0.184–0.399) for diarrhoea, severe. efforts led to the following regional Acute diarrhoea (moderate): GBD2010 incidence (per 100 000 population) disability weight of 0.202 (95% UI 0.133– estimates: AFR subregions D and E 0.299) for diarrhoea, moderate. Acute 1.4; AMR subregions A and D 93.5; diarrhoea (mild): GBD2010 disability AMR subregion B 27.2; EMR 152.6; EUR weight of 0.061 (95% UI 0.036–0.093) for subregion A 47.1; EUR subregion B diarrhoea, mild. STEC haemolytic uraemic 2.7; EUR subregion C 2.5; SEAR 66.3; syndrome: GBD2010 disability weight WPR subregion A 44.5; and WPR 0.210 (95% UI 0.139–0.298) for infectious subregion B 3.5. disease, acute episode, severe. STEC end- stage renal disease: GBD2010 disability Clinical Outcomes weight of 0.573 (95% UI 0.397–0.749) for Clinical outcomes of STEC infections end-stage renal disease, on dialysis [82]. were acute STEC diarrhoea (severe), acute STEC diarrhoea (moderate), acute Mortality STEC diarrhoea (mild), STEC haemolytic STEC haemolytic uraemic syndrome uraemic syndrome (HUS), STEC end- case fatality ratio 3.7%. STEC end-stage stage renal disease (ESRD), and STEC renal disease case fatality ratio 20% in death. We assumed that 2% of STEC countries in AMR A, EUR A and WPR infections resulted in severe diarrhoea, A; case fatality ratio 100% in other 18% of STEC infections resulted in countries [265]. APPENDICES WHO Estimates of the global burden of foodborne diseases 151

Age distribution Clinical Outcomes Acute STEC diarrhoea and STEC Clinical outcomes were botulism (mild haemolytic uraemic syndrome age (HUS) to moderate), botulism (severe), and distribution: 29% <5 years;, 20% 5–14 botulism death. We assumed that 35% years; 35% 15–54 years; 16% *55 years. (range 20-50%) of botulism cases STEC end-stage renal disease (ESRD) resulted in severe botulism [269–271]. and ESRD deaths age distribution: 41% <5 years; 18% 5–14 years; 26% 15–54 Duration years; 15% *55 years. HUS deaths age Botulism (mild to moderate): duration distribution: 11% <1 year; 47% 1–4 years; 10 days (min. 5 days–max. 20 days); 14% 5–14 years; 22% 15–64 years; 6% *65 botulism (severe): duration 30 days (min. years [267]. 15 days–max. 180 days) [269–271].

Sex distribution Disability weight Acute STEC diarrhoea, STEC haemolytic – Botulism (mild to moderate): uraemic syndrome (HUS), STEC end- GBD2010 disability weight 0.198 (95% stage renal disease (ESRD), HUS deaths UI 0.137–0.278) for multiple sclerosis, and ESRD deaths sex distribution: mild. 50% male [265]. – Botulism (severe): GBD2010 disability weight 0.445 (95% UI 0.303–0.593) A4.7 Botulism for multiple sclerosis, moderate [82]. Mortality Incidence Estimates of mortality were only Estimates of incidence were only conducted for the 55 countries in EUR conducted for the 61 EUR and other and AMR A. Severe botulism case fatality subregion A (low mortality) countries. ratio 15% (range 5-25%). Assume no Based on a literature review for articles deaths among mild to moderate botulism with national estimates of foodborne cases [188, 269, 270]. diseases including botulism, we identified national estimates of the incidence of Age distribution botulism from five countries: Canada Mild to moderate botulism, severe [175], France [174], Georgia [269], Poland botulism, and botulism death age [270] and the United States of America distribution: mode 50 years (min. age [188]. The median botulism incidence 4 years–max. age 88 years) [269–271]. from these five countries was from Canada, therefore the botulism incidence Sex distribution from Canada (0.04 per 100 000 Mild to moderate botulism, severe population, with a 90% confidence botulism, and botulism death sex interval of 0.02–0.08 per 100 000) was distribution: 48% male [269–271]. used as the incidence for all 55 countries in EUR and AMR A. Appendices 152

A4.8 Clostridium Mortality perfringens intoxication Estimates of mortality were only conducted for the 61 EUR and other Incidence subregion A (low mortality) countries. Estimates of incidence were only National estimates of Clostridium conducted for the 61 EUR and other perfringens intoxications cases and subregion A (low mortality) countries. deaths were available from Australia Based on a literature review for articles [272], France [174], Netherlands [154], with national estimates of foodborne New Zealand [252], and the United States diseases that included Clostridium of America [188]; the median case fatality perfringens intoxications, we identified ratio (CFR) from these five countries national incidence estimates for was the New Zealand (0.0030% [95%CI: Clostridium perfringens intoxications 0.0024%-0.0038%]), therefore the CFR from seven countries: Australia [272], from New Zealand was used as the Canada [175], France [174], Netherlands CFR for all EUR and other subregion [154], New Zealand [252], United A countries. Kingdom [48], and the United States of America [188]. The median C. perfringens Age distribution intoxication incidence from these Acute gastroenteritis and deaths due to seven countries was from the United Clostridium perfringens intoxication age States of America, therefore the C. distribution: 1% <5 years; 13% 5–14 years; perfringens intoxication incidence from 59% 15–54 years; 27% *55 years [273]. the United States of America (324.19 per 100 000 population with a 95% Sex distribution confidence interval of 126.14–833.44 per Acute gastroenteritis and deaths due to 100 000) was used as the C. perfringens Clostridium perfringens intoxication sex intoxication incidence for all EUR and distribution: 63% male [273]. other subregion A countries.

Clinical Outcomes A4.9 Staphylococcus Clinical outcomes were acute aureus intoxication gastroenteritis due to Clostridium Incidence perfringens intoxication and death due to C. perfringens intoxication [273]. Estimates of incidence were only conducted for the 61 EUR and other Duration subregion A (low mortality) countries. Acute gastroenteritis due to Clostridium Based on a literature review for articles perfringens intoxication: duration 1 day with national estimates of foodborne (min. 0.25 days–max. 2.5 days) [273]. diseases that included Staphylococcus aureus intoxication, we identified Disability weight national incidence estimates for S. aureus intoxication from seven countries: – Acute gastroenteritis due to Australia [272], Canada [175], France Clostridium perfringens intoxication: [174], Netherlands [154], New Zealand GBD2010 disability weight 0.061 [252], England and Wales as a proxy (95% UI 0.036–0.093) for diarrhoea, for United Kingdom [172], and the mild [82]. United States of America [188]. The median S. aureus intoxication incidence APPENDICES WHO Estimates of the global burden of foodborne diseases 153

from these seven countries was Sex distribution from Canada, therefore the S. aureus Acute gastroenteritis and deaths due to intoxication incidence from the Canada S. aureus intoxication sex distribution: (77.3 per 100 000 population with a 48% male [273]. 95% confidence interval of 50.65–118.0 per 100 000) was used as the S. aureus intoxication incidence for all EUR and A4.10 Bacillus cereus intoxication other subregion A countries. Incidence Clinical Outcomes Estimates of incidence were only Clinical outcomes were acute conducted for the 61 EUR and other gastroenteritis due to S. aureus subregion A (low mortality) countries. intoxication and death due to S. aureus Based on a literature review for articles intoxication [273]. with national estimates of foodborne diseases that included Bacillus cereus Duration intoxication, we identified national Acute gastroenteritis due to S. aureus incidence estimates for Bacillus cereus intoxication: duration 1 day (min. 0.25 intoxication from seven countries: days–max. 2.5 days) [273]. Australia [272], Canada [175], France [174], Netherlands [154], New Zealand Disability weight [252], England and Wales as a proxy for the United Kingdom [172], and the United – Acute gastroenteritis due to S. aureus States of America [188]. The median intoxication: GBD2010 disability B. cereus intoxication incidence from weight 0.061 (95% UI 0.036–0.093) these seven countries was for the United for diarrhoea, mild [82]. Kingdom (England and Wales), therefore Mortality the B. cereus intoxication incidence from Estimates of mortality were only the United Kingdom (21.4 per 100 000) conducted for the 61 EUR and other was used as the B. cereus intoxication subregion A (low mortality) countries. incidence for all EUR and other subregion National estimates of S. aureus A countries; because the available B. intoxication cases and deaths were cereus intoxication incidence estimate available from the Netherlands [154] from England and Wales did not include and the United States of America [188]; a corresponding confidence interval, the the case fatality ratio (CFR) for the average values of the intervals from the Netherlands was 0.0024% and for the countries with the next lowest and next United States of America was 0.0025%. highest B. cereus intoxication incidence We used the CFR from the United States were used (United States of America of America as the CFR for all EUR and 5.2–49.4 and the Netherlands 11.5–67.2) other subregion A countries with a 95% for a 95% confidence interval 7.9–58.3 confidence interval of 0.0012%-0.0045%. per 100 000.

Age distribution Clinical Outcomes Acute gastroenteritis and deaths due to Clinical outcomes were acute S. aureus intoxication age distribution: gastroenteritis due to Bacillus cereus 5% <5 years; 19% 5–14 years; 48% 15–54 intoxication [273]. years; 28% *55 years [273]. Appendices 154

Duration Clinical outcomes included perinatal and Acute gastroenteritis due to Bacillus non-perinatal listeriosis; we estimated cereus intoxication: duration 1 day (min. that 79.3% (min. 77.3%–max. 81.3%) 0.25 days–max. 2.5 days) [273]. of listeriosis cases were perinatal and 20.7% (min. 19.0%–max. 22.4%) were Disability weight non-perinatal. Clinical outcomes among – Acute gastroenteritis due to Bacillus perinatal listeriosis cases were neonatal cereus intoxication: GBD2010 septicaemia, neonatal meningitis, disability weight 0.061 (95% UI 0.036– neurological sequelae, stillborn, and 0.093) for diarrhoea, mild [82]. death; stillborns were estimated but not included in the final FERG estimates of Mortality deaths and DALYS. We estimated 30.7% No deaths estimated. of perinatal listeriosis cases developed neonatal septicaemia and 15.2% (min. Age distribution 13.1%–max. 17.3%) neonatal meningitis, of Acute gastroenteritis due to Bacillus whom 43.8% (min. 31.8%–max. 55.8%) cereus intoxication age distribution: had neurological sequelae. Clinical 3% <5 years; 14% 5–14 years; 53% 15–54 outcomes among non-perinatal listeriosis years; 30% *55 years [273]. cases were septicaemia, meningitis, neurological sequelae and death; it was Sex distribution estimated that 61.6% (min. 59.4%–max. Acute gastroenteritis due to Bacillus 63.8%) of non-perinatal listeriosis cases cereus intoxication sex distribution: 50% developed septicaemia and 30.7% male [273]. (min. 28.7%–max. 32.7%) meningitis, of whom 13.7% (min. 8.2%–max. 19.2%) had neurological sequelae [70]. A4.11 Listeriosis Duration Incidence For perinatal and non-perinatal listeriosis Using a FERG-commissioned systematic cases: septicaemia duration 7 days, review which screened 11 22 papers and meningitis duration 182 days, and national surveillance, listeriosis incidence neurological sequelae 7 years [70]. data were extracted from 43 papers. National listeriosis incidence estimates Disability weight were then calculated for all countries – For listeriosis septicaemia: GBD2010 using the extracted data and imputed disability weight (DW) of 0.210 (95% estimates through a multilevel random UI 0.139–0.298) for infectious disease, effects model [70]. acute episode, severe [82]. – For listeriosis meningitis: a DW of Clinical Outcomes 0.426 (95% UI 0.368–0.474) derived Clinical outcomes were determined using from multiplicative methodology and outcome probabilities from the FERG- expert elicitation (with bootstrap commissioned review and a random analysis for CI) using a combination effects meta-regression model; for each of the following DWs: (1) 0.210 for study identified in the review, a weight infectious disease, acute episode, was assigned reflecting the study quality. severe; (2) 0.126 for intellectual These weights were included as a fixed disability, severe; (3) average of effect in the meta-regression model. 0.488 for epilepsy, severe and APPENDICES WHO Estimates of the global burden of foodborne diseases 155

epilepsy, treated with recent seizures; 5–14 years; 10% 15–34 years; 6% 35–44 and (4) 0.76 for motor impairment, years; 7% 45–54 years; 13% 55–64 years; moderate. 20% 65–74 years; 20% 75–84 years; 18% – For listeriosis neurological sequelae: *85 years. a DW of 0.292 (95% UI 0.272–0.316) derived from a multiplicative Sex distribution methodology and expert elicitation The sex distribution of listeriosis cases (with bootstrap analysis for CI) using and deaths was determined from a combination of following DWs: published papers during the FERG- (1) 0.047 resulting from average of commissioned review [70]. The sex all 10 DWs involving hearing loss; distribution for listeriosis cases and (2) 0.087 resulting from average of all deaths was: 50% male. 5 DWs for vision loss; and (3) 0.303 resulting from average of all 4 DWs for stroke, long-term consequence A4.12 Non-typhoidal [70, 82]. Salmonella infection Mortality Incidence Listeriosis case fatality ratios were The incidence of diarrhoeal non-typhoidal estimated following the same approach Salmonella (NTS) was estimated for estimating clinical outcomes of separately for middle and high mortality listeriosis cases; using probabilities from countries, and low mortality countries. the FERG-commissioned review and a For the 133 middle to high mortality random effects meta-regression model. countries, we used a modification of the For each study identified in the review, a Child Health Epidemiology Reference weight was assigned reflecting the study Group (CHERG) approach [50]. To quality; these weights were included as a derive “envelopes” of diarrhoea cases, fixed effect in the meta-regression model. for children <5 years of age we used The case fatality ratio for perinatal cases estimates of diarrhoea incidence from was 14.9% (mimimum 11.3% - max. 18.5%); a CHERG systematic review [51] and 9.2% (mimimum 7.5% - max. 10.9%) for persons >5 years of age we used resulted in neonatal deaths and 5.7% a FERG-commissioned systematic (minumum 3.8% - max. 7.6%) resulted in review [52]. We then estimated the stillbirths; stillborns were not included in aetiological proportions of diarrhoeal the final FERG estimates of deaths and illnesses due to NTS and the 10 other DALYs. The case fatality ratio for non- diarrhoeal pathogens1 in children <5 perinatal cases was 25.9% (mimimum years of age using CHERG and FERG 23.8% - max. 29.0%). systematic reviews of aetiology studies among outpatients and persons in the Age distribution community [40], and the aetiological The age distribution of listeriosis cases proportion of diarrhoeal illnesses due and deaths was determined from to NTS and the other 10 diarrhoeal published papers during the FERG- pathogens in persons >5 years of age commissioned review [70]. The age 1 The 11 diarrhoeal pathogens are: non-typhoidal distribution for perinatal listeriosis cases Salmonella, Campylobacter, Shigella, norovirus, and deaths was: 100% <1 month. The age enterotoxigenic E. coli (ETEC), enteropathogenic distribution for non-perinatal cases and E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba histolytica, other diarrhoeal agents not known deaths was: 0% <1 year; 2% 1–4 years; 4% to be foodborne (rotavirus and astrovirus), and unspecified agents. Appendices 156

using an updated FERG systematic Clinical Outcomes review of aetiology studies among Clinical outcomes were acute non- inpatients, outpatients and persons typhoidal Salmonella (NTS) diarrhoea in the community [40, 274]. The NTS (severe); acute NTS diarrhoea aetiological proportions were extracted (moderate); acute NTS diarrhoea (mild); from studies, and regional median NTS and NTS death. We assumed that 2% of aetiological proportions calculated. NTS diarrhoeal cases resulted in severe We modified the CHERG approach diarrhoea, 25% of NTS diarrhoeal cases by dropping regional median NTS resulted in moderate diarrhoea, and aetiological proportion outliers that 73% of NTS diarrhoeal cases resulted in were *5 times greater than the global mild diarrhoea. median NTS aetiological proportion, and replacing missing regional NTS Duration aetiological proportions with the global In children <5 years of age, duration median. Furthermore, for children of severe diarrhoea was 8.4 days, <5 years of age, we proportionally moderate diarrhoea was 6.4 days, and decreased the aetiological proportions mild diarrhoea was 4.3 days [275]. Based for all 11 diarrhoeal pathogens in each on the assumed distribution of severe, region so that the sum of the aetiological moderate and mild diarrhoea cases, the proportions for all diarrhoeal pathogens duration of all non-typhoidal Salmonella in a region equalled 1. The resultant (NTS) diarrhoea cases in children <5 regional NTS aetiological proportions years of age was estimated to be 4.9 were multiplied by the regional estimates days (min. 4.3 days–max. 8.4 days). In of diarrhoea incidence, and the resultant persons >5 years of age, the duration of regional NTS incidence was applied to NTS diarrhoea was 2.8 days [275]. all countries in that subregion. In the 61 low mortality countries (EUR and other Disability weight subregion “A” countries), we used a – Acute non-typhoidal Salmonella literature review that identified national (NTS) diarrhoea (severe): GBD2010 incidence estimates for NTS from seven disability weight of 0.281 (95% UI countries: Australia [272], Canada [175], 0.184–0.399) for diarrhoea, severe. France [174], Netherlands [154], New – Acute NTS diarrhoea (moderate): Zealand [252], United Kingdom [48], GBD2010 disability weight of 0.202 and the United States of America [188]. (95% UI 0.133–0.299) for diarrhoea, These national estimates were based on moderate. systematic reviews, national surveillance – Acute NTS diarrhoea (mild): GBD2010 data, and expert judgment. In these disability weight of 0.061 (95% UI seven countries, we used the estimated 0.036–0.093) for diarrhoea, mild [82]. national NTS incidence (and range) for that country. For low mortality countries Mortality without a national estimate, we used the The mortality of NTS was estimated median NTS incidence from the seven separately for middle-to-high mortality national studies. The median incidence countries, and low mortality countries. was from Australia: 301.5 per 100 000 For the 133 middle-to-high mortality population (which was increased by 19% countries, we used a modification of to account for travellers using proxy the CHERG approach [50]. We received information from New Zealand), with envelopes of diarrhoeal deaths from range 171.1–541.8. WHO; because this estimate was not APPENDICES WHO Estimates of the global burden of foodborne diseases 157

available with an uncertainity interval, expert judgment. In these five countries, we used the uncertainty range from we used the estimated national NTS the GBD2010 estimate of diarrhoeal mortality (and range) for that country. deaths (81.7% to 114.6% around the point For low mortality countries without a estimate) [58]. We then estimated the national estimate, we used the median aetiological proportions of diarrhoeal NTS mortality from the five national deaths due to NTS and the other studies. The median NTS mortality was 10 diarrhoeal pathogens in children from the United States; 0.15 per 100,000 <5 years of age using a CHERG and population: range 0.08 – 0.40. FERG systematic review of aetiology studies among inpatients [40], and the Age distribution aetiological proportions of diarrhoeal In middle-to-high mortality countries deaths due to NTS and the other we estimated incidence and mortality 10 diarrhoeal pathogens in persons of non-typhoidal Salmonella (NTS) >5 years of age, using an updated diarrhoea seperately for children <5 years FERG systematic review of aetiology of age and persons >5 years of age. studies among inpatients [40, 274]. In low mortality countries, the age The NTS aetiological proportions were distribution for NTS diarrhoea cases was extracted from studies, and regional 24% <5 years; 10% 5–14 years; 11% 15–24 median NTS aetiological proportions years; 42% 25–64 years; and 13% >65 calculated. We modified the CHERG years [276]. approach by dropping regional median NTS aetiological proportion outliers that Sex distribution were >5 times greater than the global Salmonella sex distribution: 50% male. median NTS aetiological proportion, and replacing missing regional NTS aetiological proportions with the global A4-13 Invasive Non-typhoidal median. Furthermore, for children Salmonella (iNTS) infection <5 years of age, we proportionally decreased the aetiological proportions Incidence for all 11 diarrhoeal pathogens in each Rates of iNTS are highly correlated with region so that the sum of the aetiological HIV prevalence and malaria risk [277]. To proportions for all diarrhoeal pathogens estimate iNTS incidence globally, we used in a region equalled 1. The resultant age-specific estimates of incidence from regional NTS aetiological proportions a systematic review [277] to construct were multiplied by the regional estimates a random effect log linear model using of diarrhoea deaths, and the resultant covariates of country-specific HIV and regional NTS mortality was applied to malaria deaths, and the log of Gross all countries in that region. In the 61 Domestic Product. As data were sparse, low mortality countries (EUR and other we predicted incidence for all ages, subregion “A” countries), we used a which was converted to age-specific literature review that identified NTS incidence based on age profiles for iNTS mortality estimates from five countries: cases in low and high incidence settings Australia [272], France [174], Netherlands [277]. From this, we predicted iNTS [154], New Zealand [252], and the United incidence among persons not infected States of America [188]. These national with HIV [62, 278]. To estimate deaths, estimates were based on systematic we assumed that the CFR for iNTS reviews, national surveillance data, and in non-HIV infected individuals was a Appendices 158

uniform distribution with a most likely 133 middle-to-high mortality countries, value of 10% (range 5–20%) in subregion we used a modification of the CHERG B to E countries, and a most likely value approach [50]. To derive “envelopes” of of 4.3% (range 3.9–6.6%) in subregion A diarrhoea cases, for children <5 years countries [279]. of age we used estimates of diarrhoea incidence from a CHERG systematic Clinical Outcomes review [51] and for persons >5 years Clinical outcomes were invasive of age we used a FERG-commissioned Salmonella infection and death. systematic review [52]. We then estimated the aetiological proportions of Duration diarrhoeal illnesses due to Campylobacter The duration of iNTS infection was and the 10 other diarrhoeal pathogens2 assumed to be the same as the duration in children <5 years of age using a of typhoid which was estimated to be 28 CHERG and FERG systematic review of days (min. 7 days–max. 56 days). aetiology studies among outpatients and persons in the community [40] and the Disability weight aetiological proportions of diarrhoeal – iNTS infection: GBD2010 disability illnesses due to Campylobacter and the weight of 0.210 (95% UI 0.139–0.298) 10 other diarrhoeal pathogens in persons for infectious disease, acute episode, >5 years of age, using an updated severe [82]. FERG systematic review of aetiology studies among inpatients, outpatients Mortality and persons in the community [40, To estimate deaths, we assumed that 274]. The Campylobacter aetiological the CFR for iNTS in non-HIV infected proportions were extracted from studies, individuals was a uniform distribution and regional median Campylobacter with a most likely value of 10% (range aetiological proportions calculated. 5–20%) in subregion B to E countries and We modified the CHERG approach by a most likely value of 4.3% (range 3.9– dropping regional median Campylobacter 6.6%) in subregion A countries [63]. aetiological proportion outliers that were >5 times greater than the global median Age distribution Campylobacter aetiological proportion, We assessed the age distribution and replacing missing regional of invasive NTS cases and deaths in Campylobacter aetiological proportions high (Mali) and low (United States) with the global median. Furthermore, burden settings. for children <5 years of age, we proportionally decreased the aetiological Sex distribution proportions for all 11 diarrhoeal Salmonella sex distribution: 50% male. pathogens in each region so that the sum of the aetiological proportions for all 11 diarrhoeal pathogens in a region equalled A4.14 Camplyobacter infection 1. The resultant regional Campylobacter Incidence 2 The 11 diarrhoeal pathogens are: non-typhoidal Salmonella, Campylobacter, Shigella, norovirus, The incidence of diarrhoeal enterotoxigenic E. coli (ETEC), enteropathogenic Campylobacter was estimated separately E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba histolytica, other diarrhoeal agents not known for middle-to-high mortality countries, to be foodborne (rotavirus and astrovirus), and and low mortality countries. For the unspecified agents. APPENDICES WHO Estimates of the global burden of foodborne diseases 159

aetiological proportions were multiplied 2% of Campylobacter diarrhoeal cases by the regional estimates of diarrhoea resulted in severe diarrhoea, 25% of incidence, and the resultant regional Campylobacter diarrhoeal cases resulted Campylobacter incidence was applied to in moderate diarrhoea, and 73% of all countries in that region. Campylobacter diarrhoeal cases resulted In the 61 low mortality countries (EUR in mild diarrhoea. and other subregion “A” countries), we Duration used a literature review that identified national incidence estimates for In children <5 years of age, duration of Campylobacter from seven countries: severe diarrhoea was 8.4 days; moderate Australia [272], Canada [175], France diarrhoea was 6.4 days; and mild [174], Netherlands [154], New Zealand diarrhoea was 4.3 days [266]. Based [252], United Kingdom [48], and the on the assumed distribution of severe, United States of America [188]. These moderate and mild diarrhoea cases, the national estimates were based on duration of all Campylobacter diarrhoea systematic reviews, national surveillance cases in children <5 years of age was data, and expert judgment. In these estimated to be 4.9 days (min. 4.3 days– seven countries, we used the estimated max. 8.4 days). In persons >5 years of national Campylobacter incidence (and age, the duration of Campylobacter range) for that country. For low mortality diarrhoea was 2.8 days [266]. The countries without a national estimate, duration of Guillain-Barre Syndrome due we used the median Campylobacter to Campylobacter infection was assumed incidence from the seven national life-long [281]. studies. The median incidence was from Canada: 789.2 per 100 000 population Disability weight (after increasing by 20% to account for – Acute Campylobacter diarrhoea travellers according to proxy infomation (severe): GBD2010 disability weight from the United States of America) with of 0.281 (95% UI 0.184–0.399) for range of 532.3–1140.3. Using a systematic diarrhoea, severe. review that identified 63 papers, updated – Acute Campylobacter diarrhoea for papers published through 2013 for (moderate): GBD2010 disability FERG by the author with the addition of weight of 0.202 (95% UI 0.133–0.299) 9 papers, the incidence of Guillain-Barre for diarrhoea, moderate. Syndrome (GBS) in all countries was – Acute Campylobacter diarrhoea estimated at 1.4 per 100 000 population (mild): GBD2010 disability weight (min. 1.1–max. 1.8) [55]. Based on a of 0.061 (95% UI 0.036–0.093) for systematic review, we assumed that 31% diarrhoea, mild. (min. 28%–max. 45%) of GBS cases were – Guillain-Barre Syndrome due to due to Campylobacter infection [280] Campylobacter infection: GBD201 disability weight of 0.445 (95% UI Clinical Outcomes 0.303–0.593) for multiple sclerosis, Clinical outcomes were acute moderate [82]. Campylobacter diarrhoea (severe); acute Campylobacter diarrhoea Mortality (moderate); acute Campylobacter The mortality of Campylobacter was diarrhoea (mild); Guillain-Barre Syndrome estimated separately for middle-to- due to Campylobacter infection; and high mortality countries, and for low Campylobacter death. We assumed that mortality countries. For the 133 middle- Appendices 160

to-high mortality countries, we used a that identified Campylobacter mortality modification of the CHERG approach estimates from five countries: Australia [50]. We received envelopes of diarrhoeal [272], France [174], Netherlands [154], deaths from WHO; because this estimate New Zealand [252], and the United was not available with an uncertainty States of America [188]. These national interval, we used the uncertainty range estimates were based on systematic from the GBD2010 estimate of diarrhoeal reviews, national surveillance data, deaths (81.7% to 114.6% around the point and expert judgment. In these five estimate) [58]. We then estimated the countries, we used the estimated national aetiological proportions of diarrhoeal Campylobacter mortality (and range) for deaths due to Campylobacter and the 10 that country. For low mortality countries other diarrhoeal pathogens in children without a national estimate, we used the <5 years of age using a CHERG and median Campylobacter mortality from FERG systematic review of aetiology the five national studies. The median studies among inpatients [40], and the Campyloacter mortality was the mean aetiological proportions of diarrhoeal from the United States: 0.04 per 100 000 deaths due to Campylobacter and population, with a range 0–0.17). We the 10 other diarrhoeal pathogens assumed that the case fatality ratio in persons >5 years of age using an for Gullain-Barre Syndrome due to updated FERG systematic review of Campylobacter infection was 4.1% (min. aetiology studies among inpatients [40, 2.4%–max. 6%) [281]. 282]. The Campylobacter aetiological proportions were extracted from studies, Age distribution and regional median Campylobacter In middle-to-high mortality countries aetiological proportions calculated. we estimated incidence and mortality of We modified the CHERG approach by Campylobacter diarrhoea seperately for dropping regional median Campylobacter children <5 years of age and persons >5 aetiological proportion outliers that years of age. In low mortality countries were >5 times greater than the global the age distribution for Campylobacter median Campylobacter aetiological diarrhoea cases was 11% <5 years; proportion, and replacing missing 8% 5–14 years; 10% 15–24 years; 57% regional Campylobacter aetiological 25–64 years; and 14% >65 years [276]. proportions with the global median. We assumed the age distribution of Furthermore, for children <5 years Campylobacter Guillian-Barre Syndrome of age, we proportionally decreased cases and deaths were the same the aetiological proportions for all 11 as Campylobacter diarrhoea cases diarrhoeal pathogens in each region and deaths. so that the sum of the aetiological proportions for all diarrhoeal pathogens Sex distribution in a region equalled 1. The resultant Campylobacter sex distribution: regional Campylobacter aetiological 50% male. proportions were multiplied by the regional estimates of diarrhoea deaths, and the resultant regional Campylobacter A4.15 Norovirus infection mortality was applied to all countries in that region. In the 61 low mortality Incidence countries (EUR and other subregion “A” The incidence of diarrhoeal norovirus countries), we used a literature review and vomiting-only norovirus were APPENDICES WHO Estimates of the global burden of foodborne diseases 161

estimated separately. The incidence <5 years of age, we proportionally of diarrhoeal norovirus was estimated decreased the aetiological proportions separately for middle-to-high mortality for all 11 diarrhoeal pathogens in countries, and low mortality countries. each region so that the sum of the For the 133 middle-to-high mortality aetiological proportions for all diarrhoeal countries, we used a modification of pathogens in a region equalled 1. The the CHERG approach [50]. To derive resultant regional norovirus aetiological “envelopes” of diarrhoea cases, for proportions were multiplied by the children <5 years of age we used regional estimates of diarrhoea incidence, estimates of diarrhoea incidence from and the resultant regional norovirus a CHERG systematic review [51] and incidence was applied to all countries in for persons >5 years of age we used a that region. FERG-commissioned systematic review In the 61 low mortality countries (EUR [52]. We then estimated the aetiological and other subregion “A” countries), we proportions of diarrhoeal illnesses used a literature review that identified due to norovirus and the 10 other national incidence estimates for norovirus 3 diarrhoeal pathogens in children <5 from seven countries: Australia [272], years of age using a CHERG and FERG Canada [175], France [174], Netherlands systematic review of aetiology studies [154], New Zealand [252], United among outpatients and persons in the Kingdom [48], and the United States of community [40], and the aetiological America [188]. These national estimates proportions of diarrhoeal illnesses due were based on systematic reviews, to norovirus and the 10 other diarrhoeal national surveillance data, and expert pathogens in persons >5 years of age, judgment. In these seven countries, we using an updated FERG systematic used the estimated national norovirus review of aetiology studies among incidence (and range) for that country. inpatients, outpatients and persons in the For low mortality countries without a community [40, 274]; these systematic national estimate, we used the median reviews were supplemented by a FERG- norovirus incidence from the seven commissioned norovirus systematic national studies. The median incidence review [283]. The norovirus aetiological was from the United States: 6978.5 per proportions were extracted from 100 000 population, with range 4 295.0– studies, and regional median norovirus 10 282.3. aetiological proportions calculated. We modified the CHERG approach by To estimate the incidence of vomiting- dropping regional median norovirus only norovirus, based on a FERG- aetiological proportion outliers that were commissioned systematic review [57], we multiplied the incidence of diarrhoeal >5 times greater than the global median norovirus by 19% (min. 15%–max. 23%). norovirus aetiological proportion, and replacing missing regional norovirus Clinical Outcomes aetiological proportions with the global median. Furthermore, for children Clinical outcomes were acute norovirus diarrhoea (severe); acute norovirus 3 The 11 diarrhoeal pathogens are: non-typhoidal Salmonella, Campylobacter, Shigella, norovirus, diarrhoea (moderate); acute norovirus enterotoxigenic E. coli (ETEC), enteropathogenic diarrhoea (mild); acute norovirus E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba vomiting-only; and norovirus death. We histolytica, other diarrhoeal agents not known to be foodborne (rotavirus and astrovirus), and assumed that 0.5% of norovirus diarrhoea unspecified agents. cases resulted in severe diarrhoea, 8.5% Appendices 162

of norovirus diarrhoea cases resulted norovirus and the other 10 diarrhoeal in moderate diarrhoea, and 91% of pathogens in persons >5 years of age norovirus diarrhoea cases resulted in using an updated FERG systematic mild diarrhoea. review of aetiology studies among inpatients [40, 274]; these systematic Duration reviews were supplemented by a FERG- The duration of norovirus diarrhoea was commissioned norovirus systematic estimated to be 2 days (min. 1 day–max. review [192]. The norovirus aetiological 4 days). We assumed norovirus vomiting- proportions were extracted from only cases had the same duration as studies, and regional median norovirus norovirus diarrhoea cases. aetiological proportions calculated. We modified the CHERG approach by Disability weight dropping regional median norovirus – Acute norovirus diarrhoea (severe): aetiological proportion outliers that were GBD2010 disability weight of 0.281 >5 times greater than the global median (95% UI 0.184–0.399) for diarrhoea, norovirus aetiological proportion, and severe. replacing missing regional norovirus – Acute norovirus diarrhoea (moderate): aetiological proportions with the global GBD2010 disability weight of 0.202 median. Furthermore, for children (95% UI 0.133–0.299) for diarrhoea, <5 years of age, we proportionally moderate. decreased the aetiological proportions – Acute norovirus diarrhoea (mild) for all 11 diarrhoeal pathogens in and acute norovirus vomiting-only: each region so that the sum of the GBD2010 disability weight of 0.061 aetiological proportions for all diarrhoeal (95% UI 0.036–0.093) for diarrhoea, pathogens in a region equalled 1. The mild [82]. resultant regional norovirus aetiological proportions were multiplied by the Mortality regional estimates of diarrhoea deaths, The mortality of norovirus was estimated and the resultant regional norovirus separately for middle-to-high mortality mortality was applied to all countries countries, and low mortality countries. in that region. In the 61 low mortality For the 133 middle-to-high mortality countries (EUR and other subregion “A” countries, we used a modification of countries), we used a literature review the CHERG approach [50]. We received that identified norovirus mortality envelopes of diarrhoeal death from WHO; estimates from four countries: Australia because this estimate was not available [272], Netherlands [154], New Zealand with an uncertainity interval, we used the [252], and the United States of America uncertainity range from the GBD2010 [188]. These national estimates were estimate of diarrhoeal deaths (81.7% to based on systematic reviews, national 114.6% around the point estimate) [58]. surveillance data, and expert judgment. We then estimated the aetiological In these four countries, we used the proportions of diarrhoeal deaths due to estimated national norovirus mortality norovirus and the other 10 diarrhoeal (and range) for that country. For low pathogens in children <5 years of age mortality countries without a national using a CHERG and FERG systematic estimate, we used the median norovirus review of aetiology studies among mortality from the four national studies. inpatients [40], and the aetiological The median norovirus mortality was the proportions of diarrhoeal deaths due to mean from New Zealand and the United APPENDICES WHO Estimates of the global burden of foodborne diseases 163

States: 0.18 per 100 000 with a range of to Shigella and the 10 other diarrhoeal 0.11– 0.28. We assumed no deaths among pathogens in persons >5 years of age vomiting-only norovirus cases. using an updated FERG systematic review of aetiology studies among Age distribution inpatients, outpatients and persons in In middle-to-high mortality countries the community [40, 274]. The shigellosis we estimated incidence and mortality aetiological proportions were extracted of norovirus seperately for children <5 from studies, and regional median years of age and persons >5 years of shigellosis aetiological proportions age. In low mortality countries the age calculated. We modified the CHERG distribution for norovirus was 40% <5 approach by dropping regional median years; 10% 5–14 years; 30% 15–44 years; shigellosis aetiological proportion 10% 45–64 years; and 10% outliers that were >5 times greater >65 years [284]. than the global median shigellosis aetiological proportion, and replacing Sex distribution missing regional shigellosis aetiological Norovirus sex distribution: 50% male. proportions with the global median. Furthermore, for children <5 years of age, we proportionally decreased A4.16 Shigellosis the aetiological proportions for all 11 diarrhoeal pathogens in each region Incidence so that the sum of the aetiological The incidence of shigellosis was proportions for all diarrhoeal pathogens estimated separately for middle-to- in a region equalled 1. The resultant high mortality countries, and low regional shigellosis aetiological mortality countries. For the 133 middle- proportions were multiplied by the to-high mortality countries, we used a regional estimates of diarrhoea incidence, modification of the CHERG approach and the resultant regional shigellosis [50]. To derive “envelopes” of diarrhoea incidence was applied to all countries in cases, for children <5 years of age we that region. used estimates of diarrhoea incidence In the 61 low mortality countries (EUR from a CHERG systematic review [51] and other subregion “A” countries), we and for persons >5 years of age we used a literature review that identified used a FERG-commissioned systematic national incidence estimates for review [52]. We then estimated the shigellosis from five countries: Australia aetiological proportions of diarrhoeal [272], Canada [175], France [174], New illnesses due to Shigella and the 10 other Zealand [252], and the United States of diarrhoeal pathogens4 in children <5 America [188]. These national estimates years of age using a CHERG and FERG were based on systematic reviews, systematic review of aetiology studies national surveillance data, and expert among outpatients and persons in the judgment. In these five countries, we community [40], and the aetiological used the estimated national shigellosis proportion of diarrhoeal illnesses due incidence (and range) for that country. 4 The 11 diarrhoeal pathogens are: non-typhoidal For low mortality countries without a Salmonella, Campylobacter, Shigella, norovirus, enterotoxigenic E. coli (ETEC), enteropathogenic national estimate, we used the median E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba shigellosis incidence from the five histolytica, other diarrhoeal agents not known national studies. The median incidence to be foodborne (rotavirus and astrovirus), and unspecified agents. was from Canada (which was increased Appendices 164

by 8% to account for travellers, using For the 133 middle-to-high mortality proxy information from the United States countries, we used a modification of of America) which was 23.6 per 100 000 the CHERG approach [50]. We received population, with a range of 13.2–38.7. envelopes of diarrhoeal deaths from WHO; because this estimate was not Clinical Outcomes available with an uncertainity interval, Clinical outcomes were acute Shigella we used the uncertainity range from diarrhoea (severe); acute Shigella the GBD2010 estimate of diarrhoeal diarrhoea (moderate); acute Shigella deaths (81.7% to 114.6% around the point diarrhoea (mild); and Shigella death. estimate) [58]. We then estimated the We assumed that 2% of Shigella cases aetiological proportions of diarrhoeal resulted in severe diarrhoea, 25% of deaths due to Shigella and 10 other Shigella cases resulted in moderate diarrhoeal pathogens5 in children diarrhoea, and 73% of Shigella cases <5 years of age using a CHERG and resulted in mild diarrhoea. FERG systematic review of aetiology studies among inpatients [40], and the Duration aetiological proportions of diarrhoeal In children <5 years of age, duration of deaths due to Shigella and the 10 severe diarrhoea was 8.4 days, moderate other diarrhoeal pathogens in persons diarrhoea was 6.4 days, and mild >5 years of age using an updated diarrhoea was 4.3 days [266]. Based FERG systematic review of aetiology on the assumed distribution of severe, studies among inpatients [40, 274]. moderate and mild diarrhoea cases, the The shigellosis aetiological proportions duration of Shigella diarrhoea cases in were extracted from studies, and children <5 years of age was estimated regional median shigellosis aetiological to be 4.9 days (min. 4.3 days–max. proportions calculated. We modified 8.4 days). In persons >5 years of age, the CHERG approach by dropping the duration of Shigella diarrhoea was regional median shigellosis aetiological 2.8 days [266]. proportion outliers that were >5 times greater than the global median shigellosis Disability weight aetiological proportion, and replacing – Acute Shigella diarrhoea (severe): missing regional shigellosis aetiological GBD2010 disability weight of 0.281 proportions with the global median. (95% UI 0.184–0.399) for diarrhoea, Furthermore, for children <5 years severe. of age, we proportionally decreased – Acute Shigella diarrhoea (moderate): the aetiological proportions for all 11 GBD2010 disability weight of 0.202 diarrhoeal pathogens in each region (95% UI 0.133–0.299) for diarrhoea, so that the sum of the aetiological moderate. proportions for all diarrhoeal pathogens – Acute Shigella diarrhoea (mild): in a region equalled 1. The resultant GBD2010 disability weight of 0.061 regional shigellosis aetiological (95% UI 0.036–0.093) for diarrhoea, proportions were multiplied by the mild [82]. 5 The 11 diarrhoeal pathogens are: non-typhoidal Mortality Salmonella, Campylobacter, Shigella, norovirus, The mortality of shigellosis was estimated enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba separately for middle-to-high mortality histolytica, other diarrhoeal agents not known countries, and low mortality countries. to be foodborne (rotavirus and astrovirus), and unspecified agents. APPENDICES WHO Estimates of the global burden of foodborne diseases 165

regional estimates of diarrhoea deaths, [52]. We then estimated the aetiological and the resultant regional shigellosis proportions of diarrhoeal illnesses due to mortality was applied to all countries ETEC and 10 other diarrhoeal pathogens in that region. In the 61 low mortality in children <5 years of age using a countries (EUR and other subregion “A” CHERG and FERG systematic review of countries), we used a literature review aetiology studies among outpatients that identified shigellosis mortality and persons in the community [40], estimates from the United States of and the aetiological proportion of America [188]. This national estimate was diarrhoeal illnesses due to ETEC and 10 based on national surveillance data, and other diarrhoeal pathogens in persons expert judgment. We used the shigellosis >5 years of age using an updated FERG mortality from the United States of systematic review of aetiology studies America for all low mortality countries: among inpatients, outpatients and 0.013 per 100,000 population with range persons in the community [40, 274]. 0.002 – 0.085. The ETEC aetiological proportions were extracted from studies, and regional Age distribution median ETEC aetiological proportions calculated. We modified the CHERG In middle-to-high mortality countries approach by dropping regional median we estimated incidence and mortality ETEC aetiological proportion outliers that of Shigella separately for children <5 were >5 times greater than the global years of age and persons >5 years of median ETEC aetiological proportion, age. In low mortality countries the age and replacing missing regional ETEC distribution for Shigella cases was 24% aetiological proportions with the global <5 years; 23% 5–14 years; 10% 15–24 median. Furthermore, for children years; 39% 25–64 years; and 4% >65 <5 years of age, we proportionally years [276]. decreased the aetiological proportions for all 11 diarrhoeal pathogens in each Sex distribution region so that the sum of the aetiological Shigella sex distribution: 50% male. proportions for all diarrhoeal pathogens in a region equalled 1. The resultant A4.17 Enterotoxigenic Escherichia regional ETEC aetiological proportions coli (ETEC) infection were multiplied by the regional estimates of diarrhoea incidence, and the resultant Incidence regional ETEC incidence was applied to all countries in that region. In the 61 The incidence of diarrhoea due to ETEC low mortality countries (EUR and other was estimated separately for middle- subregion “A” countries), a liturature to-high mortality countries, and low review identified a national incidence mortality countries. For the 133 middle- estimates for ETEC in the United States to-high mortality countries, we used a of America that was based on national modification of the CHERG approach surveillance data, and expert judgment [50]. To derive “envelopes” of diarrhoea [188]. We used the ETEC incidence from cases, for children <5 years of age we the United States for all low mortality used estimates of diarrhoea incidence countries; 13.3 per 100,000 population from a CHERG systematic review [51] and with range 3.9 – 34.2. for persons >5 years of age we used a FERG-commissioned systematic review Appendices 166

