Improving pesticide regulation by use of impact analyses: A case study for bees

Abstract

When changes to regulatory guidance for risk assessment are proposed it is necessary to undertake an impact analysis to assess whether they bring the desired improvement to a risk assessment and reliability of the outcomes to inform decision making. In particular impact analyses should estimate the chances of getting both false negative (concluding low risk where more research is needed) and false positive outcomes
(concluding high risks where the product is of low risk). Such analyses are also used to inform on future
product development costs and workload for regulatory authorities. In this paper, we present the findings from an impact analysis conducted on the proposed EFSA bee guidance document (2013) and discuss whether the proposed guidance would provide for a cost effective and tiered approach toward the protection of bees due to the potential risks posed by the use of plant protection products. Following on from this a second impact assessment is presented based on new data generated by ECPA member companies regarding the assessment of chronic risk to bees. Critical areas are discussed and suggestions for the improvement of assess the risk assessment for plant protection products (PPP) to bees are presented.

Full text

Hazards of pestici des to bees - 13th inte rnational sympos ium of the ICP-PR Bee protection group, Oct ober 18 – 20 2017 , Valencia (Spain)
86 Julius-Kühn-Archiv, 462, 2018
1.20 OECD GD 239 Honey bee larvae in vitro testing and solvents: on the job
training
Stefan Kimmel, Magdaléna Cornement
Innovative Environmental Services (IES) Ltd, Benkenstrasse 260, 4108 Witterswil, Switzerland; corresponding
author: s.kimmel@ies-ltd.ch
DOI 10.5073/jka.2018.462.020
Abstract
With adopting the OECD guidance document 239 for in vitro bee larvae repeated exposure testing in the
laboratory, a new guidance became final without taking into account several pending issues and unsolved
problems still to take care of. Important aspects to be taken into account is the use of solvents when testing
practically insoluble compounds (e.g., during active substance testing), as well as confirmed homogeneity of
substance within the final feeding solutions.
Testing of the active ingredient as technical instead of the corresponding formulation for registration purposes
is requested from, but not only limited to US and Canadian authorities, several other authorities around the
globe seem to follow that approach. Having in mind the high sensitivity and susceptibility of the young larvae
reared in the test, this leads to quite some problems in the practicability of the test itself.
The here presented results and methodology shall share experience and lessons learned from the past years
for this specific test, further on a technical approach to make the use of solvents helpful but not harmful.
Further on, adaptions and modifications on the analytical verifications required for this study are shown and
being put on discussion. Overall a feasible way of adaption and modification for this highly discussed and still
criticized test system is presented, the improvements shall be seen as turning this setup into a more reliable
and reproducible study design helpful for assessing potential risks during the process of registration of plant
protection products and chemicals.
1.21 Improving pesticide regulation by use of impact analyses: A case study for
bees
Mark Miles1, Anne Alix2, Roland Becker3, Mike Coulson4, Axel Dinter5, Laurent Oger6, Ed
Pilling2, Amanda Sharples7, Gabe Weyman8
1Bayer Crop Science Division, Cambridge, CB4 0WB, UK. 2Dow AgroSciences, Abingdon, OX14 4RN, UK. 3BASF
SE, Limburgerhof, Germany. 4Syngenta, Jealott's Hill, RG42 6EY, UK (Exponent Harrogate, HG2 8RE, UK from
May 2017). 5FMC Agricultural Solutions, Frankfurt/Main, Germany. 6ECPA, Brussels, Belgium. 7FMC, Harrogate,
HG3 1RY, UK. 8ADAMA Thatcham, Berkshire, RG19 4LW, UK.
DOI 10.5073/jka.2018.462.021
Abstract
When changes to regulatory guidance for risk assessment are proposed it is necessary to undertake an impact
analysis to assess whether they bring the desired improvement to a risk assessment and reliability of the
outcomes to inform decision making. In particular impact analyses should estimate the chances of getting
both false negative (concluding low risk where more research is needed) and false positive outcomes
(concluding high risks where the product is of low risk). Such analyses are also used to inform on future
product development costs and workload for regulatory authorities.
In this paper, we present the findings from an impact analysis conducted on the proposed EFSA bee guidance
document (2013) and discuss whether the proposed guidance would provide for a cost effective and tiered
approach toward the protection of bees due to the potential risks posed by the use of plant protection
products. Following on from this a second impact assessment is presented based on new data generated by
ECPA member companies regarding the assessment of chronic risk to bees. Critical areas are discussed and
suggestions for the improvement of assess the risk assessment for plant protection products (PPP) to bees are
presented.
Keywords: Honeybee, risk assessment, impact analysis, pesticide

