Umtn Satn
ivciion
Agency
OHic* of
hgiaM and Tokic Submncai
WaMflgtan OC 2SHS0
July 1983
S4o! i-fj-m
alpha,aJpha,aipha - trifluoro - 2,6 -
dinitro - N,IM - dipropyi - p - toluidine
Trifluralin (TREFLAN®)
Position Document 4
-------
Treflan® Position Document 4
Special Pesticide Review Division,
Office of Pesticide Programs,
Office of Pesticides and Ibxic Substances,
U.S. Environmental Protection Agency
-------
BIBLIOGRAPHIC INFORMATION
PB82-263252
Trifluralin (TREFLAN (Trade Name)), alpha,alpha,alpha -
trifluoro -2,6 - dinitro - N,N - dipropyl - p - toluidine.
Jul 82
PERFORMER: Environmental Protection Agency, Washington, DC.
Office of Pesticides and Toxic Substances.
EPA-540/9-82-011
Contents: Elanco's chronic feeding study; Revised exposure
estimate; Revised cancer risk estimate; Ecological effects;
Comments relating to risk; Comments relating to benefits;
Comments relating to testing requirements/regulatory
options.
KEYWORDS: *Pesticides, *Toxicology, *Tr ifluralin.
Available from the National Technical Information Service,
Springfield, Va. 22161
PRICE CODE: PC A05/MF A01
i
-------
Acknowledarren ts
Ann Barton - Science Advisor, FED, QPP
Dr. Clayton Bushong - Biologist, EEB, HED, QPP
Chris Cnaisson - Biocnemist, IB, HED, OPP
Dr. Chau Cnen - Statistician, CAG
Laurence A. Cook - Attorney, CGC
Et. David Coppage - Aquatic Biologist, EEB, HED, OPP
Dr. Julian Donoso - Chemist, RCB, HED, QPP
Dr. Harry Gaede - Econanist, EAB, BFSD, OPP
Timothy A. Gardner - Section Head, SPED, QPP
Linda Garczynski - Writer/Editor, SPRD, OPP
Dr. Bernard Habeman - Pathologist, CAG
Hcmer K. Hall - Branch Cnief, SPRD, OPP
Dr. Ricnard Hill - Senior Science Mvisor, OPTS
Dr. touis Kasza - Pathologist, TB, HED, OPP
Carol E. langley - Project Manager, SPRD, QPP
Dr. Irving Mauer - Geneticist, "IB, HED, OPP
Dr. Robert McGaugry - Tbxicologist, CAG
Tem Miller - Project Manager, SPRD, OPP
Abranam Mittelman - Chemist, EFB, HED, QPP
Ricnard Mountfort - Product Manager, RD, OPP
Qnil Regelman - Chemist, EFB, IIED, CFP
Dr. David Severn - Chemist, EEB, HED, OPP
Les Tbuart - Fisheries Biologist, EEB, HED, QPP
Christine Watson - Typist, SPRD, OPP
Delores Williams - Typist, SPRD, QPP
Dr. Robert Zendzian - Pharmacologist, 1DX, HED, OPP
/
-------
Table of Contents .
*
X¦ Introduction•»»•«••••¦«#•«•#»»••»«»••#••»»•••••»«•*••*«»¦••#»• X
II* liew Information and Revisions to tne PD 1/2/3 4
)
A. Elanco's Chronic Feeding Study............ 4
B. Revised Exposure Estimate 7
1. Dietary Exposure ....... 8
2. Mixer/Applicator/loader Exposure.. 12
a¦ Factors A^jIi@d to Exposure Data..................... 12
b. Summary of Mixer/Applicator/Loader Exposure
Estimates 13
3. Been try Exposure 16
a. Exposure to Vapor or Particulate Matter in trie Air... 16
b. Dermal Exposure Fran Contact with the Soil........... 21
C. Revised Cancer Risk Estimate 25
1. Rationale for Revisions 25
5. NDPA 25
b. Trifluralin. 25
c. C7/Cg Nitrosarames................*..............*...31
2• Dietary Risk....... 31
a. Trifluralin.......... 31
b. A. 32
c. Cj/Cg Nitrosaraines 32
d. Comparison of Dietary Risk Estimates With Those
of the H) 1/2/3 32
3. Mixer/Applicator/Loader Risk....... 33
a. Calculations and Assumptions 33
b. Comparison of Risk Estimates with Those of the
PD 1/2/3 .' 36
4. Reentry Risk. 37
a. Nitrosamines 37
b. Trifluralin.......................................... 30
5. Summary of Dietary arid Worker Risk .. 38
D. Ecological Effects 40
III. Analysis of Ccnments. 43
A. Goraraents Relating to Risk.. 43
1. Worker Exposure 43
a. Exposure Estimate of N-nitrosodipropylamine
(NDPA) 43
b. Mixer/Applicator/Loader Exposure to NDPA. 44
1} Inhalation...................................... 44
2) Dermal45
c. Reentry Field Worker Exposure... 46
-------
2. Dietary Exposure 48
3. Ibxicology. 50
a. Cancer Risk Assessment 50
X5 Elanco s Qj'E'icn^s30
2) American Cyanamid1 s Garments 52
b. NDPA Oncogenic and Mutagenic Risk.................. 52
c. Spindle Effects 53
d. Spindle Effects Ihresnold 55
B. Garments Relating to Benefits.'.. . 56
1. Base Planted. Acres 56
2. Base Ccrmodity Price. #**••* 56
3. Relative Importance of Prowl as a Treflan
Alternative 57
4. Tne Relative Economic Importance of Trifluralin in
tne Current Soybean/Cotton Herbicide Market............ 58
5. The Relative Economic Importance of Trifluralin
in tne Future Soybean/Cotton Herbicide Market......... 59
C. Cootents Relating to Testing Requirements/Regulatory
Options[[[ 60
1. Benzimidazole Metabolites of Trifluralin............... 60
2. Reproduction and Te r a tolog y............................ 62
3. Mutagenicity Including Heritable Spindle Effects 62
a. DtlA/Gene Effects................................... 63
b. Spindle Effects 1\2S t ing............................ 65
4. Labe1ing Requirements6*7
I\/e Conclusions and Requirements...........................u....... 69
¦ i^ f-K/s incwf11 al C+"a+>amDntT'j-sy p'in1 a 71
* IL tQ vTiG UOHL lQCiiulo J. DtdLCiucnt OX rOOilUXo •••0«e»ce« i X
B. Testing Requirements72
1. Mutagenicity Itesting Requirements 72
2. Otner Testing Requirements » 72
Bibliograpnv of Comments
References
Appendix A. Comments by tne Scientific Advisory Panel
Appendix B. Comments by trie U.S. Department of Agriculture
Appendix C. Exposure Data on Treflan® from Mittelman (1978)
(Available on Request)
-------
Executive Suitroary
On August 30, 1979, the Agency issued a preliminary notice of determination
concerning the rebuttable presumption against registration (RPAR5 of all
pesticide products containing trifluralin (44 FR 50911), The Agency had
determined that the trifluralin contaminant, N-nitroso-di-n-propylamine (NDPA)
did meet or exceed the oncogenic risk criterion, thus necessitating an in-depth
review to examine the risks and benefits associated with the use of products
containing trifluralin.
