Presentation at European Food Safety Authority (EFSA), Parma, by Manolis Kogevinas (ISGlobal, ISEE) on pesticides risk assessment.

1. Presentation at workshop by EFSA (European Food Safety Agency) at Parma, Italy
on November 21, 2017 on pesticides risk assessment
Critical review of the EFSA approach
Manolis Kogevinas MD, PhD
President,
International Society for Environmental Epidemiology (ISEE)
Barcelona Institute for Global Health (ISGlobal)
manolis.kogevinas@isglobal.org
@KogevinasM
EFSA 2017, Parma

2. Notes on previous slide
• Excellent workshop organized by the Pesticides
Committee of EFSA on how to incorporate
epidemiology in the risk assessment they are
doing for the European Commission. EFSA
corresponds to the US FDA but focuses only on
food. The pesticides committee exceptionally also
deals with occupational exposures to pesticides.
The usual difficulties in cross-discipline
discussions occurred but overall the workshop
was excellent and helped communication.

3. Man’s mind cannot grasp the causes of events in their
completeness, but the desire to find those causes is
implanted in man’s soul. And without considering the
multiplicity and complexity of the conditions any one
of which taken separately may seem to be the cause,
he snatches at the first approximation to a cause that
seems to him intelligible and says: “This is the cause!”
Leo Tolstoy
War and Peace; Book Thirteen: 1812; Chapter 1

4. Notes on previous slide
• My summer readings. Leo Tolstoy refers to the
Napoleonic wars and how do persons and
societies take decisions. Isn’t this a perfect fit
for our science?

5.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

6.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

7. Notes on previous slide
The complex issues of risk assessment of pesticides
are also found in many other exposures, e.g.
endocrine disruption, water contaminants. Also the
many problems with RA of pesticides are not
specific to epidemiology but to all other disciplines.
What distinguishes pesticides from other exposures
is that the about 500 of them used in the EU are
regulated (however this regulation applies in
practice). By contrast regulation does not exist for
many other components of mixtures or complex
exposure circumstances

8. Pesticides and cancer
Lindane, classified as human carcinogen (Group 1) in relation to
risk of non-Hodgkin Lymphoma (IARC 2015; D Loomis, Lancet
Oncol, 2015)
Why is there only one insecticide classified as human carcinogen
by IARC/WHO?
Lack of convincing evidence for other pesticides clearly shows the
difficulties in evaluating the carcinogenicity of many chemical
agents in human populations

9. Notes on previous slide
I chaired the IARC Monograph that classified the one and only pesticide in
Group 1 (definitive carcinogens), Lindane. One of the reasons for not having
many more pesticides classified in Group1 is that not necessarily many of the
pesticides we use are actually carcinogens, though they may show acute
toxicities. However, this cannot simply be the only explanation and certainly
the most important reason for having only one pesticide in Group1 refers to
difficulties in research. Many of the most toxic chemicals we know show
multiple toxicities, for example affect neurodevelopment, growth,
cardiometabolic, and are also genotoxic. Pesticides that have been shown to
have multiple toxicities could also be expected to be genotoxic but we still
cannot prove this. This is because of the difficulties in research (not only
epidemiologic research). In the same monograph that we evaluated Lindane
we also evaluated DDT. It is really unlikely that DDT is not carcinogenic to
humans (at high and prolonged doses) but still applying IARC rules, we could
not classify DDT in group 1.

10. Pesticides and cancer
Lindane, classified as human carcinogen (Group 1) in relation to
risk of non-Hodgkin Lymphoma (IARC 2015; D Loomis, Lancet
Oncol, 2015)
Why is there only one insecticide classified as human carcinogen
by IARC/WHO?
Lack of convincing evidence for other pesticides clearly shows the
difficulties in evaluating the carcinogenicity of many chemical
agents in human populations
• standard environmental toxicity tests used to
license pesticides are performed on particular
test species and have limited predictive power
when chemicals are used widely (see also Milner
and Boyd, Science 2017)
• low level of trust in current toxicology testing
regimes because of serious difficulties to
encompass the full range of toxic effects that
could emerge when a pesticides is used at scale

