RISK ASSESSMENT OF MYCOTOXINS IN FOOD

1
FAO / WHO Workshop Applying the Risk Analysis
Framework to Address Problems of Mycotoxins on
Foods, Accra, Ghana, 23 February 2009
RISK ASSESSMENT OF MYCOTOXINS IN FOOD
Gordon S Shephard PROMEC Unit Medical Research
Council Tygerberg, South Africa
2
INTRODUCTION -1
Aspects of modern food safety systems

• Science-based
• Risk analysis
• National food safety policies
• Food legislation
• Inspection and monitoring services
• Laboratory services
• Updated and harmonized standards
• Epidemiological surveillance

3
INTRODUCTION - 2

• Food Safety Risk Analysis
• a developing discipline
• a scientific method to assess links between
  hazards in the food chain and human health risks
• several advantages

• scientific
• stakeholder participation
• transparent
• systematic
• harmonized

4
THE GENERIC COMPONENTS OF RISK ANALYSIS
Risk Communication
Risk Assessment
Risk Management
Decisions involving policy and values
Scientific inputs
Social Economic Cultural Ethical Legal Environment
al Political
FAO Food Nutrition paper 87
5
GENERIC FRAMEWORK FOR RISK MANAGEMENT
Preliminary Risk Management Activities
Identify Select Risk Management Options
Monitoring Review
Implement Risk Management Decision
FAO Food Nutrition paper 87
6
GENERIC FRAMEWORK FOR RISK MANAGEMENT
Preliminary Risk Management Activities

• Identify food safety issue
• Develop risk profile (systematic collection of
  all info)
• Establish goals of risk management
• Decide on need for RISK ASSESSMENT
• Establish risk assessment policy
• Commission RISK ASSESSMENT, if necessary
• Consider results of RISK ASSESSMENT
• Rank risks, if necessary

Implement Risk Management Decision
Identify Select Management Options
Monitoring Review
FAO Food Nutrition paper 87
7
INTERACTION OF RISK MANAGERS AND RISK ASSESSORS

• Functional separation separating the tasks
  of each
• Managers and assessors communicate, but do not
  interfere
• Managers have key role in beginning the risk
  assessment process
• Managers assemble an appropriate team of experts
• Managers articulate exact goals of risk
  assessment (purpose and scope)
• Managers ensure appropriate questions are
  answered
• Managers specify the form of the risk assessment
  (qualitative, quantitative, probabilistic, etc)

8
RISK ASSESSORS
The tasks of the assessors

• Perform the RISK ASSESSMENT
• Commissioned and guided by risk managers, they
  perform independent evaluation
• Risk assessments should be structured and
  systematic
• The risk assessment should be based on sound
  science and cognizant of data uncertainty
• Assessors should be unbiased and free of
  conflict of interest

9
PARTICIPATION BY MEMBERS

• Input into the compounds and priorities to be
  evaluated by JECFA
• Input into risk management questions addressed
  by CCCF to JECFA
• Provision of experts for the Risk Assessment
  team established by JECFA
• Provision of data for the Risk Assessment team
• Ensuring the relevance of decisions by assessors
  and managers to their respective countries

10
RISK ASSESSMENT
1. Hazard Assessment 1a. Hazard
identification 1b. Hazard characterization 2.
Exposure Assessment 3. Risk Characterization
11
(No Transcript)
12
HAZARD ASSESSMENT

• Hazard Identification from
• Epidemiological studies in human populations
• Animal toxicity studies
• In vitro assays
• Qualitative structure-activity relationships

• Hazard Characterization Parameters measured
• LD50 (Single dose 50 fatality)
• TD50 (Chronic dose 50 tumours)
• TD05 (Chronic dose 5 tumours)
• NO(A)EL (No observed (adverse) effect level)
• (Obtained from toxicity study)

13
HAZARD CHARACTERIZATION Dose Response Effects
All substances are poisons there is none which
is not a poison. The right dose differentiates a
poison from a remedy Philippus Aureolus
Theophrastus Bombastus von Hohenheim Paracelsus,
1493-1541
14
HAZARD CHARACTERIZATION Purpose of the
Dose-Response Curve