In the Position Document (PD 1/2/3), the agency proposed to cancel all
registrations for products containing trifluralin unless registrants modified
labeling of their products to reflect less than 1 ppm NDPA contamination. When
the PD 1/2/3 was issued, the principal registrant, Elanco Products Co.
(Elaneo), had already instituted manufacturing methods to reduce the NDPA
contamination level to less than 1 ppm NDPA.
The Agency also indicated in the PD 1/2/3 that the registrants would need to
perform additional testing of trifluralin for reproductive effects,
teratogenicity, and mutagenicity (including heritable spindle effects). The
Agency specified that metabolism studies would need to be conducted on
trifluralin containing NDPA to assess its ability to reach the marraalian gonad
in a metabolically active form. Mutagenicity testing was also recommended for
trifluralin1 s benzimidazole metabolites. At the time the PD 1/2/3 was issued
Elanco was conducting a chronic feeding study to assess oncogenic effects due
to administration of trifluralin with an NDPA concentration of less than O.Di
ppn, the limit of detection.
The Agency received comments on the Preliminary Notice of Determination and the
PD 1/2/3 frcm the Secretary of the U.S. Department of Agriculture (USDA), the
Scientific Advisory Panel ('SAP), and 17 other interested parties. The USDS
concurred with the Agency's proposed decision to continue registrations of
products containing trifluralin with a limit of 1 ppm NDPA contamination.
The SAP agreed with the Agency's position and had specific recarmendations for
testing requirements. Further testing was recortroended by the SAP for
oncogenicity, mutagenic effects and spindle effects, using the product as
currently produced with less than 1 ppm NDPA. The Elanco chronic feeding
study submitted by Elanco fulfills the SAP's request for oncogenicity testing.
The Agency agrees with the SAP about the need far further mutagenicity and
spindle effects testing, but has not required the specific tests specified by
the SAP. Instead, the Agency believes a better assessment of mutagenicity can
be obtained by requiring further microbial tests, a dominant lethal test, and
assessment for presence of active oanpounds in the manraalian gonad. Regarding
testing far spindle effects, the Agency will identify outside scientists to
help delineate a meaningful research program to assess risks from spindle
inhibitors since current test systems are not adequate or sensitive enough.
I
-------
When the appropriate tests have been identified the registrants will be
required to identify any mutagenicity problem related to inhibition of spindle
fiber formation or function. lastly, the SAP questioned the need for
mutagenicity"testing of benzimidazoles, potential metabolites of trifluralin,
as recommended in the PD 1/2/3. The Agency has determined that although
benzimidazoles have been reported in in vitro studies, their presence in
mammalian in vivo metabolism studies Ts uncertain, arrf, if present, the Agency
assumes they probably exist in minute quantities. Therefore, the
registrants will not be required to conduct mutagenicity studies cn
benzimidazole metabolites.
Of the seventeen other interested parties, eight concurred with the Agency's
position, and six wanted more information, which was subsequently supplied by
the Agency. Another had a question about worker exposure which is answered in
the ED 4. Elanco Products • Co. and American Cyanamid Co. contented in detail on
the PD V2/3. Both submitted comments criticizing various aspects of the
Agency's benefit analysis. The Agency has qualitatively reevaluated its
analysis and finds it to be appropriate and reasonable. The Agency has
determined that the benefits have riot changed appreciably since the PD 1/2/3
was issued.
Elanco also submitted ocntments on the Agency's dietary and worker exposure and
risk analysis; on the Agency's proposed requirements for additional
reproduction, teratology, and mutagenicity studies; on the establishment of a
maximum NDPA level in trifluralin containing products; and on the Agency's
proposed label amendment. Elanco also sufcmitted results of their trifluralin
chronic feeding study, which indicated that high doses of trifluralin, with
NDPA levels less than 0,01 ppm, are associated with a statistically significant
increase in tumors of the kidney, bladder, and thyroid of rats when compared to
controls. As a result of Elanco's comments and results from the chronic
feeding study, the Agency has recalculated and reevaluated the dietary and
worker exposure to nitrosamines and trifluralin and the risks associated with
this exposure. The Agency has determined that the overall risk associated with
exposure to Treflan® has not changed appreciably since the PD 1/2/3 was issued;
the added risk due to exposure to trifluralin is offset by the reduction of
NDPA by Elanco.