11. Notes on previous slide
Situations of other disciplines that have failed to
do good predictions on pesticides, e.g.
environmental tox (quotes are from Milner and
Boyd’s commentary in Science, that is excellent).
Important to note that exposure and effects may
differ significantly between short term tox tests
and long term application in the field

12.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
(see presentation by Laura Beane Freeman)
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

13. Notes on previous slide
Exposure misclassification is extremely
important in pesticides research; I cover this
very briefly because other presenters discussed
this issue at length

14. • Seasonal
• Often outdoors but also indoors
• Highly variable
• Type of agent and exposure
• Biological, chemical and physical
• Individual agents (active ingredients; adjuvants)
• Intensity, duration and frequency
• Multiple agents
• Multiple routes
• Not limited to farmers
(slide modified from Hans Kromhout, Univ Utrecht)
Nature of exposures in agriculture

15.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

16. Notes on previous slide
The term Exposome used to indicate in reality
modern environmental epidemiology using new
approaches both for exposure assessment and
for evaluation of mechanisms

17. The Exposome
Recognizing the disparity in current knowledge between
genes and environmental exposures, Chris Wild (2005)
defined the “exposome” representing all environmental
exposures (including those from diet, lifestyle, and
endogenous sources) from conception onwards, as a
quantity of critical interest to disease etiology.
E

18. Notes on previous slide
Big G means a lot of money and attention for
genetics. Little E means little money and
attention to the evaluation of Environmental
exposures

19. “Modern” Epidemiology
(Slide from Perry Hystad, Oregon State University)

20. Notes on previous slide
John Snow, Broad street pump. I make reference
to the fact that for a long period Environmental
Epi continued working with this principle, ie a
source, and exposure evaluated in concentric
cycles around the source. Works some times and
for some exposures, but too simplistic for most
of what affects human populations

21. “Modern” Epidemiology
(Slide from Perry Hystad, Oregon State University)

22. Notes on previous slide
Refer to the new tools that have revolutionized
environmental exposure assessment. Not all of
them are relevant for occupational pesticides
epidemiology

23. “Modern” Epidemiology

24. Notes on previous slide
• Refer to the capacity to evaluate “internal” exposome
and mechanisms. The right panel is from our study on
metabolomics of swimming in a chlorinated pool (see
Karin van Veldhoven, in press) in Barcelona,
distinguishing clearly metabolic patterns before and
after a 40 minutes swim and identifying new pathways
(well you cannot see the pathways in this graph). The
left hand panel is from Casals-Casas and Desvergne,
2011, on pathway perturbation on endocrine
disruption. A very similar approach to what
toxicologists do with AOPs (adverse outcome
pathways)

25.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

26.  Most of the evidence in human from studies examining mainly
other risk factors (hence no detailed analysis of pesticides-
hence not surprising that these data cannot be used in risk
assessment)
 Some major studies funded (a mistake to consider only
AgHealth)
 We need 100M€ (indicative amount) to do a couple of new
powerful cohort studies in different settings. Multidisciplinary,
extensive industrial hygiene, repeated biomarkers, omics, long
term follow-up
Serious underfunding of research on pesticides

27. Notes on previous slide
Self evident. Unless we get more funding to do
pesticides research we will come back in 10-15
years at EFSA complaining about problems of
research. If 100 million Euros seem a lot to you,
look at the R&D budgets of big food-chain
companies

28.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

29. The EFSA Scientific Opinion: a report on epidemiology
written by non-epidemiologists. An interesting
endeavour (for the authors) but not an EFSA document

30. Notes on previous slide
EFSA did a serious attempt to incorporate
epidemiology in the discussion of pesticides risk
assessment. This is positive. The document
however circulated is naive concerning
epidemiology and it is surprising that the working
group producing this document is basically non-
epidemiologists (with few exceptions of a couple
excellent epidemiologists participating as external
advisors). It is problematic when this type of
documents with a very unbalanced critique not
clear understanding of epidemiological methos,
become official EFSA documents