Hazards of pestici des to bees - 13th inte rnational sympos ium of the ICP-PR Bee protection group, Oct ober 18 – 20 2017 , Valencia (Spain)
Julius-Kühn-Archiv, 462, 2018 87
Introduction
When significant changes to regulatory guidance for risk assessment are proposed it is necessary
to undertake an impact analysis to assess whether they bring the desired improvement to a risk
assessment and reliability of the outcomes to inform decision making. In particular impact
analyses should estimate the chances of getting both false negative (concluding low risk where
more research is needed) and false positive outcomes (concluding high risks where the product is
of low risk). Such analyses are also used to inform on future product development costs and
workload for regulatory authorities.
In July 2013 EFSA released their guidance document on the risk assessment of pesticides on bees
and considerably updated it in 2014 based on feedback from a workshop with European Union
member states. The document demanded a more thorough approach for the testing and risk
assessment of plant protection products (PPP) for bees.
Material and methods
In 2013 industry undertook a detailed evaluation of the impact of the proposed screening and tier
I risk assessments on the pass/fail rate of currently available active substances on the EU market
using honey bee endpoints. The analysis considers 151 active substances covering 163 uses: 52
were herbicides, 52 fungicides, 51 insecticides or acaricides and 8 other uses like plant growth
regulators. Solid applications were also considered with 20 active substances representing 36 uses
(not shown). At the time not all data on all substances were available as test methods were yet to
be developed; consequently it was necessary to estimate some endpoints.
Acute contact and oral toxicity: reported LD50 values used (µg a.s./bee).
Chronic oral toxicity (LDD50 ): Estimated as 1/5th acute oral toxicity endpoint based on advice from
EFSA e.g. acute oral LD50 = 100 g a.s./bee was converted to chronic oral LDD50 = 20 g a.s./bee.
Larval toxicity (NOED, no observable effect dose): Estimated as 1/10th of adult’s acute oral LD 50
corrected for mean larval body weight (83 mg, e.g. acute oral LD50 = 100 g a.s./bee was
converted NOED = 8.3 g a.s./larva.
The screening level risk assessment according to EFSA 2013 was conducted for all 163 uses
according to the following formulae:
HQ contact = AR /LD50 contact
ETRoral = AR x SV /LD 50 oral
ETRchronic = AR x SV /LDD50 oral
ETRlarva = AR x SV /NOED
Where: AR = application rate in g a.s./ha for HQcontact and kg a.s./ha for all other ETR values.
SV = short cut value.
Following on from this a second impact assessment was conducted using the real-life endpoints
from 10 day chronic studies with honey bees using the same procedure as above based on new
data generated by ECPA member companies. Data from 85 uses including 32 herbicides, 32
fungicides and 17 insecticides were evaluated. In addition to the screening assessment a tier I
evaluation was also conducted using the tier I S.V.s and accounting for dissipation in pollen and
nectar over time using a default half-life (DT50 ) of 10 days.
Results and discussions
The number of uses passing or failing the screening risk assessment for the original impact
analysis is presented as percent in table 1. The pass / fail rate of the EFSA proposal for acute risk
(HQcontact and ETRacute adult oral) was very similar to the current risk assessment proceeded
using HQ values for both contact and oral routes of exposure and the Annex VI trigger of 50
indicating that possibly there will be no overall significant changes in the risk assessment outcome