Because neither the risks nor the benefits have changed appreciably since the
PD 1/2/3 was issued, the Agency has determined that the benefits still outweigh
the risks if specific requirements are met. The Agency is currently
requiring that all registrations for products containing trifluralin be allowed
to continue only if the following requirements are met:
(15 Within 30 days after notification, registrants (present and future) must
amend the inert ingredients statement in the Confidential Statement of
Formula to reflect a total N-nitrosamine concentration of no greater than
0.5 ppm for technical trifluralin products. The inert ingredients
2
-------
statement in the Confidential Statement of.Ftormula of all formulated
trifluralin products shall reflect a total N— nitrosamine oontairsination set
as a function of the amount of trifluralin in the end use product. The
total N-nitrosamine content in formulated products will be allowed to be
twice what would normally be found in a straight-forward dilution. For
example, a 25% formulated product could contain up to 0.25 ppm and a 501
formulated product could contain up to 0.50 ppm. This increase which can
occur beyond a simple dilution of the technical material (containing an
upper limit of 0.5 ppm) is to allow for possible nitrosamine generation to
i n**^ i11 sat* l /"vri 'nrr'^VNacis yon *i e+¦"»"* surfr" <5 mi ie"4"* si * "i ea 4*l,^ca
U ImLL XI IM ' J Kf 1 liUxd LxUI 1 mL vwv D>3 * -Li It 4 CU J, ~ if <* ¦ J 1111 J~*s ^ tri 1 » c#JV XSCI y, Jt.
Agency that the level of total N-nitrosamine as stated in the Confidential
Statement of Ebrmulas is not exceeded.
(2) Registrants must perform mutagenicity testing on trifluralin (as stated in
the FD 1/2/2) as veil as testing for the presence of metabolically active
product (trifluralin and NDFA) in the mairmalian gonad as well as germinal
testing.
(3) Registrants must perform testing of trifluralin for reproductive and
teratogenic effects, since there are not adequate data to assess risks due
to trifluralin in these areas.
(4) Registrants must conduct a field monitoring study to assess potential toxic
effects' to aquatic organisms, since, in light of new data and a
reevaluation of existing information, the Agency determined that
trifluralin could reach aquatic environments through soil runoff and,
UsCaUsc OX OiOCOriC"iltX^clwiQu clPJLJL 1 wicS ? OvLlJLQ Dc nctr illX ULL ZO I.ISO cutQ
molluses.
An economic analysis concerning the cost of the data requirements was
conducted. The estimation of the total cost of the required data and the
estimated value of future trifluralin earnings indicate that it is economically
->*¦» 4 W1 f/-y i»Vw von i cst'Tant'e i»Vvi 14 fry ynwnn•tF-vxS
1 r** r!H 1 Lit f** 1 Cjr Li K f 1>)1 f Hi li H 113 UcV C,i,tJV LI fc* mClLQ LfcryUiX^CU LlX v. a II 11" ,1,1 1 Urf; J
mm *mr mmmm mmm. .-mmm -J"— ¦— — w w
registration.
3
-------
I. Introduction
Section 12(a)(1)(A) of tne Federal Insecticide, Fungicide, and Rodenticide
Act, FIFRA, (7 U.S.C. 136 et sec.) prohibits the sale or distribution of
pesticide products which are not registered by the EPA Administrator, and.
Section 3 of tne Pet sets fortn the registration procedures. Before a
pesticide may be registered, however, tne Administrator must detetmine that its
use will not result in "unreasonable adverse effects on tne environment,"
[Sec. 3(c)(5)(C)], defin«3,in Section 2(bb) of FIFRA as "any unreasonable risk
to man or tne environment, taking into account tne economic, social, and
environmental costs and benefits of the use of any pesticideTherefore, any
decision an pesticide registration must take into account both tne risks and
tne benefits associated witn the pesticide's use.
Under Section 6(b) of FIFRA, the Mministrator may issue a notice of intent to
cancel tne registration of a pesticide-or to change its classification if it
appears that the pesticide or its labeling "does not comply with tne provisions
of (FIFRA) or, when used in accordance with widespread and commonly recognized
practice, generally causes unreasonable adverse effects on tne environment."
Tnus, tne Administrator may cancel the registration of a pesticide that no
longer satisfies the statutory standard for registration. The frtainistrator
may also change tne classification of any use of a pesticide if he determines
that such a cnange "is necessary to prevent unreasonable adverse effects on tne
environment" (FIFRA 3(d)(2)).
Tb implement its authorized functions, tne Agency designed the Rebuttable
Presumption Against Registration (RPAR) process to gather data on the risks and
benefits associated witn the uses of suspect pesticides. By allowing all
interested parties to participate by submitting information, this process
enables EPA to make balanced decisions concerning problem pesticides. The RPAR
process is set forth in 40 CFH 162.11 which describes various risk criteria and
provides tnat an REAR shall arise if tne Agency determines that any of these
criteria has been met or exceeded.
Once a rebuttable presumption nas arisen, registrants, applicants, and
interested persons may submit evidence in rebuttal, or in support, of the
presumption. These parties may also submit evidence on the economic, social,
and- environmental benefits of any use of the pesticide. If the presumptions of
risk are not rebutted, tne evidence pertaining to benefits must be evaluated
and considered together with the evidence pertaining to risk. Various risk-
reduction measures and their costs are analyzed. The Agency then determines
whether tne pesticide may be regulated so that a balance is achieved between
risks and benefits. If the statutory balance cannot be reached for any given
use, the registrations for that use must be cancelled.
4
-------
Cn August 30, 1979, tne Agency issued a preliminary notice of determination
concerning tne rebuttable presumption against registration and the continued
registration of all pesticide products containing trifluralin (Treflan® EC,
registered trademark) (44 FR 50911). The trifluralin KPAR review differed
somewhat frcm the typical RPAR review in that two of the analytical RPAR phases
(tne initial determination that the risk criteria had been exceeded and the
weighing of risks and benefits to determine the appropriate regulatory action)
were combined. The Position Document (PD) 1/2/3 presented a detailed
description of this KPAR review and proposed a decision to conclude the RPAR
process. Tne PD 1/2/3 did not accompany the Notice, but copies were provided
to all registrants and any other concerned parties.