31. The EFSA Scientific Opinion: comments submitted by ISEE
• Epistemological (toxicologic studies versus epidemiology)
• Mechanistic (favor ranking etc., rather than a integrated
assessment of knowledge)
• Missing considerations (no advocacy for the necessary
substantial ongoing stream of funding for surveillance and post-
marketing surveillance of pesticides that could strengthen our
capacity to identify real life events, nor for expanded,
diversified, well-funded, and more detailed epidemiologic
studies being now concentrated in just some centers and labs)
• Vulnerability of study populations (strengths of epidemiology
to examine real life conditions of exposure and disease,
vulnerable populations, and real life outcomes that can happen
and are seldom observed in vivo)

32. Notes on previous slide
A summary of some of the comments sent by
the ISEE Policy committee to the draft
document circualted by EFSA (full draft can be
found in EFSA’s web)

33. The EFSA Scientific Opinion: comments by ISEE
• Writing: unequal; many parts valuable; overall could be
improved
• Overall message of the report: epidemiology is not reliable, text
making systematically broad generalizations
• Many cliché on causal inference, ranking of evidence etc
• Important areas poorly covered, e.g. retrospective
exposure assessment and biomonitoring, post market
surveillance
• Scope: unclear (or at least poorly described)

34. Notes on previous slide
Specific comments on draft. Scope refers to the
fact that a framework for risk assessment is
produced by a specific panel while EFSA should
have provided centrally these guidelines, similar
for example to what IARC does. It is actually
surprising that EFSA does not provide these
guidelines centrally

35. Occupational Human carcinogens
(Group 1- IARC)
• 118 agents in Group 1
• 57 are occupational or also occur in the occupational
environment (e.g. aflatoxins, SHS, radiations etc)
• Of those, 36 were identified as Group 1 before the year
2000, and 21 after the year 2000

36. Notes on previous slide
Making a point that most occupational carcinogens we
have identified in Group1 (IARC) were identified through
epi studies before the wider use of mechanistic data in
hazard identification. This to contrast the repeated
statements by toxicologists and basic scientists that
epidemiology does not provide firm conclusions. Just
amazing how these statements are repeated and become
dogmas. Statements like “epidemiology has intrinsic
weaknesses and does not allow conclusions but still
concern” (this was mentioned by a key toxicologists at
the EFSA meeting) are repeated in one or other way. This
simply shows a very poor understanding on causal
inference.

37. Probable Occupational Human
carcinogens (Group 2A- IARC)
• 81 agents in Group 2A (probable carcinogens)
• 48 are occupational
• Of those, 20 were identified as Group 2A before the
year 2000, and 28 after the year 2000
• Use of evidence on mechanisms very important for
this group (upgrade from 2B-possible to 2A-probable)

38. Notes on previous slide
Interesting statistics on 2A (probable IARC)
carcinogens. Here we have more agents
identified in recent years (by contrast to
Group1), and most are through the combination
of human, animal and mechanistic data. This is
actually a very positive development of the last
20 years in IARC, i.e. that the evaluations take
more formally into account all available
evidence and mechanistic data are key part of
the evidence

39. There is no predefined hierarchy in study design.
RCTs are good for clinical settings and not good for environmental,
occupational and many other exposures

40. Notes on previous slide
Graph from NYT. One more repeated statement by
tox, clinicians that RCTs (randomized controlled
trials) are the gold standard. It is repeated so many
times that this also has become a dogma. RCTs
should not be taken as a gold standard because
they are not applicable in most circumstances
outside clinical settings and because they simply
cannot respond to complex questions in population
studies. Just asked the audience if they thought we
should be doing RCTs on availability of guns and
mass shootings. OK, an extreme example but that is
why we have examples to make a point!

41. The EFSA Scientific Opinion: out of scope
Conclusions of the report (p58):
‘The PPR Panel will specifically’:
1) Collect and review all sources of gaps and limitations … , of the available
epidemiological studies.
2) Based on the gaps and limitations identified in point 1, propose potential
refinements for future epidemiological studies to increase the quality, relevance
and reliability … This may include study design, exposure assessment…
3) Identify areas in which information and/or criteria are insufficient or lacking
and propose recommendations for how to conduct pesticide epidemiological
studies in order to improve and optimise the application in risk assessment. …
4) Discuss how to make appropriate use of epidemiological findings in risk
assessment of pesticides during the peer review process of draft assessment
reports, e.g. WoE as well as integrating the epidemiological information with
data from experimental toxicology, AOPs, mechanism of actions, etc.