Hazards of pestici des to bees - 13th inte rnational sympos ium of the ICP-PR Bee protection group, Oct ober 18 – 20 2017 , Valencia (Spain)
88 Julius-Kühn-Archiv, 462, 2018
for acute risk assessment for foliar applied products, i.e. the overall protection level is similar.
Consequently, based on the sample of 163 uses, 26% of all uses would require evaluation at a
higher tier for acute risks to adult bees. This would include at least 60% of insecticides.
The risk to larvae based on the calculated ETRlarvae values indicated that less than half of the uses
will pass the screening tier risk assessment with 56% of uses indicating higher tier evaluation,
including 74% of insecticides. This pass rate is similar to that based on real-life data from 22 day
repeated dose studies (Becker et al 2018).
The chronic risk to honey bees as measured by calculation of ETRchronic adult oral was remarkably
different to the acute risk. In this case only 18% of uses were observed to have passed the
screening level trigger of <0.03. For this assessment 79% of all herbicide uses failed as well as 75%
of fungicide uses and all 92% of insecticide uses. Overall this would mean that in 82% of all cases a
higher tier risk assessment would be required which may necessitate the generation of higher tier
data (e.g. field residue tests, semi-field and field tests).
The distribution of screening level ETRch ronic adult oral risk is presented in Figure 1. It can be seen
that the majority of substances do not pass the risk assessment. A tier I risk assessment only gives
a moderate improvement (less than a factor of 10). The ETRchronic for many herbicides and
fungicides require a refinement level of 2 – 3 orders of magnitude and for insecticides of 5 – 6
orders of magnitude. This means that risk assessment refinement methods such as used of
measured residues in treated crops, generation of higher endpoints pose significant challenges
even for herbicides and fungicides. As this was based on extrapolated values from acute studies
to check if these findings were realistic and predictive industry undertook a follow-up analysis
using endpoints and data obtained from several years of practical experience with 10 day adult
testing. It was found that the overall pass rate was similar to that predicted with 18% predicted by
the impact analysis and 24% based on real-life data. The number of uses passing for herbicides
based on real data was slightly higher than predicted (31% vs. 21%) and this can be accounted for
by the fact that in the impact assessment endpoints were often extrapolated from limit tests
where no toxicity was observed. The number of fungicides uses passing based on real-life data
was very close to the predicted level (28% vs. 21%). The impact analysis predicted that 8% of
insecticides would pass the screening risk assessment however in real-life the actual value was 0%.
Table 1 Risk to honey bees: Percentage number of uses passing the screening risk assessment for foliar (based
on 163 uses) from impact assessment 2013.
Chemical
group
Acute risks to adult honey bees Chronic risk to
Adult honey
bees*
Larvae**
HQ contact
(current
HQ<50)
HQ contact
(new HQ <42
or 85)
HQ oral (current
HQ<50)
ETR acute
adult oral
(<0.2)
ETR chronic adult oral
(<0.03)
ETR larvae
(<0.2)
Herbicides 96 94 94 88 21 50
Fungicides 98 100 96 92 25 58
Insecticides 47 47 40 40 8 26
Other 100 100 88 75 13 25
All 81 82 78 74 18 44
* 10 day LD50 for adults estimated as 1/5 of acute LD 50
** NOEL for larvae estimated as 1/10 of adult’s LD50 corrected for body weight (83 mg/bee)

Hazards of pestici des to bees - 13th inte rnational sympos ium of the ICP-PR Bee protection group, Oct ober 18 – 20 2017 , Valencia (Spain)
Julius-Kühn-Archiv, 462, 2018 89
Table 2 Chronic risk to honey bees: Percentage number of uses passing the screening and tier I risk
assessment for foliar (based on 81 uses) from industry data collection of real-life endpoints 2015.
Chemical group % uses passing honey bee chronic risk assessment
Screening level Tier I
Herbicides (n=32) 31% 47%
Fungicides (n=32) 28% 44%
Insecticides (n=17) 0.0% 18%
All (n=81) 24% 36%
Figure 1 Chronic to risk honey bee adult s: Distribution of exposure toxicity ratios for sprayed products. Values
below the dashed line pass the EFSA trigger of 0 .03.
Conclusions
The impact analysis and the follow-up work by Becker et al 2018 on larvae and with chronic adult
data in this paper highlight the problem of releasing new guidance without proper consideration
of the impact on all users and stakeholders. The impact analysis allowed researchers to focus on
key challenges such as appropriate triggers (Miles et al, 2018a), better ways to estimate bee
exposure taking into account real-life bee feeding behaviour by use of models (Miles et al 2018b;
Miles and Preuss, 2018) and additional consideration of the relative sensitivities of honey bees and
non-Apis bees (Dinter et al 2018).
The EFSA 2013 guidance for bees is unworkable in its current form and will lead to systemic failure
for almost all substances without providing workable higher tier options. Levels of screening and
tier I refinement needed are large; 2 to 3 orders of magnitude for low toxicity compounds and 5 to
6 orders for insecticides. New guidance should be designed to work within current regulatory
testing frameworks and be built on existing guidance. Before implementation any new guidance
with potential to impact innovation should be subject to a testing phase and modified if needed
to create workable processes.
References
European Food Safety Authority, 2013. EFSA Guidance Document on the risk assessment of plant protection products o n bees
(Apis mellifera, Bombus spp. and solitary bees). EFSA Journal 2013;11(7):3295, 268 pp., doi:10.2903/j.efsa.2013.3295
Roland Becker, Johannes Lückmann, Mark Miles, Anne Alix, Gabe We yman, Ed Pilling, Natalie Ruddle, Axel Di nter, Amanda
Sharples, Stephani e Fritz & Laurent Oger (2018). Sensiti vity of honey bee larvae to PPPs and i mpact on EFSA GD.
International commission for pl ant-pollinator relationships (ICP-PR) Bee Protecti on Group, 13th International symposium
“Hazards of pesticides to bees”. Palau de la Musica de Valencia, Valencia Spain, 18th – 20th October 2017.