• NON-GENOTOXIC COMPOUNDS (Threshold Effects)
• Estimation of safe dose for humans, i.e.
• Tolerable daily intake (TDI) for mycotoxins
• (using NOAEL, LOAL, threshold data or benchmark
  dose)
• GENOTOXIC COMPOUNDS (Non-threshold Effects)
• Estimate potency in humans
• Extrapolation of the risk from high dose
  animal studies to low dose human exposure

15
HAZARD CHARACTERIZATION NON-GENOTOXIC
MYCOTOXINS Estimation of TDI
Presence of threshold NO(A)EL Tolerable daily
intake, TDI NO(A)EL / Safety factor Generally,
Safety factor 100 10 x 10
Extrapolation from animal to human
Variation between human individuals
16
HAZARD CHARACTERIZATION GENOTOXIC MYCOTOXINS
Absence of threshold Absence of safe
dose Establish the potency of the
contaminant gt From animal data (Slope of
dose-response linear extrapolation) gt From
human epidemiological data Expressed
as cancers/100000 per year per ng/kg body
weight per day
17
EXPOSURE ASSESSMENT IN HUMAN POPULATIONS

• General Factors
• Levels of mycotoxin contamination
• Effects of various food processing methods
• Food consumption by the population
• Data are country-specific
• A distribution of contamination levels
• A distribution of consumption levels
• Mean, median or 90th percentile values
• Data for all persons or for consumers only

18
EXPOSURE ASSESSMENT - 1

• Calculation of actual dietary exposure
• Single Point Data
• Dietary Exposure Mycotoxin level x Consumption
• Body weight
• B. Deterministic Methods
• Model exposure by using the variation in
  consumption
• with a fixed contamination level or vice versa

19
EXPOSURE ASSESSMENT - 2
C. Probabilistic Methods Simulations based on
random sampling of probability distributions of
the input data (eg. Monte Carlo method)
20
RISK CHARACTERIZATION
Integration of Hazard Assessment and Exposure
Assessment into estimate of likelihood of
adverse health effects as a result of mycotoxin
exposure in a population
A. For non-genotoxic mycotoxins Compare TDI
with exposure gt actual (Probable daily
intake, PDI) gt potential (based on
contamination scenarios)
21
RISK CHARACTERIZATION Comparison of actual
exposure with TDI
Deterministic model
22
DEOXYNIVALENOL IN THE NETHERLANDSMonte Carlo
Total exposure distribution
H. Marvin, RIKILT Inst. Food Safety
23
RISK CHARACTERIZATION
B. For genotoxic mycotoxins Population risk
Exposure x Potency

• Assess level of population risk
• Does it require intervention?
• What risk level is considered negligible?
• 1 cancer in 105, 106 or 109?

24
RISK ASSESSMENT FOR MYCOTOXINS

• 1a. Hazard identification Most important
  mycotoxins are known
• 1b. Hazard characterization JECFA has set TDI
  for some.
• JECFA has set potency for aflatoxin B1 and
  M1.
• Exposure assessment Country-specific
• - more data generally needed.
• 3. Risk characterization Extent of the
  mycotoxin problem in individual
  countries.

25
CASE STUDY AFLATOXIN B1
26
SUMMARY OF AFLATOXIN CASE STUDY

• Develop a risk profile of known data for risk
  manager
• The risk manager commissions a risk assessment
• Hazard identification (data already present in
  the risk profile)
• Hazard characterization derive potency estimates
  for carcinogenic mycotoxin
• Exposure assessment
• Risk characterization combine potency with
  exposure to define population risks

27
AFLATOXINS Risk Profile Natural Occurrence

• Produced in the field and in storage
• Major producers Aspergillus flavus, A.
  parasiticus
• Commodities include staple African foods
• Nuts, maize, sorghum, pulses, coconut
• Reported in products from over 20 African
  countries
• Studies on aflatoxin biomarkers in west Africa
  indicate extensive exposure in adults, children
  and prenatal

28
AFLATOXINS Risk Profile Health Effects

• Hepatotoxic, teratogenic, mutagenic and
  carcinogenic
• Acute effects Aflatoxicosis (jaundice, oedema,
  GI haemorrhage, death)
• Immunosuppressive
• Genotoxic Covalently binds to DNA (p53 tumour
  suppressor gene)
• Carcinogenic - Causes liver cancer in animals
• - IARC Group 1 carcinogenic to humans
• Causes stunting in children, role in kwashiorkor?

