Workpackage 3: Risk Modelling of Combined Exposure

1
Workpackage 3 Risk Modelling of Combined
Exposure

• Dr Leif Busk / Jacob van Klaveren

2
Limitations of approach in 2002
Limitation of risk assessment 2002

• Comparison of results within Europe not always
  possible
• Variation and uncertainty in residue levels
  and consumption not always considered
• One chemical at a time (no cumulative
  exposure assessment)
• Short- and long-term exposure not well defined
• EU needs to harmonize the risk assessment
  procedures
• Important recommendations from EU-projects
  FOSIE, EFCOSUM, SSC

3
Start of SAFE FOODs

• Aims of WP3
• To establish an electronic platform for
  pan-European exposure modelling
• Integration of probabilistic exposure and effect
  modelling including risk prioritisation
• To perform or to compare exposure assessments to
  more than one chemical

4
Sampling residue and consumption data
Residue database
Consumption database
99, 99.9, and/or 99.99 percentile
5
Different computers in different MS
6
Communication via the internet
7
Harmonization of food codes

• Food as eaten is not as food as measured. Who
  has a conversion model?
• Uniformed approach, using CODEX coding system
• Food conversion as done in the Netherlands for
  CZ, DE, SE, IT

8
Harmonization of food codes
Convert national codes to their raw agricultural
commodity ingredients (common codex code)
Apple pie
Conversion table is needed!!!
9
How to start the MCRA, how does it work
10
Relevant steps in pan-EU assessment

• Use the food consumption and residue data as
  they are
• Select comparable age groups and comparable
  modelling input parameters
• Is there a difference in dietary habits or in
  residue levels and monitoring practices

11
Calculated intake captan using data as they are
12
Food consumption data used in calculation
13
Residue data captan

• Make one European residue database

14
EU residue database, same age group 18-74
15
Use of MCRA or E-platform is growing

• EFSA
• PPR opinion on acute dietary intake of pesticides
• Project on cumulative risk assessment
• National Food Authorities, network is growing
• Acrylamide, Iodine, Dioxine
• Mycotoxins, Natural toxins
• Flam retardents
• Sweeteners
• Pesticide Industry
• Food Industry
• ILSI (risk-benefit e.g. sodium intake and
  acrylamide)
• Unilever (nutrients and fatty acids)

16
Integration of exposure and effect models

• How to handle positive and negative effects?
• EFSA's 6th Scientific Colloquium 2006 -
  Risk-Benefit Analysis of Foods
• ADI/TDI and RDA are not appropriate for
  quantitative risk-benefit assessment
• future tools may be based on probabilistic
  modelling
• One of the main challenges of such an exercise
  is to define a common scale of measurement for
  comparing the risks and the benefits
• Positive effects decrease existing burden of
  disease (DALY)
• Negative effects often no existing burden of
  disease, but there is a safety margin that should
  be respected/increased
• How to handle cumulative effects?
• How to handle uncertainties?
• EFSA Scientific Committee, Guidance 2006
• strongly encourages the systematic evaluation of
  uncertainties in risk assessments

17
Probabilistic assessment of MoE, why?

• Some people eat more of a certain food than
  average
• Some batches of food are higher contaminated than
  average
• Food processing decreases residue concentrations
  sometimes less than average
• Some people are more sensitive than average
• There is uncertainty about almost everything

18
The IPRA model, part 1
contributions from foods
19
The IPRA model, part 2
20
The IPRA model, part 3
21
Distribution of IMoE
van der Voet Slob (2007)

• Cumulative distribution function y axis gives
  of population not exceeding IMoE on x axis
• PoCE Probability of Critical Exposure P( IMoE
  lt 1)

22
Health Impact Appreciation

• Examples of Health Impact Categorization

No Health Impact -------------------HIC-1--------
----------- Low Health Impact -------------------
HIC-2------------------- Moderate Health
Impact -------------------HIC-3-------------------
Severe Health Impact
23
Health Impact Assessment Deoxynivalenol (real
effect 100 to demonstrate principe)
100
10
population
1
0.1
1000
1
10
100
0.01
0.1
IMoE
24
Simple representation of IMoE distribution

• Main bar describes inter-individual variation in
  the study population (variation due to different
  exposure, different sensitivity)Extends from p1
  to p99 of distribution
• Error bars show uncertainty (from exposure, tox,
  extrapolation)lower (5) uncertainty limit for
  p1 and upper (95) uncertainty limit for p99

25
Comparison of fungicide and mycotoxin risks

• Decisions of fungicide use are an example of
  risk-benefit analysis
• Fungicides may have toxic effects (hazard)
• Fungicides may reduce risk of mycotoxin
  production (benefit)
• Example calculation, current health impact for
  one mycotoxin and one fungicide

Muri et al. (paper in prep.)
26
Cumulative effect antiandrogenic substances
Androgens

• Key regulators of male sexual differentiation
  during pre- and early post-natal development

Antiandrogens

• Substances that counteract the androgen action at
  some stage in this period and thereby affect the
  male reproductive system

Effects on the male reproductive system following
in utero exposure in animals

• Reduced anogenital distance (AGD)
• Retained nipples
• Reproductive organ weights changes and
  histopathology
• Malformations of external genitalia e.g.
  hypospadias, vaginal pouch, cleft phallus etc.

27
Relative Potency Factors (RPF)
Estimation of RPFs
Levator ani/bulbocavernosus muscle
28
IPRA for cumulative effects
On three reproductive developmental endpoints
29
Margin of Exposure MoE vs. IMoE

• MoE is deterministic estimate without safety
  factor
• Can be corrected for probabilistic exposure ?
  more realistic
• IMoE is also probabilistic for sensitivities,
  includes modelling of interspecies and
  intraspecies factors

• IMoE more realistic. Also useful for risk
  managers?

conservative safety assessment
more realistic expectation
30
Dissemination

• Scientific papers in a special volume of Food and
  Chemical Toxicology
• Use cases for EFSA and demos for stakeholders
• Workshop (18 February 2008)
• EFSA contribution
• Industry contribution (CIAA, Agrochemicals)
• SAFE FOODs contribution
• 65 participants, including all relevant
  stakeholders

31
Conclusions E-platform

• Yes there is a future for the E-platform
• Harmonised data collection, Pan-European intake
  assessment within hours
• Overview of difference in consumption and/or
  residue results between Member States
• Possibility to pool data or to borrow data
• Quality of data is important

32
Pan-EU modelling and proposed organisation
DG Sanco
Member States
Food Industry (CIAA)
Industry Agrochemicals
EFSA
MCRA
Re-run MCRA
Run MCRA
DK
SE
NL
IT
CZ
Data access agreements
33
Conclusions IPRA model

• Model integrates exposure and effects
• Can be used when we need to refine risk
  assessment (e.g. exceedance of ADI)
• Different effects for different chemicals can be
  compared
• Can be used for risk assessment of combined
  exposure of more than one chemical
• Will be useful for evaluating uncertainties

34
www.safefoods.nl