In that Position Eocument the Agency proposed to cancel all registrations for
products containing trifluralin unless registrants modifed labeling of their
products to reflect less than 1 ppm N-nitrosodipropylamine {NDPA)
contamination. The principal registrant, Elanco Products Company (Elanco),
had already reduced the 15DPA contamination to less than 1 ppm when the PD 1/2/3
was issued. Tnis requirement was intended to pertain to both the product
label and the confidential statement of formula maintained by the Pgency.
In addition, the Agency indicated that registrants of trifluralin
containing products would nave to do the following: certify this level to
be the upper limit of contamination, advise the Agency of quality control
procedures instituted to assure tnat tne level of NDPA did not exceed 1 ppm,
and maintain accurate quality control records cn these products.
The Agency indicated that the registrants would be required to submit the
results of tne ongoing oncogenicity study, and to perform additional testing of
trifluralin for reproductive effects, teratogenicity, and mutagenicity
(including heritable spindle effects). The Agency also stated that
mutagenicity testing was required on tne benzimidazole metabolites of
trifluralin and that metabolism studies were required cn trifluralin with NDPA
in order to assess the ability of trifluralin and/or NDPA to reach the
marrmalian gonad in a metabolically active form..
Sections 6(b) and 25(d) of FIFRA require that the Agency surmit notices of the
proposed decision to the Secretary of the U.S. Department of Agriculture (USDA)
for comment on the impact of the proposed action on the agricultural economy
and to the FIFRA Scientific Advisory Panel (SAP) for ccnment on the impact of
the proposed action cn nealtn and tne environment. In accordance with FIFRA,
tne Secretary and the SAP were invited to ccnment in writing within 30 days of
receiving the notice. The Agency is required to publish their written
comments, if submitted within 30 days of the receipt of tne Notice, and the EPA
Administrator's response to these oanraents.
Although not required to do so under the statute, the Agency decided that it
was consistent with the purposes of the RPAR process and the Agency's overall
policy of open decisionmaking to also afford registrants and other interested
5
-------
persons an opportunity to Garment cn the basis for trie proposed action while it
was under review by tne Secretary of Agriculture and the SAP. The Bosition
Document 1/2/3 was therefore made available to all interested parties for
comment.
Since tne preliminary notice of determination and notice of availability of the
FD 1/2/3 was published on August 30, 1979, the agency has received new
information and a number of axments from interested parties. The new
information, including tne results of a chronic feeding study submitted by
Elanco, resulted in revisions in the exposure and risk analyses; this is
presented in detail in Section II of this document. Responses frati the SAP,
tne USDA, and other interested parties have been analyzed and are addressed in
Section III of this document. Section IV sunraarizes tne Agency's decision
concerning pesticide products containing trifluralin. The regulatory position
takes into account all comments frcm the SAP, USDA, and other interested
parties and is a slight modification of "option 3" set forth in the PD 1/2/3.
Because tne new information did not necessitate a change in this regulatory
position, tne Agency determined tnat it was not necessary to have the SAP
review the study.
The responses frcm the SAP and the USDA are presented in their entirety in
Appendix A and Appendix B, respectively. Appendix C contains exposure data on
Treflan® from the Agency (Mittelman,l978) wnich was incorporated in the
calculations for estimating current exposure to trifluralin and nitrosamines.
Appendix C, and Appendix D which contains the PD 1/2/3, are available on
request. All comments are available for review in the public file in tne office
of the Document Control Officer, Chemical Information Division, Rxxn E 477, (TS-
793), EPA, 401 M Street, S.W.; Washington, D.C; 20460.
6
-------
II. New Information arid Revisions to the PD 1/2/3
A. Elanco's Chronic Feeding Study
In trie trifluralin Position Document 1/2/3', the Agency mentioned tnat tests
were being performed'by tne principal trifluralin registrant, Elanco, to
determine the oncogenicity of trifluralin in wnicn tne level of I EPA
contamination was belcw the limits of detection. These tests were
completed and the results were submitted to the Agency in Septentoer, 1980.
Studies were conducted cn two rodent species: B6C3F1 mice and Fischer 344 rats
(Elanco, 1980d). The- test substance, purified trifluralin containing NDPA at
levels below the 0.01 ppm analytical detection limit, was administered in the
diet over a period of two years. At the end of the study period the survivors
were sacrificed and pathological examinations were made of each animal in the
study. Hematology and blood chemistry tests were also performed.
In tne mouse study, a total of 720 animals were tested, using a control group
of 120 males and 120 females, and three treatment groups of 80 males and 80
females each. The diets of the treatment groups contained trifluralin in
proportions of 563 ppm, 2250 ppm and 4500 ppm respectively. Elanco scientists
concluded that the "primary effects of treatment...were reduced body weight and
manifestations of renal toxicity", and tnat "tnere was no evidence that
treatment induced increases of benign or malignant neoplasias" {Elanco 1980d).
The Agency considers tnis irouse study to be valid, and agrees with Elanco that
the results of the study do not show evidence of oncogenicity for trifluralin
in B6C3F1 mice (Kasza, 1981; Chen and Haberman, 1981).
In tne rat study a total of 480 rats were distributed among a control group
and three treatment groups, each group containing 60 males and 60 females.
Rats in the treatment groups were fed diets containing 813 ppm, 3250 pro and
6500 ppm of trifluralin respectively. Elanco found tnat the only significant
increase in malignant neoplasms occurred in the kidneys of male rats at all
doses; the incidences were dose-related. An increase in benign bladder
neoplasms was also noted by Elanco in all dose groups, an increase which was
significant and dose-related for female rats,'and for male and female rats
considered together, but was not significant for male rats alone. A slight
increase in benign interstitial cell tumors of the testes was found, in all
dose groups; these increases were not reported to be statistically
significant. Elanco scientists also noted a high incidence of chronic renal
disease and tne presence of renal calculi. Elanco reported the presence of
chronic renal disease, characterized as progressive glonerulonephrosis, as
being significant far tne tws highest dose groups. Renal calculi were also
found in a high proportion of rats in all treatment groups; the incidence of
renal calculi was dose-related. The incidence of renal calculi in the hignest
dose group was 42 out of 60 male and 42 out of 60 females.