Clinical Outcomes available with an uncertainty interval, Clinical outcomes were acute ETEC we used the uncertainty range from diarrhoea (severe); acute ETEC diarrhoea the GBD2010 estimate of diarrhoeal (moderate); acute ETEC diarrhoea (mild); deaths (81.7% to 114.6% around the point and death. We assumed that 0.5% of estimate) [58]. We then estimated the ETEC cases resulted in severe diarrhoea, aetiological proportions of diarrhoeal 8.5% of ETEC cases resulted in moderate deaths due to ETEC and 10 other diarrhoea, and 91% of ETEC cases diarrhoeal pathogens6 in children resulted in mild diarrhoea. <5 years of age using a CHERG and FERG systematic review of aetiology Duration studies among inpatients [40], and the In children <5 years of age, duration of aetiological proportions of diarrhoeal severe diarrhoea was 8.4 days, moderate deaths due to ETEC and 10 other diarrhoea was 6.4 days, and mild diarrhoeal pathogens in persons >5 diarrhoea was 4.3 days [266]. Based years of age using an updated FERG on the assumed distribution of severe, systematic review of aetiology studies moderate and mild diarrhoea cases, the among inpatients [40, 274]. The ETEC duration of ETEC diarrhoea cases in aetiological proportions were extracted children <5 years of age was estimated from studies, and regional median ETEC to be 4.9 days (min. 4.3 days–max. 8.4 aetiological proportions calculated. days). In persons >5 years of age, the We modified the CHERG approach duration of ETEC diarrhoea was 2.8 days by dropping regional median ETEC [266]. aetiological proportion outliers that were >5 times greater than the global Disability weight median ETEC aetiological proportion, and replacing missing regional ETEC – Acute ETEC diarrhoea (severe): aetiological proportions with the global GBD2010 disability weight of 0.281 median. Furthermore, for children (95% UI 0.184–0.399) for diarrhoea, <5 years of age, we proportionally severe. decreased the aetiological proportions – Acute ETEC diarrhoea (moderate): for all 11 diarrhoeal pathogens in each GBD2010 disability weight of 0.202 region so that the sum of the aetiological (95% UI 0.133–0.299) for diarrhoea, proportions for all diarrhoeal pathogens moderate. in a region equalled 1. The resultant ETEC – Acute ETEC diarrhoea (mild): aetiological proportions were multiplied GBD2010 disability weight of 0.061 by the regional estimates of diarrhoea (95% UI 0.036–0.093) for diarrhoea, deaths, and the resultant regional ETEC mild [82]. mortality was applied to all countries Mortality in that region. We estimated no ETEC The mortality of ETEC was estimated deaths in the 61 low mortality countries separately for middle-to-high mortality (EUR and other subregion “A” countries). countries, and low mortality countries. For the 133 middle-to-high mortality countries, we used a modification of 6 The 11 diarrhoeal pathogens are: non-typhoidal the CHERG approach [50]. We received Salmonella, Campylobacter, Shigella, norovirus, enterotoxigenic E. coli (ETEC), enteropathogenic envelopes of diarrhoeal deaths from E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba WHO; because this estimate was not histolytica, other diarrhoeal agents not known to be foodborne (rotavirus and astrovirus), and unspecified agents. APPENDICES WHO Estimates of the global burden of foodborne diseases 167

Age distribution to EPEC and the 10 other diarrhoeal In middle-to-high mortality countries pathogens in persons >5 years of age we estimated incidence of diarrhoea using an updated FERG systematic seperately for children <5 years of age review of aetiology studies among and persons >5 years of age. In low inpatients, outpatients and persons in mortality countries, no information the community [40, 274]. The EPEC was available on the age distribution of aetiological proportions were extracted EPEC cases; we therefore, used the age from studies, and regional median EPEC distribution for Campylobacter diarrhoea aetiological proportions calculated. cases as a proxy, which was 11% <5 years; We modified the CHERG approach 8% 5–14 years; 10% 15–24 years; 57% by dropping regional median EPEC 25–64 years; and 14% >65 years. aetiological proportion outliers that were >5 times greater than the global Sex distribution median EPEC aetiological proportion, and replacing missing regional EPEC ETEC sex distribution: 50% male. aetiological proportions with the global median. Furthermore, for children A4.18 Enteropathogenic <5 years of age, we proportionally Escherichia coli (EPEC) infection decreased the aetiological proportions for all 11 diarrhoeal pathogens in each Incidence region so that the sum of the aetiological The incidence of diarrhoea due to EPEC proportions for all diarrhoeal pathogens was estimated separately for middle- in a region equalled 1. The resultant to-high mortality countries, and low regional EPEC aetiological proportions mortality countries. For the 133 middle- were multiplied by the regional estimates to-high mortality countries, we used a of diarrhoea incidence, and the resultant modification of the CHERG approach regional EPEC incidence was applied [50]. To derive “envelopes” of diarrhoea to all countries in that region. In the 61 cases, for children <5 years of age we low mortality countries (EUR and other used estimates of diarrhoea incidence subregion “A” countries), we adopted the from a CHERG systematic review [51] assumption used in the national study in and for persons >5 years of age we the United States of America that EPEC used a FERG-commissioned systematic was as common as enterotoxigenic E. coli review [52]. We then estimated the [188]. The national estimate for ETEC in aetiological proportions of diarrhoeal the United States of America was based illnesses due to EPEC and the 10 other on national surveillance data, and expert diarrhoeal pathogens7 in children <5 judgment. For low mortality countries, we years of age using a CHERG and FERG used the EPEC incidence from the United systematic review of aetiology studies States of America, which was 13.33 per among outpatients and persons in the 100 000 population with range 4.00 – 34.24. community [40], and the aetiological proportion of diarrhoeal illnesses due Clinical Outcomes 7 The 11 diarrhoeal pathogens are: non-typhoidal Salmonella, Campylobacter, Shigella, norovirus, Clinical outcomes were acute EPEC enterotoxigenic E. coli (ETEC), enteropathogenic diarrhoea (severe); acute EPEC diarrhoea E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba (moderate); acute EPEC diarrhoea (mild); histolytica, other diarrhoeal agents not known to be foodborne (rotavirus and astrovirus), and and EPEC death. We assumed that unspecified agents. 0.5% of EPEC cases resulted in severe Appendices 168

diarrhoea, 8.5% of EPEC cases resulted <5 years of age using a CHERG and FERG in moderate diarrhoea, and 91% of EPEC systematic review of aetiology studies cases resulted in mild diarrhoea. among inpatients [40], and the aetiological proportions of diarrhoeal deaths due Duration to EPEC and the 10 other diarrhoeal In children <5 years of age, duration pathogens in persons >5 years of age using of severe diarrhoea was 8.4 days, an updated FERG systematic review of moderate diarrhoea was 6.4 days, and aetiology studies among inpatients [40, mild diarrhoea was 4.3 days [188]. Based 274]. The EPEC aetiological proportions on the assumed distribution of severe, were extracted from studies, and regional moderate and mild diarrhoea cases, the median EPEC aetiological proportions duration of EPEC diarrhoea cases in calculated. We modified the CHERG children <5 years of age was estimated approach by dropping regional median to be 4.9 days (min. 4.3 days–max. 8.4 EPEC aetiological proportion outliers that days). In persons >5 years of age, the were >5 times greater than the global duration of diarrhoea was 2.8 days [266]. median EPEC aetiological proportion, and replacing missing regional EPEC Disability weight aetiological proportions with the global – Acute EPEC diarrhoea (severe): median. Furthermore, for children <5 years GBD2010 disability weight of 0.281 of age, we proportionally decreased the (95% UI 0.184–0.399) for diarrhoea, aetiological proportions for all 11 diarrhoeal severe. pathogens in each region so that the – Acute EPEC diarrhoea (moderate): sum of the aetiological proportions for all GBD2010 disability weight of 0.202 diarrhoeal pathogens in a region equalled 1. (95% UI 0.133–0.299) for diarrhoea, The resultant EPEC aetiological proportions moderate. were multiplied by the regional estimates of – Acute EPEC diarrhoea (mild): diarrhoea deaths, and the resultant regional GBD2010 disability weight of 0.061 EPEC mortality was applied to all countries (95% UI 0.036–0.093) for diarrhoea, in that region. We estimated no EPEC deaths in the 61 low mortality countries mild [82]. (EUR and other subregion “A” countries). Mortality The mortality of EPEC was estimated Age distribution separately for middle-to-high mortality In middle-to-high mortality countries countries, and low mortality countries. we estimated incidence and mortality For the 133 middle-to-high mortality of EPEC separately for children <5 years countries, we used a modification of the of age and persons >5 years of age. In CHERG approach [50]. We received low mortality countries, no information envelopes of diarrhoeal deaths from WHO; was available on the age distribution of because this estimate was not available EPEC cases; we therefore used the age with an uncertainity interval, we used the distribution for Campylobacter diarrhoea uncertainity range from the GBD2010 cases as a proxy, which was 11% <5 years; estimate of diarrhoeal deaths (81.7% to 8% 5–14 years; 10% 15–24 years; 57% 114.6% around the point estimate) [58]. We 25–64 years; and 14% >65 years. then estimated the aetiological proportions of diarrhoeal deaths due to EPEC and the Sex distribution 10 other diarrhoeal pathogens in children EPEC sex distribution: 50% male. APPENDICES WHO Estimates of the global burden of foodborne diseases 169

A4.19 Cholera the global burden of cholera [285] range of estimates around the mean estimate Incidence of global cholera cases (2.8 million with Estimates of the incidence of cholera a range of 1.4 to 4.3 million) to derive a were adapted from a published range of estimates for cholera incidence. systematic review of the global burden of cholera [285] updated with 2010 Clinical Outcomes population estimates. This review Clinical outcomes were cholera (severe); classified 51 countries as cholera- cholera (moderate); cholera (mild); endemic countries based on results and cholera death. We assumed that of the systematic review and national 35% of cholera cases resulted in severe cholera reports in the WHO Weekly cholera, 40% of cholera cases resulted Epidemiological Record. The review in moderate cholera, and 25% of cholera then used WHO 2008 country-specific cases resulted in mild cholera [290, 291]. estimates of the proportion of each country’s population that lacked Duration improved sanitation [286] to estimate We assumed the duration of cholera was the proportion of the population in the 7 days (min. 3 day–max. 10 days). cholera-endemic countries that were at risk for cholera. Then a cholera incidence Disability weight was assigned to the population at risk – Cholera (severe): GBD2010 disability for cholera in the cholera endemic weight of 0.281 (95% UI 0.184–0.399) countries based on population-based for diarrhoea, severe. studies in India [287], Indonesia [288] – Cholera (moderate): GBD2010 and Mozambique [289]. The review also disability weight of 0.202 (95% UI identified an additional 18 countries that 0.133–0.299) for diarrhoea, moderate. reported cholera to WHO during 2000 – Cholera (mild): GBD2010 disability to 2008, but were judged to be not be weight of 0.061 (95% UI 0.036–0.093) endemic for cholera; a country-specific for diarrhoea, mild [82]. cholera incidence in each of these “non- endemic” countries was estimated Mortality using the annual average number of For 51 cholera-endemic and 18 cholera cholera cases reported to WHO cases in non-endemic countries, we used the case each country times a multiplier of 10 to fatality ratios (CFRs) estimated in the account for under-reporting. For all other systematic review of the global burden countries, we used a literature review of cholera [285]. This review calculated that identified national cholera incidence a variance-weighted average cholera estimates from three countries countries: CFR by region; the CFR was 1% in WPR France [174], New Zealand [252] and subregion B, 1% in SEAR B (except 1.5% in the United States of America [188]. The Bangladesh), 1.3% in EMR B, 3% in SEAR cholera incidence in the United States D, 3.2% in EMR D, and 3.8% in AFR. For of America was the median estimate all other countries, the literature review of from these three countries and was used national incidence estimates for cholera (0.093 per 100 000 population) as the identified no reported deaths; therefore cholera incidence for all countries (other we assumed no cholera deaths occurred than the cholera-endemic and non- in countries (other than the cholera- endemic countries) which did not have endemic and non-endemic countries). national incidence estimates. We used We used the global burden of cholera Appendices 170

[285] range of estimates around the proportions calculated. We modified the mean estimate of global cholera deaths CHERG approach by dropping regional (91 000, with a range of 28 000 to median cryptosporidiosis aetiological 142 000) to derive a range of estimates proportion outliers that were >5 for cholera deaths. times greater than the global median cryptosporidiosis aetiological proportion, Age distribution and replacing missing regional Cholera age distribution: 15% <5 years; cryptosporidiosis aetiological proportions 25% 5–14 years; 42% 15–34 years; 15% with the global median. Furthermore, 35–64; 3% >60 years [292, 293]. for children <5 years of age, we proportionally decreased the aetiological Sex distribution proportions for all 11 diarrhoeal Cholera sex distribution: 50% male. pathogens in each region so that the sum of the aetiological proportions for all A4.20 Cryptosporidiosis diarrhoeal pathogens in a region equalled 1. The resultant regional cryptosporidiosis Incidence aetiological proportions were multiplied by the regional estimates of diarrhoea The incidence of cryptosporidiosis incidence, and the resultant regional was estimated separately for middle- to-high mortality countries, and low cryptosporidiosis incidence was applied mortality countries. For the 133 middle- to all countries in that region. In the 61 to-high mortality countries, we used a low mortality countries (EUR and other modification of the CHERG approach subregion “A” countries), we used a [50]. To derive “envelopes” of diarrhoea literature review that identified national cases, for children <5 years of age we incidence estimates for cryptosporidiosis used estimates of diarrhoea incidence from six countries: Australia [272], from a CHERG systematic review [51] Canada [175], Netherlands [154], New and for persons >5 years of age we Zealand [252], United Kingdom [48] used a FERG-commissioned systematic and the United States of America [188]. review [52]. We then estimated the These national estimates were based on aetiological proportions of diarrhoeal systematic reviews, national surveillance illnesses due to Cryptosporidia and 10 data, and expert judgment. In these other diarrhoeal pathogens in children <5 six countries, we used the estimated years of age using a CHERG and FERG national cryptosporidiosis incidence (and systematic review of aetiology studies range) for that country. For low mortality among outpatients and persons in the countries without a national estimate, community [40], and the aetiological we used the median cryptosporidiosis proportion of diarrhoeal illnesses incidence from the six national studies. due to Cryptosporidia and 10 other The median incidence was the mean diarrhoeal pathogens in persons >5 from Australia (which was increased years of age using an updated FERG by 19% to account for travellers, using systematic review of aetiology studies proxy information from New Zealand) among inpatients, outpatients and and the Netherlands, which was 128.4 per persons in the community [40, 274]. The 100 000 population with a range cryptosporidiosis aetiological proportions of 50.3 – 601.6. were extracted from studies, and regional median cryptosporidiosis aetiological APPENDICES WHO Estimates of the global burden of foodborne diseases 171

Clinical Outcomes [50]. We received envelopes of diarrhoeal Clinical outcomes were acute deaths from WHO; because this estimate cryptosporidiosis diarrhoea (severe); was not available with an uncertainity acute cryptosporidiosis diarrhoea interval, we used the uncertainity range (moderate); acute cryptosporidiosis from the GBD2010 estimate of diarrhoeal diarrhoea (mild); and death. We assumed deaths (81.7% to 114.6% around the point that 0.5% of cryptosporidiosis cases estimate) (14). We then estimated the resulted in severe diarrhoea, 8.5% aetiological proportions of diarrhoeal of cryptosporidiosis cases resulted deaths due to Cryptosporidia and 10 in moderate diarrhoea, and 91% of other diarrhoeal pathogens8 in children cryptosporidiosis cases resulted in <5 years of age using a CHERG and mild diarrhoea. FERG systematic review of aetiology studies among inpatients [40], and the Duration aetiological proportions of diarrhoeal In children <5 years of age, duration of deaths due to Cryptosporidia and 10 severe diarrhoea was 8.4 days, moderate other diarrhoeal pathogens in persons diarrhoea was 6.4 days, and mild >5 years of age using an updated diarrhoea was 4.3 days [266]. Based FERG systematic review of aetiology on the assumed distribution of severe, studies among inpatients [40, 274]. The moderate and mild diarrhoea cases, the cryptosporidiosis aetiological proportions duration of cryptosporidiosis diarrhoea were extracted from studies, and regional cases in children <5 years of age was median cryptosporidiosis aetiological estimated to be 4.9 days (min. 4.3 days– proportions calculated. We modified the max. 8.4 days). In persons >5 years of CHERG approach by dropping regional age, the duration of diarrhoea was 2.8 median cryptosporidiosis aetiological days [266]. proportion outliers that were >5 times greater than the global median Disability weight cryptosporidiosis aetiological proportion, – Acute cryptosporidiosis diarrhoea and replacing missing regional (severe): GBD2010 disability weight cryptosporidiosis aetiological proportions of 0.281 (95% UI 0.184–0.399) for with the global median. Furthermore, diarrhoea, severe. for children <5 years of age, we – Acute cryptosporidiosis diarrhoea proportionally decreased the aetiological (moderate): GBD2010 disability proportions for all 11 diarrhoeal weight of 0.202 (95% UI 0.133–0.299) pathogens in each region so that the for diarrhoea, moderate. sum of the aetiological proportions for – Acute cryptosporidiosis diarrhoea all diarrhoeal pathogens in a region (mild): GBD2010 disability weight equalled 1. The resultant regional of 0.061 (95% UI 0.036–0.093) for cryptosporidiosis aetiological proportions diarrhoea, mild [82]. were multiplied by the regional estimates of diarrhoea deaths, and the resultant Mortality regional cryptosporidiosis mortality was The mortality of cryptosporidiosis 8 The 11 diarrhoeal pathogens are: non-typhoidal was estimated separately for middle- Salmonella, Campylobacter, Shigella, norovirus, to-high mortality countries, and low enterotoxigenic E. coli (ETEC), enteropathogenic mortality countries. For the 133 middle- E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba histolytica, other diarrhoeal agents not known to-high mortality countries, we used a to be foodborne (rotavirus and astrovirus), and modification of the CHERG approach unspecified agents. Appendices 172

applied to all countries in that region. “envelopes” of diarrhoea cases, for In the 61 low mortality countries (EUR children <5 years of age we used and other subregion “A” countries), we estimates of diarrhoea incidence from used a literature review that identified a CHERG systematic review [51] and cryptosporidiosis mortality estimates for persons >5 years of age we used from three countries: Netherlands [154], a FERG-commissioned systematic New Zealand [252] and the United review [52]. We then estimated the States of America [188]. These national aetiological proportions of diarrhoeal estimates were based on systematic illnesses due to Giardia and the 10 other reviews, national surveillance data, diarrhoeal pathogens in children <5 and expert judgment. In these three years of age using a CHERG and FERG countries, we used the estimated national systematic review of aetiology studies cryptosporidiosis mortality (and range) among outpatients and persons in the for that country. For low mortality community [40], and the aetiological countries without a national estimate, proportion of diarrhoeal illnesses due we used the median cryptosporidiosis to Giardia and the 10 other diarrhoeal mortality from the three national studies. pathogens in persons >5 years of age The median cryptosporidiosis mortality using an updated FERG systematic was from the United States: 0.015 per review of aetiology studies among 100 000 population with a range of inpatients, outpatients and persons in range 0.003 – 0.080. the community [40, 274]. The aetiological proportions were extracted Age distribution from studies, and regional median giardiasis aetiological proportions In middle-to-high mortality countries, we calculated. We modified the CHERG estimated incidence of cryptosporidiosis approach by dropping regional median seperately for children <5 years of age giardiasis aetiological proportion and persons >5 years of age. In low outliers that were >5 times greater mortality countries, the age distribution than the global median giardiasis for cryptosporidiosis was 16% <5 years; aetiological proportion, and replacing 17% 5–14 years; 13% 15–24 years; 14% missing regional giardiasis aetiological 25–34 years; 11% 35–44 years; 9% 45–54 proportions with the global median. years; 7% 55–64 years; 6% 65–74 years; Furthermore, for children <5 years 7% >75 years [294]. of age, we proportionally decreased the aetiological proportions for all 11 Sex distribution diarrhoeal pathogens in each region Cryptosporidiosis sex distribution: so that the sum of the aetiological 50% male. proportions for all diarrhoeal pathogens in a region equalled 1. The resultant A4.21 Giardiasis regional giardiasis aetiological proportions were multiplied by the Incidence regional estimates of diarrhoea incidence, The incidence of giardiasis was estimated and the resultant regional giardiasis separately for middle-to-high mortality incidence was applied to all countries in that region. countries, and low mortality countries. For the 133 middle-to-high mortality In the 61 low mortality countries (EUR countries, we used a modification of and other subregion “A” countries), we the CHERG approach [50]. To derive used a literature review that identified APPENDICES WHO Estimates of the global burden of foodborne diseases 173

national incidence estimates for giardiasis Disability weight from six countries: Australia [272], – Acute giardiasis diarrhoea (severe): Canada [175], Netherlands [154], New GBD2010 disability weight of 0.281 Zealand [252], United Kingdom [48] (95% UI 0.184–0.399) for diarrhoea, and the United States of America [188]. severe. These national estimates were based on – Acute giardiasis diarrhoea (moderate): systematic reviews, national surveillance GBD2010 disability weight of 0.202 data, and expert judgment. In these six (95% UI 0.133–0.299) for diarrhoea, countries, we used the estimated national moderate. giardiasis incidence (and range) for that – Acute giardiasis diarrhoea (mild): country. For low mortality countries GBD2010 disability weight of 0.061 without a national estimate, we used the (95% UI 0.036–0.093) for diarrhoea, median giardiasis incidence from the six mild [82]. national studies. The median incidence Mortality was the mean from Canada (which was increased by 8% to account for travellers, We estimated no giardiasis deaths. using proxy information from the United Age distribution States of America) and the United States of America, which was 384.6 per In middle-to-high mortality countries, 100 000 population, with a range of we estimated incidence of giardiasis 266.4–537.0. seperately for children <5 years of age and persons >5 years of age. In low Clinical Outcomes mortality countries, the age distribution for cryptosporidiosis was 20% <5 years; Clinical outcomes were acute giardiasis 17% 5–14 years; 10% 15–24 years; 11% diarrhoea (severe); acute giardiasis 25–34 years; 12% 35–44 years; 12% 45–54 diarrhoea (moderate); acute giardiasis years; 9% 55–64 years; 5% 65–74 years; diarrhoea (mild); and giardiasis death. 4% >75 years [295]. We assumed that 0.5% of giardiasis cases resulted in severe diarrhoea, 8.5% Sex distribution of giardiasis cases resulted in moderate Giardiasis sex distribution: 50% male. diarrhoea, and 91% of giardiasis cases resulted in mild diarrhoea. A4.22 Amoebiasis Duration In children <5 years of age, duration of Incidence severe diarrhoea was 8.4 days, moderate The incidence of diarrhoea due to diarrhoea was 6.4 days, and mild amoebiasis was estimated separately diarrhoea was 4.3 days [266]. Based for middle-to-high mortality countries, on the assumed distribution of severe, and low mortality countries. For the moderate and mild diarrhoea cases, the 133 middle-to-high mortality countries, duration of giardiasis diarrhoea cases in we used a modification of the CHERG children <5 years of age was estimated approach [50]. To derive “envelopes” to be 4.9 days (min. 4.3 days–max. 8.4 of diarrhoea cases, for children <5 days). In persons >5 years of age, the years of age we used estimates of duration of diarrhoea was 2.8 days [266]. diarrhoea incidence from a CHERG systematic review [51] and for persons >5 years of age we used a FERG- Appendices 174

commissioned systematic review [52]. Clinical Outcomes We then estimated the aetiological Clinical outcomes were acute amoebiasis proportions of diarrhoeal illnesses due to diarrhoea (severe); acute amoebiasis Entamoeba histolytica and the 10 other diarrhoea (moderate); acute amobiasis diarrhoeal pathogens9 in children <5 diarrhoea (mild); and amoebiasis death. years of age using a CHERG and FERG We assumed that 0.5% of amoebiasis systematic review of aetiology studies cases resulted in severe diarrhoea, 8.5% among outpatients and persons in the of amoebiasis cases resulted in moderate community [40], and the aetiological diarrhoea, and 91% of amoebiasis cases proportion of diarrhoeal illnesses due resulted in mild diarrhoea. to Entamoeba histolytica and the 10 other diarrhoeal pathogens in persons Duration >5 years of age using an updated FERG In children <5 years of age, duration of systematic review of aetiology studies severe diarrhoea was 8.4 days, moderate among inpatients, outpatients and diarrhoea was 6.4 days, and mild persons in the community [40, 274]. diarrhoea was 4.3 days [266]. Based The amoebiasis aetiological proportions on the assumed distribution of severe, were extracted from studies, and moderate and mild diarrhoea cases, the regional median amoebiasis aetiological duration of amoebiasis diarrhoea cases proportions calculated. We modified in children <5 years of age was estimated the CHERG approach by dropping to be 4.9 days (min. 4.3 days–max. 8.4 regional median amoebiasis aetiological days). In persons >5 years of age, the proportion outliers that were >5 duration of diarrhoea was 2.8 days [266]. times greater than the global median amoebiasis aetiological proportion, and Disability weight replacing missing regional amoebiasis – Acute amoebiasis diarrhoea (severe): aetiological proportions with the global GBD2010 disability weight of 0.281 median. Furthermore, for children (95% UI 0.184–0.399) for diarrhoea, <5 years of age, we proportionally severe. decreased the aetiological proportions – Acute amoebiasis diarrhoea for all 11 diarrhoeal pathogens in each (moderate): GBD2010 disability region so that the sum of the aetiological weight of 0.202 (95% UI 0.133–0.299) proportions for all diarrhoeal pathogens for diarrhoea, moderate. in a region equalled 1. The resultant – Acute amoebiasis diarrhoea (mild): regional amoebiasis aetiological GBD2010 disability weight of 0.061 proportions were multiplied by the (95% UI 0.036–0.093) for diarrhoea, regional estimates of diarrhoea incidence, mild [82]. and the resultant regional amoebiasis Mortality incidence was applied to all countries in that region. We estimated no amoebiasis The mortality of amoebiasis was cases in the 61 low mortality countries estimated separately for middle-to- (EUR and other subregion “A” countries). high mortality countries, and low mortality countries. For the 133 middle- 9 The 11 diarrhoeal pathogens are: non-typhoidal Salmonella, Campylobacter, Shigella, norovirus, to-high mortality countries, we used a enterotoxigenic E. coli (ETEC), enteropathogenic modification of the CHERG approach E. coli (EPEC), Cryptosporidia, Giardia, Entamoeba [50]. We received envelopes of diarrhoeal histolytica, other diarrhoeal agents not known to be foodborne (rotavirus and astrovirus), and deaths from WHO; because this estimate unspecified agents. was not available with an uncertainity APPENDICES WHO Estimates of the global burden of foodborne diseases 175

interval, we used the uncertainity range No other information on age distribution from the GBD2010 estimate of diarrhoeal for diarrhoea cases. deaths (81.7% to 114.6% around the point estimate) [58]. We then estimated the Sex distribution aetiological proportions of diarrhoeal Amoebiasis sex distibution: 50% male. deaths due to Entamoeba histolytica and the 10 other diarrhoeal pathogens in children <5 years of age using a CHERG A4.23 Congenital Toxoplasmosis and FERG systematic review of aetiology Incidence studies among inpatients [40], and the aetiological proportions of diarrhoeal Full details of how estimates of deaths due to Entamoeba histolytica congenital toxoplasmosis was estimated and the 10 other diarrhoeal pathogens are available in [76] and online in persons >5 years of age using an appendixes (available at: www.vetepi.uzh. updated FERG systematic review of ch/research/Diseaseburden/Burden_CT- aetiology studies among inpatients Appendices.pdf ). [40, 274]. The amoebiasis aetiological proportions were extracted from Clinical Outcomes studies, and regional median amoebiasis Based on data in [296, 297], the following aetiological proportions calculated. probabilities were assigned to clinical We modified the CHERG approach by outcomes: neonatal death probability dropping regional median amoebiasis 0.7% (UI 0.4%–1.2%); chorioretinitis in first aetiological proportion outliers that were year of life probability 13% (UI 12%–15%); >5 times greater than the global median chorioretinitis later in life probability 16% amoebiasis aetiological proportion, and (UI 5%–52%); chorioretinitis in first year replacing missing regional amoebiasis of life (AMR) probability 80% (UI 70%– aetiological proportions with the global 90%); chorioretinitis later in life (AMR) median. Furthermore, for children probability 10% (UI 5%–15%); intracranial <5 years of age, we proportionally calcification probability 11% (UI 7.9%– decreased the aetiological proportions 12%); hydrocephalus probability 2.0% for all 11 diarrhoeal pathogens in each (UI 1.0%–3.0%); CNS abnormalities 2.9% region so that the sum of the aetiological (UI 1.0%–6.0%). proportions for all diarrhoeal pathogens in a region equalled 1. The resultant Duration amoebiasis aetiological proportions were Lifelong (i.e. life expectancy at birth), multiplied by the regional estimates except chorioretinitis later in life, of diarrhoea deaths, and the resultant which has duration the same as the life regional amobiasis mortality was applied expectancy at age 10 years (mean age of to all countries in that region. We onset is 10 years) estimated no amoebiasis deaths in the 61 low mortality countries (EUR and Disability weight other subregion “A” countries). Suitable DWs were selected from GBD2010 [82]. These were Age distribution – chorioretinitis 0.033, The incidence of amoebiasis diarrhoea – intracranial calcification 0.01, was estimated separately for children <5 – hydrocephalus 0.36. and years of age and persons >5 years of age. – other CNS abnormalities 0.36. Appendices 176

Mortality [298], for example). Incidence estimates A value of 0.7% (UI 0.4%–1.2%) was with uncertainty limits were made using used, as these are the proportions of age-stratified seroconversion rates on a cases that die in the neonatal peiod. In country by country basis, and summed addition, there are approximately 2.4% over regions to derive global estimates. (2.3%–6.3%) fetal loss (stillbirths) after 24 weeks, but these were not assigned as Clinical Outcomes fatal cases. Mild chorioretinitis p = 4.5%; moderate chorioretinitis p = 0.25%–0.69%; severe Age distribution chorioretinitis 0.01%. Acute infectious In AMR: 90% onset at birth, 10% at age 10 disease: p = 26%; post-acute syndromes years. Other regions: 86% onset at birth, p = 2.9%. These were estimated from 14% at age 10 years. data in [212, 299–301], which derived from cohort and cross-sectional Sex distribution studies of individuals with confirmed There is no evidence that male and acquired toxoplamsosis. female infants have different risks Duration of having congenital toxoplasmosis. Therefore, the sex ratio at birth was used Eye lesions: lifelong, i.e. life expectancy to determine the sex distribution. at age of incident case. Acute 4 weeks; post-acute syndromes 8 weeks.

A4.24 Acquired toxoplasmosis Disability weight Mild chorioretinitis = 0.004; moderate Incidence chorioretinitis = 0.033; severe Generally there is an increase in sero- chorioretinitis = 0.191. Acute infectious positivity with age, and estimates of disease = 0.053, post-acute incidence were made from age-stratified syndromes = 0.254. sero conversion data as the difference in prevalence between age t and age t+1. Mortality Where there were insufficient data points, None a model was constructed based on the assumptions that individuals that convert Age distribution remain seropositive for life and live under Five different age distributions were a constant infection pressure. In this used which was driven by the country- model, the prevalence p(t) at age t can specific data. be described by: p(t) = 1–exp (-ßt) where ß is the incidence. This model has been Sex distribution widely used for infectious diseases (see Male = 0.5

AGE DISTRIBUTION <5 5-14 15-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1Mean 0.26 0.34 0.19 0.11 0.055 0.027 0.013 0.0057 0.0015 0.00016 G2 Mean 0.21 0.3 0.2 0.13 0.078 0.043 0.023 0.011 0.0029 0.0002 G3 Mean 0.09 0.18 0.19 0.15 0.14 0.12 0.071 0.041 0.02 0.0059 G4Mean 0.075 0.14 0.14 0.14 0.14 0.13 0.11 0.064 0.039 0.016 G5Mean 0.072 0.11 0.17 0.32 0.21 0.086 0.021 0.0049 0.00098 0.00019 APPENDICES WHO Estimates of the global burden of foodborne diseases 177

A4.25 Cystic echinococcosis CNS lesions: moderate motor and/or cognitive impairments. Incidence Non-treatment-seeking: mild abdominal For cystic echinococcosis (CE), due pelvic problems, mild chronic to infection with the larval stage of respiratory disease. Echinococcus granulosus, a systematic review was conducted to collect and CNS: mild motor or synthesize data on both the frequency cognitive impairments. and clinical manifestations of CE Duration globally [75]. In addition to information acquired via the systematic review, World Lifelong for non-treatment-seeking. Organisation for Animal Health (OIE) Median of 2 years for treated cases. and European Food Safety Authority (EFSA) databases were queried to obtain Disability weight officially reported numbers of human – Mild abdominal pelvic problems: 0.012 cases by country. Individual government – Moderate abdominal pelvic websites and reports were also searched problems: 0.123. for relevant CE frequency data. Such data – Mild chronic respiratory disease: 0.015. included official hospital discharge data – Moderate chronic respiratory and notified cases in countries where disease: 0.192. the disease is notifiable. Where no data – Mild motor or cognitive impairment: are available, but the disease is believed 0.054. to be endemic then the incidence – Moderate motor or cognitive was imputed. impairment: 0.221. Mortality Clinical Outcomes For treatment-seeking: 2%; 1% for non- Treatment-seeking: moderate abdominal treatment-seeking. pelvic problems, chronic respiratory disease moderate (chronic obstructive Age distribution pulmonary disease – COPD).

AGE DISTRIBUTION <5 5-14 15-24 25-34 35-44 45-54 55-64 65+ (YEARS) G1Mean 4.6 10.4 17.1 21.2 18.1 12.8 8.7 7.1

Sex distribution from countries such as Kyrgyzstan [302] Male = 42.8% and Poland [303].

Clinical Outcomes A4.26 Alveolar echinococcosis Abdominopelvic problems followed by recovery after treatment, or death. Incidence Full details of the methodology of Duration estimating the incidence of alveolar Europe: 10 years. Other: 8 years. echinococcosis (AE) can be found in [72]. In addition, this data has been Disability weight updated because of subsequent reports – 0.123 Appendices 178

Mortality to the total numbers of epilepsy by Following abdominopelvic problems: dividing by 0.58 (58% of epilepsy cases western and central Europe and north being idiopathic – see appendix of [83]. America 2–5%; eastern Europe: 10–30%; Population at risk was estimated by using elsewhere: 100% seven assumptions: – (1) Countries with negligible pig Age distribution populations (less than 30 000 pigs f Europe: 0–9 years, 0%; 10–19 years, (FAO data) were assumed to have 2.7%; 20–29 years, 8.8%; 30–39 years, zero risk due to there being no 13.6%; 40–49 years, 18.7%; 50–59 years , opportunity to transmit T. solium. This 18.4%; 60–69 years, 20.6%; 70–79 years, excluded countries where the Muslim 12.1%; 80 years and over, 5.1%. population was over 90% and a few f Eastern Europe: 0–9 years, 1.7%; 10–19 non-Muslim countries (for example years , 5.1%; 20–29 years, 12.8%; 30–39 Ethiopia) where the pig population years, 14.5%; 40–49 years, 20.5%; 50–59 was very low. years , 17.9%; 60–69 years, 14.5%; 70–79 – (2) For countries that raise pigs and years, 12.0%; 80 years and over, 1.0%. have more than 80% of the population f Central Asia: 0–9 years, 2.7%; 10–19 living with unimproved sanitation, years, 10.3%; 20–29 years, 33.3%; 30–39 population at risk was estimated as years, 25.1%; 40–49 years, 14.1%; 50–59 the proportion of the population that years , 10%; 60years and over, 4.5% . was not Muslim. f China: 0–9 years, 1.4%; 10–19 years , 7%; – (3) For countries that raise pigs and 20–29 years, 10.2%; 30–39 years, 23%; have less than 80% of the population 40–49 years, 24.5%; 50–59 years , 16%; living with unimproved sanitation, 60 years and over, 17.9%. population at risk was estimated as Sex distribution the proportion of the population that was not Muslim, multiplied by the Europe: male 44%, central Asia, proportion of the population that male 36%, China male 47%. lived with unimproved sanitation. – (4) For the United States of America, A4.27 Taenia solium neurocysticercosis it was assumed that transmission does not occur, and hence nearly all Incidence cases are in immigrants, mainly from Taenia solium neurocysticercosis (NCC) Latin America. Thus a weighted mean is known to cause epilepsy and other of the population at risk from the neurological sequelae [73]. A systematic entire Latin America was applied to review revealed that NCC may be the population of hispanic immigrants, responsible for approximately 29.0% born outside of the United States (95% UI 22.9%–35.5% of the burden of of America but now resident in the epilepsy in at-risk populations in low United States of America. and middle income, pork consuming – (5) For the ex-Soviet states, the risk societies [74]. Consequently, the number of cysticercosis was assumed to be of prevalent cases of epilepsy used in close to zero due to lack of evidence the GBD2010 [58, 81–83] were utilized to for cysticercosis or taeniasis in public estimate the prevalent cases of epilepsy- health surveillance data. associated NCC. The total numbers – (6) Indonesia is predominantly Muslim of cases of idiopathic epilepsy were (87%). However, the predominantly available by country and were corrected non-Muslim provinces of Papua and APPENDICES WHO Estimates of the global burden of foodborne diseases 179

West Papua are known to be highly A4.28 Chlonorchiosis endemic regions for T. solium and hence the combined population of Incidence these provinces was used as the Incidence estimates and clinical sequelae population at risk, with the remainder for foodborne trematodiasis were mainly of Indonesia having zero risk. based on the results of two systematic – (7) There is no FAO data on pig review articles [77, 78]. The reviews populations in Sudan or South Sudan, identified available qualitative and the latter being predominantly non- quantitative information on prevalence, Muslim and with little improved incidence, mortality and remission rates, sanitation. However, extensive sex- and age-distributions and the searches for information about pigs progression of foodborne trematodiasis in Sudan revealed that the domestic into different sequelae. From these pig population in both countries is data, simplified disease models were negligible and hence there is virtually developed and quantitative data zero risk if cysticercosis. summarized by meta-analyses. As Due to the absence of available data information on incidence, remission, and on all cysticercosis sequelae, only the duration of foodborne trematodiasis frequency of NCC-associated epilepsy was particularly scant, zero remission was estimated in this study. was assumed and entered into the DisMod 3 software [304], together with Clinical Outcomes the available prevalence and mortality Epilepsy-associated NCC [74] estimates. DisMod 3 computed internally consistent and complete sets of sex-, Duration age- and country-specific prevalence, incidence, remission, duration and No data. mortality for foodborne trematodiasis Disability weight and associated sequelae. However, unlike the original study, which computed – GBD2010 for epilepsy [58, 81–83] incidence rates only for countries Mortality reporting national prevalence rates, and GBD2010 for epilepsy [58, 81–83] otherwise considered the incidence rate to be zero [77], the present study also Age distribution imputed incidence rates for countries GBD2010 for epilepsy [58, 81–83] where no records of national prevalence or incidence rates were available, but at Sex distribution least one autochthonous human infection could be identified in the systematic GBD2010 for epilepsy [58, 81–83] review. Hierarchical random-effects models with incidence information from other countries as input data were applied in this additional imputation process [79].

Clinical Outcomes Abdominal pelvic discomfort, carcinoma. Appendices 180

Duration Disability weight Lifelong due to low treatment coverage – Only for severe infections: 0.123. in affected populations, longevity of Mortality parasites in humans, high re-infection rates, supposedly high susceptibility 1% case fatality. of clinical cases, and irreversibility of Age distribution pathology after several years of infection.