42. Notes on previous slide
Point made to EFSA to show that the report they produced
were out of scope. The PPR (pesticides) panel does not have a
single epidemiologist and they still see their job as making
proposals on how we should be doing epidemiological studies.
Getting feedback from other disciplines is very positive and
desirable. Having however people who do not understand and
are trained in epi methods defining the protocols we should
use in our studies is another business. Interesting discussion
at questions times, after my presentation where the Chair of
the panel said that 6 out of 8 panel members were actually
epidemiologists because they had co-authored sometime in
their life epidemiological papers. Oh well, this is how cell
biologists understand epidemiology…

43.  We have a problem with health risk assessment of pesticides
(not only pesticides; not only epidemiology)
 The most complex issue in epidemiological studies of
pesticides is exposure assessment. There are solutions to this
 “Exposome” approaches open new possibilities for research
and advanced risk assessment bridging toxicology and
epidemiology
 We need more funding on pesticides research
 The EFSA Scientific Opinion could be significantly improved
 EFSA needs to standardize protocols
Main messages

44.  We suggest a balanced panel, addressing the overall
production of pesticide science, aiming to enhance
the integration and advancement of knowledge.
 We urge EFSA to develop and apply standardized
protocols for risk assessment rather than ask each
panel to improvise and produce position papers on
issues that are not within their area of knowledge as
is the case with the pesticides paper.
ISEE’s comments to draft

45. Notes on previous slide
Comments by ISEE Policy committee to EFSA.
Contrary to other EU or International
organizations, EFSA has not developed as much
formal protocols and depends very much on
ideas and approaches developed by each Panel
(mostly small panels, specialized on specific
topics that are not always sufficiently
multidisciplinary).

46.  We urge EFSA to consider in developing these protocols
relevant guidelines for systematic review of evidence that
already exist e.g. in WHO.
 We urge EFSA to consider that older reports such as the
WHO “guidelines for guidelines” have been modified to
enable the application of a wider more holistic perspective
concerning the types of evidence to be used
ISEE’s comments to draft

47. Notes on previous and subsequent
slides
• Comments by ISEE Policy committee to EFSA.

48.  Science is one and epidemiology works integrated with other
sciences;
 Epidemiology as any other science is advancing, and to
further contribute to the assessment of the health effects of
pesticides requires of independent and rigorous research
well-funded, as well as the input from post marketing
surveillance;
 We have to take advantage of our understanding of the
“exposome” and need to consider it on the real
vulnerabilities of population, only provided through
population (epidemiologic) research.
The role of epidemiology

49. • Transformational change in the breadth and depth of
exposure assessment that would improve integration with
and responsiveness to toxicology and epidemiology
• Questions as to whether or how the data now being
generated can be used to improve risk-based decision-
making
• We need to invest in common understanding and
exchange of ideas and link modern exposure assessment,
molecular epidemiology/exposome with toxicological
approaches on mode of Action/Adverse Outcome
Pathways. Both are in combination essential to establish
evidence based risk assessments and policies
New approaches for risk assessment

50. thank you, on behalf of the
International Society for Environmental
Epidemiology (ISEE) and EPICOH
Acknowledgments
ISEE Council: Beate Ritz, President elect; Greg Wellenius and
Chang-Chuan Chan, Councilors;
ISEE Policy Committee: Michal Krzyzanowski, Chair,
Carlos Santos Burgoa, Tony Fletcher, Silvia Medina, Erik Lebret
EPICOH: Roel Vermeulen, Chair; Aaron Blair, Hans Kromhout

51. thank you, on behalf of the
International Society for Environmental
Epidemiology (ISEE) and EPICOH
Acknowledgments
ISEE Council: Beate Ritz, President elect; Greg Wellenius and
Chang-Chuan Chan, Councilors;
ISEE Policy Committee: Michal Krzyzanowski, Chair,
Carlos Santos Burgoa, Tony Fletcher, Silvia Medina, Erik Lebret
EPICOH: Roel Vermeulen, Chair; Aaron Blair, Hans Kromhout