ElanoD does not interpret the increase in kidney and bladder neoplasms to be
evidence of tne oncogenicity of trifluralin. Elanco states that,the fact that
botn urinary tract tumors and renal calculi were found in this study, but not
in previous rat studies, indicates tnat tne test compound, trifluralin, was not
7
-------
directly responsible for tne excess tumors, and that other causal factors may
have been present (Elanco, 198Od; Elanco 1980e). In particular, Elanco
implicates tne presence of renal calculi as a potential cause of the tumors.
Renal calculi, described as '"small mineral deposits in tne renal pelvis
epitfielium and/cr tne calices" (Elanco, 1980d) and characterized as "similar to
kidney stones" (Elanco, 19806), were found in a high proportion of test
animals. Since renal calculi have bean shown to cause mechanical damage
leading to neoplastic cnanges like tnose observed in the study, Elanco
postulates tnat tne excess tumors may have resulted frcm the presence of tne
calculi, or sane interaction between the test substance and calculi. Elanco
concludes that tne results of the rat chronic feeding study do not permit a
conclusive evaluation of tne oncogenicity of trifluralin, since tne presence of
renal calculi provides an alternate explanation for the formation of the
urinary tract tumors. Elanco scientists also suggest that tne neoplastic
response in the urinary tract seen in this study may be related to traits
peculiar to Fischer 344 rats, since studies with trifluralin using other
strains did not result in similar findings.
Agency scientists have reviewed tne Elanco rat study and have found it to be a
valid test for tne oncogenicity of trifluralin (Kasza, 1981; Chen and tiaberman,
1981; Barton, 1981). The Agency has arrived at a different interpretation of
the results, however, regarding the types and significance of tumors showing
increases in frequency as a result of treatment witn trifluralin. Significant
increases were found by Agency scientists in two types of tumors: tumors of
the urinary (bladder and kidney) transitional epithelium in both male and
female rats and follicular tumors of tne thyroid in male rats (Chen and
Haberman, 1981; Barton, 1981). Information on tne incidence of these tumors is
summarized in Table 1 (Haberman, 1981; Chen and Haberman, 1981).
In its analysis of the reported incidence of urinary tract tumors, the Agency
considers that bladder papillomas, bladder carcinomas and renal pelvis
carcinomas should be counted together, since they are similar tumors of a
continuous epithelial tissue found in both tne bladder and renal pelvis and
exposed to tne same substances in formation of tne urine (Kasza, 1981). These
neoplasms have been grouped together as tumors of the urinary transitional
epithelium (Kasza, 1981; Chen and Haberman, 1981; Barton, 1981). In male rats,
there were such tumors in 3 of 60 low dose animals, 3 of 60 in the middle dose
group, and 6 of 60 at the highest dose. Compared to an absence of sucn tumors
in 60 control animals, this increase is -significant for the nigh dose at a
value of p = 0.014 by tne Fisher exact test. For female rats, no urinary tract
tumors were found in tne low dose groups of 60 animals each, but 1 tumor was
found in the middle dose group aid 5 tumors were found in the high dose group.
Compared to the absence of sucai tumors in the 60 control animals, this increase
at tne high dose level is significant at a value of p = 0.029 by tne Fisher
exact test.
The Agency differs with Elanco about the significance of tne presence of renal
calculi in many of the treated animals. There are basically three aspects of
tne cnronic feeding study results which, in tne judgement of Agency scientists,
8
-------
Table 1
Turner Incidence^An Rats in the Elanco Chronic Feeding Study of Trifluralin^
Tumor Type and Site
Sex
0
Dose (in ppm)
813 3250
6500
Statistical Significance
(Fisher Exact Test)
(High Dose vs. Control)
Transitional Cell Papilloma
Male
0/60
3/60
3/60
6/60
0.014
and Carcinoma
(Kidney and Bladder)
Female
0/60
0/60
1/60
5/60
0.029
Follicular Adenoma and
Male^
5/60
2/60
8/60
13/60
0.036^
Carcinoma
(Thyroid Gland)
Combined Tumors^
Male^/
5/60
5/60
9/60
17/60
0.004
(Kidney, Bladder and/cr
4
Thyroid)
a/ Nuirtoer of animals with tumcrs/nunber of animals in treatment or control group.
b/ Haberman, 1981.
c/ Each rat had at least one tumor.
d/ Increase in tumor incidence is significant for males only.
e/ Dose response is significant at a level p< 0.01 when computed by the Cochran-Armitage Test.
-------
are not consistent with Elanco's conclusions that the presence of renal calculi
are likely to have produced the observed increase in urinary tract tumors.
First, the tumors are found in both the bladder and renal pelvis, but calculi
were not observed in the bladder in this study. Since the mechanism proposed
by Elanoo is a progression frcm the formation of calculi to hyperplasia, then
to neoplasia, such a mechanism would not explain the occurrence of tumors in
trie bladder, where no calculi were found (Kasza, 1981).
Second, "trie "calculi1 reported in the renal pelvis were microscopic
calcifications rather thai stones. Evidence in the literature provides
substantial confirmation of the concept that stones can cause tumors of the
urinary tract in rats. Tne evidence that crystals or microscopic
calcifications can cause urinary tract tumors in rats or in humans is more
equivocal." (Kasza, 1981). Sinoe stones and microsoopic 'calcifications differ
in their tumor-causing ability, Agency scientists do not oonsider the studies
cited by Elanoo concerning stones to be directly relevant to the question of
the tumorigenicity of the microscopic calcifications found in the present
study.