29
IARC EVALUATION OF AFLATOXINS Evidence for
Carcinogenicity
Sufficient in humans AFB1 Natural mixtures
of AFs Inadequate in humans AFM1 (group
2B) Sufficient in animals AFG1 and AFM1 Limited
in animals AFB2 Inadequate in animals AFG2
Group 1
30
AFLATOXIN B1 RISK ASSESSMENT
Hazard identification Human epidemiological
studies - Correlation between human liver
cancer and aflatoxin exposure (Synergism
with hepatitis B) Animal studies - LD50 in
male rat 7200 µg/kg (oral) - Hepatotoxicity
- Hepatocarcinogenicity
31
AFLATOXIN B1 RISK ASSESSMENT
Hazard characterization (JECFA, 1998) -
Quantitative assessment of genotoxic carcinogen
(JECFA, 1998) - No threshold level and hence no
NOEL - Risk affected by presence of hepatitis B
surface antigen - Potency estimates from
epidemiological data HBsAg 0.3 cancers/year
per 100 000 population per ng AFB1/kg b.w.
per day HBsAg- 0.01 cancers/year per 100 000
population per ng AFB1/kg b.w. per day
32
UNCERTAINTY IN POTENCY ESTIMATES

• Epidemiological data comes from areas of high AF
  and high HBsAg levels (confounding factors)
• Reliability and precision of estimates of AF
  exposure in study populations are unknown
• Shape of the dose-response relationship is
  unknown leading to uncertainty in mathematical
  interpolation
• Exclusion of studies showing no association
  between AF and liver cancer
• Current AF levels do not represent levels at
  time of cancer induction
• Early method limitations in detecting hepatitis
  B
• Non-confirmation of primary liver cancer by
  histology
• Differences between animal species due to
  differences in activation and detoxification

33
AFLATOXIN B1 RISK ASSESSMENT Exposure Assessment
Maize in Africa
Maize consumption 400 g/person/day Max AFB1
level in commercial maize 20 µg/kg or 20
ng/g Dietary exposure Mycotoxin level x
Consumption Body weight 20 x 400/60
ng/kg b.w./day 133.3 ng/kg b.w./day
34
AFLATOXIN B1 RISK ASSESSMENT Risk
Characterization Maize in Africa
Population risk Exposure x Potency HBsAg Popu
lation risk 133.3 x 0.3 40 cancers/year per
105 population HBsAg- Population risk
133.3 x 0.01 1.3 cancers/year per 105
population
35
AFLATOXIN B1 RISK ASSESSMENT Risk
Characterization Maize in Africa
Overall population risk If P Hepatitis B virus
prevalence rate (as fraction) Then, average
Potency 0.01x(1-P) 0.3xP African maize
example If P 0.25 (25 of population is
HBsAg) Average population risk 11 cancers/year
per 105 population
36
Thank You for Your Attention
37
EFFECT OF DIFFERENT REGULATORY LEVELS IN EU - 1
AFB1 potency in EU (1 HBsAg) 0.0129 Exposure
in EU Peanuts, maize their products Compare 2
levels of regulation At 1998 levels of
contamination At 20 ppb limit All products
greater than this removed Exposure in European
diet 19 ng / person / day At 10 ppb limit All
products greater than this removed Exposure in
European diet 18 ng / person / day (From 49
JECFA, 1998)
38
EFFECT OF DIFFERENT REGULATORY LEVELS IN EU - 2
Compare 2 levels of regulation At 1998 levels of
contamination At 20 ppb limit All products
greater than this removed Risk Potency x
Exposure 0.0129 x 19 / 60 0.0041 cancers /
105 population At 10 ppb limit All products
greater than this removed Risk 0.0039 cancers
/ 105 population Difference between the 2 levels
2 per 109 population 1 cancer per year in
the EU (From 49 JECFA, 1998)
39
ACTUAL LIVER CANCER INCIDENCE IN EU
(1995) Males 10 per 100 000 or 100 000 per
billion Females 3.3 per 100 000 or 33 000
per billion