Potential Effect of Adjuvants on Residues in Relation to Routine 0 - - PDF document

Message From :

BLEEKE, MARIAN S [AG/1000] [/O=MONSANTO/OU=NA-1000-01/CN=RECIPIENTS/CN=198145] Sent : 5/31/2016 3:59:50 PM To: LEVINE, STEVEN L [AG/1000] [/O=MONSANTO/OU=NA-1000-01/cn=Recipients/cn=594761]; LEOPOLD, VINCENT A [AG/1000] [/O=MONSANTO/OU=NA-1000-01/cn=Recipients/cn=VALEOP] Subject : RE: Emailing: CPDA-2012-2013-Accomplishments.pdf Attachments : CLA Presentation Adjuvants_PMRA Nov 20 2012.pdf Steve,

We worked extensively on this in the CLA residue expert work group. EPA was primarily looking at three types of

adjuvants surfactants, spreader/stickers, and oils. Registrants collected lots of data from side-by-side trials with and without adjuvants. We analyzed the data, EPA also analyzed the data, there were lots of discussions. Bottom line was that there was a modest statistical increase in residues with adjuvants, conclusion was that it was not significant from a risk assessment perspective, and label

restrictions weren't needed (or trials re-run). EPA did highly encourage running residue trials with adjuvants, which is

now standard procedure.

I've attached a ppt, presentation from 20:1.2 on the topic.

Probably more detail than you want: - but the last three slides

show the EPA memo and ChemSAC recommendations.

Marian

From : LEVINE, STEVEN L [AG/1000] Sent : Tuesday, May 31, 2016 10:11 AM

To: LEOPOLD, VINCENT A [AG/1000]; BLEEKE, MARIAN S [AG/1000] Subject : RE: Emailing: CPDA-2012-2013-Accomplishments.pdf Thanks.

From : LEOPOLD, VINCENT A [AG/1000] Sent : Tuesday, May 31, 2016 10:10 AM

Confidential - Produced Subject to Protective Order

MONGLY05272935

To: LEVINE, STEVEN L [AG/1000]; BLEEKE, MARIAN S [AG/1000] Subject : RE: Emailing: CPDA-2012-2013-Accomplishments.pdf No, don't believe so.

From : LEVINE, STEVEN L [AG/1000] Sent : Tuesday, May 31, 2016 10:07 AM

To: BLEEKE, MARIAN S [AG/1000]! LEOPOLD, VINCENT A [AG/1000] Subject : Emailing: CPDA-2012-2013-Accomplishments.pdf Marian and Vincent, Did you ever get involved with this issue?

CPDA scored a decisive victory with EPA's decision that the use of tank-mix adjuvants does not increase pesticide

active ingredient residues in plants to levels that would be of concern. As such, EPA will no longer require a label statement prohibiting the use of a tank-mix adjuvant that has not been tested in residue field trials with the registered active ingredient. This decision was articulated in an August 13, 2012 internal memorandum and followed a four year effort, led by CPDA, to alleviate EPA's concern that the use of tank mix adjuvants on plants grown under actual agricultural practices could result in as much as a two-fold increase in pesticide residue levels Confidential - Produced Subject to Protective Order

MONGLY05272936

Potential Effect of Adjuvants on

Residues in Relation to Routine

§1

Residue Trial Variability

CLA Residue Expert Working Group

in collaboration with CPDA

Health Canada - PMRA

Nov 20, 2012

Confidential - Produced Subject to Protective Order

MONGLY05272937

Adjuvants

M

i n agriculture

• Surfactants (non-ionic - N IS, ionic, blends)
• Oils (modified seed oil - MSO, crop oil concentrate
• COC)
• Stickers (e.g. polymers)

Function:

• penetration
• dispersion
• emulsifying agents
• aids

... etc.

Confidential - Produced Subject to Protective Order

Use of adjuvants is common practice for

farmers

• mmercial Applicatio

MONGLY05272938

Potential for adjuvants to affect resmidues

Do adjuvants increase residue levels?

• Limited data submitted to EPA suggested this

might be the case

• EPA concerned that US tolerances could be set

too low if based on trial data w/o adjuvants

• Prompted EPA to require "Adjuvant Prohibition"

statements on registrants' product labels

Confidential - Produced Subject to Protective Order

MONGLY05272939

Compilation of Data for Adjuvant

Investigation

437 side-by-side trials "with adjuvant.-Wi'thout

adjuvant," were compared (excluding values <

LOQ)

Data included fungicides, insecticides,

herbicides; 25

.

were predominantly oils & surfactants ; NAFTA and European trials

Confidential - Produced Subject to Protective Order

MONGLY05272940

Independent statistical analyses of data

• 1.2
• 1
• 0.8
• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 0.8

1

1.2 1.4 1.6

Log of Ratios (Residue with Adjuvant / Residue without Adjuvant)

Confidential - Produced Subject to Protective Order

Analysis to Determine Effects of Adjuvants

• n pesticide Residues

Final Report

George Cas el la, PhD Consulting Statistician

Gainesville', FL*

May 23, 2012

MONGLY05272941

Industry's Analysis Result

.......... ................................................................................................................................................................................................