Third, if calcification, hyperplasia and neoplasia are sequential stages in
the development of tuners, a correlation between the dose—response curves for
eacn of those oonditons would be expected. However, the dose-response curves
lot GuJlcxLxCo uxonSf nyp&L^los uo- cuno n€opia5 is qg Tow snow txjnsistsriL
correlations" (Kasza, 1981).
In view of these serious inconsistencies between the experimental results and
Elanco's proposed mechanism for oncogenicity, trie Agency has concluded that
Elanco nas not established the role of the observed micro-calcifications in the
formation of urinary tract tumors with sufficient certainty for the Agency to
alter its conclusion tnat trifluralin treatment is associated witn the
production of kidney and bladder tumors in rats.
"TYia Jopncv also fi Qianificant ®vidpnce nf an increased incidence of tumors
<** I (U V i Y mi tm * i>M» *3 tHJ i IJnk Jb VU| I V< W V .ftiS. .r I iWC W ^ ifli I JrVtVi IVV Va> t w
of the follicular epithelium of the thyroid in male rats (Chen and Haberman,
1981j Barton, 1981). The Elanco study found five different tuuor types in tne
follicular epithelium: follicular adenoma, follicular papillary adenana,
follicular cystadenoma, papillary crystadencma, and follicular carcincma. The
Agency does not consider these lesions to be different tumors for the purpose
of oncogenicity testing of a pesticide, since they occur in a single cell type
and are differentiated only by having cystic or papillary components (Kasza,
1981). In male rats, tnyroid follicular epithelial tumors were found in 2 of
60 low dose animals, 8 of 60 at the middle dose, and 13 of 60 at the high
dose. Compared to an incidence of 5 tumors in 60 control animals, tne results
in the hign dose groi?> are significant at p=Q.036 by the Fisher exact test, and
significantly dose-related at p<0.01 by the Cbcnran-Armitage trend test. No
significant increase or dose-related trend was found in female rats. The Agency
concludes that trifluralin treatment is also associated with tne production of
thyroid tumors in male rats.
9
-------
B. Revised Exposure Estimates
In addition to trie cnronic feeding study, otner information has cane to tne
Agency's attention since 1979 which results in changes in or amendments to the
exposure estimates presented in tne PD 1/2/3.
Trifluralin itself has been shown to be associated with tumor production in
laboratory animals. Therefore, it became necessary for tne Agency to estimate
dietary exposure to the general population and to estimate exposure to
mixer/applicator/loaders and also to workers reentering fields treated with
trifluralin. The TO 1/2/3 had only addressed exposure and risk associated with
NDPA.
The Agency also received further information on tne nitrosamine contamination
in Treflan® EC. The NDPA exposure estimates in the PD 1/2/3 were based on a
contamination of 5 ppm. However, since that time Elanco has altered the
procedures in trie manufacturing process and has reduced this level of
contamination. Elanco (1980a, 1900b) reported tne results of 635 analyses of
Treflan® EC far both NDPA and total nitrosamine contamination. The samples,
representirtg production nans from September 1, 1979, to May 30, 1980, were
found to contain average NDPA residues of 0.10 ppm (0.00 to 0.962) and an
average of 0.12 ppm for total nitrosamine residues (0.00 to 0.96). The total
nitrosamine residues included both NDPA (Cg) and otner nitrosamines
identified as having either seven or eight carbon atcms. These will be
referred to as C^/Cg nitrosamines in this document and are assumed to be
present at an average level of 0.02 ppm. This figure was obtained by
substracting the average NDPA contamination of 0.10 ppm from the average total
nitrosamine contamination of 0.12 ppm (Regelman, 1981a). Quality control in
tne manufacture of Treflan® EC is expected to maintain nitrosamine impurities
at less than 1 ppm. Elanco (1981) has identified tne chemical structure of tne
Cj nitrosamine and has partially identified that of the Cg nitrosamine;
this information is classified as confidential, however, and will not be
presented in this document.
Since 1979 wnen tne PD 1/2/3 was issued, Treflan® EC has been registered for
weed control in barley and grain sorghum. Dietary exposures for these two
crops have been estimated and have been included in Table 2 which presents
estimates for exposure to trifluralin, F3DPA, and C7/Cg nitrosamines.
Worker exposures for these two crops could not be calculated because
information cn tne hours spent by tne workers during application, mixing and
loading, and during reentry activities is not readily available. As shown in
Table 2, tne increase in dietary expDSure for barley and sorghum is extremely
small. Based on exposure estimates frcm other uses, it is reasonable to assune
there would be an extremely snail increase in worker exposure as well.
10
-------
1. Dietary Exposure
In trie PD 1/2/3, trie dietary exposure estimates for NDPA were based on
tolerances for trifluralin. The risk from the dietary exposure to NDPA was
stated to be a maximum number based on tne assumption tnat trifluralin residues
were present in concentrations equal to tneir established tolerances which
ranged froti 0.05 to 2.00 ppm. It was stated in tne PD 1/2/3 (p.81) that NDPA
residues were probably mucn lower tnan tnose estimated.
Because new data indicated that tnere oould 'be risk associated with
trifluralin itself (Elanco, 1980d) and the Cy/Cg nitrosamines as well
(Elanoo, 1978a), trie agency decided to base the new dietary estimate on studies
wnicn measured actual residues of trifluralin in order to obtain a more
accurate estimate of dietary exposure. These studies had been submitted by'
Elanco in support of petitions for tolerances for the many food commodities
treated with Treflan®. In using the residue data (summarized by Regelman,
1981a), trie Agency is assuming that the studies are representative of "real
world" conditions and that, therefore, the exposure estimate based on these
data will reflect more accurately tne actual dietary exposure to trifluralin
and, thus, NDPA, and C^/Cg nitrosamines.
This dietary estimate will include, therefore, potential exposure to
trifluralin, NDPA, and C7/Cg nitrosamine residues in Treflan® EC-treated
crops, since the risk is associated with tne active ingredient as well as with
tne various nitrosamine impurities.