Hazards of pestici des to bees - 13th inte rnational sympos ium of the ICP-PR Bee protection group, Oct ober 18 – 20 2017 , Valencia (Spain)
90 Julius-Kühn-Archiv, 462, 2018
Mark Miles, Zhenglei Gao & Thomas Preus s (2018a). Linking protection goals to trigger values using compound specific
properties: Chronic risks to bees. International commission for plant-pollinator relationships (ICP-PR) Bee Protection
Group, 13th Internatio nal symposium “Hazards of pesticides to bees”. Palau de la Musica de Valencia, Valencia Spain, 18th –
20th October 2017.
Mark Miles, Zhenglei Gao & Thomas Preuss (2018a). Si mple modelling approaches to refine exposure for bee risk assessment
based on worst case assumptions. SETAC E urope 27th annual meeting “Envi ronmental quality through transdisciplinary
collaboration”, Brussels 7th – 11th May 2018.
Mark Miles, Thomas Preuss (2018a). Simple modelling approaches to refine exposure for bee risk assessment. International
commission for plant-pollinator relationships (ICP-PR) Bee Protection Group, 13th International symposium “Hazards of
pesticides to bees”. Palau de la Musica d e Valencia, Valencia Spain, 18th – 20th October 2017.
Axel Dinter, Anne Alix, Roland Becker, Peter Campbell, Mark Miles, Ed Pilling, Natalie Ruddle, Amanda S harples, Gabe Weyman,
Laurent Oger (2018). Non-Apis (Bombus ter restris) versus honeybee (Apis mellif era) acute oral and contact sensitivity –
Preliminary results of ECPA company data evaluation. International commission for plant-pollinator relationships (ICP-PR)
Bee Protection Group, 13th International symposium “Hazards of pesticides to bees”. Palau de la Musica de Valencia,
Valencia Spain, 18th – 20th October 2017.
1.22 Weight differences of honey bees after administration of sublethal doses of
dimethoate
Annette Kling1, Stephan Schmitzer2
1 Eurofins Agroscience Services Ecotox GmbH, Eutinger Str. 24, 75223 Niefern-Öschelbronn, Germany,
AnnetteKling@eurofins.com
2 ibacon GmbH, Arheilger Weg 17, 64380 Rossdorf, Germany, stephan.schmitzer@ibacon.com
DOI 10.5073/jka.2018.462.022
Abstract
Background: The aim of this work was to assess honey bee body weight as a possible further parameter to
detect effects in a 10 day chronic feeding study according to OECD 2451 following exposure to sublethal
concentrations of a plant protection product (i.e. dimethoate). This investigation is based on the assumption
that weight differences might be caused by chronic feeding of dimethoate. Two set of tests in two different
laboratories (Lab 1 and Lab 2) were conducted in order to investigate possible weight changes of complete
adult honey bees and/or parts of their body (honey stomach and intestine) following treatment of dimethoate.
Bees were weighed before and after chronic feeding of sub-lethal concentrations of dimethoate.
Results: Differences in the number of bees which lost weight following treatment of sublethal concentrations
of dimethoate was found in Lab 1, but could not confirmed in Lab 2.
The difference in weight between the control group and the dimethoate treatment could only be detected as
a statistical significant difference in one lab at the highest concentration (0.4 mg/kg).
Assessment of weight changes of parts of the bee body (honey stomach and intestine) shows a very high
variation (CV) which makes interpretation of the data of the total body weight questionable.
Conclusion: The results of the two laboratories were contradictory and no conclusive assessment can be done
following the two sets of experiments. Assessment of bee body weight within a 10-day chronic feeding study
is considered questionable for the detection of sublethal effects. Further work with other active ingredients is
needed to clarify if body weight change of honey bees can be used as a parameter for sublethal effects.
Keywords: chronic toxicity, sublethal effects, weight differences, honey bees
Introduction
Testing of chronic effects of Plant Protection Products (PPP) on adult honey bees by continuous
feeding of contaminated sugar solution over a period of 10 days is an integral part of the current
risk assessment for honey bees.
According to the OECD Guideline 245 mortality and food consumption have to be assessed in
order to detect possible side-effects of PPP to honey bees. Additionally, sublethal parameters like
behavioural abnormalities should be quantitatively recorded.
The tests were performed in two independent contract laboratories providing bees of two
different breeding lines of Apis mellifera carnica. In each laboratory an experiment was conducted