0.

d

m

75% of the ratios are below 1.4

(the log of the ratios are below 0.16)..

87% of the ratios are below

(corresponding to log

.8).

• 1.2
• 1
• 0.8
• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 0.8

1

1.2 1.4 1.6

Log of Ratios (Residue with Adjuvant / Residue without Adjuvant)

Confidential - Produced Subject to Protective Order

MONGLY05272942

Industry's Analysis Conclusion

• Range of variability seen is typical of residue

data : sometimes residues are less with adjuvants , sometimes more . Were a few unusually high values (plums)

• Existing labels in the market which are "silent"

with respect to use of adjuvants are not a

• concern. Existing tolerances and dietary risk

assessments are adequate

Confidential - Produced Subject to Protective Order

MONGLY05272943

Abt's Analys is for EPA

ample, rcr

t

LL._F

T non-

en

1 :

ceI

• bs

m

I I

€I

Log Residue Pct

n

• Confidential - Produced Subject to Protective Order

I

Ln

a

I

Ln va

m

t-A r

cu

i;

n

• Ii, st

I

Mean

MONGLY05272944

AN's Analysis Conclusion

D

MALL UT non-cans o:bs

11 c

n:

• bs

ln'wa - -

r- 1 04 '.

lnR

wa = -2,.1:09:+ 0.96 9.1

VFM

The average percent differences between pesticide residues wi

and without adjuvant were positive (implying that adjuvant

increases residue), ranging from 12% %. The average percent difference for the FRUIT category was approximately 60% , while the average percent difference for this category when we eliminated

all experiments with plums was only approximately 28% .

Mean

Confidential - Produced Subject to Protective Order

,ample. rcrol:

r

:0

MONGLY05272945

EPA's Bootstrapp hig Analys is

r

a

Sample: n observations with

r

pl r

rw

i

ent from the:

na:1

data, w:

e:::re n is V

:n

fa f ate the MIRL

i

highest

e i

.u a Mean

O) with appropriate

r u in,

• w

M

It pl

each: observation of the. sample co::

step

ire

denotes the ppe:rcent increase d : ue to adjuva nt.

R L of : th is new sam pl e s

it

.

appropriate rou nd

i ng s Denote t: e MR: I'

• f th is

r

f

r

ed

sample as M RL

.

Define an:

I nd icator va riable , Z .

If MR.... _

r

use Z

Repeat steps I and:' 4 many, B, times.

Confidential - Produced Subject to Protective Order

MONGLY05272946

EPA's Bootstrapping Conclusion

"The

previous finding by mathematical derivation. The

result of the simulation implies that there is

difference in MRL estimate with adjuvant and

adj uvant. " without

MONGLY05272947

Question we should like to answer:

Is the routine variation seen in residue trials

greater than the observed variation found in the comparison of side-by-side trials with

adjuvant vs without adjuvant?

Field Res idue

• Conduct controlled stud ies

side -by-s ide trials

range of data set s izes us i ng vs no adjuvant - very expensive

Confidential - Produced Subject to Protective Order

MONGLY05272948

Dr, Davold Miller's suggestion.

A more appea li ng a lternative to this would be to te field

tria ls that we re a lready com pleted (say 1 6 fie ld tria ls for potato)

and d ivide

into two (equa l sized ) g roups (random ly) of 8

each

esti mate the tolerance for each g roup . Th is could be

repeated h u nd reds of times (by compute r), each ti me random ly

d raw i ng e igh t fie ld tra il va l ues fo r each of the g rou ps.

If it

could

demonstrated that the resu lti ng tolerance d i ffered by,

say, 75-100%,

could then say that the adj uva nt vs. non-

adjuva nt tria ls exam i ned

d iffe red

each othe r in

a manne r sim ila r to what one woul d find in repeated fiel d tra ils."

Confidential - Produced Subject to Protective Order

MONGLY05272949

Question we should like to answer:

Is the routine variation seen in residue trials

greater than the observed variation found in the comparison of side-by-side trials with

adjuvant vs without adjuvant?