All petitions for tolerances which had previously been submitted to the Agency
were reviewed and virtually all reported residues of trifluralin were either at
or below tne limit of detection of tne method used (0.01 ppm). The reported
residues (Table 2) are approximately one fifth of the established tolerances.
Assuming the data to be representative of treated crops generally, tne Agency
assigned a value of 0.01 ppm trifluralin to all food canrtodities in which NDR
(no detectable residues) or <0.01 ppm were reported. It is reasonable to
assume chat actual residues of trifluralin are at or below this level.
Residues were found in some of the food commodities studied, but west were
below the limit of detection. Only in carrots, peppermint oil and spearmint
oil were residues consistently found above the limit of detection. In carrots,
residues ranged from 0.49 to 0.86 ppm. Elanoo observed that these residues
were not uniformly distributed' throughout tne carrot, concentrating mostly
within the outer 1/16 inch (74 percent). Assuming that washing or peeling of
the outer skin would reduce exposure to trifluralin below 0.86 ppm the Agency
used an average value of 0.65 ppn in the dietary estimate. In both peppermint
oil and spearmint oil, residues ranged from 0.57 to 1.71 ppm. Since residues
could routinely be present at these levels tne Agency chose to use the
tolerance level of 2,.00 ppm in the dietary estimate.
11
-------
Table 2
Dietary Exposure to Trifluralin, NDPA and C-j/Cg Nltrosamlnes
Dally Intake (rng/kq diet)
Trifluralin
Residues®/
Fraction
Trifluralin
NDPA!^
c7/c^
of Food
of Crop
-------
Table 2 (continued)
Dietary Exposure to Trifluralin, NDPA and Ci/Cr Nitrosamines
Daily Intake (mg/kg dictTT.
Trifluralin Fraction Trifluralin NDPAiy C-j/C^Z
Residues^/ of Food of Crop
Commodity
(ppm)
in Diet
Treated
(xlO-6)
(xlO-12) (xlO-12)
Seed/Pod Vegetables, Beans
<0.01
.0204
0.756
154.22
34.66
6.93
Soybeans
<0.01
.0092
0.377
34.68
7.79
1.56
Peas
<0.01
.0069
0.138
9.52
2.14
0.43
Sorghum, grain
<0.01
.0003
0.0206/
0.06
0.01
0.00
Spearmint Oil
2.003,/
.0003
0.077
46.20
10.38
2.08
Stone Fruits
<0.01
.0125
1.000
125.00
28>09
5.62
Sugar, Cane & Beet
<0.01
.0364
0.117
42.59
9.57
1.91
Sunflower Seeds
<0.01
.0003
0.650
1.95
0.44
0.09
Wheat Grain and Straw
<0.01
.1036
0.011
11.40
2.56
0.51
TOTALS
2992.62
672.50
134.50
a/ Regelman, 1981a.
b/ Assumes that NDPA levels continue to remain <0.10 ppm in Treflan EC, that C7/C8 levels continue
to reirain <0.02 ppm in Treflan EC, and that current usage does not increase.
Daily Intake (Trifluralin) = Trifluralin Residues x Fraction of Food x Fraction of Crop
in Diet Treated
c/ The ratio of trifluralin to NDPA in Treflan EC (44.5% trifluralin) is assumed to be 4,450,000:1
(445,000 ppn:0.10 ppm). Daily Intake (NDPA) = Daily Intake (trifluralin) f 4,450,000
d/ The ratio of trifluralin to C7/C8 nitrosamines in Treflan EC (44.5% trifluralin) is assumed
to be 22,250,000:1 (445,000 ppm:0.02 ppm).
Daily Intake (C-j/C%) = Daily Intake (trifluralin) -f 22,250,000
n
e/ langley, C., 1981.
f/ Residues ranged from 0.49-0.86 ppm.
g/ Residues were assumed to be present at the tolerance level.
-------
Fear purposes of this dietary estimate, the Agency assumed that NDPA was
present in Treflan® EC at a level of 0.10 ppm (in contrast to tne 5.0 ppra level
assumed in tne PD 1/2/3), and tnat all other nitrosamine impurities (C^/Cg)
were present at a level of 0.02 ppm.
Treflan® EC contains 44,5 percent trifluralin or 445,000 ppm. At tne levels
of 0,10 ppn and 0.02 ppm in Treflan® EC, the ratio of trifluralin to NDPA and
to tne Oj/Cq nitrosamine contaminants would be 4,450,000 : 1 and
22,250,000 ; 1, respectively (445,000 ; 0.10 and 445,000; 0.02). The Agency
used these ratios to estimate exposure bo nitrosamines based on trifluralin
residue data. In order to .estimate possible exposure to other levels of
nitrosamine contamination, one oould adjust these ratios appropriately.
Trifluralin is recoverable Utilizing the FDA market basket survey procedures
but no such residues have been reported. There are a few FDA reports of
trifluralin found by its surveillance and compliance programs, for the period
1975-1979 (USEPA, 1979, and FDA, 1981). I Jo residues were found in surveys by
the Animal and Plant Health Inspection Service {AMIS} of USDA. Furthermore,
there are no known reports of residues of nitrosamines found in any of these
surveys. However, since there are*no tolerances for nitrosamines, these
surveys would not normally analyze for such impurities.
For these reasons, the Agency used the following information to estimate
potential dietary exposure to trifluralin, NDPA and C7/CR nitrosamines
(Regelman, 1981a).
o the percentage of crop acreage treated;
o actual trifluralin residue data?
o food factors;
o estimated trifluralin to nitrosamine ratios (noted above).
Where trifluralin results were reported as ICR (no detectable residues) at the
limit of detection of 0.005 to 0.01 ppm, or as <0.01 ppm, tne Agency assumed
that residues were present at tne 0.01 ppm limit of detection. Table 2
sunraarizes these data.