AGE DISTRIBUTION 0-1 1-4 5-9 10-14 15-19 20-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1Mean 5.0 8.1 11.5 13.4 22.7 12.4 13.9 8.1 2.9 1.1 0.6 0.2 0.0 G2 Mean 4.3 7.5 10.4 12.0 18.8 13.5 16.7 8.0 3.8 2.4 1.4 0.8 0.3 G3 Mean 4.1 7.2 9.1 13.0 32.9 18.6 12.9 1.6 0.3 0.1 0.1 0.1 0.1 G4Mean 6.8 8.8 11.2 12.4 23.5 12.2 13.4 7.6 2.6 0.9 0.5 0.2 0.0 G5Mean 5.1 8.2 11.4 13.2 22.6 12.5 14.0 8.0 2.9 1.1 0.6 0.3 0.1

Sex distribution the original study, which computed Male 65–68%. incidence rates only for countries reporting national prevalence rates and otherwise considered the incidence rate A4.29 Fasciolosis to be zero [77], the present study also imputed incidence rates for countries, Incidence where no records of national prevalence Incidence estimates and clinical sequelae or incidence rates were available, but at for foodborne trematodiasis were mainly least one autochthonous human infection based on the results of two systematic could be identified in the systematic review articles [77, 78]. The reviews review. Hierarchical random-effects identified available qualitative and models with incidence information from quantitative information on prevalence, other countries as input data were applied incidence, mortality and remission rates, in this additional imputation process [79]. sex- and age-distributions and the progression of foodborne trematodiasis Clinical Outcomes into different sequelae. From these data, Abdominal pelvic discomfort. simplified disease models were developed and quantitative data summarized Duration by meta-analyses. As information on Lifelong due to low treatment coverage incidence, remission and duration of in affected populations, longevity of foodborne trematodiasis was particularly parasites in humans, high re-infection scant, zero remission was assumed, and rates, supposedly high susceptibility entered into the DisMod 3 software [304], of clinical cases, and irreversibility of together with the available prevalence and pathology after several years of infection mortality estimates. DisMod 3 computed internally consistent and complete Disability weight sets of sex-, age- and country-specific – 0.123 prevalence, incidence, remission, duration and mortality for foodborne trematodiasis Mortality and associated sequelae. However, unlike Zero. APPENDICES WHO Estimates of the global burden of foodborne diseases 181

Age distribution

AGE DISTRIBUTION 0-1 1-4 5-9 10-14 15-19 20-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1 Mean 49.1 26.9 11.6 4.6 2.3 1.4 1.8 1.0 0.6 0.3 0.2 0.1 0.0 G2 Mean 19.5 33.1 29.5 10.4 4.3 1.4 1.0 0.4 0.2 0.1 0.0 0.0 0.0 G3 Mean 75.3 14.8 5.1 1.8 0.9 0.6 0.7 0.4 0.2 0.1 0.1 0.0 0.0 G4 Mean 21.9 52.6 16.1 1.3 0.5 0.7 1.5 1.3 1.1 0.8 0.7 0.7 0.7 G5 Mean 59.1 22.0 9.4 3.6 1.8 1.1 1.3 0.7 0.4 0.2 0.2 0.1 0.0

Sex distribution incidence rates only for countries Male = 49.5% reporting national prevalence rates and otherwise considered the incidence rate to be zero [77], the present study also A4.30 Opisthorchosis imputed incidence rates for countries, where no records of national prevalence Incidence or incidence rates, but at least one Incidence estimates and clinical sequelae autochthonous human infection, could for foodborne trematodiasis were mainly be identified in the systematic review. based on the results of two systematic Hierarchical random-effects models review articles [77, 78]. The reviews with incidence information from other identified available qualitative and countries as input data were applied in quantitative information on prevalence, this additional imputation process [79]. incidence, mortality and remission rates, sex- and age-distributions and the Clinical Outcomes progression of foodborne trematodiasis Abdominal pelvic discomfort, carcinoma. into different sequelae. From these data, simplified disease models were Duration developed and quantitative data Lifelong due to low treatment coverage summarized by meta-analyses. As in affected populations, longevity of information on incidence, remission and parasites in humans, high re-infection duration of foodborne trematodiasis rates, supposedly high susceptibility was particularly scant, zero remission of clinical cases, and irreversibility of was assumed and entered into the pathology after several years of infection DisMod 3 software [304], together with the available prevalence and mortality Disability weight estimates. DisMod 3 computed internally – 0.123 consistent and complete sets of sex-, age- and country-specific prevalence, Mortality incidence, remission, duration and Overall case fatality rate: 9.2%. mortality for foodborne trematodiasis and associated sequelae. However, unlike the original study, which computed Appendices 182

Age distribution

AGE DISTRIBUTION 0-1 1-4 5-9 10-14 15-19 20-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1 Mean 4.2 11.8 13.9 12.4 10.7 8.7 13.4 10.3 7.1 3.9 2.4 0.9 0.2 G2 Mean 4.3 13.4 15.8 15.4 12.7 8.5 10.9 8.1 5.5 2.3 2.1 0.9 0.2 G3 Mean 2.3 6.8 8.4 9.6 11.6 9.6 13.7 11.6 11.1 6.2 5.6 2.9 0.5 G4 Mean 4.2 11.7 13.8 12.4 10.7 8.7 13.4 10.3 7.2 4.0 2.4 1.0 0.2

Sex distribution countries reporting national prevalence Male=55% rates and otherwise considered the incidence rate to be zero [77], the present study also imputed incidence rates for A4.31 Paragonimosis countries, where no records of national prevalence or incidence rates, but at least Incidence one autochthonous human infection, could Incidence estimates and clinical sequelae be identified in the systematic review. for foodborne trematodiasis were mainly Hierarchical random-effects models with based on the results of two systematic incidence information from other countries review articles [77, 78]. The reviews as input data were applied in this additional identified available qualitative and imputation process [79]. quantitative information on prevalence, incidence, mortality and remission rates, Clinical Outcomes sex- and age-distributions and the Lifelong due to low treatment coverage progression of foodborne trematodiasis in affected populations, longevity of into different sequelae. From these data, parasites in humans, high re-infection simplified disease models were developed rates, supposedly high susceptibility and quantitative data summarized of clinical cases, and irreversibility of by meta-analyses. As information on pathology after several years of infection. incidence, remission and duration of foodborne trematodiasis was particularly Duration scant, zero remission was assumed and Lifelong. entered into the DisMod 3 software [304], together with the available prevalence and Disability weights mortality estimates. DisMod 3 computed – Pulmonary: 0.132. internally consistent and complete sets of – Cerebral paragonimosis: 0.42. sex-, age- and country-specific prevalence, incidence, remission, duration and mortality Mortality for foodborne trematodiasis and associated 10% case fatality for cerebral cases only sequelae. However, unlike the original study, which computed incidence rates only for Age distribution

AGE DISTRIBUTION 0-1 1-4 5-9 10-14 15-19 20-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1 Mean 8.7 11.5 12.3 10.1 9.2 6.7 12.6 12.2 7.9 4.6 2.8 1.2 0.2 G2 Mean 8.9 16.1 16.2 12.5 9.5 7.3 11.0 7.7 5.1 3.0 1.8 0.8 0.2 G3 Mean 2.6 9.0 11.4 11.5 10.9 9.8 16.5 12.5 8.1 4.2 2.4 0.9 0.2 G4 Mean 2.0 8.1 11.5 11.4 10.7 10.6 20.1 13.4 7.2 3.3 1.2 0.5 0.2 G5 Mean 8.6 11.7 12.4 10.2 9.2 6.7 12.6 12.0 7.8 4.5 2.8 1.2 0.2 Sex distribution Male = 55.9% APPENDICES WHO Estimates of the global burden of foodborne diseases 183

A4.32 Intestinal flukes reporting national prevalence rates and otherwise considered the incidence rate Incidence to be zero [77], the present study also Incidence estimates and clinical sequelae imputed incidence rates for countries, for foodborne trematodiasis were mainly where no records of national prevalence based on the results of two systematic or incidence rates, but at least one review articles [77, 78]. The reviews autochthonous human infection, could identified available qualitative and be identified in the systematic review. quantitative information on prevalence, Hierarchical random-effects models incidence, mortality and remission rates, with incidence information from other sex- and age-distributions, and the countries as input data were applied in progression of foodborne trematodiasis this additional imputation process [79]. into different sequelae. From these data, simplified disease models were Clinical Outcomes developed and quantitative data Abdominal pelvic discomfort. summarized by meta-analyses. As information on incidence, remission and Duration duration of foodborne trematodiasis Lifelong due to low treatment coverage was particularly scant, zero remission in affected populations, longevity of was assumed and entered into the parasites in humans, high re-infection DisMod 3 software [304], together with rates, supposedly high susceptibility the available prevalence and mortality of clinical cases, and irreversibility of estimates. DisMod 3 computed internally pathology after several years of infection. consistent and complete sets of sex-, age- and country-specific prevalence, Disability weight incidence, remission, duration and – Heavy infections only: 0.123. mortality for foodborne trematodiasis and associated sequelae. However, unlike Mortality the original study, which computed Zero. incidence rates only for countries Age distribution

AGE DISTRIBUTION 0-1 1-4 5-9 10-14 15-19 20-24 25-34 35-44 45-54 55-64 65-74 75-84 85+ (YEARS) G1 Mean 24.4 32.7 22.1 8.3 4.3 1.9 2.5 1.8 1.0 0.5 0.3 0.2 0.0 G2 Mean 18.3 38.4 24.2 9.8 4.7 1.9 1.3 0.5 0.3 0.2 0.1 0.1 0.0 G3 Mean 70.1 10.4 6.0 2.8 1.8 1.3 2.1 1.5 1.2 1.2 0.9 0.6 0.2 G4 Mean 56.4 26.4 9.1 3.1 1.6 0.9 1.1 0.6 0.4 0.2 0.1 0.1 0.0 G5 Mean 24.8 40.8 18.6 7.7 4.5 1.7 1.0 0.4 0.2 0.1 0.1 0.1 0.0 G6 Mean 46.4 27.7 13.3 4.8 2.5 1.3 1.6 1.0 0.6 0.4 0.2 0.1 0.0

Sex distribution Male = 55%. Appendices 184

A4.33 Ascaris spp. Clinical Outcomes The clinical outcomes were death Incidence in severe cases, severe wasting, Age-stratified prevalence was used pelvic abdominal disconfort and to estimate the burden of disease in clinical ascariasis as described in GBD2010 [81] and these data supplied GBD2010 [82]## by IHME were used to estimate incidence of ascariasis. The estimated numbers of Duration prevalent cases were available for every Duration of each incidence case was set country and each age group. at a mean of 1 year Prevalence at age t = P(t) Prevalence Disability weight at age t+1 = P(t+1) Incidence = b = proportion of population infected – For severe wasting = 0.127 between t and t= t+1 P(t+1) = b*(1– – For mild abdominal pelvic discomfort P(t))–m * P(t) As proportion acquiring = 0.012 new infections are at a rate of b*(1–P(t)) – For clinical ascariosis = 0.296 and proportion losing infections at Mortality rate –m * P(t) Therefore incidence Incidence of mortaility due to ascariasis (proportion) of new infections is used the GBD2010 mortality figures. In given by b=(P(t+1)+m*p(t))/1–P(t) b = ascaris-endemic countries this ranged proportion that are infected in time t from a low of 0.000095 per 100 000 = 1 year m = proportion that lose their per annum in Dominica to a high of 0.159 infection = death rate = 1/life expectancy per 100 000 per annum in Equitorial Approximate life expectancy of Ascaris Guinea. Upper-income countries had = 1 year Therefore b=(P(t+1)+P(t))/ zero mortality. Globally there were 0.031 (1–P(t)) Incidence per 100 000 per deaths per 100 000 year = b *100 000 It is well known [88] that the infection pressure or incidence Age distribution varies with age with ascaris. In particular children have a higher incidence. But Fatalities, Ascariasis and mild abdomino by using this step equation all that is pelvic problems <1 year, 9%; 1–4 years, required is the different 56.7%; 5–14 years, 16%; 15–24 years, 3%; (proportion infected) at age t and 25–34 years 2.1%; >35 years, 13.2%. t+1 which can be calculated from the Severe wasting: <1 year, 19%; 1–4 data provided by GBD2010. Assuming years, 81%. the duration for ascariasis and other manifestations (mild abdominopelvic Sex distribution discomfort and severe wasting) are of Male =55.1% (fatalities), Male = 50.1% the same duration, then this can be used (ascaris), Male=50.2% (mild abdominal to estimate the incidence of all sequelae pelvic discomfort), from the stratified prevalence data. There Male= 51% (severe wasting). is some evidence that ascaris induces some degree of protective . But this acts to decrease the abundance of infection rather than the prevalence and so can be discounted in this exercise. APPENDICES WHO Estimates of the global burden of foodborne diseases 185

A4.34 Trichinella Age distribution According to [71], the majority of cases Incidence were between 20 and 50 years of age, Incidence was estimated from the results with a median of 33.1. A generalized Beta of a FERG-commisioned systematic distribution was fitted to the estimates to review. Full details can be found in [71] define the full distribution of cases from age 0 to 90 years [84]. Clinical Outcomes In the absence of data on the probability Sex distribution of occurrence of the major clinical According to [71], 51% of cases symptoms of acute trichinellosis, it was were male. assumed, as a worst case scenario, that all patients would develop diarrhoea, A4.35 Aflatoxin facial oedema, myalgia and fever/ headache [84]. Incidence Duration A population-attributable fraction (PAF) approach was used to estimate Based on the systematic review by [71], the incidence of aflatoxin-related disease duration ranged from 21.5 to 70 hepatocellular carcinoma (HCC). We days. These values were divided by 365 assumed a multiplicative model for to express the duration in years. the effects of aflatoxin exposure and Disability weight (HBV) infection. The excess risk due to aflatoxin exposure Because no specific DW for acute is estimated as HCCa- = b * a for HBV- trichinellosis is available, DWs were negative individuals and HCCa+ = b * derived for each of the outcomes h * a for HBV-positive individuals, where separately. The four clinical symptoms a = exposure to aflatoxin (ng/(kg bw were, respectively, matched to the * day)), b = aflatoxin cancer potency GBD2010 health states: factor in HBV- individuals ((ng/(kg bw * – Diarrhoea: moderate – DW = 0.202. day)-1) and h = relative risk for aflatoxin – Disfigurement: level 2, with itch or exposure in HBV+ individuals compared pain – DW = 0.187. with HBV- individuals. We used potency – Musculoskeletal problems: factors as derived by JECFA [111]: b = 0.01 generalized, moderate – DW = 0.292. [0.002–0.03](ng/(kg bw * day)-1 and b * – Infectious disease: acute episode, h = 0.30 [0.005–0.50](ng/(kg bw * day)- severe (DW = 0.210) [82]. 1. Uncertainty in the potency factors was modelled as a Gamma distribution with These four DWs were then aggregated the most likely value as the mean and the using the multiplicative method, which range representing an approximate 95% defines the aggregated DW as 1–∏i(1– confidence interval. DWi)=0.637 [84]. To account for differences in background Mortality rates between different populations, we Mortality was estimated from the results estimated PAFs by country, and applied of a FERG-commisioned systematic them to HCC incidence, based on [305]. review. Full details can be found in [71] We assumed PAFs for incidence and deaths were equal and calculated PAFs based on published studies on HCC Appendices 186

mortality. For the study population that Disability Weight was the basis of the JECFA potency Not applicable; population-attributable estimates in Guangxi, China [306], the fractions as described above were PAF was estimated as PAFa = ((1–p) * directly applied to WHO YLD HCCa- + p * HCCa+)/HCC, where p estimates [308]. = prevalence of HBV infection and HCC = total incidence of HCC by all causes. Mortality We used the HCC death rate for the Population-attributable fractions as Guangxi cohort, standardized to the described above were directly applied to global population (121.5 per 100 000) and WHO mortality estimates [308]. calculated average exposure (607 ng/(kg bw * day) based on [[307], Table I]. HBV Age distribution prevalence was 23% based on [306], We compared age distributions of resulting in PAFa = 0.383. Background HCC in the populations of Beijing and death rate of HCC by all causes in the Qidong from [114], and hypothesized Guangxi population was calculated as that the main difference in HCC risk HCC0,s = (HCCs–HCCa,s) / (1–ps + h * factors between these two cities is ps), with the subscript s referring to the aflatoxin exposure, since every other risk study population; resulting in HCC0,s factor is the same, and they are both = 9.77 per 100 000. predominantly Han (i.e. same ethnicity). To calculate attributable incidence in Hence, the difference in age distributions all countries, we estimated relative was presumed to be the contribution of risks due to aflatoxin exposure as aflatoxin. This resulted in an average age RRa,c = 1 + b * ac / HCC0,s, and PAFs at onset of 49. per country as PAFa,c = (RRa,c–1) / RRa,c, with the subscript c indicating Sex distribution country. Attributable incidence was then In absence of information on the calculated as HCCa,c = HCCc * PAFa,c. sex distribution of aflatoxin-induced Aflatoxin exposure by country was based hepatocellular carcinoma, a 50:50 age on [110], with uncertainty represented by distribution was assumed. a uniform distribution over the reported range. A Bayesian log-normal random effects model [79, 151] was used to A4.36 Cyanide in cassava extrapolate available PAFs to countries Incidence without data. A total of 2376 konzo cases have been Clinical outcomes reported in 5 countries (Cameroon, Hepatocellular carcinoma. Central African Republic, Democratic Republic of Congo, Mozambique Duration and United Republic of Tanzania), corresponding to 149 cases per year for Not applicable; population-attributable 122 million people [86]. Based on these fractions as described above were cases and dividing the average annual directly applied to WHO YLD number of case for each country by the estimates [308]. corresponding country population gives an observed incidence of 0.043 to 0.179 per 100 000. APPENDICES WHO Estimates of the global burden of foodborne diseases 187

The degree of underestimation is difficult Disability Weight to estimate, as konzo occurs in remote No specific DW exists for konzo rural areas, often under conditions of war, paraparesis. WHO [89] defined three and the disease is not notifiable. The only severity levels for konzo: previous calculation of underestimation was that of Tylleskar [90] in the DRC in 1. Mild = Able to walk without support 1994, when he estimated that there may 2. Moderate = uses one or two sticks or have been at least twice as many cases crutches to walk as those reported. The underestimation in 3. Severe = not being able to walk the DRC is now likely to be much greater, due to war and displacement. It was These three severity levels can be decided to account for the uncertainty matched with the GBD2010 health states: in the under reporting by applying an – Motor impairment, mild: DW = 0.012. expansion factor ranging from 1 to 10 to – Motor impairment, moderate: the observed cases. Therefore, the annual DW = 0.076. total of new cases would range from 149 – Motor impairment, severe: to 1490 in the 5 countries and the mean DW = 0.377 [82]. annual incidence rate would be 0.9 per 100 000 (0.04 to 1.8 per 100 000). Information on the distribution of konzo severity levels is available from 9 studies We restricted our estimates of konzo [86][1]. Out of a total of 753 cases, disease to the 5 African countries in 476 (63%) were mild, 203 (27%) were which the disease has been reported, moderate and 74 (10%) were severe. together with Angola, based on a report to the World Congress on Neurology The resulting weighted DW suggesting that cases have occurred in equalled 0.065. that country [92]. The incidence of konzo disease was assumed to be null in other Mortality countries around the world. Information on case fatality was provided in 4 studies [94–96, 309]. Out of a total Clinical outcomes of 340 cases, 73 deaths were observed, Konzo disease is a paraparesis occurring yielding an average case fatality ratio in populations exposed to cyanogenic of 21%. glycoside in a context of bitter cassava consumption associated with a low intake Age distribution of protein-rich food. The age and sex distribution observed by [90] was generalized to the whole konzo Duration affected population. The age distribution The onset of paraparesis is abrupt, for fatal cases was adapted from [309]. usually within minutes or hours, with occasional progression during the first Sex distribution days of the illness. After that time, the The age and sex distribution observed by paraparesis is non-progressive and [90] was generalized to the whole konzo permanent. As a result, duration was affected population. The sex distribution defined as lifelong for non-fatal cases. for fatal cases was adapted from [309]. For fatal cases, it was assumed that death occurred one to seven years after onset, with an average of three years after onset, following [93]. Appendices 188

A4.37 Dioxin Age distribution Hypothyroidy due to prenatal exposure: Incidence Onset = birth. Incidence rates were generated for Hypothyroidy due to postnatal exposure: 50 countries and specified as lower and Onset = 20 years. upper bounds (hypothyroidy-postnatal & male infertility) or point estimates Male infertility: Onset = 20 years. (hypothyroidy-prenatal). Incidence rates for the remaining 144 countries were Sex distribution imputed using a Bayesian log-normal In absence of information on the random effects model [151]. sex distribution of dioxin-induced hypothyroidy, a 50:50 age distribution Clinical outcomes was assumed. Hypothyroidy due to prenatal exposure; For male infertility, the entire burden was hypothyroidy due to postnatal assigned to males. exposure; or male infertility due to prenatal exposure. A4.38 Peanut allergens Duration Hypothyroidy was assumed to be Incidence lifelong; the male infertility impact was Data on clinically confirmed peanut assumed to be present in the 20–44 age [Arachis hypogaea] allergy in children group, in accordance with [83]. were available from six countries (Canada, Denmark, Iceland, Sweden, Disability Weights Turkey and UK). Average incidences – Hypothyroidy due to prenatal ranged from 0 to 22.6 per 100 000 [102]. exposure: 0.019; corresponding to To reflect this uncertainty, the incidence GBD 2013 health state Hypothyroidy rate of clinical peanut allergy in subregion [142]. Note that no corresponding DW “A” countries was modelled as a Uniform was available in GBD2010 or WHO distribution ranging from 0/100 000 Global Health Estimates (GHE). to 22.6/100 000. Given the lack of – Hypothyroidy due to postnatal data, no estimates were generated for exposure: 0.019; corresponding to other countries. GBD 2013 health state Hypothyroidy [142]. Note that no corresponding Clinical outcomes DW was available in GBD2010 or The symptoms of peanut allergy vary WHO GHE. from mild to severe, from swollen lips, – Male infertility: 0.056; corresponding shortness of breath, to an anaphylactic to WHO GHE health state Infertility: shock, which is potentially fatal. However, primary [310]. Note that this is higher because of the very short duration of than the corresponding GBD2010 acute peanut allergy, we decided not health state. to include acute peanut allergy in the burden assessment. The considered Mortality clinical outcome was therefore living No mortality was assumed. with peanut allergy and the anxiety of a possible allergic reaction. APPENDICES WHO Estimates of the global burden of foodborne diseases 189

Duration Mortality It is assumed that peanut allergy is a The limited data on the mortality rate of lifelong disease. It is important to note peanut-induced anaphylaxis show values that the duration of allergic symptoms is ranging from 0 to 0.006 deaths per very short. 100 000 person-years [102]. To reflect this uncertainty, the mortality rate of Disability Weight peanut-induced anaphylaxis in subregion Mullins et al. [103] reported that “A” countries was modelled as a Uniform 52% of cases referred to a specialist distribution ranging from 0/100 000 allergy medical practice in Australia to 0.006/100 000. Given the lack of suffered from mild symptoms (skin and data, no estimates were generated for subcutaneous tissue involvement only), other countries. 42% from moderate symptoms (features suggestive of respiratory, cardiovascular Age distribution or gastrointestinal involvement), and The onset of peanut allergy is early in 6% from severe symptoms (cyanosis, life (median age 18–24 months, [107, hypotension, confusion, collapse, loss 108], with continued prevalence in older of consciousness, incontinence). We age groups. All incident cases of peanut propose the DW for clinically relevant allergy were therefore assumed to peanut allergy be a weighted average develop early in life, i.e. before the age accounting for this severity distribution. of five. GBD2010 DWs [82] for the health states Deaths due to peanut allergen were “Asthma: controlled” (DW = 0.009) assumed to occur at all ages, with an are considered applicable for mild and average age of 37 years [102][1]. moderate cases (94%), and “Generic uncomplicated disease: anxiety about the Sex distribution diagnosis” (DW = 0.054) for severe cases In the absence of information on the sex (6%), leading to a severity-weighted distribution of peanut allergy, a 50:50 DW of 0.012 for clinically relevant age distribution was assumed. peanut allergy. Appendices 190

APPENDIX 5. Disease models

Aflatoxin Dioxin Disease Model Disease Model

INC MRT YLD YLL INC INC INC Dioxin-induced Thyroid impairment Thyroid impairment HCC HCC HCC HCC impaired male due to prenatal dioxin due to postnatal fertility exposure dioxin exposure

RATIO—local RATIO—local cases/100k births to cases/100k births to PROB—local PROB—local PROB—local PROB—local cases/100k population cases/100k population PAF aflatoxin PAF aflatoxin PAF aflatoxin PAF aflatoxin correction factor correction factor

Paragonimus spp. Opisthorchis spp. Disease Model Disease Model

INC INC INC INC INC symptomatic cerebral paragonimosis- symptomatic opisthorchiosis- paragonimosis paragonimosis related mortality opisthorchiosis related mortality

Intestinal flukes Fasciola spp. Disease Model Disease Model

INC INC symptomatic intestinal symptomatic trematodosis fasciolosis

Clonorchis sinensis Ascaris spp. Disease Model Disease Model

INC INC INC symptomatic clonorchiosis-related INC ascariosis-related clonorchiosis mortality ascariosis mild abdominopelvic problems

INC INC ascariosis-related ascariosis-related severe wasting mortality APPENDICES WHO Estimates of the global burden of foodborne diseases 191

Trichinella spp. Taenia solium Disease Model Disease Model

INC INC INCIDENCE MORTALITY YLD YLL Acute clinical Trichinellosis death Epilepsy Epilepsy Epilepsy Epilepsy trichinellosis

RATIO—global RATIO—global RATIO—global total:idiopathic total:idiopathic total:idiopathic PROB—local PROB—local epilepsy epilepsy epilepsy Underreporting Underreporting correction factor correction factor PROB—global PROB—global PROB—global PROB—global NCC-associated NCC-associated NCC-associated NCC-associated epilepsy epilepsy epilepsy epilepsy

PROB—local PROB—local PROB—local PROB—local Taenia solium Taenia solium Taenia solium Taenia solium population at risk population at risk population at risk population at risk

Echinococcus multilocularis Echinococcus granulosus Disease Model Disease Model

INC PROB—global PROB—global pulmonary CE alveolar pulmonary CE echinococcosis PROB—global PROB—global INC hepatic CE INC hepatic CE CE cases seeking CE cases not seeking treatment treatment PROB—global PROB—global PROB—global CNS CE CNS CE case-fatality ratio PROB—global PROB—global death death

Staphylococcus aureus Clostridium perfringens Disease Model Disease Model

INC PROB—global INC PROB—global S. aureus case fatality ratio C. perfringens case fatality ratio

Clostridium botulinum Bacillus cereus Disease Model Disease Model

PROB—global PROB—global INC severe disease case fatality ratio B. cereus

INC C. botulinum PROB—global moderate/mild disease Appendices 192

Toxoplasma gondii (acquired) Toxoplasma gondii (congenital) Disease Model Disease Model

PROB—global PROB—global INC Chorioretinitis - Congenital toxoplasmosis mild Acute illness INC Acquired PROB—global Intracranial toxoplasmosis calcifications PROB—global PROB—global PROB—local PROB—local Chorioretinitis - Post-acute Non-genotype 2 Genotype 2 moderate illness PROB—global Hydrocephalus

PROB—global PROB—global PROB—global Chorioretinitis PROB—global Chorioretinitis - Chorioretinitis early in life early in life CNS severe abnormalities

PROB—global PROB—global PROB—global Chorioretinitis Chorioretinitis later in life later in life Neonatal death

Salmonella Typhi Salmonella Paratyphi Disease Model Disease Model

INC INC INC INC Typhoid fever Typhoid cysts Paratyphoid fever Paratyphoid cysts

INC INC Typhoid deaths Paratyphoid deaths

Mycobacterium bovis Listeria monocytogenes Disease Model Disease Model

INC INC INC PROB—global tuberculosis tuberculosis deaths listeriosis non-perinatal cases

PROB—global PROB—global PROB—global PROB—global death septicemia CNS infection perinatal cases PROB—local PROB—local attributable proportion attributable proportion M. bovis M. bovis PROB—global PROB—globalPROB—global PROB—global neurological death septicemia CNS infection sequela

PROB—global neurological sequela

Brucella spp. Hepatitis A virus Disease Model Disease Model

PROB—global INC INC chronic infection Hepatitis A Hepatitis A deaths

INC PROB—global Brucella orchitis

PROB—global case fatality ratio APPENDICES WHO Estimates of the global burden of foodborne diseases 193

Giardia spp. Entamoeba histolytica Disease Model Disease Model

INC PROB—local INC PROB—local <5 AF INC Giardia <5 diarrhea <5 AF INC Entam <5

INC INC INC diarrhea 5-14 Giardia-associated diarrhea 5-14 diarrhea INC PROB—local INC PROB—local diarrhea 15-54 AF INC Entam 5+ diarrhea 15-54 AF INC Giardia 5+ National studies approach INC INC diarrhea 55+ diarrhea 55+ CHERG approach

INC PROB—local diarrhea deaths <5 AF MRT Entam <5 INC diarrhea deaths 5+M PROB—local INC AF MRT Entam 5+ diarrhea deaths 5+F CHERG approach

Cryptosporidium spp. Vibrio cholerae Disease Model Disease Model

INC PROB—local INC diarrhea <5 AF INC Crypto <5 cholera INC diarrhea 5-14 INC INC PROB—local Cryptosporidium- diarrhea 15-54 AF INC Crypto 5+ associated diarrhea PROB—local INC death diarrhea 55+ INC Death due to INC PROB—local Cryptosporidium- diarrhea deaths <5 AF MRT Crypto <5 associated diarrhea INC diarrhea deaths 5+M PROB—local National studies approach INC AF MRT Crypto 5+ diarrhea deaths 5+F CHERG approach

Shigella spp. Salmonella enterica Disease Model Disease Model

INC INC INC PROB—local INC PROB—local Salmonella- Death due to diarrhea <5 AF INC Shigella <5 diarrhea <5 AF INC Salmo <5 associated Salmonella- diarrhea associated INC INC diarrhea diarrhea 5-14 diarrhea 5-14 National studies INC approach INC PROB—local INC PROB—local Shigella-associated diarrhea 15-54 AF INC Shigella 5+ diarrhea 15-54 AF INC Salmo 5+ diarrhea INC INC INC INC diarrhea 55+ diarrhea 55+ Invasive non- Invasive non- INC typhoidal typhoidal Death due to PROB—local INC PROB—local INC salmonellosis <5 salmonellosis 5+ Shigella-associated diarrhea deaths <5 AF MRT Shigella <5 diarrhea deaths <5 AF MRT Salmo <5 diarrhea INC INC diarrhea deaths 5+M diarrhea deaths 5+M PROB—local National studies approach PROB—local PROB—global PROB—global AF MRT Salmo 5+ INC AF MRT Shigella 5+ INC case fatality case fatality diarrhea deaths 5+F diarrhea deaths 5+F ratio <5 ratio 5+ CHERG approach CHERG approach Appendices 194

STEC ETEC Disease Model Disease Model

PROB—global PROB—global PROB—global HUS Death ESRD Death ESRD INC PROB—local Survivor diarrhea <5 AF INC ETEC <5 INC diarrhea 5-14 PROB—global PROB—global PROB—global INC O157 HUS ESRD INC PROB—local ETEC -associated diarrhea 15-54 AF INC ETEC 5+ INC diarrhea STEC INC diarrhea 55+ INC PROB—global PROB—global PROB—global Death due to Non O157 HUS ESRD PROB—local INC ETEC -associated diarrhea deaths <5 AF MRT ETEC <5 diarrhea INC PROB—global diarrhea deaths 5+M National studies approach PROB—global PROB—global PROB—local HUS Death ESRD Death ESRD Survivor INC AF MRT ETEC 5+ diarrhea deaths 5+F CHERG approach

EPEC Campylobacter spp. Disease Model Disease Model

INC INC INC PROB—local INC PROB—local Campylobacter- Death due to diarrhea <5 AF INC EPEC <5 diarrhea <5 AF INC Campy <5 Camylobacter- associated associated INC INC diarrhea diarrhea diarrhea 5-14 diarrhea 5-14 INC National studies approach INC PROB—local EPEC -associated INC PROB—local diarrhea 15-54 AF INC EPEC 5+ diarrhea diarrhea 15-54 AF INC Campy 5+ INC INC INC GBS diarrhea 55+ INC diarrhea 55+ Death due to PROB—global INC PROB—local EPEC -associated INC PROB—local diarrhea deaths <5 AF MRT EPEC <5 diarrhea deaths <5 AF MRT Campy <5 attributable fraction diarrhea Campylobacter INC INC diarrhea deaths 5+M National studies approach diarrhea deaths 5+M PROB—local PROB—local PROB—global INC AF MRT EPEC 5+ INC AF MRT Campy 5+ GBS case fatality diarrhea deaths 5+F diarrhea deaths 5+F ratio CHERG approach CHERG approach

Cyanide in cassava Norovirus Disease Model Disease Model

INC INC PROB—local cassava cyanide diarrhea <5 AF INC induced konzo Norovirus <5 INC diarrhea 5-14 INC INC PROB—local PROB—global Norovirus- diarrhea 15-54 AF INC konzo incidence Norovirus 5+ associated diarrhea expansion factor INC diarrhea 55+ INC Death due to INC PROB—local PROB—global PROB—global AF MRT Norovirus- PROB—global diarrhea deaths <5 associated konzo konzo Norovirus <5 identity factor diarrhea survival ratio case-fatality ratio INC diarrhea deaths 5+M PROB—local National studies approach AF MRT INC Norovirus 5+ diarrhea deaths 5+F CHERG approach

Peanut allergen Disease Model

INC INC Death due to peanut- Peanut-induced allergy induced allergy APPENDICES WHO Estimates of the global burden of foodborne diseases 195

APPENDIX 6. Derivation of Disability Weights

Figure A6.1 FERG hazards, causally related health states and corresponding disability weights (DWs). The fourth column describes how the various DWs were derived from the Global Burden of Disease Studies (GBD) and the World Health Organization Global Health Estimates (WHO/ GHE).

HAZARD HEALTH STATE DW MAPPING Diarrhoeal hazards Norovirus Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Campylobacter spp. Diarrhoeal disease 0.101 Weighted average of 73% Diarrhoea: mild (DW=0.061); 25% Diarrhoea: moderate (DW=0.202); and 2% Diarrhoea: severe (DW=0.281) Guillain-Barré syndrome 0.445 Proxy health state of Multiple sclerosis: moderate Enteropathogenic E. coli Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Enterotoxigenic E. coli Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Shiga toxin-producing E. coli Diarrhoeal disease 0.091 Weighted average of 80% Diarrhoea: mild (DW=0.061); 18% Diarrhoea: moderate (DW=0.202); and 2% Diarrhoea: severe (DW=0.281) Haemolytic uraemic syndrome 0.210 Proxy health state of Infectious disease: acute episode, severe End-stage renal disease 0.573 Mapped health state of End-stage renal disease: on dialysis Non-typhoidal S. enterica Diarrhoeal disease 0.101 Weighted average of 73% Diarrhoea: mild (DW=0.061); 25% Diarrhoea: moderate (DW=0.202); and 2% Diarrhoea: severe (DW=0.281) Invasive salmonellosis 0.210 Proxy health state of Infectious disease: acute episode, severe Shigella spp. Diarrhoeal disease 0.101 Weighted average of 73% Diarrhoea: mild (DW=0.061); 25% Diarrhoea: moderate (DW=0.202); and 2% Diarrhoea: severe (DW=0.281) Vibrio cholerae Diarrhoeal disease 0.194 Weighted average of 25% Diarrhoea: mild (DW=0.061); 40% Diarrhoea: moderate (DW=0.202); and 35% Diarrhoea: severe (DW=0.281) Cryptosporidium spp. Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Entamoeba histolytica Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Giardia spp. Diarrhoeal disease 0.074 Weighted average of 91% Diarrhoea: mild (DW=0.061); 8.5% Diarrhoea: moderate (DW=0.202); and 0.5% Diarrhoea: severe (DW=0.281) Appendices 196

HAZARD HEALTH STATE DW MAPPING Invasive enteric hazards Hepatitis A virus Hepatitis 0.108 Weighted average of 50% Infectious disease: acute episode, mild (DW=0.005); and 50% Infectious disease: acute episode, severe (DW=0.210) Brucella spp. Acute brucellosis 0.132 Weighted average of 50% Infectious disease: acute episode, moderate (DW=0.053); and 50% Infectious disease: acute episode, severe (DW=0.210) Chronic brucellosis 0.079 Proxy health state of Musculoskeletal problems: legs, moderate Orchitis 0.097 Mapped health state of Epididymo-orchitis Listeria monocytogenes, Sepsis 0.210 Proxy health state of Infectious disease: acute perinatal episode, severe Central nervous system 0.426 Weighted average; see [70] infection Neurological sequelae 0.292 Weighted average; see [70] Listeria monocytogenes, Sepsis 0.210 Proxy health state of Infectious disease: acute acquired episode, severe Central nervous system 0.426 Weighted average; see [70] for details infection Neurological sequelae 0.292 Weighted average; see [70] for details Mycobacterium bovis Tuberculosis 0.331 Mapped health state of Tuberculosis: without HIV infection Salmonella Paratyphi Paratyphoid fever 0.210 Proxy health state of Infectious disease: acute episode, severe Liver abscesses and cysts 0.254 Proxy health state of Infectious disease: post- acute consequences (fatigue, emotional lability, insomnia) Salmonella Typhi Typhoid fever 0.210 Proxy health state of Infectious disease: acute episode, severe Liver abscesses and cysts 0.254 Proxy health state of Infectious disease: post- acute consequences (fatigue, emotional lability, insomnia) Toxoplasma gondii, congenital Intracranial calcification 0.010 Proxy health state; see [296] for details Hydrocephalus 0.360 Weighted average; see [296] for details Chorioretinitis, 1st year of life 0.033 Proxy health state of Distance vision: moderate impairment Chorioretinitis, later in life 0.033 Proxy health state of Distance vision: moderate impairment Central nervous system 0.360 Weighted average; see [296] for details abnormalities Toxoplasma gondii, acquired Chorioretinitis, mild 0.004 Proxy health state of Distance vision: mild impairment Chorioretinitis, moderate 0.033 Proxy health state of Distance vision: moderate impairment Chorioretinitis, severe 0.191 Proxy health state of Distance vision: severe impairment Acute illness 0.053 Mapped health state of Infectious disease: acute episode, moderate Post-acute illness 0.254 Mapped health state of Infectious disease: post-acute consequences (fatigue, emotional lability, insomnia) APPENDICES WHO Estimates of the global burden of foodborne diseases 197

HAZARD HEALTH STATE DW MAPPING Enteric intoxications Bacillus cereus (1) Acute intoxication 0.061 Proxy health state of Diarrhoea: mild Clostridium botulinum (1) Moderate/mild botulism 0.198 Proxy health state of Multiple sclerosis: mild Severe botulism 0.445 Proxy health state of Multiple sclerosis: moderate Clostridium perfringens (1) Acute intoxication 0.061 Proxy health state of Diarrhoea: mild Staphylococcus aureus (1) Acute intoxication 0.061 Proxy health state of Diarrhoea: mild Cestodes Echinococcus granulosus, cases Pulmonary cystic 0.192 Proxy health state of COPD and other chronic seeking treatment echinococcosis respiratory diseases: moderate Hepatic cystic echinococcosis 0.123 Proxy health state of Abdominopelvic problem: moderate Central nervous system cystic 0.221 Proxy health state of Motor plus cognitive echinococcosis impairments: moderate Echinococcus granulosus, cases Pulmonary cystic 0.015 Proxy health state of COPD and other chronic not seeking treatment echinococcosis respiratory diseases: mild Hepatic cystic echinococcosis 0.012 Proxy health state of Abdominopelvic problem: mild Central nervous system cystic 0.054 Proxy health state of Motor plus cognitive echinococcosis impairments: mild Echinococcus multilocularis Alveolar echinococcosis 0.123 Proxy health state of Abdominopelvic problem: moderate Taenia solium Epilepsy: treated, seizure free 0.072 Mapped health state of Epilepsy: treated, seizure free Epilepsy: treated, with recent 0.319 Mapped health state of Epilepsy: treated, with seizures recent seizures Epilepsy: severe 0.657 Mapped health state of Epilepsy: severe Epilepsy: untreated 0.420 Mapped health state of Epilepsy: untreated Nematodes Ascaris spp. Ascariasis infestation 0.030 Mapped health state of Intestinal nematode infections: symptomatic Mild abdominopelvic 0.012 Mapped health state of Abdominopelvic problems due to ascariasis problem: mild Severe wasting due to 0.127 Mapped health state of Severe wasting ascariasis

Trichinella spp. Acute clinical trichinellosis 0.637 Aggregate of Diarrhoea: moderate (DW = 0.202); Disfigurement: level 2, with itch or pain (DW = 0.187); Musculoskeletal problems: generalized, moderate (DW = 0.292); and Infectious disease: acute episode, severe (DW = 0.210) [84] Trematodes Clonorchis sinensis Abdominopelvic problems 0.123 Proxy health state of Abdominopelvic problem: due to heavy clonorchiosis moderate Fasciola spp. Abdominopelvic problems 0.123 Proxy health state of Abdominopelvic problem: due to heavy fasciolosis moderate Intestinal flukes (2) Abdominopelvic problems 0.123 Proxy health state of Abdominopelvic problem: due to heavy intestinal fluke moderate infections Opisthorchis spp. Abdominopelvic problems 0.123 Proxy health state of Abdominopelvic problem: due to heavy opisthorchiosis moderate Appendices 198

HAZARD HEALTH STATE DW MAPPING Paragonimus spp. Central nervous system 0.420 Proxy health state of Epilepsy: untreated problems due to heavy paragonimosis Pulmonary problems due to 0.132 Proxy health state of Asthma: uncontrolled heavy paragonimosis Organic pollutants Dioxin Infertility 0.056 (3) Mapped health state of Infertility: primary Hypothyroidy due to prenatal 0.019 (4) Mapped health state of Hypothyroidy exposure Hypothyroidy due postnatal 0.019 (4) Mapped health state of Hypothyroidy exposure Toxins and allergens Aflatoxin Hepatocellular carcinoma: 0.294 Mapped health state of Cancer: diagnosis and diagnosis and primary therapy primary therapy Hepatocellular carcinoma: 0.484 Mapped health state of Cancer: metastatic metastatic Hepatocellular carcinoma: 0.508 Mapped health state of Cancer: terminal phase terminal phase with with medication medication Hepatocellular carcinoma: 0.519 Mapped health state of Cancer: terminal phase terminal phase without without medication medication Cyanide in cassava Konzo 0.065 Weighted average of 63% Motor impairment: mild (DW=0.012); 27% Motor impairment: moderate (DW=0.076); and 10% Motor impairment: severe (DW=0.377) Peanut (1) Living with peanut-induced 0.012 Weighted average of 94% Asthma: controlled allergy (DW=0.009); and 6% Generic uncomplicated disease: anxiety about diagnosis (DW=0.054) (1) Excluded from global burden assessments.

(2) Includes Echinostoma spp., Fasciolopsis buski, Heterophyes spp., Metagonimus spp. and other foodborne intestinal trematode species.

(3) Note the higher values used in WHO/GHE [310] compared with GBD2010 [82].

(4) Value taken from the GBD 2013 disability weights [142]. APPENDICES WHO Estimates of the global burden of foodborne diseases 199

APPENDIX 7: Attribution – Expert Elicitation Results

Table A7.1 Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by Campylobacter spp., non-typhoidal Salmonella spp., Shiga-toxin producing Escherichia coli (STEC), Brucella spp. and Shigella spp. through each exposure pathway.