Use OECD MRL Calculato r

• Sub -sample ex ist i ng , actual res idue

determine

variability

estimations data set s izes

• a more reasonable approach

Confidential - Produced Subject to Protective Order

MONGLY05272950

Test i ng of OECD Calculator

subsampling of real data sets

51 parent residue data sets from JMPR and EFSA

Each parent set had at least 20 residue values (incl

<LOQ values)

Subsampling without replacement - 10 times for each parent set

3 subsample sizes (4, 8, and 16)

MRL calculated for each subsample set (510 MRLs

for each subsample size)

22 parent residue data sets with at

residue values for subsample size of 20

Confidential - Produced Subject to Protective Order

MONGLY05272951

Process used - step by step

One subset size (eg, 8)

Residue Level

Parent Set 1

51 parent sets

subsampling

.......................................... ........................................... ........................................... ........................................... .......................................... ........................................

Residue Level

Parent Set 51 10 L 10

MRLs MRLs

L

Variability 1

subset size 8

MRL

Variability 51

subset size 8 10

MRL10

Confidential - Produced Subject to Protective Order

MONGLY05272952

MRL Varmlabollity - Percent Doifference*

*% D ifference = (Max

L- M i n

L) / M i n

L x 100

MRL Variation - Percent Difference

Unrounded Rounded

20 16

8 4

20 16

8 4

218 244 1190 3501 186 275 1400 3233

11

10 29 32 25 43 68 75 204 421

71

76 218 445 22

51 51 51

22

51 51 51 Confidential - Produced Subject to Protective Order

MONGLY05272953

MRL Variabollmity - Percent Difference

Subset size 20 Subset size 16 Subset size 8

................................................. 4\1

Subset size 4

Unrounded MRL

Variability

Percent Difference

Confidential - Produced Subject to Protective Order Subset size 20 Subset size 16 Subset size 8

.................................................................

Subset size 4

Rounded MRL

Variability

Percent Difference

.............................. .............................. .............................. .............................. .............................. .............................. ..............................

500 1000 1500 2000 2500 3000 3500

MONGLY05272954

CLA-REWG Conclusions

• Calculato r

The range of MRLs generated by the OECD MRL

from randomly produced subsets drawn from 51 actual

data sets is

251400% for subset s ize

• The smaller the data set, the larger the var iat ion
• Th is wide

is

representat ion rout i ne va riat ion

i n

superv ised res idue trial

• The observed differences are more than adequate to

accommodate variability

residues adjuvants

Confidential - Produced Subject to Protective Order

MONGLY05272955

EPA Decision on Effect of Adjuvants

MEMORANDUM

SUBJECT .

Regulatory Implications Related t Recent

,

s; c:^sr ent by l l.l.D on. Adjuvants t

z

{

i

l

i w

Lois Rossi, Director.

l

{

It rE ?

{

• Ike< iaaranon Division

Richard Keigwia, Direc tor,

Pcstici:d

Rcevaa uation Division

TO-

All Staff RD and.11RD

HFl

has examined i ,dustr;y`s analysis of the effect of adjuvants on pesticide r sidue: and

aa:e to the conclusion that existing tolerances and. risk assessments will be adequate and

to :receiv°ee questions about its

aph

ations from a regulatory and labeling standpoint.

Thu&, the purpose of this memo is to provide general guidance on these points going forward

et

• retation. is more flexible than the position taken by

.

in the past. RD and PRI) are iikel rote$ tiv°e of public l;:ealtlh, eve a if special field trials involving adjuvant use were not conducted.

The attached analysis from l.-ll D s C III.MS .C describes that because of mare significant

variability inherent in the performance of field trials and the selec t ion of a protective tolerance

level., it is not necessary to reqr.irc that adjuvants be u

d in crop field trials. Because this

l

'PA. w 'ill no longer require an a juvant restrict on on labels. Confidential - Produced Subject to Protective Order

MONGLY05272956

Summary of ChemSAC Recommendat ions

• ChemSAC does not see a need, from a risk-

perspective ,

require residue trials

• Sufficient conservatism in risk assessment

ensures that use of residue data will not under

estimate exposure , even if adjuvants were not

used in the trials

• Monitoring data used in refined risk assessments

reflect real-world farm practices , including use

and effects of adjuvants

Confidential - Produced Subject to Protective Order

MONGLY05272957

Summary of ChemSAC Recommendations

• ChemSAC does not see a need, from a tolerance-

standpoint , to require adjuvants in residue

trials

• Existing tolerances are sufficient to address

impact of adjuvants

Addition of adjuvants to previously registered

products that have restrictions would not be expected to result in tolerance exceedances

• Registrants encouraged to conduct trials with

adjuvants , even though they are no longer needed

to fulfill specific requirement

Confidential - Produced Subject to Protective Order

MONGLY05272958

THANK YOU

Confidential - Produced Subject to Protective Order

MONGLY05272959