Julius Kühn-Institut
Bundesforschungsinstitut für Kulturp anzen
Julius-Kühn-Archiv
Pieter A. Oomen, Jens Pistorius (Editors)
Hazards of pesticides to bees
13th International Symposium of the
ICP-PR Bee Protection Group
18. - 20. October 2017, València (Spain)
- Proceedings -

Bibliograsche Information der Deutschen Nationalbibliothek
Die Deutsche Nationalbibliothek verzeichnet diese Publikation. In der Deutschen Nationalbibliograe: detail-
lierte bibliograsche. Daten sind im Internet über http://dnb.d-nb.de abrufbar.
ISSN 1868-9892
ISBN 978-3-95547-064-7
DOI 10.5073/jka.2018.462.000
History ICPPR-Bee Protection Group conferences
1st Symposium, Wageningen, the Netherlands, 1980
2nd Symposium, Hohenheim, Germany, 1982
3rd Symposium, Harpenden, UK, 1985
4th Symposium, Řež, Czech Republic, 1990
5th Symposium, Wageningen, the Netherlands, 1993
6th Symposium, Braunschweig, Germany, 1996
7th Symposium, Avignon, France, 1999
8th Symposium, Bologna, Italy, 2002
9th Symposium, York, UK, 2005
10th Symposium, Bucharest, Romania, 2008
11th Symposium, Wageningen, the Netherlands, 2011
12th Symposium, Ghent, Belgium, 2014
13th Symposium València, Spain, 2017
14th Symposium scheduled, Bern, 2019
Organising committee 13th conference
Dr. Jens Pistorius (Julius Kühn-Institut, Germany)
Dr. Anne Alix (Dow Agrosciences, United Kingdom)
Dr. Carmen Gimeno (Trialcamp, Spain), local organiser
Dr. Gavin Lewis (JSC, United Kingdom)
Dr. Pieter Oomen (Wageningen, The Netherlands)
Dr. Veronique Poulsen (ANSES, France)
Dr. Guy Smagghe (Ghent University, Belgium)
Dr. Thomas Steeger (US Environmental Protection Agency, USA)
Dr. Klaus Wallner (Hohenheim University, Germany)
Editors
Dr. Pieter A. Oomen, Wageningen, The Netherlands
Dr. Jens Pistorius, Braunschweig
Group photo of all symposium participants, standing in front, from left:
Thomas Steeger (new board member),
Jens Pistorius (new chairman),
Françoise & Pieter Oomen with award (editor & former chairman),
Guy Smagghe (organiser, symposium host and new board member),
Job & Margreet van Praagh with award,
Anne Alix (secretary of the board)
Foto
Pieter A. Oomen (Bumble bee Bombus lapidarius on thistle)
The proceedings of the symposia (such as these) are being published by the Julius Kühn Archive in Germany
since the 2008 symposium in Bucharest, Romania. These proceedings are also accessible on internet, e.g. the
former symposium proceedings published by JKI can be found on https://ojs.openagrar.de/index.php/JKA/
issue/archive (Issues 423, 437, 450). Furthermore, proceedings of former meetings have meanwhile been
digitalized and can be found on https://www.openagrar.de/receive/openagrar_mods_00032635.
Alle Beiträge im Julius-Kühn-Archiv sind unter einer
Creative Commons - Namensnennung - Weitergabe unter gleichen Bedingungen -
4.0 Lizenz veröentlicht.
Printed in Germany by Arno Brynda GmbH, Berlin.