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL OTHER (DOMESTIC AND CONTACT WILD) DIARRHOEAL DISEASE Campylobacter spp. 0.57 0.18 0.04 0.09 0.00 0.06 AFR D (0.31–0.77) (0.00–0.42) (0.00–0.22) (0.01–0.29) (0.00–0.12) (0.00–0.16) 0.57 0.17 0.04 0.09 0.00 0.06 AFR E (0.29–0.77) (0.00–0.42) (0.00–0.23) (0.00–0.30) (0.00–0.12) (0.00–0.16) 0.73 0.10 0.00 0.11 0.00 0.00 AMR A (0.38–0.91) (0.00–0.37) (0.00–0.20) (0.00–0.32) (0.00–0.11) (0.00–0.02) 0.68 0.11 0.03 0.08 0.00 0.06 AMR B (0.41–0.82) (0.00–0.33) (0.00–0.21) (0.00–0.27) (0.00–0.11) (0.00–0.16) 0.67 0.12 0.03 0.08 0.00 0.06 AMR D (0.37–0.81) (0.01–0.36) (0.00–0.21) (0.00–0.29) (0.00–0.15) (0.00–0.16) 0.67 0.11 0.03 0.07 0.00 0.06 EMR B (0.38–0.82) (0.01–0.35) (0.00–0.27) (0.00–0.29) (0.00–0.15) (0.00–0.15) 0.67 0.11 0.03 0.07 0.00 0.06 EMR D (0.41–0.82) (0.00–0.34) (0.00–0.22) (0.00–0.27) (0.00–0.20) (0.00–0.15) 0.76 0.08 0.01 0.06 0.01 0.00 EUR A (0.44–0.93) (0.00–0.31) (0.00–0.13) (0.00–0.35) (0.00–0.09) (0.00–0.08) 0.66 0.11 0.03 0.12 0.03 0.00 EUR B (0.34–0.87) (0.00–0.39) (0.00–0.21) (0.00–0.40) (0.00–0.13) (0.00–0.05) 0.66 0.11 0.03 0.12 0.03 0.00 EUR C (0.34–0.87) (0.00–0.38) (0.00–0.23) (0.00–0.39) (0.00–0.19) (0.00–0.02) 0.57 0.13 0.11 0.05 0.03 0.02 SEAR B (0.27–0.81) (0.00–0.36) (0.00–0.36) (0.00–0.35) (0.00–0.21) (0.00–0.06) 0.51 0.11 0.11 0.07 0.03 0.02 SEAR D (0.03–0.79) (0.00–0.39) (0.01–0.41) (0.00–0.44) (0.00–0.32) (0.00–0.10) 0.68 0.13 0.00 0.11 0.00 0.00 WPR A (0.40–0.89) (0.00–0.33) (0.00–0.23) (0.00–0.32) (0.00–0.08) (0.00–0.01) 0.57 0.17 0.06 0.05 0.03 0.02 WPR B (0.25–0.82) (0.00–0.42) (0.00–0.34) (0.00–0.32) (0.00–0.15) (0.00–0.07) Non-typhoidal Salmonella enterica 0.46 0.15 0.18 0.10 0.01 0.02 AFR D (0.13–0.74) (0.00–0.43) (0.00–0.48) (0.00–0.39) (0.00–0.13) (0.00–0.06) 0.46 0.15 0.18 0.10 0.01 0.02 AFR E (0.10–0.73) (0.00–0.42) (0.00–0.48) (0.00–0.40) (0.00–0.19) (0.00–0.08) 0.73 0.10 0.05 0.02 0.00 0.00 AMR A (0.38–0.91) (0.00–0.39) (0.00–0.28) (0.00–0.22) (0.00–0.09) (0.00–0.05) 0.49 0.19 0.15 0.09 0.01 0.02 AMR B (0.09–0.74) (0.00–0.45) (0.00–0.40) (0.00–0.32) (0.00–0.12) (0.00–0.05) 0.50 0.19 0.15 0.09 0.01 0.02 AMR D (0.14–0.75) (0.00–0.46) (0.00–0.39) (0.00–0.31) (0.00–0.12) (0.00–0.05) 0.50 0.15 0.15 0.12 0.01 0.02 EMR B (0.18–0.75) (0.00–0.43) (0.01–0.38) (0.00–0.33) (0.00–0.19) (0.00–0.04) 0.50 0.15 0.15 0.12 0.01 0.02 EMR D (0.19–0.74) (0.00–0.43) (0.01–0.39) (0.00–0.32) (0.00–0.21) (0.00–0.05) 0.76 0.05 0.06 0.03 0.00 0.00 EUR A (0.47–0.94) (0.00–0.30) (0.00–0.26) (0.00–0.21) (0.00–0.11) (0.00–0.14) Appendices 200

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL OTHER (DOMESTIC AND CONTACT WILD) 0.62 0.10 0.11 0.07 0.02 0.00 EUR B (0.31–0.84) (0.00–0.37) (0.01–0.32) (0.00–0.32) (0.00–0.12) (0.00–0.01) 0.62 0.10 0.10 0.07 0.02 0.00 EUR C (0.32–0.84) (0.00–0.36) (0.00–0.32) (0.00–0.32) (0.00–0.12) (0.00–0.01) 0.58 0.06 0.10 0.11 0.02 0.00 SEAR B (0.23–0.84) (0.00–0.32) (0.00–0.38) (0.00–0.40) (0.00–0.20) (0.00–0.03) 0.54 0.06 0.10 0.15 0.02 0.00 SEAR D (0.00–0.85) (0.00–0.37) (0.00–0.42) (0.00–0.59) (0.00–0.29) (0.00–0.06) 0.74 0.09 0.04 0.01 0.00 0.00 WPR A (0.45–0.93) (0.00–0.31) (0.00–0.28) (0.00–0.22) (0.00–0.08) (0.00–0.04) 0.57 0.10 0.12 0.08 0.02 0.00 WPR B (0.25–0.82) (0.00–0.33) (0.00–0.35) (0.00–0.37) (0.00–0.21) (0.00–0.01) Shiga toxin-producing E. coli 0.42 0.21 0.16 0.10 0.05 0.00 AFR D (0.19–0.66) (0.04–0.46) (0.00–0.33) (0.00–0.30) (0.00–0.25) (0.00–0.03) 0.43 0.21 0.17 0.10 0.05 0.00 AFR E (0.14–0.66) (0.04–0.46) (0.01–0.34) (0.00–0.34) (0.00–0.19) (0.00–0.03) 0.59 0.13 0.07 0.07 0.00 0.00 AMR A (0.19–0.84) (0.00–0.41) (0.00–0.32) (0.00–0.31) (0.00–0.13) (0.00–0.27) 0.53 0.17 0.11 0.08 0.04 0.00 AMR B (0.24–0.73) (0.01–0.44) (0.01–0.29) (0.00–0.32) (0.00–0.21) (0.00–0.03) 0.53 0.15 0.11 0.09 0.04 0.00 AMR D (0.24–0.75) (0.00–0.43) (0.01–0.29) (0.00–0.32) (0.00–0.17) (0.00–0.03) 0.53 0.15 0.11 0.10 0.04 0.00 EMR B (0.24–0.76) (0.02–0.43) (0.00–0.29) (0.00–0.37) (0.00–0.18) (0.00–0.03) 0.52 0.14 0.11 0.10 0.04 0.00 EMR D (0.26–0.75) (0.01–0.42) (0.01–0.30) (0.00–0.37) (0.00–0.17) (0.00–0.03) 0.60 0.11 0.08 0.07 0.03 0.00 EUR A (0.26–0.83) (0.01–0.37) (0.00–0.33) (0.00–0.33) (0.00–0.19) (0.00–0.14) 0.49 0.12 0.10 0.09 0.08 0.00 EUR B (0.15–0.75) (0.00–0.42) (0.01–0.32) (0.00–0.38) (0.00–0.35) (0.00–0.01) 0.49 0.12 0.10 0.09 0.08 0.00 EUR C (0.15–0.75) (0.00–0.42) (0.01–0.32) (0.00–0.36) (0.00–0.35) (0.00–0.01) 0.41 0.12 0.07 0.23 0.06 0.00 SEAR B (0.10–0.70) (0.00–0.47) (0.00–0.31) (0.00–0.53) (0.00–0.26) (0.00–0.01) 0.40 0.13 0.06 0.23 0.06 0.00 SEAR D (0.08–0.71) (0.00–0.47) (0.00–0.35) (0.00–0.53) (0.00–0.26) (0.00–0.02) 0.57 0.14 0.07 0.07 0.00 0.00 WPR A (0.25–0.82) (0.00–0.36) (0.00–0.35) (0.00–0.29) (0.00–0.16) (0.00–0.24) 0.43 0.12 0.07 0.22 0.06 0.00 WPR B (0.12–0.73) (0.00–0.44) (0.00–0.35) (0.00–0.46) (0.00–0.27) (0.00–0.01) Brucella spp. 0.44 0.50 0.01 0.01 0.01 AFR D na (0.10–0.68) (0.26–0.81) (0.00–0.08) (0.00–0.10) (0.00–0.06) 0.44 0.50 0.01 0.01 0.01 AFR E na (0.06–0.70) (0.22–0.83) (0.00–0.12) (0.00–0.11) (0.00–0.06) 0.75 0.19 0.01 0.01 0.01 AMR A na (0.28–0.93) (0.00–0.62) (0.00–0.04) (0.00–0.09) (0.00–0.12) 0.44 0.50 0.01 0.01 0.01 AMR B na (0.09–0.69) (0.24–0.81) (0.00–0.08) (0.00–0.12) (0.00–0.08) APPENDICES WHO Estimates of the global burden of foodborne diseases 201

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL OTHER (DOMESTIC AND CONTACT WILD) 0.44 0.50 0.01 0.01 0.01 AMR D na (0.09–0.72) (0.18–0.81) (0.00–0.12) (0.00–0.11) (0.00–0.06) 0.51 0.43 0.01 0.01 0.01 EMR B na (0.08–0.80) (0.11–0.81) (0.00–0.07) (0.00–0.08) (0.00–0.11) 0.44 0.50 0.01 0.01 0.01 EMR D na (0.07–0.70) (0.20–0.83) (0.00–0.14) (0.00–0.15) (0.00–0.06) 0.66 0.23 0.01 0.01 0.02 EUR A na (0.23–0.90) (0.01–0.60) (0.00–0.04) (0.00–0.05) (0.00–0.35) 0.45 0.50 0.01 0.01 0.01 EUR B na (0.09–0.71) (0.20–0.81) (0.00–0.07) (0.00–0.08) (0.00–0.06) 0.44 0.50 0.01 0.01 0.01 EUR C na (0.10–0.73) (0.18–0.81) (0.00–0.06) (0.00–0.06) (0.00–0.06) 0.51 0.43 0.01 0.01 0.01 SEAR B na (0.07–0.81) (0.10–0.81) (0.00–0.07) (0.00–0.07) (0.00–0.07) 0.45 0.50 0.01 0.01 0.01 SEAR D na (0.07–0.70) (0.22–0.82) (0.00–0.08) (0.00–0.07) (0.00–0.06) 0.71 0.18 0.01 0.01 0.02 WPR A na (0.28–0.92) (0.00–0.58) (0.00–0.09) (0.00–0.26) (0.00–0.30) 0.51 0.43 0.01 0.01 0.01 WPR B na (0.07–0.80) (0.12–0.81) (0.00–0.07) (0.00–0.07) (0.00–0.07) Shigella spp. 0.15 0.50 0.27 0.00 0.00 AFR D na (0.00–0.52) (0.06–0.81) (0.03–0.62) (0.00–0.19) (0.00–0.13) 0.15 0.50 0.26 0.00 0.00 AFR E na (0.00–0.51) (0.08–0.80) (0.05–0.61) (0.00–0.19) (0.00–0.16) 0.12 0.69 0.10 0.00 0.00 AMR A na (0.00–0.46) (0.33–0.93) (0.00–0.41) (0.00–0.21) (0.00–0.06) 0.14 0.51 0.27 0.00 0.00 AMR B na (0.00–0.52) (0.10–0.81) (0.03–0.61) (0.00–0.18) (0.00–0.06) 0.14 0.51 0.27 0.00 0.00 AMR D na (0.00–0.52) (0.11–0.80) (0.02–0.60) (0.00–0.20) (0.00–0.02) 0.14 0.51 0.28 0.00 0.00 EMR B na (0.00–0.52) (0.11–0.81) (0.03–0.61) (0.00–0.17) (0.00–0.02) 0.14 0.51 0.28 0.00 0.00 EMR D na (0.00–0.52) (0.11–0.81) (0.02–0.61) (0.00–0.18) (0.00–0.02) 0.07 0.54 0.12 0.01 0.00 EUR A na (0.00–0.46) (0.14–0.90) (0.00–0.52) (0.00–0.20) (0.00–0.55) 0.11 0.44 0.31 0.02 0.02 EUR B na (0.00–0.50) (0.10–0.75) (0.04–0.60) (0.00–0.20) (0.00–0.21) 0.19 0.43 0.26 0.01 0.05 EUR C na (0.00–0.51) (0.07–0.70) (0.02–0.53) (0.00–0.20) (0.00–0.22) 0.36 0.30 0.26 0.04 0.01 SEAR B na (0.01–0.68) (0.01–0.65) (0.01–0.59) (0.00–0.21) (0.00–0.03) 0.34 0.25 0.29 0.04 0.01 SEAR D na (0.01–0.69) (0.00–0.64) (0.01–0.65) (0.00–0.26) (0.00–0.06) 0.13 0.66 0.12 0.00 0.00 WPR A na (0.00–0.50) (0.25–0.91) (0.00–0.42) (0.00–0.22) (0.00–0.19) 0.36 0.28 0.27 0.04 0.01 WPR B na (0.01–0.70) (0.00–0.65) (0.01–0.60) (0.00–0.22) (0.00–0.03) Appendices 202

Table A7.2 Subregional estimates (median and 95% uncertainty interval) of the proportion of Diarrhoeal Disease illnesses caused by four hazards: enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), Cryptosporidium spp. and Giardia spp. through each exposure pathway.

ANIMAL CONTACT SUB- HUMAN TO HUMAN FOOD (DOMESTIC AND WATER OTHER REGION CONTACT WILD) Enteropathogenic E. coli AFR D 0.29 (0.02–0.62) 0.00 (0.00–0.33) 0.16 (0.00–0.51) 0.45 (0.12–0.76) 0.00 (0.00–0.01) AFR E 0.29 (0.01–0.62) 0.00 (0.00–0.32) 0.16 (0.00–0.51) 0.46 (0.10–0.76) 0.00 (0.00–0.01) AMR A 0.72 (0.20–0.97) 0.00 (0.00–0.31) 0.11 (0.00–0.53) 0.00 (0.00–0.57) 0.00 (0.00–0.01) AMR B 0.29 (0.01–0.62) 0.00 (0.00–0.34) 0.16 (0.00–0.50) 0.46 (0.12–0.76) 0.00 (0.00–0.01) AMR D 0.30 (0.03–0.61) 0.00 (0.00–0.33) 0.15 (0.00–0.47) 0.47 (0.13–0.74) 0.00 (0.00–0.01) EMR B 0.31 (0.06–0.62) 0.00 (0.00–0.35) 0.14 (0.00–0.44) 0.46 (0.11–0.70) 0.00 (0.00–0.01) EMR D 0.31 (0.05–0.62) 0.00 (0.00–0.37) 0.14 (0.00–0.44) 0.45 (0.10–0.70) 0.00 (0.00–0.01) EUR A 0.64 (0.17–0.90) 0.05 (0.00–0.38) 0.17 (0.00–0.58) 0.03 (0.00–0.31) 0.00 (0.00–0.21) EUR B 0.48 (0.06–0.81) 0.08 (0.00–0.41) 0.26 (0.00–0.65) 0.08 (0.00–0.43) 0.00 (0.00–0.01) EUR C 0.48 (0.06–0.81) 0.09 (0.00–0.42) 0.26 (0.00–0.65) 0.08 (0.00–0.42) 0.00 (0.00–0.02) SEAR B 0.29 (0.01–0.62) 0.09 (0.00–0.34) 0.29 (0.01–0.62) 0.27 (0.01–0.58) 0.00 (0.00–0.02) SEAR D 0.29 (0.01–0.67) 0.09 (0.00–0.38) 0.27 (0.00–0.65) 0.27 (0.00–0.63) 0.00 (0.00–0.05) WPR A 0.69 (0.16–0.94) 0.00 (0.00–0.34) 0.18 (0.00–0.66) 0.00 (0.00–0.30) 0.00 (0.00–0.02) WPR B 0.30 (0.01–0.62) 0.14 (0.00–0.40) 0.23 (0.00–0.59) 0.26 (0.02–0.55) 0.00 (0.00–0.01) Enterotoxigenic E. coli AFR D 0.33 (0.09–0.65) 0.00 (0.00–0.33) 0.13 (0.00–0.44) 0.45 (0.12–0.71) 0.00 (0.00–0.01) AFR E 0.33 (0.06–0.64) 0.00 (0.00–0.33) 0.13 (0.00–0.45) 0.45 (0.09–0.71) 0.00 (0.00–0.01) AMR A 0.36 (0.12–0.63) 0.04 (0.00–0.32) 0.15 (0.00–0.37) 0.42 (0.11–0.66) 0.00 (0.00–0.19) AMR B 0.34 (0.08–0.65) 0.00 (0.00–0.34) 0.12 (0.00–0.42) 0.46 (0.11–0.70) 0.00 (0.00–0.13) AMR D 0.36 (0.07–0.68) 0.00 (0.00–0.32) 0.13 (0.00–0.43) 0.47 (0.10–0.72) 0.00 (0.00–0.01) EMR B 0.34 (0.07–0.65) 0.00 (0.00–0.31) 0.13 (0.00–0.42) 0.49 (0.10–0.72) 0.00 (0.00–0.01) EMR D 0.35 (0.05–0.66) 0.00 (0.00–0.31) 0.12 (0.00–0.41) 0.48 (0.12–0.73) 0.00 (0.00–0.01) EUR A 0.42 (0.09–0.73) 0.05 (0.00–0.31) 0.26 (0.01–0.60) 0.18 (0.00–0.53) 0.00 (0.00–0.08) EUR B 0.43 (0.05–0.73) 0.05 (0.00–0.34) 0.31 (0.02–0.66) 0.14 (0.00–0.47) 0.00 (0.00–0.18) EUR C 0.43 (0.06–0.72) 0.05 (0.00–0.34) 0.31 (0.02–0.66) 0.14 (0.00–0.47) 0.00 (0.00–0.20) SEAR B 0.38 (0.03–0.73) 0.05 (0.00–0.32) 0.09 (0.00–0.51) 0.39 (0.02–0.71) 0.00 (0.00–0.02) SEAR D 0.37 (0.02–0.73) 0.06 (0.00–0.34) 0.09 (0.00–0.52) 0.38 (0.03–0.73) 0.00 (0.00–0.11) WPR A 0.38 (0.10–0.72) 0.04 (0.00–0.29) 0.20 (0.00–0.53) 0.33 (0.00–0.61) 0.00 (0.00–0.01) WPR B 0.38 (0.03–0.72) 0.04 (0.00–0.29) 0.08 (0.00–0.50) 0.39 (0.04–0.71) 0.00 (0.00–0.20) Cryptosporidium spp. AFR D 0.15 (0.00–0.44) 0.06 (0.00–0.27) 0.38 (0.01–0.72) 0.35 (0.01–0.68) 0.01 (0.00–0.16) AFR E 0.15 (0.00–0.47) 0.05 (0.00–0.26) 0.36 (0.01–0.72) 0.37 (0.01–0.71) 0.01 (0.00–0.17) AMR A 0.16 (0.01–0.44) 0.10 (0.01–0.42) 0.30 (0.03–0.64) 0.37 (0.08–0.72) 0.00 (0.00–0.09) AMR B 0.11 (0.01–0.38) 0.20 (0.02–0.47) 0.35 (0.07–0.66) 0.26 (0.05–0.61) 0.00 (0.00–0.09) AMR D 0.16 (0.01–0.44) 0.21 (0.03–0.49) 0.34 (0.07–0.66) 0.20 (0.03–0.59) 0.00 (0.00–0.08) EMR B 0.09 (0.00–0.41) 0.14 (0.00–0.46) 0.31 (0.02–0.65) 0.36 (0.05–0.69) 0.01 (0.00–0.17) EMR D 0.08 (0.00–0.36) 0.13 (0.00–0.43) 0.32 (0.01–0.66) 0.38 (0.06–0.71) 0.01 (0.00–0.17) APPENDICES WHO Estimates of the global burden of foodborne diseases 203

ANIMAL CONTACT SUB- HUMAN TO HUMAN FOOD (DOMESTIC AND WATER OTHER REGION CONTACT WILD) EUR A 0.10 (0.00–0.39) 0.14 (0.00–0.44) 0.30 (0.01–0.65) 0.38 (0.03–0.70) 0.01 (0.00–0.09) EUR B 0.11 (0.00–0.39) 0.16 (0.00–0.46) 0.28 (0.01–0.64) 0.37 (0.02–0.68) 0.01 (0.00–0.08) EUR C 0.09 (0.00–0.40) 0.15 (0.00–0.48) 0.29 (0.01–0.64) 0.36 (0.05–0.70) 0.01 (0.00–0.09) SEAR B 0.10 (0.00–0.37) 0.13 (0.00–0.46) 0.31 (0.01–0.66) 0.38 (0.02–0.71) 0.01 (0.00–0.09) SEAR D 0.10 (0.00–0.42) 0.13 (0.00–0.46) 0.30 (0.01–0.66) 0.37 (0.03–0.71) 0.01 (0.00–0.15) WPR A 0.10 (0.00–0.40) 0.12 (0.00–0.46) 0.29 (0.01–0.66) 0.39 (0.03–0.72) 0.01 (0.00–0.09) WPR B 0.10 (0.00–0.45) 0.10 (0.00–0.45) 0.29 (0.01–0.66) 0.39 (0.04–0.73) 0.01 (0.00–0.10) Giardia spp. AFR D 0.11 (0.00–0.43) 0.03 (0.00–0.27) 0.43 (0.01–0.75) 0.33 (0.05–0.69) 0.02 (0.00–0.18) AFR E 0.11 (0.00–0.43) 0.03 (0.00–0.25) 0.44 (0.04–0.75) 0.32 (0.04–0.67) 0.02 (0.00–0.19) AMR A 0.11 (0.00–0.39) 0.14 (0.00–0.41) 0.25 (0.00–0.64) 0.42 (0.05–0.75) 0.00 (0.00–0.12) AMR B 0.12 (0.00–0.42) 0.18 (0.00–0.47) 0.32 (0.01–0.67) 0.30 (0.04–0.65) 0.00 (0.00–0.09) AMR D 0.12 (0.00–0.42) 0.18 (0.00–0.46) 0.36 (0.01–0.69) 0.26 (0.03–0.63) 0.00 (0.00–0.10) EMR B 0.13 (0.00–0.50) 0.02 (0.00–0.15) 0.45 (0.03–0.77) 0.32 (0.03–0.71) 0.01 (0.00–0.19) EMR D 0.13 (0.00–0.47) 0.02 (0.00–0.25) 0.39 (0.02–0.73) 0.35 (0.03–0.71) 0.01 (0.00–0.18) EUR A 0.11 (0.00–0.44) 0.02 (0.00–0.15) 0.47 (0.02–0.79) 0.32 (0.03–0.72) 0.01 (0.00–0.14) EUR B 0.12 (0.00–0.47) 0.02 (0.00–0.15) 0.44 (0.02–0.77) 0.34 (0.02–0.73) 0.01 (0.00–0.12) EUR C 0.12 (0.00–0.48) 0.02 (0.00–0.15) 0.44 (0.02–0.77) 0.34 (0.04–0.74) 0.01 (0.00–0.13) SEAR B 0.13 (0.00–0.48) 0.02 (0.00–0.23) 0.41 (0.02–0.74) 0.35 (0.02–0.72) 0.01 (0.00–0.17) SEAR D 0.13 (0.00–0.48) 0.02 (0.00–0.22) 0.41 (0.02–0.76) 0.35 (0.03–0.72) 0.01 (0.00–0.16) WPR A 0.12 (0.00–0.45) 0.02 (0.00–0.31) 0.46 (0.02–0.78) 0.29 (0.01–0.68) 0.01 (0.00–0.18) WPR B 0.14 (0.00–0.49) 0.02 (0.00–0.29) 0.43 (0.02–0.75) 0.30 (0.03–0.69) 0.01 (0.00–0.19) Appendices 204

Table A7.3 Subregional estimates (median and 95% uncertainty interval) of the proportion of Diarrhoeal Disease illnesses caused by Salmonella Typhi, Vibrio cholerae, Entamoeba histolytica, norovirus, and hepatitis A virus through each exposure pathway.

HUMAN-TO-HUMAN SUBREGION FOOD WATER OTHER CONTACT Salmonella Typhi AFR D 0.24 (0.00–0.58) 0.22 (0.00–0.54) 0.51 (0.13–0.82) 0.00 (0.00–0.09) AFR E 0.24 (0.00–0.58) 0.22 (0.00–0.53) 0.51 (0.16–0.81) 0.00 (0.00–0.10) AMR A 0.26 (0.00–0.64) 0.11 (0.00–0.48) 0.57 (0.14–0.87) 0.00 (0.00–0.37) AMR B 0.23 (0.00–0.59) 0.21 (0.00–0.53) 0.52 (0.14–0.82) 0.00 (0.00–0.10) AMR D 0.23 (0.00–0.56) 0.21 (0.00–0.52) 0.53 (0.18–0.81) 0.00 (0.00–0.09) EMR B 0.24 (0.00–0.58) 0.21 (0.00–0.53) 0.52 (0.15–0.82) 0.00 (0.00–0.10) EMR D 0.24 (0.00–0.58) 0.21 (0.00–0.53) 0.52 (0.15–0.83) 0.00 (0.00–0.10) EUR A 0.10 (0.00–0.53) 0.23 (0.00–0.72) 0.41 (0.00–0.83) 0.01 (0.00–0.66) EUR B 0.08 (0.00–0.43) 0.47 (0.16–0.78) 0.35 (0.04–0.62) 0.02 (0.00–0.21) EUR C 0.08 (0.00–0.43) 0.47 (0.15–0.78) 0.35 (0.03–0.62) 0.02 (0.00–0.21) SEAR B 0.43 (0.11–0.82) 0.12 (0.00–0.49) 0.40 (0.01–0.70) 0.00 (0.00–0.03) SEAR D 0.40 (0.01–0.81) 0.13 (0.00–0.54) 0.42 (0.00–0.80) 0.00 (0.00–0.10) WPR A 0.33 (0.00–0.84) 0.11 (0.00–0.55) 0.48 (0.00–0.86) 0.00 (0.00–0.36) WPR B 0.49 (0.10–0.84) 0.13 (0.00–0.51) 0.33 (0.01–0.66) 0.00 (0.00–0.03) Vibrio cholerae AFR D 0.21 (0.01–0.57) 0.02 (0.00–0.31) 0.72 (0.29–0.94) 0.00 (0.00–0.03) AFR E 0.21 (0.01–0.56) 0.02 (0.00–0.30) 0.72 (0.33–0.94) 0.00 (0.00–0.04) AMR A 0.30 (0.01–0.95) 0.02 (0.00–0.43) 0.59 (0.00–0.93) 0.00 (0.00–0.37) AMR B 0.25 (0.00–0.58) 0.02 (0.00–0.27) 0.70 (0.33–0.95) 0.00 (0.00–0.34) AMR D 0.25 (0.00–0.57) 0.02 (0.00–0.29) 0.69 (0.34–0.94) 0.00 (0.00–0.29) EMR B 0.23 (0.01–0.64) 0.02 (0.00–0.30) 0.69 (0.25–0.94) 0.00 (0.00–0.03) EMR D 0.23 (0.01–0.65) 0.02 (0.00–0.31) 0.70 (0.23–0.94) 0.00 (0.00–0.03) EUR A 0.31 (0.00–0.85) 0.03 (0.00–0.44) 0.44 (0.00–0.86) 0.01 (0.00–0.57) EUR B 0.46 (0.01–0.86) 0.11 (0.00–0.47) 0.36 (0.00–0.77) 0.00 (0.00–0.36) EUR C 0.46 (0.02–0.86) 0.11 (0.00–0.47) 0.36 (0.00–0.76) 0.00 (0.00–0.38) SEAR B 0.36 (0.04–0.78) 0.14 (0.00–0.50) 0.45 (0.02–0.79) 0.00 (0.00–0.02) SEAR D 0.25 (0.00–0.75) 0.08 (0.00–0.50) 0.58 (0.04–0.91) 0.00 (0.00–0.02) WPR A 0.25 (0.01–0.92) 0.04 (0.00–0.64) 0.56 (0.00–0.93) 0.00 (0.00–0.05) WPR B 0.29 (0.01–0.74) 0.13 (0.00–0.49) 0.51 (0.04–0.83) 0.00 (0.00–0.30) Norovirus AFR D 0.15 (0.01–0.40) 0.68 (0.37–0.89) 0.07 (0.00–0.38) 0.04 (0.00–0.23) AFR E 0.15 (0.00–0.40) 0.68 (0.38–0.89) 0.07 (0.00–0.37) 0.04 (0.00–0.24) AMR A 0.23 (0.04–0.50) 0.50 (0.18–0.79) 0.22 (0.00–0.49) 0.00 (0.00–0.22) AMR B 0.14 (0.00–0.42) 0.72 (0.36–0.90) 0.06 (0.00–0.40) 0.04 (0.00–0.24) AMR D 0.15 (0.00–0.46) 0.72 (0.36–0.89) 0.06 (0.00–0.41) 0.04 (0.00–0.23) EMR B 0.15 (0.00–0.40) 0.72 (0.43–0.89) 0.07 (0.00–0.30) 0.04 (0.00–0.22) EMR D 0.15 (0.00–0.40) 0.72 (0.42–0.89) 0.06 (0.00–0.32) 0.04 (0.00–0.23) EUR A 0.26 (0.00–0.73) 0.43 (0.00–0.83) 0.17 (0.00–0.58) 0.00 (0.00–0.36) EUR B 0.23 (0.01–0.57) 0.32 (0.02–0.67) 0.33 (0.00–0.65) 0.04 (0.00–0.34) APPENDICES WHO Estimates of the global burden of foodborne diseases 205

HUMAN-TO-HUMAN SUBREGION FOOD WATER OTHER CONTACT EUR C 0.23 (0.01–0.57) 0.33 (0.02–0.67) 0.33 (0.01–0.63) 0.04 (0.00–0.33) SEAR B 0.12 (0.00–0.48) 0.53 (0.13–0.83) 0.21 (0.00–0.53) 0.00 (0.00–0.42) SEAR D 0.15 (0.00–0.55) 0.46 (0.00–0.79) 0.29 (0.00–0.72) 0.00 (0.00–0.35) WPR A 0.22 (0.01–0.52) 0.48 (0.12–0.77) 0.22 (0.00–0.51) 0.00 (0.00–0.32) WPR B 0.15 (0.00–0.55) 0.46 (0.00–0.79) 0.28 (0.01–0.68) 0.00 (0.00–0.34) Hepatitis A AFR D 0.36 (0.07–0.63) 0.40 (0.10–0.68) 0.17 (0.00–0.49) 0.04 (0.00–0.10) AFR E 0.29 (0.07–0.57) 0.36 (0.08–0.64) 0.30 (0.06–0.59) 0.02 (0.00–0.06) AMR A 0.42 (0.06–0.77) 0.46 (0.04–0.78) 0.01 (0.00–0.19) 0.10 (0.00–0.32) AMR B 0.31 (0.03–0.60) 0.46 (0.16–0.74) 0.11 (0.00–0.39) 0.09 (0.00–0.21) AMR D 0.32 (0.03–0.61) 0.35 (0.11–0.65) 0.26 (0.04–0.57) 0.04 (0.00–0.09) EMR B 0.35 (0.04–0.61) 0.42 (0.17–0.69) 0.15 (0.02–0.34) 0.09 (0.00–0.20) EMR D 0.32 (0.02–0.59) 0.36 (0.11–0.66) 0.22 (0.00–0.49) 0.08 (0.00–0.23) EUR A 0.42 (0.02–0.75) 0.46 (0.10–0.79) 0.01 (0.00–0.17) 0.10 (0.00–0.32) EUR B 0.35 (0.12–0.59) 0.35 (0.18–0.61) 0.20 (0.01–0.36) 0.08 (0.00–0.19) EUR C 0.34 (0.08–0.60) 0.42 (0.17–0.69) 0.14 (0.00–0.35) 0.09 (0.00–0.24) SEAR B 0.34 (0.05–0.60) 0.35 (0.14–0.65) 0.23 (0.04–0.55) 0.04 (0.00–0.09) SEAR D 0.29 (0.04–0.56) 0.37 (0.13–0.64) 0.29 (0.06–0.56) 0.02 (0.00–0.06) WPR A 0.42 (0.03–0.76) 0.46 (0.10–0.79) 0.01 (0.00–0.16) 0.10 (0.00–0.29) WPR B 0.34 (0.02–0.64) 0.36 (0.06–0.66) 0.21 (0.01–0.47) 0.08 (0.00–0.20) Entamoeba histolytica AFR D 0.30 (0.00–0.68) 0.37 (0.00–0.73) 0.25 (0.00–0.63) 0.04 (0.00–0.21) AFR E 0.30 (0.00–0.68) 0.37 (0.00–0.72) 0.24 (0.00–0.62) 0.04 (0.00–0.22) AMR A 0.25 (0.00–0.70) 0.34 (0.00–0.76) 0.33 (0.00–0.74) 0.00 (0.00–0.19) AMR B 0.21 (0.00–0.62) 0.38 (0.02–0.76) 0.32 (0.00–0.70) 0.00 (0.00–0.20) AMR D 0.17 (0.00–0.58) 0.37 (0.04–0.76) 0.37 (0.01–0.73) 0.00 (0.00–0.20) EMR B 0.24 (0.00–0.62) 0.42 (0.01–0.76) 0.24 (0.00–0.62) 0.04 (0.00–0.22) EMR D 0.28 (0.00–0.66) 0.39 (0.00–0.75) 0.25 (0.00–0.65) 0.04 (0.00–0.22) EUR A 0.33 (0.00–0.71) 0.49 (0.03–0.83) 0.15 (0.00–0.51) 0.01 (0.00–0.16) EUR B 0.30 (0.00–0.66) 0.42 (0.02–0.76) 0.20 (0.00–0.59) 0.04 (0.00–0.20) EUR C 0.26 (0.00–0.64) 0.42 (0.02–0.76) 0.23 (0.00–0.61) 0.04 (0.00–0.19) SEAR B 0.26 (0.00–0.65) 0.38 (0.00–0.75) 0.28 (0.00–0.68) 0.04 (0.00–0.18) SEAR D 0.25 (0.00–0.63) 0.37 (0.00–0.72) 0.29 (0.01–0.69) 0.04 (0.00–0.19) WPR A 0.25 (0.00–0.62) 0.41 (0.00–0.74) 0.26 (0.01–0.62) 0.04 (0.00–0.25) WPR B 0.27 (0.00–0.63) 0.41 (0.00–0.73) 0.24 (0.01–0.62) 0.05 (0.00–0.23) Appendices 206

Table A7.4 Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by Toxoplasma gondii, Echinococcus multilocularis, Echinococcus granulosus and Ascaris spp. through each exposure pathway.

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL AIR OTHER (DOMES TIC CONTACT AND WILD) PARASITIC DISEASE Toxoplasma gondii 0.48 0.01 0.11 0.36 AFR D na na na (0.24–0.76) (0.00–0.20) (0.00–0.37) (0.07–0.57) 0.42 0.01 0.16 0.38 AFR E na na na (0.20–0.70) (0.00–0.19) (0.02–0.41) (0.05–0.58) 0.60 0.01 0.19 0.19 AMR A na na na (0.30–0.81) (0.00–0.28) (0.01–0.42) (0.00–0.46) 0.52 0.01 0.23 0.22 AMR B na na na (0.27–0.77) (0.00–0.20) (0.01–0.45) (0.00–0.46) 0.53 0.01 0.23 0.22 AMR D na na na (0.27–0.77) (0.00–0.21) (0.02–0.44) (0.00–0.45) 0.52 0.01 0.11 0.34 EMR B na na na (0.27–0.80) (0.00–0.20) (0.01–0.29) (0.02–0.56) 0.53 0.01 0.23 0.22 EMR D na na na (0.29–0.77) (0.00–0.20) (0.02–0.43) (0.00–0.42) 0.61 0.01 0.19 0.18 EUR A na na na (0.35–0.82) (0.00–0.21) (0.02–0.36) (0.00–0.40) 0.45 0.01 0.15 0.37 EUR B na na na (0.23–0.76) (0.00–0.20) (0.02–0.35) (0.01–0.58) 0.53 0.01 0.23 0.22 EUR C na na na (0.31–0.78) (0.00–0.20) (0.03–0.41) (0.01–0.41) 0.52 0.01 0.23 0.22 SEAR B na na na (0.26–0.77) (0.00–0.19) (0.03–0.45) (0.00–0.43) 0.43 0.01 0.27 0.26 SEAR D na na na (0.09–0.73) (0.00–0.22) (0.03–0.58) (0.00–0.56) 0.60 0.01 0.19 0.18 WPR A na na na (0.33–0.81) (0.00–0.21) (0.02–0.37) (0.00–0.43) 0.53 0.01 0.23 0.22 WPR B na na na (0.29–0.77) (0.00–0.20) (0.04–0.43) (0.00–0.43) Echinococcus granulosus 0.21 0.51 0.18 0.09 0.00 0.00 AFR D na (0.07–0.42) (0.25–0.72) (0.01–0.34) (0.00–0.20) (0.00–0.06) (0.00–0.01) 0.20 0.52 0.18 0.09 0.00 0.00 AFR E na (0.05–0.40) (0.27–0.73) (0.00–0.35) (0.00–0.19) (0.00–0.06) (0.00–0.06) 0.20 0.52 0.17 0.09 0.00 0.00 AMR A na (0.03–0.40) (0.30–0.75) (0.00–0.31) (0.00–0.20) (0.00–0.14) (0.00–0.01) 0.20 0.52 0.18 0.09 0.00 0.00 AMR B na (0.02–0.43) (0.28–0.73) (0.00–0.34) (0.00–0.22) (0.00–0.14) (0.00–0.01) 0.21 0.51 0.18 0.09 0.00 0.00 AMR D na (0.05–0.41) (0.29–0.72) (0.01–0.35) (0.00–0.23) (0.00–0.13) (0.00–0.01) 0.21 0.51 0.17 0.09 0.00 0.00 EMR B na (0.05–0.43) (0.28–0.73) (0.00–0.32) (0.00–0.19) (0.00–0.14) (0.00–0.06) 0.21 0.52 0.18 0.09 0.00 0.00 EMR D na (0.06–0.41) (0.28–0.72) (0.00–0.32) (0.00–0.18) (0.00–0.14) (0.00–0.01) 0.21 0.51 0.18 0.09 0.00 0.00 EUR A na (0.04–0.40) (0.29–0.72) (0.00–0.33) (0.00–0.20) (0.00–0.14) (0.00–0.01) APPENDICES WHO Estimates of the global burden of foodborne diseases 207

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL AIR OTHER (DOMES TIC CONTACT AND WILD) 0.21 0.52 0.18 0.09 0.00 0.00 EUR B na (0.06–0.40) (0.27–0.73) (0.00–0.33) (0.00–0.19) (0.00–0.15) (0.00–0.01) 0.21 0.51 0.18 0.09 0.00 0.00 EUR C na (0.04–0.40) (0.26–0.73) (0.00–0.35) (0.00–0.21) (0.00–0.15) (0.00–0.01) 0.21 0.51 0.18 0.09 0.00 0.00 SEAR B na (0.03–0.44) (0.22–0.73) (0.00–0.35) (0.00–0.19) (0.00–0.13) (0.00–0.01) 0.20 0.52 0.18 0.09 0.00 0.00 SEAR D na (0.06–0.40) (0.29–0.73) (0.00–0.34) (0.00–0.19) (0.00–0.14) (0.00–0.01) 0.20 0.53 0.18 0.09 0.00 0.00 WPR A na (0.01–0.39) (0.30–0.75) (0.00–0.33) (0.00–0.20) (0.00–0.13) (0.00–0.01) 0.21 0.51 0.17 0.09 0.00 0.00 WPR B na (0.05–0.43) (0.29–0.73) (0.00–0.32) (0.00–0.21) (0.00–0.14) (0.00–0.01) Echinococcus multilocularis 0.58 0.02 0.20 0.20 0.00 0.00 AFR D na (0.00–0.87) (0.00–0.42) (0.00–0.61) (0.00–0.63) (0.00–0.03) (0.00–0.00) 0.58 0.02 0.20 0.20 0.00 0.00 AFR E na (0.00–0.87) (0.00–0.41) (0.00–0.62) (0.00–0.61) (0.00–0.03) (0.00–0.00) 0.51 0.03 0.17 0.16 0.00 0.00 AMR A na (0.13–0.79) (0.00–0.50) (0.01–0.40) (0.01–0.38) (0.00–0.11) (0.00–0.03) 0.58 0.02 0.20 0.20 0.00 0.00 AMR B na (0.00–0.87) (0.00–0.38) (0.00–0.62) (0.00–0.61) (0.00–0.03) (0.00–0.00) 0.58 0.02 0.19 0.20 0.00 0.00 AMR D na (0.00–0.88) (0.00–0.41) (0.00–0.61) (0.00–0.60) (0.00–0.03) (0.00–0.00) 0.43 0.14 0.17 0.17 0.00 0.00 EMR B na (0.09–0.73) (0.00–0.55) (0.00–0.42) (0.00–0.42) (0.00–0.06) (0.00–0.01) 0.48 0.12 0.20 0.20 0.00 0.00 EMR D na (0.00–0.77) (0.00–0.49) (0.00–0.54) (0.00–0.53) (0.00–0.04) (0.00–0.00) 0.52 0.03 0.17 0.16 0.00 0.00 EUR A na (0.15–0.79) (0.00–0.48) (0.01–0.40) (0.00–0.39) (0.00–0.11) (0.00–0.03) 0.45 0.13 0.18 0.17 0.00 0.00 EUR B na (0.12–0.72) (0.00–0.52) (0.02–0.38) (0.00–0.37) (0.00–0.13) (0.00–0.03) 0.44 0.14 0.17 0.17 0.00 0.00 EUR C na (0.12–0.72) (0.00–0.53) (0.01–0.38) (0.00–0.37) (0.00–0.12) (0.00–0.03) 0.58 0.02 0.20 0.20 0.00 0.00 SEAR B na (0.00–0.88) (0.00–0.41) (0.00–0.61) (0.00–0.61) (0.00–0.03) (0.00–0.00) 0.58 0.02 0.20 0.20 0.00 0.00 SEAR D na (0.00–0.88) (0.00–0.37) (0.00–0.62) (0.00–0.60) (0.00–0.05) (0.00–0.00) 0.51 0.04 0.16 0.16 0.00 0.00 WPR A na (0.09–0.81) (0.00–0.52) (0.00–0.41) (0.00–0.40) (0.00–0.03) (0.00–0.01) 0.48 0.12 0.20 0.20 0.00 0.00 WPR B na (0.00–0.78) (0.00–0.49) (0.00–0.54) (0.00–0.54) (0.00–0.12) (0.00–0.03) Ascaris spp. 0.38 0.00 0.00 0.19 0.39 0.00 AFR D na (0.10–0.66) (0.00–0.09) (0.00–0.08) (0.07–0.40) (0.07–0.65) (0.00–0.06) 0.38 0.00 0.00 0.19 0.39 0.00 AFR E na (0.07–0.67) (0.00–0.09) (0.00–0.09) (0.07–0.41) (0.05–0.65) (0.00–0.06) 0.83 0.00 0.00 0.05 0.06 0.00 AMR A na (0.43–0.97) (0.00–0.29) (0.00–0.08) (0.00–0.18) (0.00–0.42) (0.00–0.06) 0.55 0.00 0.00 0.19 0.22 0.00 AMR B na (0.17–0.75) (0.00–0.13) (0.00–0.09) (0.06–0.40) (0.05–0.50) (0.00–0.04) Appendices 208

ANIMAL HUMAN- CONTACT SUBREGION FOOD TO-HUMAN WATER SOIL AIR OTHER (DOMES TIC CONTACT AND WILD) 0.37 0.00 0.00 0.18 0.41 0.00 AMR D na (0.07–0.68) (0.00–0.15) (0.00–0.08) (0.05–0.41) (0.04–0.69) (0.00–0.04) 0.55 0.00 0.00 0.20 0.22 0.00 EMR B na (0.15–0.77) (0.00–0.10) (0.00–0.07) (0.02–0.44) (0.02–0.51) (0.00–0.06) 0.55 0.00 0.00 0.20 0.21 0.00 EMR D na (0.18–0.75) (0.00–0.10) (0.00–0.09) (0.04–0.43) (0.04–0.51) (0.00–0.05) 0.85 0.00 0.00 0.05 0.06 0.00 EUR A na (0.47–0.97) (0.00–0.25) (0.00–0.09) (0.00–0.18) (0.00–0.38) (0.00–0.06) 0.55 0.00 0.00 0.19 0.22 0.00 EUR B na (0.13–0.76) (0.00–0.27) (0.00–0.10) (0.03–0.40) (0.02–0.50) (0.00–0.06) 0.55 0.00 0.00 0.19 0.22 0.00 EUR C na (0.14–0.76) (0.00–0.25) (0.00–0.12) (0.03–0.40) (0.04–0.50) (0.00–0.05) 0.54 0.00 0.00 0.20 0.22 0.00 SEAR B na (0.18–0.75) (0.00–0.14) (0.00–0.08) (0.03–0.44) (0.01–0.52) (0.00–0.05) 0.39 0.00 0.00 0.20 0.38 0.00 SEAR D na (0.11–0.68) (0.00–0.12) (0.00–0.07) (0.04–0.44) (0.04–0.65) (0.00–0.06) 0.85 0.00 0.00 0.05 0.06 0.00 WPR A na (0.47–0.97) (0.00–0.23) (0.00–0.09) (0.00–0.19) (0.00–0.37) (0.00–0.06) 0.54 0.00 0.00 0.20 0.21 0.00 WPR B na (0.16–0.77) (0.00–0.24) (0.00–0.11) (0.02–0.43) (0.02–0.49) (0.00–0.06) APPENDICES WHO Estimates of the global burden of foodborne diseases 209 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

(0.00–0.01) (0.00–0.01) (0.00–0.01) (0.00–0.01) (0.00–0.01) (0.00–0.02) (0.00–0.02) (0.00–0.03) (0.00–0.02) (0.00–0.02) (0.00–0.02) (0.00–0.02) (0.00–0.04) (0.00–0.04) 0.03 0.03 0.03 0.02 0.05 0.05 0.02 0.02 0.02 0.02 0.02 0.05 0.05 0.06 0.06 0.04 0.04 0.04 0.04 TOYS OTHER (0.00–0.19) (0.00–0.17) (0.00–0.18) (0.00–0.18) (0.00–0.19) (0.00–0.19) (0.00–0.16) (0.00–0.22) (0.00–0.23) (0.00–0.24) (0.00–0.23) (0.00–0.23) (0.00–0.24) (0.00–0.20) 0.11 0.11 0.11 0.11 0.07 0.07 0.09 0.09 0.08 0.08 0.05 0.05 0.05 0.05 0.06 0.06 0.09 0.09 0.06 0.06 0.09 0.09 0.09 0.09 0.04 0.04 (0.01–0.28) (0.01–0.38) (0.01–0.24) (0.03–0.37) (0.03–0.27) (0.00–0.25) (0.00–0.33) (0.00–0.27) (0.00–0.23) (0.00–0.36) (0.00–0.22) (0.00–0.32) (0.00–0.35) (0.00–0.20) COOKWARE, COOKWARE, GLASSWARE POTTERY OR POTTERY 0.14 0.14 0.08 0.08 0.03 0.05 0.05 0.05 0.05 0.09 0.09 0.06 0.06 0.05 0.05 0.09 0.09 0.08 0.08 0.08 0.08 0.06 0.06 0.04 0.04 0.04 0.04 (0.00–0.38) (0.00–0.38) (0.00–0.38) (0.00–0.35) (0.00–0.36) (0.00–0.35) (0.00–0.39) (0.00–0.32) (0.00–0.24) (0.00–0.33) (0.04–0.48) (0.00–0.36) (0.00–0.36) (0.00–0.30) LEAD 0.12 0.12 0.12 0.12 0.18 0.18 0.16 0.16 0.21 0.21 0.18 0.18 0.27 0.28 0.28 0.24 0.38 0.38 0.29 0.29 0.26 0.26 0.30 0.30 0.20 0.20 (0.10–0.66) (0.00–0.51) (0.00–0.51) (0.00–0.41) (0.03–0.39) (0.00–0.53) (0.00–0.37) (0.05–0.46) (0.00–0.57) (0.00–0.35) (0.00–0.54) (0.00–0.38) (0.00–0.50) (0.00–0.40) 0.11 0.11 0.11 0.11 0.12 0.12 0.13 0.13 0.12 0.12 0.14 0.14 0.10 0.10 0.10 0.10 0.10 0.10 0.07 0.07 0.06 0.06 0.09 0.09 0.04 0.04 (0.00–0.19) (0.00–0.16) (0.00–0.27) (0.00–0.32) (0.00–0.23) (0.00–0.27) (0.00–0.22) (0.00–0.27) (0.00–0.55) (0.00–0.35) (0.00–0.24) (0.00–0.28) (0.00–0.30) (0.00–0.30) 0.17 0.17 0.15 0.15 0.14 0.14 0.16 0.16 0.19 0.19 0.21 0.21 0.14 0.14 0.22 0.22 0.22 0.22 0.22 0.22 0.29 0.29 0.28 0.28 0.09 0.09 0.30 0.30 (0.11–0.54) (0.05–0.31) (0.05–0.61) (0.02–0.38) (0.03–0.23) (0.03–0.36) (0.05–0.47) (0.03–0.42) (0.06–0.45) (0.06–0.42) (0.06–0.54) (0.02–0.40) (0.06–0.48) (0.04–0.46) 0.11 0.11 0.12 0.12 0.12 0.12 0.17 0.17 0.17 0.17 0.21 0.21 0.17 0.17 0.17 0.17 0.19 0.19 0.19 0.19 0.19 0.19 0.23 0.23 0.23 0.23 0.24 (0.01–0.37) (0.01–0.49) (0.00–0.31) (0.00–0.41) (0.00–0.37) (0.00–0.37) (0.00–0.37) (0.00–0.47) (0.00–0.30) (0.00–0.46) (0.00–0.30) (0.00–0.46) (0.00–0.40) (0.00–0.40) Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by exposure to lead through to exposure caused by of illnesses of the proportion interval) (median and 95% uncertainty estimates Subregional REGION FOOD WATER SOIL AIR PAINT WPR B WPR A SEAR D SEAR B EUR C EUR B EUR A EMR D EMR B AMR D AMR B AMR A SUB AFR E AFR D Table A7.5 each pathway. Appendices 210

THIS STUDY Formal expert expert Formal elicitation (10-49) (16-94) 69

AL., 2014 VALLY ET VALLY Formal expert expert Formal elicitation . a

LAKE ET AL., 2010 Formal expert expert Formal elicitation )---

THIS STUDY Formal expert expert Formal elicitation in how they measure central tendency and uncertainty, we cannot label the we and uncertainty, tendency central measure they in how proportion foodborne was per definition 100% and cannot be readily per definition 100% and cannot be readily was foodborne proportion n, 90% CI); Gkogka et al., 2011(median, min-max); Ravel et al., 2010 (mean, 95% et al., 2011(median, min-max); Ravel n, 90% CI); Gkogka

AL., 2011

SCALLAN ET 100 (99-100)c) 36 (12-63) - (10-49) 24 38 (10-72) Derived by the by Derived authors b) ------

AL., 2010 RAVEL ET RAVEL Formal expert expert Formal elicitation HAZARDS

THIS STUDY Formal expert expert Formal elicitation no yes yes no no yes no

AL., 2011

GKOGKA ET GKOGKA Derived by the by Derived authors b) depended on the data used - 50 (10-100) 33 (0-71) ------24 ------24 ------82 (66-98) 100 (99-100) 30 (1-95) - - - NL GR EUR A CA USA AMR A NZ AU WPR A 42 (21-78) 51 (40-90) 60 (26-83) (60-91) 76 82 (75-87) 59 (19-84) 40 (6-95) 55 (30-75) (25-82) 57 2006 1996-2006 2010 2008 2010 2010 2005 2010 2010 56 (26-88) 50 (30-63) (35-82) 61 - 50 (40-60) 60 (30-81) - - -

AL., 2008

HAVELAAR ET HAVELAAR Formal expert expert Formal elicitation yes spp. 55 (32-88) 95 (55-95) (47-94) 76 80 (68-92) 94 (91-96) (38-91) 73 60 (18-83) 71 (65-75) (45-93) 74

E. coli

E. coli spp. 12 (0-20) 5.6 (5.6-8) 10 (0-39) 9 (3-16) 8 (6-12) 16 (1-44

E. coli spp. 42 (16-84) 55 (30-80) (44-93) 76 68 (54-82) 80 (73-86) (38-91) 73 56 (26-82) (70-80) 76 68 (40-89) Salmonella Percent of illness acquired through the foodborne transmission route for six national studies and this study six national studies for route transmission the foodborne through acquired of illness Percent Typhi - 80 (55-95) 10 (0-53) - 100 (76-100) 26 (0-64) - - - spp. - 84 (50-100) 66 (23-90) - 50 (40-60) (28-93) 75 - - - spp. - 10 (8.2-31) 7 (0-46) 18 (7-29) 31 (23-40) 12 (0-46) - 11 (5-20) 13 (0-50) spp. 13 (0-24) 10 (5-30) 11 (0-44) - 7 (5-10) 11 (0-39) - - -

Vibrio cholerae Vibrio

Non-typhoidal Hepatitis A 11 (0-20) 8 (5-11) 42 (2-75) - 6 (4-16) 42 (6-77) - 12 (7-20) 42 (3-76) Norovirus 17 (16-47) - 26 (0-73) 31 (14-48) 26 (19-35) 23 (4-50) 39 (8-64) 17 (5-30) 22 (1-52) Shiga toxin-producing Shiga toxin-producing Enteropathogenic Enteropathogenic

Enterotoxigenic Enterotoxigenic Country/sub region Country/sub Period Method Only domestically acquired cases acquired Only domestically Brucella Campylobacter Cryptosporidium Entamoeba histolytica Giardia Salmonella Shigella gondii Toxoplasma These estimates were derived by a synthesis of data from different public health surveillance systems and the literature. systems public health surveillance different of data from a synthesis by derived were estimates These This table presents a measure of central tendency with its associated uncertainty bound from each study. Because studies differ Because studies each study. bound from uncertainty with its associated tendency of central a measure table presents This Only ETEC cases reported as part of foodborne outbreaks were included in the study by Scallan et al. (2011). Consequently the Scallan et al. (2011). Consequently by included in the study were outbreaks as part of foodborne Only ETEC cases reported Table A7.6 Table a b c compared with the estimate in this study, which considers infections acquired from all transmission routes. all transmission from acquired infections which considers in this study, with the estimate compared columns with a single heading. Measures include: this study (median, 90% credibility interval (CI)); Havelaar et al., 2008 (mea Havelaar (CI)); interval (median, 90% credibility include: this study with a single heading. Measures columns et al., 2014 (median, 95% CI). et al., 2010 (mean, 95% CI); Valley (mean, 90% CI); Lake CI); Scallan et al., 2011 APPENDICES WHO Estimates of the global burden of foodborne diseases 211 45 1 036 1 036 6 963 6 963 9 905 12 953 (7–171) 26 270 26 270 138 863 138 863 296 156 1 722 312 1 722 312 1 237 103 1 237 103 2 141 926 2 141 926 2 183 146 2 938 407 2 938 407 2 496 078 2 496 078 2 084 229 15 780 400 15 780 400 (299–2 805) FOODBORNE (5 951–33 664) (1 423–20 493) (3 993–23 527) (11 462–53 577) DALYs (95% UI) DALYs (47 339–503 775) (119 456–724 660) (1 175 658–5 511 092) (720 029–3 491 997) (554 204–2 520 126) (1 314 295–3 981 424) (1 535 985–3 137 980) (1 190 704–3 494 201) (1 587 757–4 865 590) (11 043 288–22 251 264) 0 0 24 175 175 128 120 1 470 1 470 3 759 3 759 (0–0) (0–0) 15 156 21 374 21 374 26 170 26 170 (7–65) 37 077 37 077 28 693 28 693 34 929 34 929 24 649 24 649 199 892 199 892 (25–351) (55–374) (95% UI) (70–407) (453–5 554) (1 520–9 115) (19 957–61 262) (6 839–30 072) (15 916–79 620) (14 887–43 523) (17 070–49 768) (14 604–32 584) (10 304–50 042) (136 903–286 616) FOODBORNE DEATHS FOODBORNE DEATHS 475 763 451 763 451 256 775 256 775 (95% UI) 1 176 854 1 176 854 3 998 164 (183–990) 8 584 805 5 409 083 28 236 123 28 023 571 28 023 571 51 014 050 95 613 970 95 613 970 23 797 284 23 797 284 78 439 785 86 502 735 86 502 735 548 285 159 548 285 159 124 803 946 124 803 946 (43 875–807 547) (310 910–1 567 682) (754 108–2 523 007) (837 262–11 529 642) (3 897 252–18 531 196) (2 187 762–12 929 293) (51 731 379–177 239 714) (49 136 952–151 776 173) (10 750 919–62 931 604) (31 579 011–210 875 866) (10 261 254–68 567 590) (20 405 214–118 927 631) (70 311 254–251 352 877) (12 945 655–56 996 454) FOODBORNE ILLNESSES FOODBORNE ILLNESSES (369 733 377–888 360 956) 0.15 0.15 0.13 0.13 0.18 0.18 1.00 1.00 1.00 1.00 0.27 0.52 0.28 0.28 0.58 0.58 0.24 0.29 0.29 0.36 0.36 0.48 0.48 0.30 0.30 (95% UI) (0.11–0.30) (0.17–0.48) (0.13–0.44) (0.14–0.44) (0.10–0.46) (0.35–0.67) (0.22–0.36) (0.07–0.24) (0.08–0.27) (0.24–0.50) (0.33–0.60) (0.44–0.69) PROPORTION FOODBORNE 45 1 036 1 036 6 963 6 963 9 905 DALYs DALYs (7–171) 26 827 26 827 171 100 515 904 2 159 331 (95% UI) 5 887 541 5 887 541 9 717 390 3 733 822 3 733 822 15 105 714 7 347 635 7 347 635 5 407 736 5 407 736 4 377 930 4 377 930 55 139 959 (299–2 805) (1 423–20 493) (3 993–23 527) (12 089–72 204) (115 777–257 315) (222 446–1 552 466) (4 190 610–8 407 186) (3 242 020–7 175 522) (3 771 300–8 107 456) (1 392 438–3 686 925) (2 857 037–5 273 652) (5 496 431–8 804 408) (7 602 047–12 387 029) (46 746 114–65 120 623) (11 649 794–19 460 578) 0 0 24 175 175 120 269 269 (0–0) (0–0) 5 450 (7–65) 73 857 73 857 27 553 27 553 65 796 65 796 56 969 56 969 37 604 37 604 715 196 105 170 105 170 122 760 122 760 212 489 (111–814) (25–351) DEATHS DEATHS (95% UI) (70–407) (2 194–17 127) (27 738–55 101) (43 272–88 129) (46 317–97 036) (97 115–154 869) (18 532–44 654) (78 671–126 058) (53 851–103 026) (160 595–278 420) (603 325–846 397) 475 256 775 256 775 2 481 511 (95% UI) 3 183 394 3 183 394 3 998 164 (183–990) 5 409 083 81 082 327 81 082 327 ILLNESSES ILLNESSES 166 175 078 166 175 078 183 842 615 183 842 615 153 097 991 684 850 131 64 003 709 64 003 709 190 849 501 190 849 501 103 943 952 103 943 952 240 886 759 240 886 759 1 912 159 038 1 912 159 038 2 849 323 016) (1 413 002 730– (43 875–807 547) (837 262–11 529 642) (2 211 329–4 146 250) (1 594 572–5 376 503) (2 187 762–12 929 293) (40 716 656–171 419 480) (47 018 659–210 632 459) (97 832 995–363 915 689) (43 049 455–104 679 951) (92 227 873–300 877 905) (64 733 607–382 208 079) (160 890 532–377 471 599) (130 018 020–262 838 002) (490 930 402–1 122 947 359) 1 **

E. coli ***

E. coli spp.

E. coli spp.* Median number of foodborne Illnesses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% uncertainty with 95% uncertainty (DALYs), Years Life Deaths, and Disability Adjusted Illnesses, Median number of foodborne ** Salmonella spp. spp. PATHOGEN Non-typhoidal Enteropathogenic Enteropathogenic Intoxications Norovirus Shiga toxin-producing Shiga toxin-producing Enterotoxigenic Enterotoxigenic Diarrhoeal Disease Campylobacter Cryptosporidium Entamoeba histolytica Giardia enterica, Shigella cholerae Vibrio Bacillus cereus botulinum Clostridium perfringens Clostridium APPENDIX 8. ENTERIC, PARASITIC, HAZARD CLASSES: INDIVIDUAL FOR TABLES DATA CHEMICAL Table A8.1 intervals, 2010. intervals, Appendices 212 1 575 1 575 118 340 124 884 124 884 607 775 607 775 855 730 855 730 1 353 767 1 353 767 1 794 575 1 794 575 9 107 557 9 107 557 3 720 565 3 720 565 (702–3 244) 25 175 035 25 175 035 FOODBORNE DALYs (95% UI) DALYs (49 634–754 680) (458 364–826 115) (43 153–2 910 416) (886 443–3 107 172) (268 879–2 100 120) (383 684–3 672 726) (1 169 040–9 130 956) (4 891 985–17 483 327) (17 547 264–37 021 003) 25 3 175 3 175 1 957 1 957 27 731 27 731 29 391 12 069 12 069 10 545 52 472 52 472 146 981 (10–55) (95% UI) 350 686 (661–45 545) (3 784–29 521) (7 169–77 320) (7 894–14 472) (1 339–20 428) (14 948–50 463) (16 454–128 350) (81 052–274 835) FOODBORNE DEATHS FOODBORNE DEATHS (240 030–524 042) ely listed under diarrhoeal disease agents. ely listed 14 169 14 169 121 268 284 972 284 972 393 239 393 239 1 741 120 1 741 120 (95% UI) 1 073 339 1 073 339 7 570 087 7 570 087 13 709 836 13 709 836 25 569 838 25 569 838 (6 112–91 175) 581 902 722 581 902 722 922 031 380) (400 741 151– (99 852–150 239) (167 455–384 321) (143 815–9 099 394) (658 463–1 639 524) (536 650–4 310 983) (2 333 263–18 743 406) (3 630 847–38 524 946) (10 019 370–58 282 758) FOODBORNE ILLNESSES FOODBORNE ILLNESSES *** 61 EUR and subregion A (low mortality) countries only, and excluding WPR and excluding only, mortality) countries A (low EUR and subregion *** 61 1.00 1.00 1.00 0.37 0.37 0.47 0.34 0.34 0.48 0.48 0.30 0.30 0.29 0.29 (95% UI) (0.17–0.52) (0.19–0.58) (0.19–0.58) (0.14–0.49) (0.30–0.61) (0.28–0.64) (0.23–0.36) PROPORTION FOODBORNE 1 575 1 575 DALYs DALYs 118 340 607 775 607 775 264 073 264 073 (95% UI) 2 367 164 2 367 164 3 895 547 3 895 547 4 580 758 4 580 758 10 292 017 10 292 017 23 070 841 23 070 841 (702–3 244) 78 730 084 78 730 084 (49 634–754 680) (458 364–826 115) (100 540–6 187 148) (875 236–5 066 375) (2 401 034–5 790 874) (1 599 296–10 408 164) (3 805 373–22 027 716) (13 388 154–39 912 033) (64 963 913–97 740 062) 25 3 175 3 175 4 145 63 312 63 312 93 961 93 961 10 545 33 325 (10–55) 371 002 371 002 144 890 DEATHS DEATHS (95% UI) 1 092 584 1 092 584 (1 557–95 894) (7 894–14 472) (1 339–20 428) (12 309–71 278) (38 986–94 193) (29 602–221 677) (218 593–631 271) (53 519–309 903) (892 999–1 374 238) 14 169 14 169 121 268 832 633 832 633 596 824 596 824 (95% UI) 1 073 339 1 073 339 4 826 477 4 826 477 77 929 723 77 929 723 ILLNESSES ILLNESSES 20 984 683 46 864 406 (6 112–91 175) 2 000 626 631 2 000 626 631 (99 852–150 239) 2 942 534 533) (1 494 986 030– (596 824–596 824) (658 463–1 639 524) (337 929–19 560 440) (1 782 796–10 323 273) (7 751 285–44 883 794) (14 417 704–111 771 902) (36 606 712–149 676 316) **

Typhi

. spp PATHOGEN TOTAL Note that non-typhoidal Salmonella enterica was split over diarrhoeal and invasive disease, whereas in Table 7 it was exclusiv 7 it was in Table whereas disease, diarrhoeal and invasive split over was non-typhoidal Salmonella enterica that Note Notes: * = Includes Guillain-Barré Syndrome cases and deaths; ** = 61 EUR and other subregion A (low mortality) countries only; mortality) countries A (low EUR and other subregion ** = 61 cases and deaths; Syndrome * = Includes Guillain-Barré Notes: A countries invasive non-typhoidal invasive Hepatitis A 1 Invasive enteric diseases enteric Invasive Staphylococcus aureus Staphylococcus Brucella monocytogenes Listeria bovis Mycobacterium Salmonella enterica Salmonella enterica A Paratyphi Salmonella enterica APPENDICES WHO Estimates of the global burden of foodborne diseases 213 2 4 31 18 25 32 36 43 30 0.2 0.2 132 0.4 0.4 229 0.5) (2–11) DALYS (8–37) (0.09– (17–51) (0.7–7) (10–51) (23–71) (19–58) (17–80) (95% UI) (22–46) (71–254) (0.2–0.8) (160–323) 2 3 0 0.2 0.2 0.3 0.3 0.5 0.5 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.4 (0– 0.02 0.05 0.05 (1–4) (2–4) 0.08) (0–0) 0.002 (0.2–1) 0.005) DEATHS DEATHS (95% UI) (0.007– (0.1–0.7) (0.1–0.4) (0.2–0.7) (0.2–0.5) (0.2–0.6) (0.3–0.9) (0.02–0.1) SES 5% uncertainty 5% uncertainty 11 17 741 125 410 410 372 346 407 407 NES NES 1 814 1 140 1 257 1 257 (714– 1 390 (752– (297– 7 968 (459– 1 728) 12 911) 2 576) 3 653) (5–23) 2 206) (11–37) 3 065) (1 022– (5 373– (95% UI) (57–269) ILL (156–915) (188–828) (146–847) (149–997) 2 5 4 4 4 10 33 58 0.1 0.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3) 0.05 0.05 (1–7) (0–1) DALYS (4–17) (17–54) (0.1–10) (95% UI) (0.2–12) (0.005– (0.4–17) (0.3–10) (19–134) (0.02–3) (0.005–1) (0.02–0.1) 0 0 0.3 0.3 0.9 0.9 (0– 0.1) (0– 0.1) 0.1) 0.2) 0.02 0.05 0.05 0.06 0.06 0.04 0.04 0.04 0.04 0.04 0.04 0.03) 0.001 (0–0) 0.002 (0.01– 0.003 0.003) 0.005) (0.3–2) DEATHS DEATHS (95% UI) (0.002– (0.003– (0.003– (0.004– (0.1–0.5) (0–0.03) (0–0.001) (0.02–0.1) SES 3 32 0.2 0.2 163 163 166 555 876 229 354 898 689 (19– NES NES (43– (189– 1 305 (170– (2–6) (359– 6 302 6 302 6 441) 3 855) 2 549) 6 428) 2 430) (2 501– 17 289) (2–170) (95% UI) (8–395) (8–1 519) (30–391) ILL (0–1 598) (0.01–0.5) 3 6 71 71 25 33 0.1 0.1 42 66 49 30 0.2 0.2 363 363 250 0.9) DALYS (0.01– (5–83) (9–83) (11–147) (95% UI) (0.3–17) (0.02–1) (15–146) (10–104) (0.9–29) (81–598) (15–230) (0.5–109) (177–649) 1 5 4 0 0.7 0.3 0.3 0.9 0.9 0.6 0.6 0.4 0.4 0.4 0.4 0.2) 0.4) 0.03 0.09 0.09 (1–9) (2–9) (0–0) 0.002 (0.01– (0.1–1) (0.2–2) (0.2–2) (0.2–3) DEATHS DEATHS (95% UI) (0.07–1) (0.004– (0–0.01) (0.1–0.9) (0.007–2) SES 17 19 78 159 841 841 872 256 594 908 NES NES (62– 1 152 1 152 (88– (113– (177– 1 075 1 075 (169– 1 084 7 074 7 074 (229– 3 521) 5 631) (200– 2 775) 3 372) 3 927) 2 288) 4 758) (2–95) 19 537) (2 570– (95% UI) (0.7–52) (10–474) (16–903) ILL (27–1 188) 8 9 4 0 0 10 23 0.2 0.2 0.2 0.2 0.3 0.3 0.1) 0.8) 0.6) 0.03 (1–8) 0.01) 0.01) (0–0) (0–0) DALYS (7–19) 0.005 0.005 (6–13) (5–14) 0.004 0.004 (0.03– (0.03– (0.001– (16–30) (95% UI) (0.002– (0.009– (0.1–0.8) 0 0 0 0 0 0.1 0.1 0.2 0.2 0.3 0.3 (0– 0.2) 0.05 0.05 0.05 0.05 0.09) (0–0) (0–0) (0–0) (0–0) (0–0) 0.003 0.003 0.003 (0.03– (0.08– 0.009) 0.006) DEATHS DEATHS (0.001– (95% UI) (0–0.01) (0.1–0.3) (0.2–0.4) (0.02–0.1) SES 3 8 6 0 21 18 19 54 186 522 522 NES NES 0.03 687) 1 652 1 652 0.06) 2 483 (363– (0–0) (630– (2–13) (0.01– (3–16) (9–51) 4 136) 3 294) (9–28) (4–70) (1 439– (0.9–8) (95% UI) (16–123) (118–275) ILL 2 4 57 35 33 38 54 20 90 0.2 0.2 0.6 0.6 108 354 DALYS (9–76) (0.1–2) (6–117) (11–89) (18–131) (95% UI) (26–87) (0.3–14) (0.7–69) (0.1–0.5) (56–130) (0.4–20) (41–250) (218–544) 1 2 4 0 0.7 0.3 0.3 0.6 0.6 0.4 0.4 0.4 0.4 0.4 0.4 (0– 0.2) 0.2) 0.02 0.04 0.04 (2–6) (0–0) 0.002 (0.1–1) (0.1–1) (0.3–1) (0.6–1) 0.004) (0.2–2) (0.7–4) DEATHS DEATHS (0.01–1) (95% UI) (0.07–1) (0.002– (0.004– SES 9 65 737 627 627 670 394 346 430 NES NES (52– (55– (80– (116– 1 610 1 610 1 873 1 873 (147– (133– 4 971 4 971 2 796 2 796 (744– 1 222) (488– 1 287) 2 193) 1 056) 16 387 16 387 7 376) 5 608) 4 648) (1 685– 34 176) 14 052) (7 729– 10 849) (95% UI) (37–97) (0.4–28) ILL (79–3 110) 1 7 5 5 9 0 13 16 44 0.3 0.3 0.3 0.3 0.3 0.3 0.6 0.6 (1–12) (1–12) (0–0) DALYS (8–18) (4–12) (3–23) (8–35) (0.2–2) (0.3–5) (95% UI) (0.03–1) (0.05–1) (30–63) (0.1–0.9) 0 0 0.1 0.1 0.1 0.1 0.3 0.3 0.5 0.5 (0– 0.2) 0.3) 0.07 0.02 0.05 0.05 0.06 0.06 0.01) 0.02) 0.05) 0.001 (0–0) (0–0) 0.007 0.004 0.004 (0.06– (0.04– 0.009) DEATHS DEATHS (0.001– (95% UI) (0.002– (0.003– (0.3–0.7) (0.2–0.6) (0.01–0.1) (0.01–0.1) (0.04–0.1) SES 31 16 114 212 189 278 (16– 309 NES NES (35– (62– 0.02 1 281 1 389 2 491 2 491 1 002 1 002 0.05) (378– (299– (898– 1 249) 7 900 (490– 6 186) 1 443) 1 209) (11–81) 1 990) 3 295) 3 207) (9–30) 13 850) (95% UI) (4 497– (0.008– (32–355) (35–730) ILL 5 13 81 89 43 70 112 0.8 0.8 109 140 687 687 307 307 0.05 0.05 508) (160– 1 106) DALYS (369– (3–37) (0.2–3) (8–124) (41–112) (95% UI) (24–185) (42–147) (0.9–39) (46–216) (35–252) (50–282) (0.02–0.1) 1 1 1 2 2 5 9 0 0 0.2 0.2 0.5 0.5 0.8 0.8 (0– 0.4) 0.4) AFR AMR EMR EUR SEAR WPR GLOBAL 0.05 0.05 (3–8) (0–0) (5–14) (0.1–2) (0.04– (0.4–1) 0.002) (0.5–2) (0.3–3) DEATHS DEATHS (0.6–3) (0.6–3) (0.5–4) (95% UI) (0.009– SES 5 43 523 796 796 982 425 896 205 454 809 NES NES (98– (45– 2 221 1 749 1 749 (172– (175– (125– (312– (491– (2–9) 1 215) (335– 9 830 2 574) 2 865) 3 868) 8 482) 2 994) 2 480) 5 060) 21 567) (13–101) (95% UI) (3 969– (35–813) ILL (156–976)

E. coli

E. coli spp. Median rates of foodborne illnesses, deaths and Disability Adjusted Life Years (DALYs) per 100 000 persons, by region, with 9 region, per 100 000 persons, by (DALYs) Years Life deaths and Disability Adjusted illnesses, of foodborne Median rates E. coli spp.**

Salmonella spp. spp. PATHOGEN* Shiga toxin-producing Shiga toxin-producing Norovirus Invasive enteric diseases enteric Invasive Enteropathogenic Enteropathogenic Diarrhoeal Disease Non-typhoidal Campylobacter Cryptosporidium Entamoeba histolytica Enterotoxigenic Giardia enterica Shigella cholerae Vibrio Table A8.2 intervals, 2010. intervals, Appendices 214 2 2 9 12 26 54 20 366 538) (255– DALYS (7–12) (4–31) (6–53) (0.7–11) (13–45) (95% UI) (17–133) (0.6–42) 5 0.2 0.2 0.2 0.2 0.8 0.8 0.4 0.4 0.4 0.4 0.3) 0.4) 0.03 0.05 0.05 (3–8) (0.1–1) (0.02– (0.05– (0.2–2) DEATHS DEATHS (95% UI) (0.1–0.2) (0.2–0.7) (0.01–0.7) SES 2 6 4 25 110 0.2 0.2 199 NES NES (1–2) (2–6) 8 369 8 369 (8–63) 13 318) (2–132) (5 723– (95% UI) (0.09–1) (34–272) (53–560) ILL 1 2 5 8 6 33 93 0.6 0.6 (1–3) (1–4) DALYS (2–10) (1–20) (5–87) (0.1–19) (95% UI) (0.3–15) (45–175) 1 0.1 0.1 0.1 0.1 0.1 0.1 0.5 0.5 0.2) 0.3) 0.01 0.3) 0.3) 0.04 0.04 0.04 0.04 0.07) 0.09) (0.02– (0.02– (0.03– (0.05– DEATHS DEATHS (0.6–2) (95% UI) (0.07–1) (0.002– (0.006– SES 1 2 51 18 77 0.2 0.2 0.8 0.8 NES NES 6 491 6 491 (3–47) (0.4–1) 17 528) (0.5–2) (3–164) (95% UI) (2 630– (0.5–61) (0.1–0.4) (12–203) ILL 1 9 13 29 48 0.8 0.8 128 96) 622 1 145) DALYS (306– (2–16) (0–21) (7–83) (6–26) (6–151) (95% UI) (0.006– (29–361) 1 2 9 0.2 0.2 0.2 0.2 0.4 0.4 0.3) 0.01 0.03 (0–1) due to a lack of data for global estimation. a lack of data for due to (0.1–1) (4–17) (0.03– (0.1–3) DEATHS DEATHS (0.4–5) (0–0.6) (95% UI) (0.1–0.5) SES 2 2 2 58 0.1 0.1 250 494 494 NES NES (57– 302) (0–2) 8 068 1 590) (0.02– (0.3–2) (0.9–4) (11–167) (95% UI) (3 294– 20 663) (50–725) ILL

Bacillus cereus 1 1 2 3 32 0.1 0.1 0.5 0.5 0.9 0.9 0.6) (1–2) (2–5) DALYS (0.01– (0.7–1) (0.3–3) (0.4–11) (95% UI) (24–45) (0.05–2) and 0.5 0.5 (0– 0.2) 0.07 0.02 0.02 0.02 0.04 0.04 0.03) 0.06) 0.06) 0.007 0.002 (0.01– (0.03– 0.008) DEATHS DEATHS (95% UI) (0.006– (0.006– (0–0.03) (0.3–0.6) (0.04–0.1) SES 1 4 11 0.2 0.2 0.2 0.2 0.2 0.2 0.8 0.8 NES NES 2 506 (1–33) 4 168) (3–28) (0.1–5) (0.6–1) (1 455– (95% UI) (0.03–1) (0.1–0.3) (0.2–0.3) ILL 1 8 9 9 11 37 23 470 728) DALYS (286– (5–18) (1–28) (3–14) (0–21) (2–74) (3–60) (5–122) (95% UI) 6 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.5 0.5 0.5 0.5 0.3) 0.3) 0.9) 0.4) 0.03 (4–9) (0.01– (0.05– (0.08– (0.06– DEATHS DEATHS (0–0.6) (95% UI) (0.06–2) (0.04–2) SES 1 1 17 73 33 0.1 0.1 237 237 NES NES (0–2) 16 865 (2–55) (0.7–2) (0.8–2) (8 051– 34 712) (95% UI) (9–240) (10–187) (17–772) ILL 1 1 3 5 3 61 0.9 0.9 0.4 0.4 botulinum, Cl. perfringens, S. aureus, Clostridium (1–5) (1–11) DALYS (0.3–3) (0.2–4) (95% UI) (0.3–12) (0.9–16) (40–93) (0.2–0.8) 0.8 0.8 0.2) 0.3) 0.01 0.07 0.07 0.02 0.02 0.06 0.06 0.01) 0.07) 0.05) 0.007 (0.03– (0.5–1) DEATHS DEATHS (95% UI) (0.003– (0.003– (0.005– (0.006– (0.03–0.1) (0.01–0.2) SES 2 3 12 10 0.1 0.1 0.7 0.3 0.3 0.2) NES NES 7 937 7 937 (1–37) (0.1–1) (2–32) (0.05– (3–33) (4 515– (0.4–7) 13 899) (95% UI) (0.4–0.9) ILL 1 1 12 23 53 30 169 169 1 001 (562– DALYS 1 543) (0–21) (3–36) (7–60) (19–42) (95% UI) (12–155) (0.1–34) (71–306) 3 14 0.7 0.2 0.2 0.5 0.5 0.5 0.5 0.5) 0.5) AFR AMR EMR EUR SEAR WPR GLOBAL 0.02 0.03 (1–5) (0.1–1) (8–21) (0.04– (0.2–2) DEATHS DEATHS (0–0.6) (95% UI) (0.002– (0.3–0.7) SES 7 3 25 25 0.1 0.1 108 232 NES NES (0–2) (4–9) (5–73) 10 304 (12–37) 22 108) (4 279– (95% UI) (24–317) (0.4–110) ILL (60–643) Typhi enterica spp. PATHOGEN* Invasive non-typhoidal Invasive Hepatitis A Notes: * Table does not include four foodborne intoxications caused by caused by intoxications foodborne does not include four Table Notes: * cases and deaths Syndrome ** Includes Guillain-Barré . TOTAL Brucella monocytogenes Listeria bovis Mycobacterium Salmonella Salmonella enterica A Paratyphi Salmonella enterica APPENDICES WHO Estimates of the global burden of foodborne diseases 215

5 * .28) 1.16 1.16 1.19 1.19 1.19 1.19 1.74 1.74 2.21 1.87 1.87 2.14 2.14 1.83 0.52 0.32 0.32 2.06 2.06 0.24 0.02 4.04 4.04 (95% UI) (1.09–1.29) (1.52–3.29) (1.26–2.92) (0.86–1.60) (1.57–10 (0.64–2.57) (0.95–2.93) (0.65–3.93) (0.33–0.78) (0.23–0.35) (0.87–4.43) (0.24–0.24) (0.38–9.49) (0.02–0.02) RATIO <5: RATIO

5 5 * * 1.21 1.91 1.75 1.75 1.52 1.52 1.23 0.18 0.18 0.19 0.19 1.49 1.49 N/A 0.01 0.35 0.86 0.86 0.96 0.96 0.84 0.84 (95% UI) (0.81–1.31) (0.18–8.71) (1.21–3.08) (0.18–0.18) (0.60–1.16) (0.01–0.01) (1.03–2.29) (0.14–0.22) (0.63–2.10) (0.44–1.83) (0.42–2.72) (0.22–0.54) (0.60–3.26) RATIO <5: RATIO

5 * SES DEATHS DALYS 1.11 1.11 2.11 2.11 2.61 2.61 0.18 0.18 0.21 0.21 0.41 0.01 0.82 0.82 3.20 0.38 0.38 0.26 0.26 0.45 0.45 0.66 0.66 0.48 0.48 NES NES (95% UI) (0.13–1.70) (0.18–0.18) (0.06–1.16) (0.41–0.41) (0.32–1.28) (0.01–0.01) (0.19–0.73) (0.15–0.23) (0.35–1.96) (0.69–8.01) (0.84–5.22) (0.34–3.47) (0.85–11.78) (0.08–2.38) RATIO <5: RATIO 5 016 5 989 911 012 911 012 43 984 122 904 767 975 767 975 104 794 104 794 750 578 750 578 404 144 (95% UI) NUMBER 1 016 047 1 016 047 1 638 925 1 638 925 1 390 973 1 390 973 7 205 002 7 205 002 6 900 776 6 900 776 (3 799 471– 10 747 526) 13 355 093) (4 790 026– (1 945–13 791) (2 654–15 877) (20 149–85 551) (40 408–256 055) (188 009–749 866) (457 215–1 575 768) (540 003–956 663) (419 834–1 204 273) (42 484–2 865 643) (581 491–2 820 499) (730 924–3 844 771) (576 408–1 405 079) 0 65 5 YEARS OF AGE <5: RATIO 524 524 * 1 673 1 673 1 936 1 936 (0–0) 7 436 11 933 11 933 6 060 15 807 15 807 14 647 14 647 25 807 25 807 20 952 20 952 123 026 123 026 107 500 107 500 (95% UI) NUMBER (24–204) (218–1 110) (638–4 149) (2 734–11 511) (654–45 081) (4 930–9 974) (11 201–61 642) (6 382–18 887) (8 762–21 942) (8 758–42 535) (7 305–25 447) (69 907–163 979) (69 306–230 318) SES DEATHS DALYS ILL NES NES 389 106 648 933 648 933 836 948 (95% UI) NUMBER 5 458 601 2 253 036 2 253 036 21 182 632 21 182 632 8 693 968 8 693 968 46 811 878 46 811 878 17 828 477 17 828 477 34 049 173 34 049 173 42 883 268 89 056 582 60 293 254 60 293 254 112 061 441) (3 337 657– (5 378 578– 95 312 884) 24 195 602) 49 059 198) (8 375 340– 50 963 825) (10 186 959– 327 209 075 327 209 075 (18 488 275– (18 350 672– 103 801 449) 206 532 318) 643 705 133) 189 066 838) (46 054 795– (20 306 649– (179 670 939– (264 273–1 332 530) (142 279–9 006 169) (536 302–1 794 298) (774 628–8 639 265) (2 145 370–16 561 005) 1 988 6 969 6 969 92 213 92 213 20 677 20 677 331 395 185 057 185 057 819 280 844 376 844 376 (95% UI) 1 149 675 1 149 675 NUMBER 2 180 916 8 547 149 8 547 149 1 383 499 1 383 499 1 303 490 1 303 490 2 004 543 (687–44 999) (3 278–17 751) (8 552–44 101) (64 847–518 497) (15 997–444 002) (138 538–672 643) (911 878–2 279 897) (406 822–1 776 252) (609 216–2 792 992) (668 837–2 446 758) (309 576–1 909 450) (1 085 765–4 219 254) (1 084 856–3 389 584) (5 903 945–12 254 175) 0 21 63 896 1 989 3 697 3 697 8 863 8 863 8 992 8 992 (0–0) 12 531 13 861 13 861 91 621 22 156 23 727 23 727 14 056 (7–463) (30–170) (95% UI) NUMBER (90–4 852) (678–5 683) (1 546–7 506) (4 251–19 347) (8 754–23 670) (6 562–30 779) (3 250–20 925) (11 944–37 473) (7 045–26 784) (11 866–45 950) (62 442–132 707) ses, Deaths, and Disability Adjusted Life Years (DALYs) by age group, with 95% uncertainty with 95% uncertainty age group, by (DALYs) Years Life ses, Deaths, and Disability Adjusted AGE GROUP: <5 YEARS OF AGEAGE GROUP: AGE nes nes SES DEATHS DALYS ILL NES NES 4 144 114 518 339 905 (95% UI) NUMBER 5 986 213 4 336 215 15 516 627 15 516 627 17 312 780 17 312 780 8 480 759 8 480 759 15 274 234 15 274 234 18 773 028 18 773 028 47 988 357 47 988 357 (1 593 697– ILL 38 352 806 38 352 806 12 738 924) (6 514 539– 34 582 700 34 582 700 216 839 210 41 696 874) (2 569 532– (6 767 766– 64 795 160) (8 075 497– 54 104 398) (21 144 875– 59 592 939) 38 649 748) 30 849 576) (19 595 826– (22 436 891– 102 663 926) 309 926 253) (148 937 428– (1 527–93 225) (46 636–235 152) (217 805–728 708) (1 675 945–9 422 681) (5 416 319–38 620 351)

E. coli , non-

E. coli spp. Median number of Ill E. coli spp.**

. spp spp. spp. PATHOGEN* Shiga toxin-producing Shiga toxin-producing Enteropathogenic Enteropathogenic Invasive enteric diseases enteric Invasive Norovirus typhoidal Enterotoxigenic Enterotoxigenic Diarrhoeal Disease Campylobacter Cryptosporidium Entamoeba histolytica Giardia Salmonella enterica Shigella cholerae Vibrio Brucella intervals, 2010. intervals, Table A8.3 Appendices 216

5 * 0.31 0.31 0.01 0.36 0.36 0.36 0.36 0.34 0.34 0.44 0.44 0.76 (95% UI) (0.01–0.01) (0.29–0.32) (0.36–0.36) (0.36–0.36) (0.24–0.49) (0.39–0.46) (0.56–1.01) RATIO <5: RATIO

5 5 * * 0.11 0.11 0.19 0.19 0.19 0.19 0.01 0.26 0.26 0.26 0.26 0.50 0.50 (95% UI) (0.19–0.19) (0.01–0.01) (0.17–0.20) (0.08–0.16) (0.26–0.26) (0.26–0.26) RATIO <5: RATIO (0.36–0.68)

5 * SES DEATHS DALYS 0.19 0.19 0.19 0.19 0.10 0.10 0.01 0.26 0.26 0.26 0.26 NES NES 0.63 (95% UI) (0.19–0.19) (0.01–0.01) (0.07–0.13) (0.17–0.20) (0.26–0.26) (0.26–0.26) (0.32–1.18) RATIO <5: RATIO due to a lack of data for global estimation. ** Includes Guillain-Barré ** Includes Guillain-Barré global estimation. a lack of data for due to 87 569 87 569 941 278 941 278 627 953 627 953 600 639 600 639 (95% UI) NUMBER 1 373 635 1 373 635 2 730 232 2 730 232 14 250 088 (9 419 295– 22 483 691) (36 830–561 221) (452 917–816 737) (197 302–1 541 909) (684 718–2 373 326) (857 835–6 703 954) (269 448–2 538 627)

Bacillus cereus 5 YEARS OF AGE <5: RATIO * 2 851 and 9 588 23 351 41 689 41 689 10 470 10 470 24 692 232 916 (95% UI) NUMBER (1 200–18 271) (7 836–14 372) (6 036–65 109) (3 007–23 454) (12 655–42 246) (13 072–101 973) (152 283–368 498) SES DEATHS DALYS ILL NES NES 12 936 12 936 120 398 239 467 239 467 (95% UI) NUMBER 6 014 372 6 014 372 1 383 306 11 544 593 11 544 593 (3 057 415– 32 440 565) 350 711 509 (5 716–80 766) 673 777 073) (199 599 319– (99 119–149 188) (142 115–321 539) (426 364–3 425 041) (1 853 758–14 891 483) 7 134 411 592 411 592 30 750 30 750 421 523 421 523 989 159 227 507 227 507 (95% UI) NUMBER 10 831 919 15 271 603) (7 587 557– (5 477–9 496) (11 700–198 862) (71 530–557 578) (112 767–1 130 290) (203 340–733 940) (311 001–2 424 250)

Clostridium botulinum, Cl. perfringens, S. aureus botulinum, Cl. perfringens, S. aureus Clostridium 76 330 2 480 4 380 4 700 4 700 10 783 10 783 116 613 116 613 (58–101) (95% UI) NUMBER (126–2 138) (778–6 067) (1 132–12 211) (2 268–8 188) (3 381–26 377) (80 862–165 379) AGE GROUP: <5 YEARS OF AGEAGE GROUP: AGE SES DEATHS DALYS ILL 869 869 NES NES 1 240 1 240 45 549 45 549 357 814 357 814 1 555 715 (95% UI) NUMBER 2 165 243 2 165 243 ILL (732–1 049) (393–10 502) 221 451 463 221 451 463 315 075 166) (25 019–62 638) (153 244 508– (110 286–885 942) (479 504–3 851 923) (573 433–6 084 381) invasive non- invasive Paratyphi A Paratyphi Typhi PATHOGEN* Notes: * Table does not include four foodborne intoxications due to due to intoxications foodborne does not include four * Table Notes: Syndrome cases and Deaths Syndrome TOTAL typhoidal Hepatitis A monocytogenes Listeria bovis Mycobacterium Salmonella enterica, Salmonella enterica Salmonella enterica APPENDICES WHO Estimates of the global burden of foodborne diseases 217 7 2 5 8 8 9 9 4 4 12 41 29 46 0.6 0.6 0.4 0.01) (2–6) (2–11) (6–9) DALYS (5–13) (3–15) 0.008 0.008 (8–18) (6–19) (6–19) (0.1–9) (0.7–7) (0.2–5) (31–52) (95% UI) (24–36) (35–60) (0.004– (0.2–0.8) 0 0 0 0.1 0.1 0.1 0.1 0.5 0.5 0.4 0.4 0.2) 0.01 0.01 0.01 0.01 0.02 0.05 0.05 0.08 0.08 0.08 0.08 0.02) 0.02) 0.08) 0.06) 0.04) 0.04) (0–0) (0–0) (0–0) 0.007 DEATHS DEATHS (95% UI) (0.007– (0.002– (0.005– (0.005– (0.006– (0.006– (0.004– (0.3–0.5) (0.4–0.7) (0.07–0.1) (0.02–0.1) (0.09–0.1) (0.04–0.2) SES 3 5 6 0.1 0.1 0.7 0.5 0.5 125 178 178 179 179 0.6 0.6 149 149 149 149 410 410 976 407 407 NES NES 0.06 0.06 (4–7) (2–4) (5–11) 0.09) 1 752) (520– inty intervals, 2010. inty intervals, (0.04– (0.5–1) (0.4–5) (95% UI) (57–269) (0.3–0.7) (121–334) (107–216) ILL (108–217) (188–828) (0.01–0.2) (149–997) (120–334) 1 2 8 6 41 16 10 10 97 29 24 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.8) (1–4) (0–1) DALYS (4–13) 0.004 0.004 (3–10) (0.08– (3–20) (3–20) 0.007) (0–34) (18–32) (0.2–4) (0.001– (95% UI) (24–35) (23–60) (0.02–3) (78–120) (0.005–1) 0 0 0 0.2 0.2 0.3 0.3 0.6 0.6 0.4 0.4 0.4 0.4 0.01 0.01 0.01) 0.01) 0.001 0.06) 0.06) (0–0) (0–0) (0–0) 0.003 0.005 0.005 0.005 0.005 0.005 0.005 0.004 0.004 (0.2–1) DEATHS DEATHS (0.001– (0.001– (0–0.8) (95% UI) (0.002– (0.002– (0–0.01) (0.1–0.3) (0–0.03) (0–0.03) (0.3–0.4) (0.3–0.4) (0–0.003) SES 2 4 4 11 32 117 116 0.3 0.3 0.3 0.3 213 213 0.4 0.4 737 229 354 0.01 NES NES (1–2) (124– (3–5) (3–6) 0.02) (8–14) 2 566) (2–170) (0–0.8) (95% UI) (0.004– (65–177) (65–176) (8–1 519) (57–358) (57–358) (0.2–0.7) ILL (0–1 598) (0.09–0.7) 2 3 8 8 6 0 11 12 12 10 37 38 0.1 0.1 0.6 0.6 0.04 0.04 DALYS (3–15) 0.006 0.006 (4–19) (6–10) (3–36) (4–23) (4–23) (0.2–2) (0.8–5) (95% UI) (0.3–17) (28–50) (28–50) (0.9–29) (0–0.03) (0.01–0.1) (0–0.001) (0.01–0.9) 0 0 0 0 0 0.1 0.1 0.4 0.4 0.4 0.4 0.2) 0.01 0.01 0.4) 0.03 0.06 0.06 0.09 0.09 0.01) 0.01) 0.03) 0.08) 0.08) 0.08) (0–0) (0–0) (0–0) (0–0) 0.008 0.008 0.006 0.006 0.006 0.006 (0.05– (0.03– DEATHS DEATHS (0.001– (0.001– (95% UI) (0.002– (0.002– (0.002– (0.004– (0.3–0.5) (0.3–0.5) (0–0.001) (0.01–0.4) SES 8 9 0 78 0.7 137 137 137 137 0.8 0.8 159 0.4 0.4 255 255 256 584 NES NES 0.01) (133– (6–11) (0–0) (7–13) 0.002 0.004 0.004 1 946) (0.5–1) (0.2–2) (0.001– (95% UI) (0.1–0.9) (10–474) (88–461) (88–461) (16–903) (56–245) (55–244) ILL (27–1 188) (0–0.004) 1 2 2 8 6 0 0.2 0.2 0.2 0.2 0.2 0.2 0.8 0.8 0.6 0.6 0.6 0.6 0.4 0.4 0.1) 0.7) 0.01 0.6) 0.03 0.04 0.04 (1–3) 0.01) 0.07) (1–10) (0–0) DALYS (5–14) (4–10) (0.02– (0.03– (0.04– (0.3–1) (0.3–1) (0.3–2) (0.4–8) (95% UI) (0.008– (0.009– (0.1–0.4) (0.3–0.8) 0 0 0 0 0 0 0.03 0.04 0.04 0.01) 0.01) 0.03) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) 0.002 0.002 0.002 0.003 0.003 0.005 0.005 0.005 0.005 0.009 0.009 0.003) DEATHS DEATHS (0.001– (95% UI) (0.002– (0.002– (0.003– (0.01–0.2) (0–0.009) (0–0.006) (0–0.006) (0–0.009) (0.02–0.2) SES 1 8 8 0 0 0 21 77 54 119 119 0.3 0.3 0.8 0.8 0.4 0.4 0.4) NES NES 0.07 0.06 0.06 (4–11) (0–0) (0–0) (0–0) (4–12) (0.03– (4–70) (0.5–2) (0.4–2) (21–181) (95% UI) (16–123) (79–188) (80–189) (0.2–0.7) (0.3–0.5) ILL (0.04–0.1) 5 2 7 7 4 0 0 11 0 19 10 10 0.7 0.8 0.8 0.6 0.6 0.01 (3–7) 0.05) (0–0) (0–0) (0–0) DALYS (6–18) (4–15) (4–15) (4–13) (2–28) (0.1–2) (11–30) (95% UI) (0.2–17) (0.2–17) (0.3–14) (0.004– (0.4–20) 0 0 0 0 0 0 0 0.2) 0.2) 0.2) 0.2) 0.07 0.02 0.02 0.02 0.02 0.02 0.04 0.04 0.06) 0.06) 0.04) 0.04) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) 0.009 0.009 0.009 0.009 DEATHS DEATHS (95% UI) (0.002– (0.002– (0.002– (0.009– (0.009– (0.004– (0.004– (0.004– (0–0.001) (0.01–0.3) SES 1 0 0 0 0.7 0.7 0.6 0.6 195 196 737 670 346 200 200 NES NES (133– 1 989 5 118) (0–0) (0–0) 0.002 (560– 2 193) (0.7–3) (95% UI) (0.3–18) (0.3–18) (72–282) (72–282) (118–292) (0.4–0.9) ILL (119–295) (79–3 110) (52–1 287) (0–0.002) (0–0.003) 1 7 7 5 9 0 0 15 19 10 20 0.3 0.3 0.3 0.3 0.3 0.3 0.6 0.6 0.01) (3–9) (0–0) (0–0) DALYS (7–13) (5–17) 0.009 0.009 (3–57) (3–57) (0.1–3) (9–26) (0.2–2) (0.5–3) (15–25) (15–26) (95% UI) (0.005– (0.03–1) (0.05–1) 0 0 0 0 0 0 0.1 0.1 0.1 0.1 0.01 0.01 0.02) 0.02) 0.02) 0.03) 0.001 0.04) 0.04) (0–0) (0–0) (0–0) (0–0) (0–0) 0.007 0.009 0.009 0.009 0.009 0.009 0.009 0.004 0.004 DEATHS DEATHS (95% UI) (0.002– (0.003– (0.003– (0.004– (0.004– (0.004– (0–0.04) (0–0.001) (0.08–0.1) (0–0.009) (0.08–0.2) SES 1 1 3 4 0 0 114 0.3 0.3 212 159 130 130 160 733 733 309 NES NES 0.06 0.06 (3–7) 0.07) (3–4) (222– (0–0) (0–0) 2 077) (0.1–3) (0.04– (0.7–2) (0.8–2) (95% UI) (92–261) (32–355) (92–263) (60–793) (60–793) ILL (16–1 209) (62–1 249) 5 8 9 9 0 0 12 21 21 13 177 175 175 0.8 0.8 0.6 0.6 0.04 0.04 0.001 (0–0) (0–0) DALYS (4–17) (4–17) (4–15) (3–37) (6–22) (6–62) (11–36) (0.2–3) (95% UI) (0.2–18) (0.9–39) (131–247) (129–241) (0.01–0.1) (0–0.002) 2 2 0 0 0 0 0 0 0.2 0.2 0.2 0.2 0.2) 0.4) 0.4) AFR AMR EMR EUR SEAR WPR GLOBAL 0.02 0.02 0.03 0.03 0.05 0.05 (1–3) (1–3) 0.07) 0.07) 0.05) 0.05) (0–0) (0–0) (0–0) (0–0) (0–0) (0–0) 0.007 (0.01– (0.01– (0.04– DEATHS DEATHS (0.001– (95% UI) (0.005– (0.005– (0.009– (0.009– (0.07–0.7) SES 2 0 0 0 15 14 0.7 170 170 170 170 229 796 796 205 230 809 NES NES (98– (172– 1 995 0.02) (0–0) (0–0) (549– 0.006 0.006 (11–19) 2 574) 3 868) 5 866) (11–37) (0.8–4) (95% UI) (0.002– (0.2–23) (35–813) (68–288) (68–288) ILL (133–387) (0–0.001) (132–386) Median rate per 100 000 of foodborne Illnesses, Deaths and Disability Adjusted Life Years (DALYs) by region, with 95% uncerta with 95% region, by (DALYs) Years Life Deaths and Disability Adjusted Illnesses, per 100 000 of foodborne Median rate spp.# spp. spp.# spp. sporidium

acquired coccus coccus , congenital PATHOGEN Trematodes Cestodes Invasive infectious infectious Invasive disease Nematodes Enteric protozoa# Enteric Crypto spp.# Entamoeba Giardia Toxoplasma gondii Toxoplasma gondii, Echino granulosus Echino multilocularis solium Taenia Ascaris Trichinella Clonorchis sinensis Table A8.4 Appendices 218 1 3 2 15 96 104 132) (2–3) (2–3) (88– DALYS (11–21) (0.8–3) (95% UI) (82–122) 0 0 Taenia 0.7 0.8 0.8 0.02 0.03) (0–0) (0–0) 0.004 0.004 (0.02– 0.005) DEATHS DEATHS (95% UI) (0.6–1) (0.002– (0.5–0.9) SES 2 0.2 0.2 0.2 0.2 0.3 0.3 337 337 NES NES (1–3) 1 325 (851– 2 237) (95% UI) (0.1–0.4) (0.2–0.3) (0.2–0.4) ILL (265–553) 3 8 55 0.8 0.8 156 158 195) (2–3) (6–11) DALYS (128– (0.1–7) (95% UI) (39–76) (127–193) 1 1 0 0 0.01 0.02 0.02) 0.03) (0–0) (0–0) (0.02– (0.5–1) DEATHS DEATHS (95% UI) (0.6–1) (0.008– SES 1 7 0.2 0.2 346 NES NES 0.09 0.09 (5–10) 1 089 (0.8–1) (429– (95% UI) 3 088) (0.2–0.3) (188–512) ILL (0.01–0.8) 8 0.1 0.1 69 80 0.4) 114) 0.05 0.05 (61– 0.06 0.06 DALYS (6–10) (0.05– (95% UI) (54–89) (0.02–0.1) (0.02–0.3) 0 0 0 0.5 0.5 0.6 0.6 0.06 0.06 0.07) (0–0) (0–0) (0–0) cysticercosis. The YLD component of the DALY for cysticercosis is cysticercosis for of the DALY YLD component The cysticercosis. (0.05– DEATHS DEATHS (95% UI) (0.5–1) ses are defined as the numbers of new cases in 2010. For For cases in 2010. defined as the numbers of new ses are (0.4–0.6) nes nes SES Neodiplostomidae and Plagiorchiidae (depending on data availability). on data availability). (depending and Plagiorchiidae Neodiplostomidae 0.6 0.6 404 NES NES 0.02 649) 0.02) 0.05) 0.04) (220– 0.008 0.008 0.006 0.006 (461– 1 007 1 007 2 491) [5].Ill (95% UI) (0.002– (0.002– (0.006– (0.4–0.9) ILL 11 12 0.3 0.3 0.05 0.05 0.06 0.06 0.03) DALYS 0.009 0.009 (8–24) (95% UI) (8–24) (0.003– (0.2–0.5) (0.02–0.1) (0.02–0.2) 0 0 0 0 0.05 0.05 0.2) 0.05 0.05 (0–0) (0–0) (0–0) (0–0) DEATHS DEATHS (0.03– (95% UI) (0.03–0.2) SES 128 210 NES NES 0.05 0.05 0.01) 328) 0.02) 0.08) 0.001 0.006 0.006 0.006 0.006 (136– (0.04– (95% UI) (0.002– (0.003– (89–199) ILL (0–0.004) 5 0 35 44 76) 0.07 0.02 (3–7) (30– 0.05) (0–0) DALYS (95% UI) (25–58) (0.006– (0.03–0.2) 0 0 0 0 0.1 0.1 0.05 0.05 0.4) (0–0) (0–0) (0–0) (0–0) DEATHS DEATHS (0.05– (95% UI) (0.03–0.3) SES 0 0.6 0.6 398 NES NES 0.03) (0–0) 0.002 0.009 0.009 (933– 2 390 5 535) (95% UI) (0.003– (0.4–0.8) ILL (253–535) (0–0.006) 5 4 0 51 53 113) 0.08 0.08 (2–7) (3–7) (42– (0–0) DALYS (41–112) (95% UI) (0.02–0.3) 0 0 0 0 0.1 0.1 0.1 0.1 0.2) (0.1– (0–0) (0–0) (0–0) DEATHS DEATHS (95% UI) (0.1–0.2) (0–0.001) SES 0 0.5 0.5 0.6 0.6 293 293 NES NES (195– 0.03) (0–0) 0.009 0.009 1 035) 1 060 (507– (95% UI) 2 790) (0.003– (0.3–0.9) (0.4–0.8) ILL 0 208 232 0.02 0.02 317) 0.02) 0.05) 0.06) (0–0) DALYS (176– 0.006 0.006 (95% UI) (0.007– (0.002– (0.005– (159–283) 2 2 0 0 0 0 AFR AMR EMR EUR SEAR WPR GLOBAL (1–3) (0–0) (0–0) (0–0) (0–0) (2–3) DEATHS DEATHS (95% UI) SES 0 0 418 418 NES NES 644) (277– (0–0) 0.002 0.003 2 428 (934– (95% UI) 6 426) ILL (0–0.007) (0–0.008) (0–0.003) spp.

. spp. cluding spp this is estimated from GBD2010 [9] regional incidence data and modified as the actual number of cases of epilepsy attributed to attributed data and modified as the actual number of cases epilepsy incidence GBD2010 [9] regional from this is estimated TOTAL PATHOGEN TOTAL (ex (ex TOTAL protozoa) enteric Intestinal flukes * flukes Intestinal Notes: * Includes selected species of the families Echinostomatidae, Fasciolidae, Gymnophallidae, Heterophyidae, Nanophyetidae, Heterophyidae, Gymnophallidae, Fasciolidae, Echinostomatidae, species of the families * Includes selected Notes: in detail elsewhere reported diseases, but are parasitic foodborne for the picture complete included to are protozoa # Enteric Fasciola Fasciola Opisthorchis Paragonimus solium GBD2010 data. from estimated prevalence-based, APPENDICES WHO Estimates of the global burden of foodborne diseases 219 550 90 041 90 041 26 270 26 270 155 165 39 950 312 461 312 461 259 618 259 618 138 863 138 863 565 816 296 156 829 071 829 071 522 863 522 863 605 278 605 278 492 354 492 354 605 738 605 738 (95% UI) 3 158 826 2 788 426 2 788 426 2 024 592 2 024 592 (285–934) (11 462–53 577) (9 083–640 716) (118 920–198 147) (16 996–322 953) (411 113–1 301 619) (47 339–503 775) (168 510–422 935) (119 456–724 660) (58 050–209 097) (431 520–635 232) (410 668–1 301 114) (354 029–891 032) (561 297–1 264 567) (239 400–1 034 790) FOODBORNE DALYS FOODBORNE DALYS (2 411 585–4 122 032) (2 137 613–3 606 582) (1 652 243–2 483 514) 4 0 0 0 0 482 684 684 1 012 7 771 1 470 1 470 (2–5) 1 008 7 533 3 759 3 759 5 558 5 770 5 770 (0–0) (0–0) (0–0) (0–0) 28 114 36 500 (95% UI) (384–2 781) (150–3 974) (333–1 300) (333–1 300) (388–2 783) (453–5 554) (1 520–9 115) (243–15 896) (2 593–11 958) (4 728–6 988) (6 383–8 845) (21 059–36 915) (25 652–50 063) FOODBORNE DEATHS FOODBORNE DEATHS SES 8 375 8 375 4 472 4 472 18 924 18 924 10 635 10 635 31 620 48 823 43 076 43 076 218 569 218 569 370 710 370 710 430 864 (95% UI) 10 228 111 8 584 805 28 236 123 12 285 286 28 023 571 28 023 571 12 280 767 12 280 767 67 182 645 67 182 645 10 280 089 10 280 089 ILLNES ILLNES 56 996 454) (12 945 655– 120 556 797) (35 794 977– (656–17 005) (2 977–5 997) FOODBORNE (6 888–24 100) (31 893–93 213) (21 515–45 059) (25 881–371 177) (14 498–24 200) (167 886–281 872) (282 937–478 123) (334 389–774 703) (7 351 013–14 844 411) (3 897 252–18 531 196) (8 287 414–22 980 491) (7 403 516–14 904 324) (8 292 732–22 984 630) (10 261 254–68 567 590) 0.17 0.17 0.15 0.15 0.21 0.21 0.14 0.14 0.28 0.28 0.85 0.85 0.48 0.48 0.49 0.49 0.49 0.46 0.46 0.46 0.46 (95% UI) (0.13–0.47) (0.15–0.29) (0.31–0.59) (0.31–0.59) (0.01–0.76) (0.07–0.27) (0.65–0.93) (0.39–0.59) (0.06–0.28) (0.09–0.29) (0.40–0.58) (0.40–0.59) PROPORTION FOODBORNE – DALYS FOODBORNE – DALYS 0.13 0.13 0.15 0.15 0.19 0.19 0.21 0.21 SES 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.72 0.28 0.28 0.47 0.45 0.45 0.45 0.45 0.49 0.49 0.49 (95% UI) (0.12–0.28) (0.31–0.59) (0.31–0.59) (0.15–0.30) (0.14–0.44) (0.35–0.79) (0.07–0.24) (0.08–0.27) (0.04–0.75) (0.40–0.58) (0.40–0.59) (0.40–0.59) PROPORTION FOODBORNE – ILLNES- 550 DALYS 90 041 90 041 171 100 155 165 183 573 183 573 526 515 526 515 515 904 522 863 522 863 687 823 687 823 1 318 104 3 721 581 3 721 581 1 317 535 1 153 779 1 153 779 2 159 331 (95% UI) 1 684 414 1 684 414 2 788 426 2 788 426 2 024 592 2 024 592 (285–934) 2 940 604 ses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% uncertainty intervals, 2010 intervals, with 95% uncertainty (DALYs), Years Life ses, Deaths, and Disability Adjusted (115 777–257 315) (118 920–198 147) (58 050–209 097) (431 520–635 232) (359 756–835 537) (772 676–1 733 114) (88 082–1 590 846) (409 190–1 106 320) (222 446–1 552 466) (1 183 025–2 701 170) (1 182 187–2 700 572) (2 137 613–3 606 582) (1 652 243–2 483 514) (2 942 173–5 416 945) (1 392 438–3 686 925) (1 978 442–4 686 855) (1 236 005–2 452 060) nes nes 4 0 0 0 0 17 118 (2–5) 2 227 2 227 1 409 1 409 2 225 7 533 2 224 2 224 5 770 5 770 (0–0) (0–0) (0–0) (0–0) 5 450 28 114 27 553 27 553 33 925 33 925 48 269 48 269 DEATHS DEATHS (95% UI) (701–2 620) (701–2 620) (862–5 942) (865–5 946) (749–19 627) (2 194–17 127) (4 728–6 988) (6 383–8 845) (21 059–36 915) (10 184–27 346) (22 933–53 614) (18 532–44 654) (36 956–70 978) SES NES NES 4 472 4 472 18 451 18 924 18 924 10 635 10 635 31 620 98 900 218 569 218 569 188 079 188 079 370 710 370 710 596 838 (95% UI) 20 817 916 20 710 906 ILL 183 842 615 183 842 615 26 845 649 26 845 649 26 840 692 26 840 692 64 003 709 64 003 709 103 943 952 103 943 952 104 679 951) 356 688 438 517 872 930) (47 018 659– 210 632 459) (43 049 455– 262 838 002) (251 929 267– (2 977–5 997) (130 018 020– (11 384–29 619) (6 888–24 100) (21 515–45 059) (14 498–24 200) (67 858–188 748) (167 886–281 872) (282 937–478 123) (156 848–1 770 405) (482 828–2 169 206) (16 337 908–29 091 701) (25 371 434–47 937 154) (25 375 461–47 940 713) (16 235 987–28 955 435) # , , spp.# Median number of total and foodborne Ill and foodborne Median number of total spp. spp. spp. spp.# Enteric protozoa Enteric Cestodes Trematodes Invasive infectious infectious Invasive disease Nematodes Intestinal flukes * flukes Intestinal PATHOGEN Cryptosporidium Cryptosporidium Entamoeba histolytica Giardia gondii Toxoplasma congenital gondii Toxoplasma acquired Echinococcus granulosus Echinococcus multilocularis solium Taenia Ascaris Trichinella sinensis Chlonorchis Fasciola TableA8.5 Appendices 220 188 346 (95% UI) 7 161 689 7 161 689 1 048 937 1 048 937 6 639 989 6 639 989 (151 906–235 431) (743 700–1 438 588) FOODBORNE DALYS FOODBORNE DALYS (5 611 955–8 414 684) (6 078 248–9 074 283) 250 1 498 1 498 51 909 45 927 45 927 (95% UI) (160–371) (1 230–1 813) (34 763–59 933) (40 020–66 992) FOODBORNE DEATHS FOODBORNE DEATHS SES 16 315 139 238 (95% UI) 91 148 998 23 220 595 ILLNES ILLNES (58 576 614– 153 950 104) FOODBORNE (11 273–22 860) (95 610–195 078) (18 215 499–38 081 817) 0.61 0.76 (95% UI) (0.65–0.81) (0.52–0.67) PROPORTION FOODBORNE – DALYS FOODBORNE – DALYS SES 1.00 1.00 1.00 1.00 0.22 0.22 0.48 0.48 (95% UI) (0.16–0.31) (0.38–0.56) PROPORTION FOODBORNE – ILLNES- DALYS 188 346 (95% UI) 8 777 198 1 048 937 1 048 937 11 794 391 11 794 391 (151 906–235 431) (743 700–1 438 588) (7 620 016–12 511 566) (10 164 903–15 675 990) 250 1 498 1 498 59 724 59 724 94 620 DEATHS DEATHS (95% UI) (160–371) (1 230–1 813) (48 017–83 616) (77 140–125 670) SES NES NES 16 315 139 238 (95% UI) ILL 48 405 537 48 405 537 79 049 913) 407 149 528 407 149 528 (43 376 746– 585 323 226) (301 670 420– (11 273–22 860) (95 610–195 078) spp. spp. TOTAL TOTAL (excluding (excluding TOTAL protozoa) enteric PATHOGEN Opisthorchis Paragonimus APPENDICES WHO Estimates of the global burden of foodborne diseases 221

Table A8.6 Median number of foodborne Ill nes ses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% Uncertainty Intervals, 2010

FOODBORNE ILL NES SES FOODBORNE DEATHS FOODBORNE DALYS CHEMICAL (95% UI) (95% UI) (95% UI) 21 757 19 455 636 869 Aflatoxin (8 967–56 776) (7 954–51 324) (267 142–1 617 081) 1 066 227 18 203 Cyanide in cassava (105–3 016) (22–669) (1 769–53 170) 193 447 0 240 056 Dioxin (155 963–1 085 675 (0–0) (192 608–1 399 562) 107 167 28 99 717 Peanut allergens* (6 262–210 093) (2–56) (5 827–195 489) 338 611 19 736 1 012 362 TOTAL (185 705–1 238 725 (8 210–51 700) (562 087–2 822 481)

Notes: * = Only the burdens for AMR A, EUR A and WPR A were assessed.

Table A8.7 Median rate per 100 000 foodborne (FB) Ill nes ses, Deaths, and Disability Adjusted Life Years (DALYs) by region, with 95% uncertainty intervals, 2010.

CHEMICAL REGION CYANIDE IN DIOXIN TOTAL CASSAVA FB Illnes ses (95% UI) 0.4 (0.1–1) 0.1 (0.01–0.4) 0.2 (0.07–7) 0.7 (0.3–8) AFR FB Deaths (95% UI) 0.4 (0.1–1) 0.03 (0.003–0.08) 0 (0–0) 0.4 (0.1–1) FB DALYs (95% UI) 15 (5–40) 2 (0.2–6) 0.2 (0.07–8) 18 (7–49) FB Illnes ses (95% UI) 0.08 (0.02–0.6) 0 (0–0) 0.2 (0.05–6) 0.2 (0.1–7) AMR FB Deaths (95% UI) 0.08 (0.02–0.6) 0 (0–0) 0 (0–0) 0.08 (0.02–0.6) FB DALYs (95% UI) 2 (0.4–15) 0 (0–0) 0.2 (0.07–9) 2 (0.6–24) FB Illnes ses (95% UI) 0.2 (0.04–0.5) 0 (0–0) 2 (1–35) 2 (1–35) EMR FB Deaths (95% UI) 0.1 (0.04–0.4) 0 (0–0) 0 (0–0) 0.1 (0.04–0.4) FB DALYs (95% UI) 4 (1–13) 0 (0–0) 2 (2–43) 7 (3–51) FB Illnes ses (95% UI) 0.02 (0.01–0.03) 0 (0–0) 1 (0.7–13) 1 (0.7–13) EUR FB Deaths (95% UI) 0.02 (0.01–0.03) 0 (0–0) 0 (0–0) 0.02 (0.01–0.03) FB DALYs (95% UI) 0.5 (0.3–0.8) 0 (0–0) 1 (0.9–19) 2 (1–19) FB Illnes ses (95% UI) 0.2 (0.08–0.6) 0 (0–0) 9 (8–32) 10 (8–32) SEAR FB Deaths (95% UI) 0.2 (0.08–0.5) 0 (0–0) 0 (0–0) 0.2 (0.07–0.5) FB DALYs (95% UI) 7 (2–17) 0 (0–0) 12 (10–41) 19 (13–54) FB Ill nes ses (95% UI) 0.6 (0.1–2) 0 (0–0) 0.05 (0.005–4) 0.8 (0.1–5) WPR FB Deaths (95% UI) 0.5 (0.09–2) 0 (0–0) 0 (0–0) 0.5 (0.09–2) FB DALYs (95% UI) 16 (3–63) 0 (0–0) 0.07 (0.007–6) 16 (3–65) FB Ill nes ses (95% UI) 0.3 (0.1–0.8) 0.02 (0.002–0.04) 3 (2–16) 3 (3–17) GLOBAL FB Deaths (95% UI) 0.3 (0.1–0.7) 0.003 (0–0.01) 0 (0–0) 0.3 (0.1–0.8) FB DALYs (95% UI) 9 (4–24) 0.3 (0.03–0.8) 3 (3–20) 13 (7–39) Appendices 222 APPENDICES 223

REFERENCES

1. Devleesschauwer, B., Havelaar, 8. Mangen, M.J.J., Batz, M.B., A.H., Maertens de Noordhout, C., Kasbohrer, A., Hald, T., Morris, Haagsma, J.A., Praet, N., Dorny, P., J.G., Taylor, M. & Havelaar, A.H. Duchateau, L., Torgerson, P.R., Van 2010. Integrated approaches for Oyen, H. & Speybroeck, N. 2014. the public health prioritization of Calculating disability-adjusted life foodborne and zoonotic pathogens. years to quantify burden of disease. Risk Analysis, 30(5): 782–797. International Journal of Public 9. EFSA [European Food Safety Health, 59(3): 565–569. Authority]. 2008. Scientific Opinion 2. Murray, C.J. 1994. Quantifying the of the Panel on Biological Hazards burden of disease: the technical on a request from EFSA on basis for disability-adjusted life Overview of methods for source years. Bulletin of the World Health attribution for human illness from Organization, 72(3): 429–445. foodborne microbiological hazards (Question No. EFSA-Q-2008-005). 3. Schroeder, S.A. 2012. Incidence, Adopted on 9 July 2008. EFSA prevalence, and hybrid approaches Journal, 764: 1–43. Available at to calculating disability-adjusted life http://www.efsa.europa.eu/sites/ years. Population Health Metrics, default/files/scientific_output/ 10(1): Art. 19. files/main_documents/biohaz_op_ 4. WHO [World Health Organization]. ej764_source_attribution_en%2C3. 2013. WHO methods and data pdf Accessed 2015-10-16. sources for global burden of disease 10. Pires, S.M., Evers, E.G., van Pelt, W., estimates 2000–2011. Global Health Ayers, T., Scallan, E., Angulo, F.J., Estimates Technical Paper. WHO/ Havelaar, A., Hald, T. & the Med- HIS/HSI/GHE/2013.4. Available at Vet-Net Workpackage team. 2009. http://www.who.int/healthinfo/ Attributing the Human Disease global_burden_disease/GlobalCOD_ Burden of Foodborne Infections method_2000-2011.pdf?ua=1 to Specific Sources. Foodborne Accessed 2015-10-16. Pathogens and Disease, 6(4): 417– 5. Ezzati, M., Lopez, A.D., Rodgers, A., 424. Vander Hoorn, S. & Murray, C.J.L. 11. Havelaar, A.H., Braunig, J., 2002. Selected major risk factors Christiansen, K., Cornu, M., Hald, and global and regional burden of T., Mangen, M.J.J., Molbak, K., disease Lancet, 360(9343): 1347– Pielaat, A., Snary, E., Van Pelt, W., 1360. Velthuis, A. & Wahlstrom, H. 2007. 6. Murray, C.J.L., Ezzati, M. & 10 Towards an integrated approach others. 2012. GBD2010: design, in supporting microbiological food definitions, and metrics. Lancet, safety decisions. Zoonoses and 380(9859): 2063–2066. Public Health,54(3-4): 103–117. 7. Pires, S.M. 2013. Assessing the 12. EFSA Panel on Biological Hazards. applicability of currently available 2013. Scientific Opinion on the methods for attributing foodborne evaluation of molecular typing disease to sources, including food methods for major food-borne and food commodities. Foodborne microbiological hazards and their Pathogens and Disease, use for attribution modelling, 10(3):206–213. 224

outbreak investigation and scanning 19. Guo, C.F., Hoekstra, R.M., surveillance: Part 1 (evaluation of Schroeder, C.M. and 15 others. 2011. methods and applications). EFSA Application of Bayesian techniques Journal, 11(12): 3502–3586. to model the burden of human salmonellosis attributable to U.S. 13. Pires, S.M., Vieira, A.R., Hald, T. & food commodities at the point of Cole, D. 2014. Source attribution processing: adaptation of a Danish of human salmonellosis: An model. Foodborne Pathogens and overview of methods and estimates. Disease, 8(4): 509–516. Foodborne Pathogens and Disease, 11(9): 667–676. 20. Little, C.L., Pires, S.M., Gillespie, I.A., Grant, K. & Nichols, G.L. 2010. 14. Hald, T., Vose, D., Wegener, H.C. Attribution of human Listeria & Koupeev, T. 2004. A Bayesian monocytogenes infections in approach to quantify the England and Wales to ready-to-eat contribution of animal-food sources food sources placed on the market: to human salmonellosis. Risk adaptation of the Hald Salmonella Analysis, 24(1): 255–269. Source Attribution Model. 15. Evers, E.G. H., Van Der Fels-Klerx, J., Foodborne Pathogens and Disease, Nauta, M.J., Schijven, J.F. & Havelaar, 7(7): 749–756. A.H. 2008. Campylobacter source 21. Pires, S.M., Vieira, A.R. Perez, attribution by exposure assessment. E.., Wong, D.L.F. & Hald, T. 2012. International Journal of Risk Attributing human foodborne illness Assessment and Management, 8(1- to food sources and water in Latin 2): 174–190. America and the Caribbean using 16. Wilson, D.J., Gabriel, E., data from outbreak investigations. Leatherbarrow, A.J.H., Cheesbrough, International Journal Of Food J., Gee, S., Bolton, E., Fox, A., Microbiology, 152(3; Special Fearnhead, P., Hart, C.A. & Diggle, Issue): 129–138. P.J. 2008. Tracing the source of 22. Domingues, A.R., Pires, S.M., Halasa, Campylobacteriosis. PLOS Genetics, T. & Hald, T. 2012. Source attribution 4(9): Art. e1000203. of human salmonellosis using a 17. Muellner, P., Jones, G., Noble, meta-analysis of case-control A., Spencer, S.E.F., Hathaway, studies of sporadic infections. S. & French, N.P. 2009. Source Epidemiology and Infection, attribution of food-borne zoonoses 140(6): 959–969. in New Zealand: A modified Hald 23. Domingues, A.R., Pires, S.M., Halasa, Model. Risk Analysis, 29(7): 970– T. & Hald, T. 2012. Source attribution 984. of human Campylobacteriosis using 18. Pires, S.M., Vigre, H ., Makela, P. a meta-analysis of case-control & Hald, T. 2010. Using outbreak studies of sporadic infections. data for source attribution Epidemiology and Infection, of human salmonellosis and 140(6): 970–981. Campylobacteriosis in Europe. Foodborne Pathogens and Disease, 7(11): 1351–1361. 225

24. Painter, J.A., Hoekstra, R.M., 31. Tyshenko, M.G., El Saadany, S., Ayers, T., Tauxe, R.V., Braden, C.R., Oraby, T., Darshan, S., Aspinall, W., Angulo, F.J., Griffin, P.M. 2013. Cooke, R., Catford, A. & Krewski, Attribution of foodborne ILL NES- D. 2011. Expert elicitation for the SES, hospitalizations, and Deaths judgment of disease risk to food commodities by using uncertainties. Journal of Toxicology outbreak data, United States, and Environmental Health-Part 1998–2008. Emerging Infectious A-Current Issues, 74(2-4; Special Diseases,19(3): 407–415. issue): 261–285. 25. Boysen, L., Rosenquist, H., Larsson, 32. Hoffmann, S., Fischbeck, P., J.T., Nielsen, E.M., Sorensen, G., Krupnick, A. & McWilliams, M. 2007. Nordentoft, S. & Hald, T. 2014. Using expert elicitation to link Source attribution of human foodborne illnesses in the United Campylobacteriosis in Denmark. States to foods. Journal of Food Epidemiology and Infection, Protection, 70(5): 1220–1229. 142(8): 1599–1608. 33. Havelaar, A.H., Galindo, A.V., 26. De Knegt, L.V., Pires, S.M. & Hald, Kurowicka, D. & Cooke, R.M. 2008. T. 2015. Attributing foodborne Attribution of foodborne pathogens salmonellosis in humans to animal using structured expert elicitation. reservoirs in the European Union Foodborne Pathogens and Disease, using a multi-country stochastic 5(5): 649–659. model. Epidemiology and infection, 34. Rice, J. 2007. Mathematical statistics 143(6): 1175–1186. and data analysis. 3rd ed. Thomson 27. Cairncross, S., Hunt, C., Boisson, S., Higher Education, Belmond, Bostoen, K., Curtis, V., Fung, I.C.H. & CA, USA. Schmidt, W.P. 2010. Water, sanitation 35. Ravel, A., Davidson, V.J., Ruzante, and hygiene for the prevention of J.M., Fazil, A. 2010. Foodborne diarrhoea. International Journal of proportion of gastrointestinal Epidemiology, 39(Suppl.1): 193–205. illness: estimates from a Canadian 28. Hunter, P.R. 2009. Household water Expert Elicitation survey. Foodborne treatment in developing countries: Pathogens and Disease, 7(12): 1463– Comparing different intervention 1472. types using meta-regression. 36. Davidson, V.J., et al., Food- Environmental Science and specific attribution of selected Technology, 43(23): 8991–8997. gastrointestinal ILL NES SES: 29. Aspinall, W. 2010. A route to more estimates from a canadian expert tractable expert advice. Nature, elicitation survey. Foodborne 463(7279): 294–295. Pathogens and Disease, 2011. 8(9): p. 983–995. 30. Tyshenko, M.G., El Saadany, S., Oraby, T., Darshan, S., Catford, A., 37. Vally, H., Glass, K., Ford, L., Hall, G., Aspinall, W., Cooke, R., & Krewski, Kirk, M.D., Shadbolt, C., Veitch, M., D. 2012. Expert judgement and Fullerton, K.E., Musto, J. & Becker, N. re-elicitation for prion disease 2014. Proportion of illness acquired risk uncertainties. International by foodborne transmission for nine Journal of Risk Assessment and enteric pathogens in Australia: Management, 16(1-3): 48–77. an expert elicitation. Foodborne 226

Pathogens and Disease, 11(9): for 30 specified pathogens and 727–733. unspecified agents, circa 2006. Foodborne Pathogens and Disease, 38. Van der Fels-Klerx, H.J., Cooke, 10(7): 639–648. R.M., Nauta, M.N., Goossens, L.H. & Havelaar, A.H. 2005. A structured 44. Havelaar, A.H., Haagsma, J.A., expert judgment study for a model Mangen, M.J.J., Kemmeren, J.M., of Campylobacter transmission Verhoef, L.P.B., Vijgen, S.M.C., during broiler-chicken processing. Wilson, M., Friesema, I.H.M., Risk Analysis, 25(1): 109–124. Kortbeek, L.M., van Duynhoven, Y.T.H.P. & van Pelt, W. 2012. Disease 39. Cooke, R.M., Nauta, M., Havelaar, burden of foodborne pathogens in A.H. & van der Fels, I. 2006. the Netherlands, 2009. International Probabilistic inversion for chicken Journal of Food Microbiology, processing lines. Reliability 156(3): 231–238. Engineering & System Safety, 91(10- 11): 1364–1372. 45. Ford, L., Kirk, M., Glass, K. & Hall, G. 2014. Sequelae of Foodborne Illness 40. Pires, S.M., Fischer Walker, C.L., Caused by 5 Pathogens, Australia, Lanata, C.F., Devleesschauwer, B., Circa 2010. Emerging Infectious Hall, A., Kirk, M.D., Duarte, A.S.R., Diseases, 20(11): 1865–1871. Black, R.E., & Angulo, F.J. Submitted. Aetiology-specific estimates of the 46. Kirk, M., Ford, L., Glass, K. & Hall, G. global and regional incidence and 2014.Foodborne Illness, Australia, mortality of diarrhoeal diseases Circa 2000 and Circa 2010. commonly transmitted through Emerging Infectious Diseases, food. PLOS ONE, vol 10, iss 12, 20(11): 1857–1864. DOI: 10.1371/journal.pone.0142927 47. Vaillant, V., de Valk, H., Baron, E., 41. Scallan, E., Griffin, P.M., Angulo, F.J., Ancelle, T., Colin, P., Delmas, M.C., Tauxe, R.V. & Hoekstra, R.M. 2011. Dufour, B., Pouillot, R., Le Strat, Foodborne illness acquired in the Y., Weinbreck, P., Jougla, E. & United States– unspecified agents. Desenclos, J.C. 2005. Foodborne Emerging Infectious Diseases, infections in France. Foodborne 17(1): 16–22. Pathogens and Disease, 2(3): 221– 232. 42. Scallan, E., Hoekstra, R.M., Angulo, F.J., Tauxe, R.V., Widdowson, M.A., 48. Tam, C.C., Rodrigues, L.C., Viviani, Roy, S.L., Jones, J.L. & Griffin, P.M. L., Dodds, J.P., Evans, M.R., Hunter, 2011. Foodborne illness acquired in P.R., Gray, J.J., Letley, L.H., Rait, G., the United States– major pathogens. Tompkins, D.S. & O’Brien, S.J. [IID2 Emerging Infectious Diseases, Study Executive Committee]. 2012. 17(1): 7–15. Longitudinal study of infectious intestinal disease in the UK (IID2 43. Thomas, M.K., Murray, R., Flockhart, study): incidence in the community L., Pintar, K., Pollari, F., Fazil, and presenting to general practice. A., Nesbitt, A. & Marshall, B. Gut, 61(1): 69–77. 2013. Estimates of the burden of foodborne illness in Canada 227

49. Lake, R.J., Cressey, P.J., Campbell, Barre Syndrome worldwide a D.M. & Oakley, E. 2010. Risk ranking systematic literature review. Neuro- for foodborne microbial hazards epidemiology, 32(2): 150–163. in New Zealand: burden of disease 56. Poropatich, K.O., Walker, C.L. & estimates. Risk Analysis, Black, R.E. 2010. Quantifying the 30(5): 743–752. association between Campylobacter 50. Lanata, C.F., Fischer-Walker, C.L., infection and Guillain-Barre Olascoaga, A.C., Torres, C.X., Aryee, syndrome: a systematic review. M.J., Black, R.E. 2013. Global causes Journal of Health Population and of diarrhoeal disease mortality Nutrition, 28(6): 545–552. in children <5 years of age: a 57. Hall, A. 2013. Estimating the systematic review. PLOS ONE, global burden of foodborne 8(9): Art e72788 norovirus. 5th Meeting of the 51. Walker, C.L.F., Rudan, I., Liu, L., Nair, WHO Foodborne Disease Burden H., Theodoratou, E., Bhutta, Z.A., Epidemiology Reference Group. O’Brien, K.L., Campbell, H. & Black, Geneva, Switzerland, 7–12 April 2013. R.E. 2013. Childhood Pneumonia WHO, Geneva. and Diarrhoea. 1. Global burden 58. Lozano, R., Naghavi, M., Foreman, K., of childhood pneumonia and Lim, S., Shibuya, K. and 184 others. diarrhoea. Lancet, 381(9875): 2012. Global and regional mortality 1405–1416. from 235 causes of death for 20 52. Fischer Walker, C.L. & Black, age groups in 1990 and 2010: a R.E. 2010. Diarrhoea morbidity systematic analysis for the Global and mortality in older children, Burden of Disease Study 2010. adolescents, and adults. Lancet, 380(9859): 2095–2128 Epidemiology and Infection, 59. Crump, J.A., Luby, S.P. & Mintz, E.D. 138(9): 1215–1226. 2004. The global burden of typhoid 53. Ali, M., Lopez, A.L., You, Y.A., Kim, fever. Bulletin of the World Health Y.F., Sah, B., Maskery, B. & Clemens, Organization, 2004. 82(5): 346–353. J. 2012. The global burden of 60. Ao, T.T., Feasey, N.A., Gordon, M.A., cholera. Bulletin of the World Health Keddy, K.H., Angulo, F.J. & Crump, Organization, 90(3): 209–218. J.A. 2015. Global Burden of Invasive 54. Majowicz, S.E., Scallan, E., Jones- Nontyphoidal Salmonella Disease, Bitton, A., Sargeant, J.M., Stapleton, 2010. Emerging Infectious Diseases, J., Angulo, F.J., Yeung, D.H. & 21(6): 941–949. Kirk, M.D. 2014. Global incidence 61. Biggs, H.M., Lester, R., Nadjm, B., of human Shiga toxin-producing Mtove, G., Todd, J.E., Kinabo, G.D., Escherichia coli infections and Philemon, R., Amos, B., Morrissey, Deaths: a systematic review and A.B., Reyburn, H., Crump, J.A. 2014. knowledge synthesis. Foodborne Invasive Salmonella infections Pathogens and Disease, 11(6): in areas of high and low malaria 447–455. transmission intensity in Tanzania. 55. McGrogan, A., Madle, G.C., Clinical Infectious Diseases, Seaman, H.E., de Vries, C.S. 2009. 58(5): 638–647. The epidemiology of Guillain- 228

62. Mtove, G., Amos, B., von Seidlein, 68. Devleesschauwer, B., Haagsma, J.A., L., Hendriksen, I., Mwambuli, A., Bellinger, D., Cole, D., Döpfer, D., Kimera, J., Mallahiyo, R., Kim, Fazil, A., Fèvre, E., Lake, R., Maertens D.R., Ochiai, R.L., Clemens, J.D., de Noordhout, C., McDonald, Reyburn, H., Magesa, S. & Deen, S.A., Pires, S.M., Speybroeck, J.L. 2010. Invasive salmonellosis N., Thomas, K., Torgerson, P.R., among children admitted to a Wu, F., Havelaar, A.H. & Praet, N. rural Tanzanian hospital and a 2015. Methodological framework comparison with previous studies. for World Health Organization PLOS ONE, 5(2): Art e9244. estimates of the global burden of foodborne disease. PLOS ONE. 63. Koch, K., Kristensen, B., Holt, H.M., Ethelberg, S., Molbak, K. & 69. Dean, A.S., Crump, L., Greter, Schonheyder, H.C. 2011. International H., Hattendorf, J., Schelling, travel and the risk of hospitalization E. & Zinsstag, J. 2012. Clinical with non-typhoidal Salmonella manifestations of human brucellosis: bacteremia. A Danish population- a systematic review and meta- based cohort study, 1999–2008. analysis. PLOS Neglected Tropical BMC Infectious Diseases, 11: Diseases, 6(12): Art e1929. Article 277. 70. Maertens de Noordhout, C., 64. Muller, B., Durr, S., Alonso, S., Devleesschauwer, B., Angulo, F.J., Hattendorf, J., Laisse, C.J.M., Verbeke, G., Haagsma, J., Kirk, M., Parsons, S.D.C., van Helden, Havelaar, A. & Speybroeck, N. 2014. P.D.& Zinsstag, J. 2013. Zoonotic The global burden of listeriosis: Mycobacterium bovis-induced a systematic review and meta- tuberculosis in humans. Emerging analysis. Lancet Infectious Diseases, Infectious Diseases, 19(6): 899–908. 14(11): 1073–1082. 65. WHO. 2014. Global tuberculosis 71. Murrell, K.D. & Pozio, E. 2011. report 2013, Available from: Worldwide occurrence and impact http://apps.who.int/iris/bitstream/ of human trichinellosis, 1986–2009. 10665/91355/1/9789241564656_eng. Emerging Infectious Diseases, pdf. WHO, Geneva. 17(12): 2194–2202. 66. Durr, S., Muller, B., Alonso, S., 72. Torgerson, P.R., Keller, K., Magnotta, Hattendorf, J., Laisse, C.J.M., van M. & Ragland, N. 2010. The global Helden, P.D. & Zinsstag, J. 2013. burden of alveolar echinococcosis. Differences in primary sites of PLOS Neglected Tropical Diseases, infection between zoonotic and 4: e722. human tuberculosis: results from 73. Carabin, H., Ndimubanzi, P.C., a worldwide systematic review. Budke, C.M., Nguyen, H., Qian, Y., PLOS Neglected Tropical Diseases, Cowan, L.D., Stoner, J.A., Rainwater, 7(8): Art e2399. E. & Dickey, M. 2011. Clinical 67. Dean, A.S., Crump, L., Greter, H., manifestations associated with Schelling, E. & Zinsstag, J. 2012. neurocysticercosis: a systematic Global burden of human brucellosis: review. PLOS Neglected Tropical a systematic review of disease Diseases, 5(5): Art e1152. frequency. PLOS Neglected Tropical Diseases, 6(10): Art. e1865. 229

74. Ndimubanzi, P.C., Carabin, H., 80. Mas-Coma, S., Bargues, M.D. Budke, C.M., Nguyen, H., Qian, &Valero, M.A. 2005. Fascioliasis Y.J., Rainwater, E., Dickey, M., and other plant-borne trematode Reynolds, S. & Stoner, J.A. 2010. A zoonoses. International Journal of systematic review of the frequency Parasitology, 35(11-12): 1255–1278. of neurocysticercosis with a focus 81. Murray, C.J.L., Vos, T., Lozano, R., on people with epilepsy. PLOS Naghavi, M., Flaxman, A.D., Michaud, Neglected Tropical Diseases, C., Ezzati, M., Shibuya, K., Salomon, 4(11): Art e870. J.A. and 350 others. 2012. Disability- 75. Budke, C.M., Carabin, H., adjusted life years (DALYs) for 291 Ndimubanzi, P.C., Nguyen, H., diseases and injuries in 21 regions, Rainwater, E., Dickey, M., Bhattarai, 1990–2010: a systematic analysis for R., Zeziulin, O. & Qian, M.B. 2013. A the Global Burden of Disease Study systematic review of the literature 2010. Lancet, 380(9859): 2197–2223. on cystic echinococcosis frequency 82. Salomon, J.A., Vos, T., Hogan, D.R. worldwide and its associated clinical and 124 others. 2012. Common manifestations. American Journal values in assessing health outcomes of Tropical Medicine and Hygiene, from disease and injury: disability 88(6): 1011–1027. weights measurement study for 76. Torgerson, P.R. & Mastroiacovo, the Global Burden of Disease Study P. 2013. The global burden of 2010. Lancet, 380(9859): 2129–2143. congenital toxoplasmosis: a 83. Vos, T., Flaxman, A.D., Naghavi, systematic review. Bulletin of M. and 352 others. 2012. Years the World Health Organization, lived with disability (YLDs) for 91(7): 501–508. 1160 sequelae of 289 diseases and 77. Furst, T., Keiser, J. & Utzinger, J. injuries 1990–2010: a systematic 2012. Global burden of human food- analysis for the Global Burden borne trematodiasis: a systematic of Disease Study 2010. Lancet, review and meta-analysis. Lancet 380(9859): 2163–2196. Infectious Diseases, 12(3): 210–221. 84. Devleesschauwer, B., Praet, N., 78. Furst, T., Sayasone, S., Odermatt, Speybroeck, N., Torgerson, P.R., P., Keiser, J. & Utzinger, J. 2012. Haagsma, J.A., De Smet, K., Manifestation, diagnosis, and Murrell, K.D., Pozio, E. & Dorny, management of foodborne P. 2015. The low global burden trematodiasis. British Medical of trichinellosis: evidence and Journal, 344: Art e4093. implications. International Journal for Parasitology, 45(2-3): 95–99. 79. McDonald, S.A., Devleesschauwer, B., Speybroeck, N., Hens, N., Praet, 85. Cardoso, A.P., Mirione, E., Ernesto, N., Torgerson, P.R., Havelaar, A.H., M., Massaza, F., Cliff, J., Haque, M.R. Wu, F., Tremblay, M., Amene, E.W. & Bradbury, J.H. 2005. Processing of & Dopfer, D. 2015. Data-driven cassava roots to remove cyanogens. methods for imputing national- Journal of Food Composition and level incidence in global burden Analysis, 18(5): 451–460. of disease studies. Bulletin of the World Health Organization, 93(4): 228–236. 230

86. Cliff, J. 2011. Incidence and 93. Tshala-Katumbay, D. 2001. On the prevalence estimates of cassava site of the lesion in konzo. p. 72, in: cyanide-induced diseases. WHO/ Report of the Faculty of Medicine, FERG report, September 2011. p. 75. 2001, University of Uppsala, Sweden. 87. Cliff, J., Muquingue, H., Nhassico, 94. Banea, M., Bikangi, N., Nahimana, D., Nzwalo, H. & Bradbury, J.H. 2011. G., Nunga, M., Tylleskar, T. & Rosling, Konzo and continuing cyanide H. 1992. High prevalence of konzo intoxication from cassava in associated with a food shortage Mozambique. Food and Chemical crisis in the region of Bandundu, Toxicology, 49(3): 631–635. Zaire. Annales De La Societe Belge De Medecine Tropicale, 72(4): 88. Howlett, W.P., Brubaker, G.R., Mlingi, 295–309. N. & Rosling, H. 1990. Konzo, An Epidemic Upper Motor Neuron 95. Ministry of Health Mozambique. Disease Studied In Tanzania. 1984. Mantakassa: an epidemic of Brain, 113(1):223–235. spastic paraparesis associated with chronic cyanide intoxication in a 89. WHO, 1996. Konzo a distinct type cassava staple area of Mozambique. of upper motoneuron disease. 2. Nutritional factors and Weekly Epidemiological Record, hydrocyanic acid content of cassava 71(30): 225–232. products. Bulletin of the World 90. Tylleskar, T., The causation of konzo. Health Organization, 62(3): Studies on a paralytic disease 485–492. in Africa. PhD Thesis. Uppsala 96. Tylleskar, T., Banea, M., Bikangi, N., University, Umea, Sweden. Fresco, L., Persson, L.A. & Rosling, 91. Tylleskar, T., Legue, F.D., Peterson, H. 1991. Epidemiologic evidence S., Kpizingui, E. & Stecker, P. 1994. from Zaire for a dietary aetiology Konzo in The Central African of konzo, an upper motor-neuron Republic. Neurology, 44(5): disease. Bulletin of the World Health 959–961. Organization, 69(5): 581–589. 92. Bettencourt, M.S., Paquisse, M.M. & 97. Bonmarin, I., Nunga, M. & Perea, Zangulo, A. 2011. Tropical spastic W.A. 2002. Konzo outbreak, in paraparesis; a major neurologic the south-west of the Democratic problem in Caungulia Angola Republic of Congo, 1996. Journal of related to of cassava: first report Tropical Pediatrics, 48(4): 234–238. from Angola. Poster presentation 98. Ngudi, D.D. 2005. Konzo and at the 20th World Congress of cassava toxicity: a study of Neurology, Marrakesh, Morocco.– associated nutritional factors in the See also Cassava Cyanide Diseases Popokabaka District, Democratic & Network http:// Republic of Congo. PhD Thesis, ipbo.vib-ugent.be/wp-content/ Faculty of Bioscience Engineering, uploads/2015/02/CCDN-News-18- University of Ghent, Belgium. December-2011.pdf Accessed 2015- 10-20. 231

99. Casadei, E.A.A. 1984. “Mantakassa” 106. Skolnick, H.S., Conover-Walker, uma epidemia de neuropatia M.K., Koerner, C.B., Sampson, H.A., tropical associada com intoxicação Burks, W. & Wood, R.A. 2001. The por mandioca na Província de natural history of peanut allergy. Nampula, Moçambique. Revista Journal of Allergy and Clinical Médica de Moçambique, 2: 1–34. Immunology, 107(2):367–374. 100. Cliff, J., Nicala, D., Saute, F., Givragy, 107. Sicherer, S.H., Furlong, T.J., Munoz- R., Azambuja, G., Taela, A., Chavane, Furlong, A., Burks, A.W. & Sampson, L. & Howarth, J. 1997. Konzo H.A. 2001. A voluntary registry associated with war in Mozambique. for peanut and tree nut allergy: Tropical Medicine & International Characteristics of the first 5149 Health, 2(11):1068–1074. registrants. Journal of Allergy and Clinical Immunology, 108(1): 128–132. 101. Howlett, W.P., Brubaker, G.M., Mlingi, N. & Rosling, H. 1992 A geographical 108. Green, T.D., LaBelle, V.S., Steele, P.H., cluster of konzo in Tanzania. Journal Kim, E.H., Lee, L.A., Mankad, V.S., of Tropical and Geographical Williams, L.W., Anstrom, K.J. & Burks, Neurology,. 2: 102–108. A.W. 2007. Clinical characteristics of Peanut-Allergic Children: recent 102. Ezendam, J. & van Loveren, H. 2012. changes. Pediatrics, 120(6): Parameters needed to estimate the 1304–1310. global burden of peanut allergy: Systematic literature review. 109. Moneret-Vautrin, D.A., Rance, F., European Journal of Food Research Kanny, G., Olsewski, A., Gueant, & Review, 2(2): 46–48. J.L., Dutau, G. & Guerin, L. 1998. Food allergy to peanuts in France– 103. Mullins, R.J., Dear, K.B. & Tang, M.L. evaluation of 142 observations. 2009. Characteristics of childhood Clinical and Experimental Allergy, peanut allergy in the Australian 28(9): 1113–1119. Capital Territory, 1995 to 2007. Journal of Allergy and Clinical 110. Liu, Y. & Wu, F. 2010. Global burden Immunology, 123(3): 689–693. of aflatoxin-induced hepatocellular carcinoma: a risk assessment. 104. Flokstra-de Blok, B.M.J., van der Environmental Health Perspectives, Velde, J.L., Vlieg-Boerstra, B.J., 118(6): 818–824. Oude Elberink, J.N.G., DunnGalvin, A., Hourihane., J.O., Duiverman, 111. JECFA [Joint FAO/WHO Expert E.J. & Dubois, A.E.J. 2010. Health- Committee on Food Additives]. related quality of life of food allergic 1998. Aflatoxins. Safety Evaluation patients measured with generic of Certain Food Additives and and disease-specific questionnaires. Contaminants. WHO, Geneva. Allergy, 65(8): 1031–1038. 112. IARC [International Agency for 105. Hourihane, J.O., Roberts, S.A. Research on Cancer]. No date. & Warner, J.O. 1998. Resolution GLOBOCAN 2012: All Cancers of peanut allergy: case-control (excluding non-melanoma skin study. British Medical Journal, cancer): Estimated Incidence, 316(7140: 1271–1275. Mortality and Prevalence Worldwide in 2012 [Online]. [Accessed 14 July 2015]. Available from: http:// globocan.iarc.fr/Pages/fact_sheets_ cancer.aspx 232

113. WHO. 2010. Health Statistics and from: http://ec.europa.eu/food/fs/ Information Systems. Estimates sc/scf/out90_en.pdf for 2000–2012. Estimated Deaths 119. JECFA [Joint FAO/WHO by cause, sex and WHO Member Expert Committee on Food State, 2010; Estimated YLDs by Additives]. 2001. Polychlorinated cause, sex and WHO Member State, dibenzodioxins, polychlorinated 2010; and Estimated YLLs by cause, dibenzofurans, and coplanar sex and WHO Member State, 201.; polychlorinated biphenyls. Available from: http://www.who.int/ Safety evaluation of certain food healthinfo/global_burden_disease/ additives and contaminants. 2001. estimates/en/index2.html. WHO, Geneva. 114. Kensler, T.W., Qian, G.S., Chen, J.G. & 120. JECFA. 2005. Joint FAO/ Groopman, J.D. 2003. Translational WHO Expert Committee on strategies for cancer prevention Food Additives. Summary and in liver. Nature Reviews Cancer, Conclusions of the sixty-fourth 3(5): 321–329. meeting, Rome, 2005. 115. WHO. 2014. Dioxins and their effects 121. US EPA [United States on human health In: World Health Environmental Protection Agency]. Organization Fact Sheet, No. 225. 2012. EPA’s reanalysis of key 116. van den Berg, M., van Birgelen, issues related to dioxin toxicity A., Birnbaum, L. and 27 others. and response to NAS comments, 2000. Consultation on assessment Volume 1 (CAS No. 1746–01– of the health risk of dioxins; re- 6), EPA/600/R-10/038F. US evaluation of the tolerable daily Environmental Protection Agency, intake (TDI): Executive Summary. Washington, D.C., USA. Food Additives And Contaminants: 122. Patterson, D.G., Needham, L.L., Part A– Chemistry Analysis Control Pirkle, J.L., Roberts, D.W., Bagby, Exposure & Risk Assessment. J., Garrett, W.A., Andrews, J.S., 17(4): 223–240. Falk, H., Bernert, J.T., Sampson, 117. EC SCF [European Commission E.J. & Houk, V.N. 1988. Correlation Scientific Committee on Food]. between serum and adipose-tissue 2000. Opinion of the SCF on the levels of 2,3,7,8-Tetrachlorodibenzo- Risk Assessment of Dioxins and Para-Dioxin in 50 persons from Dioxin-like PCBs in Food. SCF/ Missouri. Archives of Environmental CS/CNTM/Dioxins/8 Final, 2000. Contamination and Toxicology, Available at http://ec.europa. 17(2): 139–143. eu/food/fs/sc/scf/out78_en.pdf 123. LaKind, J.S., Berlin, C.M., Sjodin, A., Accessed 2015-10-21. Turner, W., Wang, R.Y., Needham, 118. EC SCF. 2001. Opinion of the EC L.L., Paul, I.M., Stokes, J.L., Naiman, Scientific Committee on Food on D.Q., Patterson, D.G. 2009. Do the Risk Assessment of Dioxins and human milk concentrations of Dioxin-like PCBs in Food. Update persistent organic chemicals really based on new scientific information decline during lactation? Chemical available since the adoption of the concentrations during lactation SCF opinion of the 22th November and milk/serum partitioning. 2000. Report No. CS/CNTM/ Environmental Health Perspectives, Dioxin/20 final, 2001. Available 117(10): 1625–1631. 233

124. Nakamura, T., Nakai, K., Matsumura, 129. UNEP [United Nations Environment T., Suzuki, S., Saito, Y., Satoh, H. Programme]. 2013. Human exposure 2008. Determination of dioxins and to POPs across the globe: POPs polychlorinated biphenyls in breast levels and human health implication: milk, maternal blood and cord blood Results of the WHO/UNEP human from residents of Tohoku, Japan. milk survey. In: [Proceedings of the] Science of the Total Environment, Conference of the Parties to the 394(1): 39–51. Stockholm Convention on Persistent Organic Pollutants, Sixth Meeting. 125. Wittsiepe, J., Furst, P., Schrey, P., Geneva, 28 April–10 May 2013. Doc. Lemm, F., Kraft, M., Eberwein, G., No. UNEP/POPS/COP.6/INF/33. Winneke, G., Wilhelm, M. 2007. Available from: http://www.unep. PCDD/F and dioxin-like PCB in org/chemicalsandwaste/portals/9/ human blood and milk from German POPs/docs/UNEP-POPS-COP.6- mothers. Chemosphere, 67(9; INF-33.English.pdf Accessed Special Issue): S286–S294. 2015-10-21. 126. UNEP [United Nations Environment 130. Crump, K.S. 1984. A new method Programme]. 2013. Human exposure for determining allowable daily to POPs across the globe: POPs intakes. Fundamental and Applied levels and human health implication: Toxicology, 4(5): 854–871. Results of the WHO/UNEP human milk survey. In: [Proceedings of the] 131. EFSA, 2009. Scientific opinion. Use Conference of the Parties to the of the benchmark dose approach Stockholm Convention on Persistent in risk assessment. Guidance of the Organic Pollutants, Sixth Meeting. Scientific Committee (Question No Geneva, 28 April–10 May 2013. Doc. EFSA-Q-2005–232). EFSA Journal, No. UNEP/POPS/COP.6/INF/33. 1150: 1–72. Available from: http://www.unep. 132. Slob, W. 2002. Dose-response org/chemicalsandwaste/portals/9/ modeling of continuous endpoints. POPs/docs/UNEP-POPS-COP.6- Toxicological Sciences, 66(2): INF-33.English.pdf Accessed 298–312. 2015-10-21. 133. Zeilmaker, M.J.,Devleesschauwer, 127. Slob, W., Bakker, M.I., te Biesebeek, B., Mengelers, M.J.B., Hoekstra, J., J.D. & Bokkers, B.G.H. 2014. Brandon, E.F.A., Bokkers, B.G.H1. Exploring the uncertainties in 2015. The disease burden of dioxins: cancer risk assessment using A global perspective. RIVM Report. the Integrated Probabilistic Risk Assessment (IPRA) approach. Risk 134. Baccarelli, A., Giacomini, S.M., Analysis, 34(8): 1401–1422. Corbetta, C., Landi, M.T., Bonzini, M., Consonni, D., Grillo, P., Patterson, 128. van der Voet, H. & Slob, W. 2007. D.G. Jr., Pesatori, A.C. & Bertazzi, Integration of probabilistic exposure P.A. 2008. Neonatal thyroid assessment and probabilistic hazard function in Seveso 25 years after characterization. Risk Analysis, maternal exposure to dioxin. Plos 27(2): 351–371. Medicine, 5(7): 1133–1142 (Article Number: e161). 234

135. Aoki, Y., Belin, R.M., Clickner, R., Toxicology Program, Research Jeffries, R., Phillips, L., Mahaffey, K.R. Triangle Park, North Carolina, USA. 2007. Serum TSH and total T-4 in 140. Van den Berg, M., Birnbaum, the United States population and L.S., Denison, M. and 16 others. their association with participant 2006. The 2005 World Health characteristics: National Health Organization re-evaluation of and Nutrition Examination Survey human and mammalian toxic (NHANES 1999-2002). Thyroid, equivalency factors for dioxins and 17(12): 1211–1223. dioxin-like compounds. Toxicological 136. Gray, L.E., Ostby, J.S. & Kelce, W.R. Sciences, 93(2): 223–241. 1997. A dose-response analysis 141. Haagsma, J.A., Polinder, S., Cassini, of the reproductive effects of A., Colzani, E., Havelaar, A.H. 2014. a single gestational dose of Review of disability weight studies: 2,3,7,8-tetrachlorodibenzo-p-dioxin comparison of methodological in male Long Evans Hooded rat choices and values. Population offspring. Toxicology and Applied Health Metrics, 12: Article No. 20. Pharmacology, 146(1): 11–20. 142. Salomon, J.A., Haagsma, J.A., Davis, 137. Cooper, T.G., Noonan, E., von A., de Noordhout, C.M., Polinder, Eckardstein, S., Auger, J., Baker, S., Havelaar, A.H., Cassini, A., H.W.G., Behre, H.M., Haugen, T.B., Devleesschauwer, B., Kretzschmar, Kruger, T., Wang, C., Mbizvo, M.T. & M., Speybroeck, N., Murray, C.J.L. & Vogelsong, K.M. 2010. World Health Vos, T. 2015. Disability weights for Organization reference values the Global Burden of Disease 2013 for human semen characteristics. study. Global health, Human Reproduction Update, 3(11): Art. e712-23. 16(3): 231–245. 143. Nithiuthai, S., Anantaphruti, M.T., 138. NTP [National Toxicology Waikagul, J. & Gajadhar, A. 2004. Program]. 2006. Toxicology and Waterborne zoonotic helminthiases. Carcinogenesis Studies of 2,3,7,8– Veterinary Parasitology, 126(1-2; Tetrachlorodibenzo-p-dioxin Special Issue): SI167–193. (TCDD) (CAS No. 1746–01–6) in Female Harlan Sprague-Dawley 144. Blakely, T., Hales, S., Kieft, C., Wilson, Rats (Gavage Studies). National N. & Woodward, A. 2005. The Toxicology Program technical global distribution of risk factors by report series 526. 180 p. National poverty level. Bulletin of the World Toxicology Program, Research Health Organization, 83(2):118–126. Triangle Park, North Carolina, USA. 145. Cooke, R.M. & Goossens, L.H.J. 139. NTP [National Toxicology 2000. Procedures Guide for Program]. 2006. Toxicology Structural Expert Judgement in and Carcinogenesis Studies of Accident Consequence Modelling. 2,3,4,7,8–pentachlorodibenzofuran Radiation Protection Dosimetry, (PeCDF) (CAS No. 57117–31–4) in 90(3): 303–309. Female Harlan Sprague-Dawley 146. Cooke, R.M. 1991. Experts in Rats (Gavage Studies). National Uncertainty: Opinion and Subjective Toxicology Program technical Probability in Science. Oxford report series Issue: 525. National University Press. 235

147. Cooke, R.M. & Goossens, L.L.H.J. Netherlands. Epidemiology and 2008. TU Delft expert judgment Infection, 138(11): 1650–1656. data base. Reliability Engineering 154. Havelaar, A.H., Haagsma, J.A., and System Safety, 93(5): 657–674. Mangen, M.J.J., Kemmeren, J.M., 148. Bamber, J.L. & Aspinall, W.P. 2013. Verhoef, L.P.B., Vijgen, S.M.C., An expert judgement assessment Wilson, M., Friesema, I.H.M., of future sea level rise from the ice Kortbeek, L.M., van Duynhoven, sheets. Nature Climate Change, Y.T.H.P. & van Pelt, W. 2012. Disease 3(4): 424–427 burden of foodborne pathogens in the Netherlands, 2009. International 149. Cooke, R.M., El Saadany, S. & Huang, Journal of Food Microbiology, X. 2008. On the performance of 156(3): 231–238. social network and likelihood- based expert weighting schemes. 155. Kretzschmar, M., Mangen, Reliability Engineering and System M.J.E.J., Pinheiro, P. and 8 others Safety, 93(5): 745–756. (BCoDE Consortium). 2012. New Methodology for Estimating the 150. R Core Team. 2014. R: A language Burden of Infectious Diseases in and environmnet for statistical Europe. PLOS Medicine, 9(4): Article computing. [15 September 2015]. Number: e1001205 Available from: https://www.r- project.org/. 156. Hald, T., Aspinall, W., Devleesschauwer, B., Cooke, R., 151. Devleesschauwer, B., Haagsma, J.A., Corrigan, T., Havelaar, A., Gibb, Bellinger, D., Cole, D., Döpfer, D., H., Torgerson, P., Kirk, M., Angulo, Fazil, A., Fèvre, E., Lake, R., Maertens F.J., Lake, R., Speybroeck, N. & de Noordhout, C., McDonald, Hoffmann, S. 2015. World Health S.A., Pires, S.M., Speybroeck, Organization estimates of the N., Thomas, K., Torgerson, P.R., relative contributions of food to the Wu, F., Havelaar, A.H. & Praet, N. burden of disease due to selected 2015. Methodological framework foodborne hazards: a structured for World Health Organization expert elicitation. PLOS ONE estimates of the global burden of (in press). foodborne disease. PLOS ONE, vol 10, iss 12 DOI: 10.1371/journal. 157. Murray, C.J., Salomon, J.A. & pone.0142498. Mathers, C. 2000. A critical examination of summary measures 152. Devleesschauwer, B., Havelaar, of population health. Bulletin of A.H., Maertens de Noordhout, C.M., the World Health Organization, Haagsma, J.A., Praet, N., Dorny, P., 78(8): 981–994. Duchateau, L., Torgerson, P.R., Van Oyen, H., Speybroeck, N. 2014. DALY 158. Prüss-Üstün, A., Mathers, C., calculation in practice: a stepwise Corvalán, C. & Woodward, A. (eds). approach. International Journal of 2003. Quantifying environmental Public Health, 59(3): 571–574. health impacts. Introduction and methods: Assessing the 153. Haagsma, J.A., Siersema, P.D., De environmental burden of disease Wit, N.J. & Havelaar, A.H. 2010. at national and local levels. [WHO] Disease burden of post-infectious Environmental Burden of Disease irritable bowel syndrome in The Series, No. 1. 236

159. Murray, C.J. 2007. Towards good Cawthorne, A., Mathers, C., Stein, practice for health statistics: lessons C., Devleesschauwer, B. on behalf from the Millennium Development of the World Health Organization Goal health indicators. Lancet, Foodborne Disease Burden 369(9564): 862–873. Epidemiology Reference Group. 2015. World Health Organization 160. Gelman, A. 2006. Prior distributions global estimates and regional for variance parameters in comparisons of the burden of hierarchical models (comment on foodborne disease, 2010. article by Browne and Draper). PLOS Medicine, DOI: 10.1371/journal. Bayesian Analysis, 1(3): 515–534. pmed.1001923 161. Plummer, M. 2003. JAGS: A program 168. Kirk, M.D., Pires, S.M., Black, for analysis of Bayesian graphical R.E., Caipo, M., Crump, J.A., models using Gibbs sampling. In: K. Devleesschauwer, B., Döpfer, D., Hornik, F. Leisch and A. Zeileis (eds). Fazil, A., Fischer-Walker, C.L., Proceedings of the 3rd International Hald, T., Hall, A.J., Keddy, K.H., Workshop on Distributed Statistical Lake, R., Lanata, C.F., Torgerson, Computing (DSC 2003). Vienna, P.R., Havelaar, A.H. & Angulo, F.J. 20–22 March 2003. 10 p. 2015. World Health Organization 162. Plummer, M. rjags: Bayesian estimates of the global and regional graphical models using MCMC. disease burden of 22 foodborne R package version 3–12. 2014; bacterial, protozoal and viral Available from: http://CRAN.R- diseases, 2010: a data synthesis project.org/package=rjags. PLOS Medicine, DOI:10.1371/journal. 163. Rubin, D. 1976. Inference and pmed.1001921 missing data. Biometrika, 63(3): 169. Mead, P.S., Slutsker, L., Dietz, V., 581–592. McCaig, L.F., Bresee, J.S., Shapiro, 164. Bhaskaran, K. & Smeeth, L. 2014. C., Griffin, P.M. & Tauxe, R.V. 1999. What is the difference between Food-related illness and death in the missing completely at random and United States. Emerging Infectious missing at random? International Diseases, 5(5): 607–625. Journal of Epidemiology, 170. Scallan, E., Hoekstra, R.M., Angulo, 43(4): 1336–1339. F.J., Tauxe, R.V., Widdowson, M.A., 165. WHO. 2013. Global tuberculosis Roy, S.L., Jones, J.L., & Griffin, P.M. report 2013. Doc. WHO/HTM/ 2011. Foodborne illness acquired in TB/201311. the United States– major pathogens. Emerging Infectious Diseases, 166. United Nations Population Division. 17(1): 7–15. 2013. World population prospects– the 2012 revision. United Nations, 171. Scallan, E., Griffin, P.M., Angulo, F.J., New York, USA. Tauxe, R.V. & Hoekstra, R.M. 2011. Foodborne illness acquired in the 167. Havelaar, A.H., Kirk, M.D., Torgerson, United States– unspecified agents. P.R., Gibb, H.J., Hald, T., Lake, Emerging Infectious Diseases, R.J., Praet, N., Angulo, F.J., 17(1): 16–22. Bellinger, D.C., de Silva, N.R., Gargouri, N., Speybroeck, N., 237

172. Adak, G.K., Long, S.M. &O’Brien, the Netherlands, 2009. International S.J. 2002. Trends in indigenous Journal of Food Microbiology, foodborne disease and Deaths, 156(3): 231–238. England and Wales: 1992 to 2000. 179. Scallan, E., Hoekstra, R.M., Mahon, Gut, 51(6): 832–841. B.E., Jones, T.F. & Griffin, P.M. 2015. 173. Hall, G., Kirk, M.D., Becker, N., An assessment of the human Gregory, J.E., Unicomb, L., health impact of seven leading Millard, G., Stafford, R. & Lalor, foodborne pathogens in the United K. 2005. Estimating foodborne States using disability adjusted gastroenteritis, Australia. Emerging life years. Epidemiology and Infectious Diseases, 11(8): 1257–1264. Infection,143(13): 2795–2804. 174. Vaillant, V., de Valk, H., Baron, E., 180. Ao, T.T., Feasey, N.A., Gordon, M.A., Ancelle, T., Colin, P., Delmas, M.C., Keddy, K.H., Angulo, F.J. & Crump, Dufour, B., Pouillot, R., Le Strat, Y., J.A. 2015. Global Burden of Invasive Weinbreck, P., Jougla, E., Desenclos, Nontyphoidal Salmonella Disease, J.C., 2005. Foodborne infections in 2010. Emerging Infectious Diseases, France. Foodborne Pathogens and 21(6): 941–949. Disease, 2(3): 221–232. 181. Aspinall, WP., Cooke, RM., Havelaar, 175. Thomas, M.K., Murray, R., Flockhart, AH., Hoffman, S., Hald, T. 2015. L., Pintar, K., Pollari, F., Fazil, Science-based global attribution A., Nesbitt, A. & Marshall, B. of foodborne diseases: Findings of 2013. Estimates of the burden WHO expert elicitation. of foodborne illness in Canada PLOS ONE. In press. for 30 specified pathogens and 182. WHO. 2008. WHO initiative to unspecified agents, circa 2006. estimate the global burden of Foodborne Pathogens and Disease, foodborne diseases: First formal 10(7): 639–648. meeting of the Foodborne Disease 176. Lake, R.J., Cressey, P.J., Campbell, Burden Epidemiology Reference D.M., Oakley, E. 2010. Risk ranking Group (FERG): implementing for foodborne microbial hazards in strategy, setting priorities and New Zealand: Burden of Disease assigning the tasks. 2008. WHO, estimates. Risk Analysis, 30(5): Geneva, Switzerland. Available at 743–752. http://www.who.int/foodsafety/ publications/burden_nov07/en/ 177. Gkogka, E., Reij, M.W., Havelaar, A.H., Accessed 2015-10-22. Zwietering, M.H., & Gorris, L.G.M. 2011. Risk-based estimate of effect 183. Porter, C.K., Gormley, R., of foodborne diseases on public Tribble, D.R., Cash, B.D., Riddle, health, Greece. Emerging Infectious M.S. 2011. The incidence and Diseases, 17(9): 1581–1590. gastrointestinal infectious risk of functional gastrointestinal 178. Havelaar, A.H., Haagsma, J.A., disorders in a healthy US adult Mangen, M.J.J., Kemmeren, J.M., population. American Journal of Verhoef, L.P.B., Vijgen, S.M.C., , 106(1): 130–138. Wilson, M., Friesema, I.H.M., Kortbeek, L.M., van Duynhoven, 184. Haagsma, J.A., Siersema, P.D., De Y.T.H.P. & van Pelt, W. 2012. Disease Wit, N.J. & Havelaar, A.H. 2010. burden of foodborne pathogens in Disease burden of post-infectious 238

irritable bowel syndrome in The Engineering & System Safety, Netherlands. Epidemiology and 96(10): 1292–1310. Infection, 138(11): 1650–1656. 191. Gibb, H., Devleesschauwer, B., 185. The MAL-ED Network Investigators. Bellinger, D., Bolger, P.M., Zeilmaker, The MAL-ED Study: A Multinational M., Barchowsky, A., Oberoi, S., and Multidisciplinary Approach Wu, F., Ezendam, J., Zang, J., to Understand the Relationship Carrington, C., Cliff, J., Verger, P., Between Enteric Pathogens, Pitt, J., Adegoke, G., Afshari, R., Malnutrition, Gut Physiology, Baines, J., Bokkers, B., Mengelers, Physical Growth, Cognitive M., van Loveren, H., Rainis, H., Development, and Immune O’Leary, K. & Liu, Y. 2015. World Responses in Infants and Children Health Organization estimates of Up to 2 Years of Age in Resource- the global and regional disease Poor Environments. Clinical burden of four foodborne chemicals Infectious Diseases. 2014;59 and toxins, 2010: A data synthesis. (suppl 4):S193-S206 F1000 Research. 186. Zinsstag, J., Schelling, E., Waltner- 192. Ahmed, S.M., Hall, A.J., Robinson, Toews, D. & Tanner, M. 2011. From A.E., Verhoef, L., Premkumar, “one medicine” to “one health” and P., Parashar, U.D., Koopmans, systemic approaches to health and M. & Lopman, B.A. 2014. Global well-being. Preventive Veterinary prevalence of norovirus in cases of Medicine, 101(3-4): 148–156. gastroenteritis: a systematic review and meta-analysis. Lancet Infectious 187. Gkogka, E., Reij, M.W., Havelaar, A.H., Diseases, 14(8): 725–730. Zwietering, M.H., & Gorris, L.G.M. 2011. Risk-based estimate of effect 193. Kotloff, K.L., Nataro, J.P., of foodborne diseases on public Blackwelder, W.C. and 40 others. health, Greece. Emerging Infectious 2013. Burden and aetiology of Diseases, 17(9): 1581–1590. diarrhoeal disease in infants and young children in developing 188. Scallan, E., Hoekstra, R.M., Angulo, countries (the Global Enteric F.J., Tauxe, R.V., Widdowson, M.A., Multicenter Study, GEMS): a Roy, S.L., Jones, J.L. & Griffin, P.M. prospective, case-control study 2011. Foodborne illness acquired in Lancet, 382(9888): 209–222. the United States– major pathogens. Emerging Infectious Diseases, 194. Hall, A. 2013. Estimating the global 17(1): 7–15. burden of foodborne norovirus. Presentation at Foodborne Disease 189. Cooke, R.M. 2015. Commentary: Epidemiology Reference Group Messaging climate change meeting– FERG 5. April 2013, WHO, uncertainty. Nature Climate Change, Geneva, Switzerland. 5(1): 8–10. 195. Crump, J.A. & Heyderman, R.S. 2014. 190. Flandoli, F., Giorgi, E., Aspinall, W.P. Invasive Salmonella infections in & Neri, A. 2011. Comparison of a Africa. Transactions of the Royal new expert elicitation model with Society of Tropical Medicine and the Classical Model, equal weights Hygiene, 108(11): 673–675. and single experts, using a cross- validation technique. Reliability 239

196. Feasey, N.A., Dougan, G., Kingsley, proportion of gastrointestinal illness: R.A., Heyderman, R.S. & Gordon, estimates from a Canadian Expert M.A. 2012. Invasive non-typhoidal Elicitation Survey. Foodborne Salmonella disease: an emerging Pathogens and Disease, 7(12): and neglected tropical disease in 1463–1472. Africa. Lancet, 379(9835): 2489– 203. Pires, S.M., Vieira, A.R., Hald, T. & 2499. Cole, D. 2014. Source attribution 197. Majowicz, S.E., Musto, J., Scallan, E., of human salmonellosis: An Angulo, F.J., Kirk, M., O’Brien, S.J., overview of methods and estimates. Jones, T.F., Fazil, A. & Hoekstra, R.M. Foodborne Pathogens and Disease, and the International Collaboration 11(9): 667–676. on Enteric Disease Burden. 2010. 204. Verhoef, L., Hewitt, J., Barclay, The Global Burden of Nontyphoidal L., Ahmed, S.M., Lake, R., Hall, Salmonella Gastroenteritis. Clinical A.J., Lopman, B., Kroneman, A., Infectious Diseases, 50(6): 882–889. Vennema, H., Vinje, J. & Koopmans, 198. Ram, P.K., Crump, J.A., Gupta, S.K., M. 2015. Norovirus genotype Miller, M.A., Mintz, E.D. 2008. Part profiles associated with foodborne II. Analysis of data gaps pertaining transmission. Emerging Infectious to Shigella infections in low and Diseases, 21(4): 592–599. medium human development index 205. Haagsma, J.A., Polinder, S., Stein, countries, 1984–2005. Epidemiology C.E., Havelaar, A.H. 2013. Systematic and Infection, 136(5): 577–603. review of foodborne burden of 199. Crump, J.A. 2014. Updating and disease studies: Quality assessment refining estimates of typhoid fever of data and methodology. burden for public health action. International Journal of Food Lancet Global Health, 2(10): Microbiology, 166(1): 34–47. e551–e553. 206. Budke, C.M., Deplazes, P. & 200. Plass, D., Mangen, M.J.J., Kraemer, Torgerson, P.R. 2006. Global A., Pinheiro, P. and 10 others. 2014. socioeconomic impact of cystic The disease burden of hepatitis B, echinococcosis. Emerging Infectious influenza, measles and salmonellosis Diseases, 12(2): 296–303. in Germany: first results of the 207. Bresson-Hadni, S., Vuitton, D.A., Burden of Communicable Diseases Bartholomot, B., Heyd, B., Godart, in Europe Study. Epidemiology and D., Meyer, J.P., Hrusovsky, S., Becker, Infection, 142(10): 2024–2035. M.C., Mantion, G., Lenys, D. & Miguet, 201. Vally, H., Glass, K., Ford, L., Hall, G., J.P. 2000. A twenty-year history of Kirk, M.D., Shadbolt, C., Veitch, M., alveolar echinococcosis: analysis Fullerton, K.E., Musto, J. & Becker, N. of a series of 117 patients from 2014. Proportion of illness acquired eastern France. European Journal by foodborne transmission for nine of Gastroenterology & Hepatology, enteric pathogens in Australia: An 12(3): 327–336. Expert Elicitation. 2014. Foodborne 208. Torgerson, P.R., Schweiger, A., Pathogens and Disease, 11(9): Deplazes, P., Pohar, M., Reichen, 727–733. J., Ammann, R.W., Tarr, P.E., Halkik, 202. Ravel, A., Davidson, V.J., Ruzante, N., Mullhaupt, B. 2008. Alveolar J.M. & Fazil, A. 2010. Foodborne echinococcosis: From a deadly 240

disease to a well-controlled 214. Torrey, E.F., Bartko, J.J., Lun, Z.R. infection. Relative survival and & Yolken, R.H. 2007. to economic analysis in Switzerland Toxoplasma gondii in patients with over the last 35 years. Journal of schizophrenia: A meta-analysis. Hepatology, 49(1): 72–77. Schizophrenia Bulletin, 33(2): 729–736. 209. Bruno, E., Bartoloni, A., Zammarchi, L., Strohmeyer, M., Bartalesi, F., 215. Torrey, E.F., Bartko, J.J. & Yolken, Bustos, J.A., Santivanez, S., Garcia, R.H. 2012. Toxoplasma gondii and H.H. & Nicoletti, A. and the COHEMI other risk factors for schizophrenia: Project Study Group. 2013. Epilepsy an update. Schizophrenia Bulletin, and neurocysticercosis in Latin 38(3): 642–647. America: A systematic review and 216. Yonghua, Z., Yue, C., Lei, Z. & meta-analysis. PLOS Neglected Qi, G. 2011. Analysis of probable Tropical Diseases, 7(10): Article correlation between Toxoplasma Number e2480. gondii and schizophrenia: a sero- 210. Montano, S.M., Villaran, M.V., epidemiological longitudinal Ylquimiche, L., Figueroa, J.J., investigation from 2002 to 2007 Rodriguez, S., Bautista, C.T., in Suzhou and Wuxi regions, Gonzalez, E., Tsang, V.C.W., Jiangsu, China. Tropical Medicine Gilman, R.H., Garcia, H.H. and the & International Health, 16(Suppl.1; Cysticercosis Working Group. 2005. Special Issue): 367–368. Neurocysticercosis– Association 217. Dorny, P. & Praet, N. 2007. Taenia between seizures, serology, and saginata in Europe. Veterinary brain CT in rural Peru. Neurology, Parasitology, 149(1-2): 22–24. 65(2): 229–234. 218. Pereira, K.S., Sciimidt, F.L., Guaraldo, 211. Thiebaut, R., Leproust, S., Chene, A.M.A., Franco, R.M.B., Dias, V.L. G., Gilbert, R. and the SYROCOT & Passos, L.A.C. 2009. Chagas’ Study Group. 2007. Effectiveness of Disease as a foodborne illness. prenatal treatment for congenital Journal of Food Protection, toxoplasmosis: a meta-analysis of 72(2): 441–446. individual patient’s data. Lancet, 369(9556): 115–122. 219. Alarcón de Noya, B., Noya González, O. & Robertson, L.J. 212. Wong, W.K., Upton, A. & Thomas, 2016. Trypanosoma cruzi as a M.G. 2013. Neuropsychiatric foodborne pathogen. Springer symptoms are common in International Publishing. immunocompetent adult patients with Toxoplasma gondii acute 220. Abanyie, F., Harvey, R.R., Harris, lymphadenitis. Scandinavian Journal J.R. and 24 others, and the of Infectios Disease, 45(5): 357–361. Multistate Outbreak Investigation Team. 2015. 2013 213. Demar, M., Ajzenberg, D., Maubon, D. multistate outbreaks of Cyclospora and 11 others. 2007. Fatal outbreak cayetanensis infections associated of human toxoplasmosis along with fresh produce: focus on the the maroni river: Epidemiological, Texas investigations. Epidemiology clinical, and parasitological aspects. and Infection, Clinical Infectious Diseases, http://dx.doi.org/10.1017/ 45(7): E88–E95. S0950268815000370 8 p. 241

221. Hall, R.L., Jones, J.L., Hurd, S., viverrini. Parasitology International, Smith, G., Mahon, B.E. & Herwaldt, 61(1; Special Issue): 38–45. B.L. 2012. Population-based 228. WHO, 2015. Global burden of active surveillance for Cyclospora disease regions used for WHO- infection– United States, Foodborne CHOICE analyses. Cost effectiveness Diseases Active Surveillance and strategic planning. World Network (FoodNet), 1997–2009. Health Organization. Clinical Infectious Diseases, 54(Suppl.5): S411–S417 229. Wild, C.P. & Hall, A.J. 2000. Primary prevention of hepatocellular 222. Adams E, Yoder J, Gould LH, carcinoma in developing countries. Hlavsa MC, Gargano J. (2015) Mutation Research– Reviews in Giardiasis outbreaks in the United Mutation Research, 462(2-3): States, 1971-2011. In: Proceedings 381–393. of the International Conference on Emerging Infectious Diseases, 230. Haagsma, J.A., Polinder, S., August 24-26, 2015, Atlanta Georgia. Stein, C.E. & Havelaar, A.H. 2013. P105. Available from: Systematic review of foodborne www.iceid.org/images/2015/ burden of disease studies: ICEID2015_web.pdf Quality assessment of data and methodology. International Journal 223. Robertson, L.J., van der Giessen, of Food Microbiology, 166(1): 34–47. J.W.B., Batz, M.B., Kojima, M. & Cahill, S. 2013. Have foodborne parasites 231. Polinder, S., Haagsma, J.A., Stein, C. finally become a global concern? & Havelaar, A.H. 2012. Systematic Trends in Parasitology, 29(3): review of general Burden of disease 101–103. studies using disability-adjusted life years. Population Health Metrics, 224. FAO/WHO. 2014. Multicriteria- 10: Art 21. based ranking for risk management of food-borne parasites. . Food 232. Mathers, C.D., Vos, T., Lopez, A.D., and Agriculture Organization of Salomon, J., Ezzati, M. (ed.) 2001. the United Nations/World Health National Burden of Disease Studies: Organization. Rome, Italy. 302 p. A Practical Guide. Edition 2.0. Global Programme on Evidence for Health 225. Craig, P.S. and C. Echinococcosis Policy. WHO, Geneva. Available Working Group. 2006. Epidemiology at http://www.who.int/healthinfo/ of human alveolar echinococcosis nationalburdenofdiseasemanual.pdf in China. Parasitology International, Accessed 2015-10-22. 55(Suppl.): S221–S225. 233. Etienne, C.F. 2013. Supplement 226. Al Shahrani, D., Frayha, H.H., on Caribbean burden of illness Dabbagh, O., Al Shail, E. 2003. study. Foreword. Journal Of Health, First case of neurocysticercosis in Population and Nutrition,31(4; Saudi Arabia. Journal of Tropical Suppl.1): 1–2. Pediatrics, 49(1): 58–60. 234. Valenciano, M., Bergeri, I., Jankovic, 227. Petney, T., Sithithaworn, P., Andrews, D., Milic, N., Parlic, M. & Coulombier, R., Kiatsopit, N., Tesana, S., Grundy- D. 2004. Strengthening early Warr, C. & Ziegler, A. 2012. The warning function of surveillance ecology of the Bithynia– first in the Republic of Serbia: intermediate hosts of Opisthorchis 242

lessons learned after a year of Available from: http://www. implementation. Euro surveillance, ugansociety.org/wp-content/.../ 9(5): 24–26. FERG-Presentation-April-2015.ppt 235. Fabiana, A., Donia, D., Gabrieli, R., 241. Pariyo, G.W., Ekirapa-Kiracho, E., Petrinca, A.R., Cenko, F., Bebeci, Okui, O., Rahman, M.H., Peterson, D., Altan, A.M.D., Buonomo, E. & S., Bishai, D.M., Lucas, H. & Peters, Divizia, M. 2007. Influence of enteric D.H. 2009. Changes in utilization viruses on gastroenteritis in Albania: of health services among poor Epidemiological and molecular and rural residents in Uganda: analysis. Journal of Medical Virology, are reforms benefitting the 79(12): 1844–1849. poor? International Journal for Equity in Health, 8: Article No. 39. 236. Sejdini, A., Mahmud, R., Lim, DOI: 10.1186/1475-9276-8-39 Y.A.L., Mahdy, M., Sejdini, F., Gjoni, V., Xhaferraj, K. & Kasmi, G. 2011. 242. Devleesschauwer, B., Ale, A., Intestinal parasitic infections among Duchateau, L., Dorny, P., Lake, R., children in central Albania Annals of Dhakal, P., Pun, S.B., Pandey, B.D. & Tropical Medicine and Parasitology, Speybroeck, N. 2013. Understanding 105(3): 241–250. the burden of disease in Nepal: a call for local evidence. Journal 237. UNICEF. No date. Albania: Children of Nepal Health Research in Albania Health Access. Available Council,11(24): 221–224. from: http://www.unicef.org/albania/ children_24931.html Accessed 243. Kang’ethe, E. 2011 Situation 2015-10-22. analysis: Improving food safety in the maize value chain in Kenya. 238. Kumagai, Y., Gilmour, S., Ota, E., Report prepared for FAO. 89 p. Momose, Y., Onishi, T., Bilano, Available from: http://www.fao.org/ V.L.F., Kasuga, F., Sekizaki, T. & fileadmin/user_upload/agns/pdf/ Shibuya, K. 2015. Estimating the WORKING_PAPER_AFLATOXIN_ burden of foodborne diseases in REPORTDJ10thOctober.pdf Japan. Bulletin of the World Health Accessed 2015-10-22. Organization, 93(8): 540–549. 244. Bennett, G. & Jessani, N.E. 2011. 239. Tangcharoensathien, The Knowledge Translation V., Limwattananon, S., Toolkit. Bridging the Know–Do Patcharanarumol, W. & Gap: A Resource for Researchers. Thammatacharee, J. 2014. SAGE/IDRC 2011. Available Monitoring and evaluating progress from: http://www.idrc.ca/ towards universal health coverage in EN/Resources/Publications/ Thailand. Plos Medicine, 11(9): Article Pages/IDRCBookDetails. No. e1001726. DOI: 10.1371/journal. aspx?PublicationID=851 pmed.1001726 Accessed 2015-10-22. 240. Agaba, F. 2015. FERG: The WHO 245. IHME [Institute for Health Metrics Initiative To Estimate The Global and Evaluation]. Online. Country Burden of foodborne diseases. Profiles, 2013. Available from http:// Uganda. The MOH/WHO/FAO/ www.healthdata.org/results/ FERG Pilot Country Study. country-profiles. Accessed [accessed 20 August 2015.]. 2015-10-22. 243

246. Codex Alimentarius Commission. Client Report FW11006. A report 2007. Principles and guidelines for the New Zealand Ministry of for the conduct of microbiological Agriculture and Forestry. ESR, risk management (MRM). CAC/ Christchurch. Available from: http:// GL 63–2007. Available from: www. www.foodsafety.govt.nz/elibrary/ codexalimentarius.org/input/ industry/estimates-burden- download/standards/10741/ foodborne-disease-2011.pdf CXG_063e.pdf Accessed 253. Zhong, Z.J., Yu, S., Wang, X.C., Dong, 2015-10-22. S.C., Xu, J., Wang, Y.F., Chen, Z.L., 247. Ezzati, M., Lopez, A.D., Rodgers, A., Ren, Z.H. & Peng, G.N. 2013. Human Vander Hoorn, S. & Murray, C.J.L. brucellosis in the People’s Republic 2002. Selected major risk factors of China during 2005–2010. and global and regional burden of International Journal of Infectious disease Lancet, 360(9343): Diseases, 17(5): E289–E292. 1347–1360. DOI: 10.1016/j.ijid.2012.12.030 248. OIE [World Organisation for Animal 254. Dean, A.S., Crump, L., Greter, Health]. No date; online. World H., Hattendorf, J., Schelling, Organisation for Animal Health E. & Zinsstag, J. 2012. Clinical Information Database (WAHID). manifestations of human brucellosis: 2013. Available from: http://www. a systematic review and meta- oie.int/wahis_2/public/wahid.php/ analysis. PLOS Neglected Tropical Wahidhome/Home Accessed Diseases, 6(12): Article e1929. 2015-10-22. 255. Al Dahouk, S., Neubauer, H., Hensel, 249. EFSA [European Food Safety A., Schoneberg, I., Nockler, K., Authority]. 2012. The European Alpers, K., Merzenich, H., Stark, Union summary report on trends K., Jansen, A. 2007. Changing and sources of zoonoses, zoonotic epidemiology of human brucellosis, agents and food-borne outbreaks in Germany, 1962–2005. Emerging the European Union in 2010. EFSA Infectious Diseases, 13(12): Journal, 10(3): 2597. 1895–1900. 250. Pappas, G., Papadimitriou, P., 256. Al-Tawfiq, J.A. 2008. Therapeutic Akritidis, N., Christou, L. & Tsianos, options for human brucellosis. E.V. 2006. The new global map of Expert Review of Anti-Infective human brucellosis. Lancet Infectious Therapy, 6(1): 109–120. Diseases, 6(2): 91–99. 257. Buzgan, T., Karahocagil, M.K., Irmak, 251. Dean, A.S., Crump, L., Greter, H., H., Baran, A.I., Karsen, H., Evirgen, Schelling, E. & Zinsstag, J. 2012. O. & Akdeniz, H. 2010. Clinical Global burden of human brucellosis: manifestations and complications A systematic review of disease in 1028 cases of brucellosis: a frequency. PLOS Neglected Tropical retrospective evaluation and review Diseases, 6(10): Article No. e1865. of the literature. International DOI: 10.1371/journal.pntd.0001865 Journal of Infectious Diseases, 14(6): e469–478. 252. Cressey, P. & Lake, R. 2011. Estimated incidence of foodborne illness 258. Muller, B., Durr, S., Alonso, S., in New Zealand: application of Hattendorf, J., Laisse, C.J.M., overseas models and multipliers. Parsons, S.D.C., van Helden, 244

P.D.& Zinsstag, J. 2013. Zoonotic health professionals–Hepatitis A. Mycobacterium bovis-induced Available at http://ecdc.europa. tuberculosis in humans. Emerging eu/en/healthtopics/hepatitis_A/ Infectious Diseases, 19(6): 899–908. factsheets/Pages/factsheet- professionals.aspx Accessed 259. Durr, S., Muller, B., Alonso, S., 2015-10-15. Hattendorf, J., Laisse, C.J.M., van Helden, P.D. & Zinsstag, J. 2013. 265. Majowicz, S.E., Scallan, E., Jones- Differences in primary sites of Bitton, A., Sargeant, J.M., Stapleton, infection between zoonotic and J., Angulo, F.J., Yeung, D.H., Kirk, human tuberculosis: results from M.D. 2014. Global incidence of a worldwide systematic review. human Shiga toxin-producing PLOS Neglected Tropical Diseases, Escherichia coli infections and 7(8): Art e2399. Deaths: A systematic review and knowledge synthesis. Foodborne 260. Buckle, G.C., Walker, C.L.F. & Pathogens and Disease, 11(6): 447– Black, R.E. 2012. Typhoid fever 455. and paratyphoid fever: Systematic review to estimate global morbidity 266. Lamberti, L.M., Fischer Walker, C.L. & and mortality for 2010. Journal of Black, R.E. 2012. Systematic review Global Health, 2(1): 010401. of diarrhoea duration and severity in children and adults in low- and 261. Torgerson, P.R., Devleesschauwer, middle-income countries. BMC B., Praet, N., Speybroeck, N., Public Health, 12: Art. no. 276. Willingham, A.L., Kasuga, F., Rokni, M.B., Zhou, X.-N., Fèvre, E.M., Sripa, 267. ECDC [European Centre of Disease B., Furst, T., Budke, C.M., Carabin, Prevention and Control]. No date; H., Kirk, M.D., Angulo, F.J., Havelaar, online. Disease specific information: A. & de Silva, N. 2015. World Shiga/Vero toxin producing Health Organization estimates of Escherichia coli infection (STEC/ the global and regional disease VTEC). Available from: http:// burden of 11 foodborne parasitic ecdc.europa.eu/en/healthtopics/ diseases, 2010: a data synthesis. escherichia_coli/basic_facts/Pages/ PLOS Medicine, DOI:10.1371/journal. basic_facts.aspx. Accessed pmed.1001920 2015-10-22. 262. Crump, J.A., Luby, S.P. & Mintz, E.D. 268. Havelaar, A.H., Van Duynhoven, 2004. The global burden of typhoid Y.T.H.P., Nauta, M.J., Bouwknegt, fever. Bulletin of the World Health M., Heuvelink, A.E., De Wit, G.A., Organization, 82(5): 346–353. Nieuwenhuizen, M.G.M. & Van De Kar, N.C.A. 2004. Disease burden in 263. Butler, T., Islam, A., Kabir, I. & Jones, The Netherlands due to infections P.K. 1991. Patterns of morbidity with Shiga toxin-producing and mortality in typhoid fever Escherichia coli O157. Epidemiology dependent on age and gender: and Infection, 132(3): 467–484. review of 552 hospitalized patients with diarrhoea. Reviews of 269. Varma, J.K., Katsitadze, G., Infectious Diseases, 13(1): 85–90. Moiscrafishvili, M., Zardiashvili, T., Chokheli, M., Tarkhashvili, N., 264. ECDC [European Centre for Jhorjholiani, E., Chubinidze, M., Disease Prevention and Control]. Kukhalashvili, T., Khmaladze, I., No date [online]. Factsheet for 245

Chakvetadze, N., Imnadze, P. & 276. Foodborne Diseases Active Sobel, J. 2004. Foodborne botulism Surveillance Network (FoodNet). in the republic of Georgia. Emerging No date; online. Available from: Infectious Diseases, 10(9): http://www.cdc.gov/foodnet 1601–1605. Accessed 2015-10-22. 270. Czerwinski, M., Czarkowski, M.P. 277. Ao, T.T., Feasey, N.A., Gordon, M.A., & Kondej, B. 2012. [Botulism in Keddy, K.H., Angulo, F.J. & Crump, Poland in 2010]. Przegl Epidemiol, J.A. 2015. Global Burden of Invasive 66(2): 267–271. Nontyphoidal Salmonella Disease, 2010. Emerging Infectious Diseases, 271. Sobel, J., Tucker, N., Sulka, A., 21(6): 941–949. McLaughlin, J. & Maslanka, S. 2004. Foodborne botulism in the United 278. Biggs, H.M., Lester, R., Nadjm, B., States, 1990–2000. Emerging Mtove, G., Todd, J.E., Kinabo, G.D., Infectious Diseases, 10(9): 1606–1611. Philemon, R., Amos, B., Morrissey, A.B., Reyburn, H. & Crump, 272. Kirk, M., Ford, L., Glass, K. & Hall, G. J.A. 2014. Invasive Salmonella 2014.Foodborne Illness, Australia, infections in areas of high and Circa 2000 and Circa 2010. low malaria transmission intensity Emerging Infectious Diseases, in Tanzania. Clinical Infectious 20(11): 1857–1864. Diseases, 58(5):638–647. 273. Bennett, S.D., Walsh, K.A. & Gould, 279. Koch, K., Kristensen, B., Holt, L.H. 2013. Foodborne disease H.M., Ethelberg, S., Molbak, K. & outbreaks caused by Bacillus Schonheyder, H.C. 2011. International cereus, Clostridium perfringens, travel and the risk of hospitalization and Staphylococcus aureus– with non-typhoidal Salmonella United States, 1998–2008. Clinical bacteremia. A Danish population- Infectious Disease, 57(3): 425–433. based cohort study, 1999–2008. 274. Fischer Walker, C., Sack, D. & Black, BMC Infectious Diseases, 11: Art. No. R. 2010. aetiology for diarrhoea 277. DOI: 10.1186/1471-2334-11-277. in older children, adolscents, and 280. Poropatich, K., Fischer Walker, C. adults: a systematic review. PLOS & Black, R. 2010. Quantifying the Neglected Tropical Diseases, association between Campylobacter 4(8):Art. No. e768. doi:10.1371/ infection and Guillain Barre journal.pntd.0000768. syndrome: a systematic review. 275. Lamberti, L.M., Bourgeois, A.L., Journal of Health, Population and Fischer-Walker, C.L.F., Black, R.E, & Nutrition, 28(6): 545–552. Sack, D. 2014. Estimating diarrhoeal 281. ECDC. No date; online. illness and Deaths attributable Disease-specific information: to Shigellae and Enterotoxigenic Campylobacteriosis. Available Escherichia coli among older from: http://ecdc.europa.eu/en/ children, adolescents, and adults healthtopics/Campylobacteriosis/ in South Asia and Africa. PLOS basic_facts/Pages/Factsheet_ Neglected Tropical Diseases, generalpublic.aspx Accessed 8(2): Art. No. e2705. DOI: 10.1371/ 2015-10-22. journal.pntd.0002705 282. Fischer Walker, C.L., Sack, D. & Black, R.E. 2010. aetiology 246

of diarrhoea in older children, Jakarta, Indonesia: findings from adolescents and adults: a systematic 24 months surveillance. BMC review. PLOS Neglected Tropical Infectious Diseases, 5: Art. No. 89. Diseases, 4(8): Art e768. Available DOI: 10.1186/1471-2334-5-89 at http://www.plosntds.org/article/ 289. Lucas, M.E.S., Deen, J.L., von info%3Adoi%2F10.1371%2Fjournal. Seidlein, L., and 15 others. 2005. pntd.0000768 Accessed 2015-10-16. Effectiveness of mass oral cholera 283. Ahmed, S.M., Hall, A.J., Robinson, in Beira, Mozambique. A.E., Verhoef, L., Premkumar, New England Journal of Medicine, P., Parashar, U.D., Koopmans, 352(8): 757–767. M., Lopman, B.A. 2014. Global 290. Siddique, A.K., Nair, G.B., Alam, M., prevalence of norovirus in cases of and 11 others. 2010. El Tor cholera gastroenteritis: a systematic review with severe disease: a new threat to and meta-analysis. Lancet Infectious Asia and beyond. Epidemiology And Diseases, 14(8): 725–730. Infection, 138(3): 347–352. 284. Phillips, G., Tam, C.C., Conti, 291. Valcin, C.L., Severe, K., Riche, C.T., S., Rodrigues, L.C., Brown, D., and 8 others. 2013. Predictors of Iturriza-Gomara, M., Gray, J. & disease severity in patients admitted Lopman, B. 2010. Community to a cholera treatment center in incidence of norovirus-associated urban Haiti. American Journal of infectious intestinal disease in Tropical Medicine and Hygiene, England: improved estimates 89(4): 625–632. using viral load for norovirus diagnosis. American Journal of 292. Gujral, L., Sema, C., Rebaudet, Epidemiology, 171(9):1014–1022. S., Taibo, C.L.A., Manjate, A.A., Piarroux, R., Gessner, 285. Ali, M., Lopez, A.L., You, Y.A., Kim, B.D. & Jani, I.V. 2013. Cholera Y.F., Sah, B., Maskery, B. & Clemens, epidemiology in Mozambique J. 2012. The global burden of using national surveillance data. cholera. Bulletin of the World Health Journal of Infectious Diseases, Organization, 90(3): 209–218. 208(Suppl.1): S107–S114. 286. WHO. 2015. Population with 293. Qadri, F., Khan, A.I., Faruque, sustainable access to improved A.S.G., Begum, Y.A., Chowdhury, sanitation. Available from: http:// F., Nair, G.B., Salam, M.A., Sack, apps.who.int.whosis/data/. D.A. & Svennerholm, A.M. 2005. 287. Sur, D., Deen, J.L., Manna, B. and 12 Enterotoxigenic Escherichia coli others. 2005. The burden of cholera and Vibrio cholerae diarrhoea, in the slums of Kolkata, India: data Bangladesh, 2004. Emerging from a prospective, community Infectious Diseases, 11(7): 1104–1107. based study. Archives of Disease in 294. Yoder, J.S., Wallace, R.M., Collier, Childhood, 90(11): 1175–1181. S.A., Beach, M.J. & Hlavsa, M.C. 2012. 288. Agtini, MD (Agtini, MD); Soeharno, Cryptosporidiosis surveillance– R (Soeharno, R); Lesmana, M United States, 2009–2010. Morbidity (Lesmana, M); and 23 others. and mortality weekly report. 2005. The burden of diarrhoea, Surveillance summaries,61(5): 1–12. shigellosis, and cholera in North 247

295. Yoder, J.S., Gargano, J.W., Wallace, R. and the Toxoplasmosis Study R.M. & Beach, M.J. 2012. Giardiasis Group. 2010. Chorioretinal lesions in surveillance– United States, 2009– mothers of children with congenital 2010. Morbidity and mortality toxoplasmosis in the National weekly report. Surveillance Collaborative Chicago-based, summaries, 61(5): 13–23. Congenital Toxoplasmosis Study. Scientia Medica (Porto Alegre), 296. Havelaar, A.H., Kemmeren, J.M. & 20(1): 20-26. Kortbeek, L.M. 2007. Disease burden of congenital toxoplasmosis. Clinical 302. Usubalieva, J., Minbaeva, G., Infectious Diseases, 44(11): Ziadinov, I., Deplazes, P. & Torgerson, 1467–1474. P.R. 2013. Human Alveolar Echinococcosis in Kyrgyzstan. 297. Kortbeek, L.M., Hofhuis, A., Nijhuis, Emerging Infectious Diseases, C.D.M. & Havelaar, A.H. 2009. 19(7): 1095–1097. Congenital toxoplasmosis and DALYs in the Netherlands. Memorias 303. Nahorski, W.L., Knap, J.P., Pawlowski, do Instituto Oswaldo Cruz, Z.S., and 18 others. 2013. Human 104(2): 370–373. Alveolar Echinococcosis in Poland: 1990-2011. PLOS Neglected Tropical 298. Hens, N., Aerts, M., Faes, C., Shkedy, Diseases, 7(1): Art. No. e1986. Z., Lejeune, O., van Damme, P. DOI: 10.1371/journal.pntd.0001986 & Beutels, P. 2010. Seventy-five years of estimating the force 304. IHME [Institute for Health Metrics of infection from current status and Evaluation]. DisMod 3. Available data. Epidemiology and Infection, from: http://winthrop.ihme. 138(6): 802–812. washington.edu 299. Balasundaram, M.B., Andavar, R., 305. IARC [International Agency for Palaniswamy, M. & Venkatapathy, Research on Cancer]. No date. N. 2010. Outbreak of Acquired GLOBOCAN 2012: All Cancers Ocular Toxoplasmosis involving (excluding non-melanoma skin 248 patients. Archives of cancer): Estimated cancer Ophthalmology, 128(1): 28–32. incidence, mortality and prevalence worldwide in 2012 [Online]. [16 300. Neves, E.S., Bicudo, L.N., Curi, March 2012]. Available from: http:// A.L., Carregal, E., Bueno, W.F., globocan.iarc.fr/Pages/fact_sheets_ Ferreira, R.G., Amendoeira, M.R., cancer.aspx Benchimol, E. & Fernandes, O. 2009. Acute acquired toxoplasmosis: 306. Yeh, F.S., Yu, M.C., Mo, C.C., Luo, clinical-laboratorial aspects and S., Tong, M.J. & Henderson, B,E. ophthalmologic evaluation in a 1989. Hepatitis B virus, aflatoxins, cohort of immunocompetent and hepatocellular carcinoma in patients. Memorias do Instituto southern Guangxi, China. Cancer Oswaldo Cruz, 104(2): 393–396 Research, 49(9): 2506–2509. 301. Noble, A.G., Latkany, P., 307. Wu-Williams, A.H., Zeise, L. & Kusmierczyk, J., Mets, M., Rabiah, Thomas, D. 1992. Risk assessment P., Boyer, K., Jalbrzikowski, J., for aflatoxin B1: a modeling Wroblewski, K., Karrison, T., Swisher, approach. Risk Analysis, 12(4): C.N., Mieler, W.F., Meier, P., McLeod, 559–567. 248

308. WHO. No date [online]. Global 310. WHO. 2013. WHO methods and Health Estimates by Cause, Age, Sex data sources for global burden and Country, 2000–2012. Available of disease estimates 2000–2011. from: http://www.who.int/entity/ Global Health Estimates Technical healthinfo/global_burden_disease/ Paper, 2013-4. Doc. ID: WHO/HIS/ HSI/GHE/2013.4. Available at http:// 309. Tshala-Katumbay, D., Banea- www.who.int/healthinfo/statistics/ Mayambu, J.-P., Kazadi-Kayembe, GlobalDALYmethods_2000_2011.pdf T., Nunga-Matadi, R., Bikangi- Accessed 2015-10-23. Nkiabungu, F., Edebol Eeg-Olofsson, K. & Tylleskär, T. 2001. African 311. Yu, S.C., Tan, F., Zhou, M.G., Liu, Journal of Neurological Science, S.W., Zhu, X.J. & Zhu, Y.L. 2014. 20(1): 8–12. Global Burden of Disease, Injury and Risk Factor Study 2010: Its Policy Implications for China. Biomedical and Environmental Sciences, 27(1): 45–48. 249

LIST OF TABLES

Table 1. FERG hazards, causally related health states and corresponding disability weights (DWs). Details on the derivation of the DWs are provided in Appendix 4...... 43 Table 2. Foodborne hazards, and structure of the expert panels...... 46 Table 3. Examples of calibration seed questions ...... 48 Table 4. Exposure routes included in the expert elicitation, per hazard ...... 51 Table 5. Modelling strategies for the hazards included in the WHO global burden of foodborne disease estimates ...... 53 Table 6. The number of experts enrolled, interviewed and finally included in the elicitation across panels ...... 63 Table 7. Median global number of foodborne illnesses, deaths, Years Lived with Disability (YLDs), Years of Life Lost (YLLs) and Disability Adjusted Life Years (DALYs), with 95% uncertainty intervals, 2010...... 73 Table 8. Median rates of foodborne Disability Adjusted Life Years (DALYs) per 100 000 population, by subregion, with 95% uncertainty intervals, 2010...... 78 Table 9. Comparisons of the total burden of parasitic diseases (foodborne and non-foodborne) with 95% uncertainty intervals, estimated by FERG and by GBD2010 [9] ...... 108 Table A7.1 Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by Campylobacter spp., non-typhoidal Salmonella spp., Shiga-toxin producing Escherichia coli (STEC), Brucella spp. and Shigella spp. through each exposure pathway...... 199 Table A7.2 Subregional estimates (median and 95% uncertainty interval) of the proportion of Diarrhoeal Disease illnesses caused by four hazards: enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), Cryptosporidium spp. and Giardia spp. through each exposure pathway...... 202 Table A7.3 Subregional estimates (median and 95% uncertainty interval) of the proportion of Diarrhoeal Disease illnesses caused by Salmonella Typhi, Vibrio cholerae, Entamoeba histolytica, norovirus, and hepatitis A virus through each exposure pathway...... 204 Table A7.4 Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by Toxoplasma gondii, Echinococcus multilocularis, Echinococcus granulosus and Ascaris spp. through each exposure pathway...... 206 Table A7.5 Subregional estimates (median and 95% uncertainty interval) of the proportion of illnesses caused by exposure to lead through each pathway...... 209 Table A7.6 Percent of illness acquired through the foodborne transmission route for six national studies and this studya...... 210 Table A8.1 Median number of foodborne Illnesses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% uncertainty intervals, 2010...... 211 Table A8.2 Median rates of foodborne illnesses, deaths and Disability Adjusted Life Years (DALYs) per 100 000 persons, by region, with 95% uncertainty intervals, 2010...... 213 Table A8.3 Median number of Ill nesses, Deaths, and Disability Adjusted Life Years (DALYs) by age group, with 95% uncertainty intervals, 2010...... 215 Table A8.4 Median rate per 100 000 of foodborne Illnesses, Deaths and Disability Adjusted Life Years (DALYs) by region, with 95% uncertainty intervals, 2010...... 217 TableA8.5 Median number of total and foodborne Illnes ses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% uncertainty intervals, 2010 ...... 219 Table A8.6 Median number of foodborne Illnes ses, Deaths, and Disability Adjusted Life Years (DALYs), with 95% Uncertainty Intervals, 2010 ...... 221 Table A8.7 Median rate per 100 000 foodborne (FB) Ill nesses, Deaths, and Disability Adjusted Life Years (DALYs) by region, with 95% uncertainty intervals, 2010...... 221 250

LIST OF FIGURES

Figure 1. Structure of the initiative to estimate the global burden of foodborne diseases ...... 6 Figure 2. Hazards for which burden of foodborne disease estimates were prepared by FERG, grouped according to TF. Hazards in grey boxes were addressed by individual TFs but were not included in the global overview. Hazards in blue boxes are pending...... 31 Figure 3. Geographical distribution of experts according to working experience (>3 years) per subregion. Several experts had experience in more than one subregion...... 64 Figure 4. Statistical accuracy versus informativeness of the experts included, when using equal weight (blue) or performance weight (red) combinations, respectively...... 65 Figure 5. Subregional estimates of the proportion of foodborne illnesses caused by Campylobacter spp., non-typhoidal Salmonella spp., Shiga-toxin producing Escherichia coli (STEC), Brucella spp. and Shigella spp. Indicated on the line plot are the 2.5th, 5th, 50th, 95th and 97.5th percentiles...... 67 Figure 6. Subregional estimates of the proportion of foodborne illnesses caused by enteropathogenic E. (EPEC), enterotoxigenic E. coli (ETEC), Cryptosporidium spp. and Giardia spp...... 68 Figure 7. Subregional estimates of the proportion of foodborne illnesses caused by typhoidal Salmonella, Vibrio cholerae, Entamoeba histolytica, norovirus, and hepatitis A virus...... 69 Figure 8. Subregional estimates of the proportion of foodborne illnesses caused by Toxoplasma gondii, Echinococcus multilocularis, Echinococcus granulosus and Ascaris spp...... 70 Figure 9. Subregional estimates of the proportion of disease caused by foodborne exposure to lead...... 71 Figure 10. Subregional estimates (medians) of the proportion of disease caused by exposure to lead through eight different exposure routes...... 71 Figure 11. Ranking of foodborne hazards, based on Disability-Adjusted Life Years at the global level, with 95% uncertainty intervals, 2010...... 76 Figure 12. The global burden of foodborne disease (DALYS per 100 000 population) by hazard groups and by subregion, 2010...... 80 Figure 13. Relative contribution of Years of Life Lost due to premature mortality (YLL) and Years Lived with Disability (YLD) to the global burden of 31 hazards in food, 2010...... 81 Figure 14. Age-distribution of disability adjusted life years for 31 hazards contributing to the global burden of foodborne disease, 2010...... 82 Figure 15. Scatterplot of the global burden of foodborne disease per 100 000 population and per incident case...... 83 Figure 16. The global burden of foodborne disease by subregion (DALYS per 100 000 population) caused by enteric hazards, 2010...... 85 Figure 17. Disability Adjusted Life Years for each pathogen acquired from contaminated food ranked from lowest to highest with 95% Uncertainty Intervals, 2010...... 86 Figure 18A. The relative contribution to the DALY incidence by each agent for each of the subregions. This includes enteric protozoa to complete the picture on foodborne parasitic diseases. However the detail is reported in the accompanying manuscript on foodborne enteric pathogens [168]...... 87 Figure 18B. Disability Adjusted Life Years for each parasite acquired from contaminated food ranked from lowest to highest with 95% Uncertainty Intervals, 2010. This includes enteric protozoa to complete the picture on foodborne parasitic diseases. However the detail is reported in the accompanying manuscript on foodborne enteric pathogens [168]...... 88 Figure 19. The relative proportion of the burden of each of the foodborne parasitic diseases contributed by YLLs and YLDs ...... 88 Figure 20. The relative contribution to the DALY incidence by each of four chemicals for each of the WHO Regions...... 90 Figure 21. The relative contributions from YLLs and YLDs for each of four chemicals...... 90 Figure 22. Disability Adjusted Life Years for each of four chemicals from contaminated food ranked, from lowest to highest, with 95% uncertainty intervals ...... 91 Figure 23. Major transmission routes of human foodborne diseases indicate two points of attribution: the reservoir level and the exposure level...... 101 Figure A6.1 FERG hazards, causally related health states and corresponding disability weights (DWs). The fourth column describes how the various DWs were derived from the Global Burden of Disease Studies (GBD) and the World Health Organization Global Health Estimates (WHO/GHE)...... 195 251

GLOSSARY

Foodborne disease Food A foodborne disease (FBD) can According to the Codex Alimentarius be defined as a disease commonly Commission, “food means any substance, transmitted through ingested food. FBDs whether processed, semi-processed comprise a broad group of illnesses, and or raw, which is intended for human may be caused by microbial pathogens, consumption, and includes drink, parasites, chemical contaminants chewing gum and any substance which and biotoxins. has been used in the manufacture, preparation or treatment of food but Burden of disease does not include cosmetics or tobacco In the context of this Initiative, the term or substances used only as drugs”. The “burden of disease” follows the principles definition includes all bottled drinks. of the Global Burden of disease Study, and includes the quantification of Source attribution morbidity, all disabling complications Source attribution (SA) is the partitioning and mortality in a single summary of the human burden of a particular measure (DALY). disease to specific sources. With regards to foodborne diseases, SA can be DALY (disability-adjusted life year) conducted at various points along the A health gap measure that combines the food distribution chain, from the animal years of life lost due to premature death reservoir to the point of consumption. (YLL) and the years lived with disability (YLD) from a disease or condition, for varying degrees of severity, making time itself the common metric for death and disability. One DALY equates to one year of healthy life lost. 252

ABBREVIATIONS

BMD Benchmark Dose BMDL Benchmark Dose lower 5% confidence bound BMDU Benchmark Dose upper confidence limit BoD Burden of Disease BW Body Weight CDC Centers for Disease Control and Prevention [of the United States of America] CE Cystic Echinococcosis CEA Comparative Exposure Assessment CFR Case fatality ratio CHERG Child Health Epidemiology Reference Group CI Confidence Interval CNS Central nervous system COPD Chronic obstructive pulmonary disease CRA Comparative Risk Assessment CSTF Country Studies Task Force CT Congenital Toxoplasmosis CTF Computational Task Force CTTF Chemicals and Toxins Task Force DALY Disability-adjusted life year DOI Declaration of interests DRC Democratic Republic of Congo DW Disability Weight EAggEC Enteroaggerative E. coli ECDC European Centre for Disease Prevention and Control EDTF Enteric Disease Task Force EFSA European Food Safety Authority EPEC Enteropathogenic Escherichia coli ESRD End-stage renal disease ETEC Enterotoxigenic Escherichia coli EU European Union FAO Food and Agriculture Organization of the United Nations 253

FBD Foodborne Diseases FDA United States Food and Drug Administration FERG Foodborne Disease Burden Epidemiology Reference Group FOS [WHO] Department of Food Safety, Zoonoses and Foodborne Diseases GBD Global Burden of Disease GBD2010 Institute of Health Metrics and Evaluation Global Burden of Disease Study, 2010. GBS Guillain-Barré Syndrome GEMS Global Environment Monitoring System GFN Global Foodborne Infections Network HALE Health-Adjusted Life Expectancy HAV hepatitis A virus HBV hepatitis B virus HCC Hepatocellular Carcinoma HUS [STEC] haemolytic uraemic syndrome IARC International Agency for Research on Cancer IHME Institute of Health Metrics and Evaluation iNTS Invasive non-typhoid salmonellosis JECFA Joint FAO/WHO Expert Committee on Food Additives KT Knowledge Translation KTPG Knowledge Translation and Policy Group LE life expectancy LOS Lipo-oligosaccharides MAR “missing at random” MAL-ED Interactions of Malnutrition & Enteric Infections: Consequences for Child Health and Development MDG Millennium Development Goal(s) NBD National Burden of Disease NCC Neurocysticercosis NGO non-governmental organization NTP National Toxicology Program NTS Non-typhoidal Salmonella enterica 254

OIE World Organisation for Animal Health PAF population attributable fraction PAHO Pan American Health Organization PCB Polychlorinated Biphenyl PCR polymerase chain reaction PDTF Parasitic Diseases Task Force RfD Reference Dose RIVM The Dutch National Institute for Public Health and the Environment SA Source attribution SATF Source Attribution Task Force SPS [Agreement on the Application of] Sanitary and Phytosanitary Measures [of the WTO] STEC Shiga toxin-producing Escherichia coli TF task force TWI Tolerable Weekly Intake UI uncertainty interval UN United Nations UNEP United Nations Environment Programme UNICEF United Nations Children’s Funds US EPA United States Environmental Protection Agency USA United States of America USDA United States Department of Agriculture WHA World Health Assembly WHO World Health Organization WTO World Trade Organization YLD years lived with disability YLL years of life lost This report presents the fi rst global and regional estimates of the burden of foodborne diseases. The large disease burden from food highlights the importance of food safety, particularly in Africa, South-East Asia and other regions. Despite the data gaps and limitations of these initial estimates, it is apparent that the global burden of foodborne diseases is considerable, and aff ects individuals of all ages, particularly children <5 years of age and persons living in low-income regions of the world. By incorporating these estimates into policy development at both national and international levels, all stakeholders can contribute to improvements in safety throughout the food chain. These results will also help to direct future research activities.

ISBN 9789241565165