Assuming a dietary intake of 1.5 kg ger day, the dietary exposures (Table 2)
for trifluralin range from 0.06 x 10 mg/kg diet/day for sorghum to_
1257.36 x 10 mg/kg diet/day for.carrots; for NDPA from 0.01 x 10" mg/kg
diet/day for sorgnum to 282 x 10 mg/kg diet/day for carrots; for C7/Cg
nitrosamines from zero for sorgnum to 56.51 x 10~12 for carrots. Using a
dietary intake of 1.5 kg/day and a body weight of 65 kg, the total dietary
exposure to trifluralin, NDPA and C^/Cg nitrosamines was calculated to be
69.06 x 10~6, 15.52 x 10~12 and 3.10 x"l0"12 mg/kg body weignt/cay,
respectively. (These values were calculated by multiplying the total daily
intakes shown in Table 2 by 1.5 and then dividing by 65.)
14
-------
In trie Treflan® PD 1/2/3, total dietary exposure to trifluralin was not
included (since no REAR criterion red been met at tnat time), but had been
estimated by Mittelman (1978) to be 170 x 10" mg/kg body weight/day (about
twios trie current estimate). Mittelman (1978) estimated the dietary exposure
to {CPA to be 1920 x 10 mgAg body weight/day (about two orders of
magnitude nigner than the current estimate). The Agency had not previously
estimated dietary exposure to C^/Cq nitrosamine impurities. Because the
current exposure estimates for NDPA are lower than those presented in the PD
1/2/3, trie potential risk associated with exposure to NDPA will also be
expected to be lower.
The dietary exposure figures in Table 2 are used to calculate trie potential
risk discussed in Section II. C. of this document.
2. Mixer/Applicator/Loader Exposure.
a. Factors Applied to Exposure Data.
The PD 1/2/3 presented estimates for both innalational and dermal exposure to
NDPA for mixer/applicator/loaders handling Treflan® EC. Because the NDPA
levels have been reduced, as explained in trie previous section, the
innalational ard dermal estimates presented in the PD 1/2/3 (fron Mittelman,
1978) have been divided by 50, reflecting a decrease in NDPA contamination in
Treflan® EC frcm 5 ppm to an average of 0.10 ppm.
In calculating tne innalational exposure estimates for the PD 1/2/3, tne Agency
assumed a breatning rate in the workers of 1.2 cubic meters per hour. Tne
Agency now has determined tnat a breathing rate of 1.8 cubic meters per hour
more closely approximates the air intake for the type of work done by
mixer/applicator/loaders and field workers. This figure is used in tne
determination of exposure estimates for all three cnemicals, trifluralin, NDPA,
and C-^/Cg nitrosamines, for workers performing reentry activities as
well. Innalational estimates for trifluralin are derived frcm Mittelman (1978)
art! assume, as stated above, a breathing rate of 1.8 cubic meters per hour.
In carder to estimate dermal exposure for trifluralin, it is necessary to
determine what tne percentage of dermal penetration would be. However, to data
exist on tne dermal absorption of trifluralin? thus an estimate, based on its
physical cnemicai properties, was obtained. Trifluralin is a solid material,
having a melting point of 48.5 to 49.0°C, and is essentially insoluble in
water, but is soluble in acetone, ethanol, and xylene. In order to penetrate
the dermis to any appreciable amount, a solid must be soluble in botn water and
organic solvents. Trifluralin would be expected to have, therefore, a maximum
absorption of no more than one percent as an upper limit. When the
emulsifiable concentrate (Treflan® EC) is added to water for application, a
macrocrystalline suspension of tne solid is formed and dermal absorption would
also not be expected to be above 1 percent (Zendzian, 1981a).
15
-------
The dermal exposure estimates for trifluralin were derived frcm Mittelman
(1978) and take into account the 1 percent dermal penetration figure. The
A r-% i -a *1 a*; *• 4 -a 4* acr ss i r ^taimnl "i f Ayg jnn It? T +¦ i e iae? C\ uridyl /Mi 4- Ira itn 25**% 1 Q /Q \
QGIIuaJ. SSX-XITlcuGS OZ5 appX XC3 LOI.J3 OTiXV * XL. IS aSoUinSG iill uuSJLIucin / '£5 /
tnat mixer/loaders wear suitable protective clothing as required on' tne
Treflan0 labels, so tnat no dermal exposure to these particular workers would
normally be expected.
In calculating tne dermal exposure estimates for IIDPA in tne PD 1/2/3 the
Agency assumed a dermal penetration of 22 percent as suggested by the CAG
(1973 5 based on a dermal absorption study in rats submitted by Elanco (1978b).
If a worker was exposed to NDPA as a vapor or particulate matter, it was
assumed only 22 percent of that would actually penetrate the skin. This
4— «i f-m 1 % in in kK i 'SkC! 1 +¦*/"¦*. —»—1» ¦ « »*¦** "S 4* »m'i 1 /3VfVN Cf+" 4 TT1 2S 4" £2k C
x. Xy LiXw xs us so xn unxs occuiiiuriv cts w€xx g lo sinrxvc qi. oGiruifix sxj^obuCc "s vjjiiQ wt.s
for NDPA.
To obtain innalational and dermal estimates for the C-,/Co nitrosamines (not
previously calculated), the NEPA values (Mittelman), 1978) were divided by 50
[as described above for NDPA) and then by 5, since tne C7/C0 nitrosamines
(0.02 ppm) are assumed cn tne average to be present in Trerlan® EC at one-
fiftn tne average concentration of 1IDPA (0.10 ppm) (Elanco, 1980b). In the
absence of actual monitoring data for tne C^/Cg nitrosamines this approach
is considered reasonable and is assumed to result in the best available
estimate (Regeiman, 1981b). Since no dermal penetration data are available for
the C-,/Cg nitrosamines tne dermal penetration is assumed to be 100 percent
reflecting an upper limit for dermal exposure (Zendzian, 1981b).
Table 3 presents the current innalational and dermal estimates for
mixer/applicator/loaders for all tnree chemicals. The footnotes in Table 3
indicate where all the various factors discussed in this section, have been
applied to tne data. Mittelman's data (1978) nave been included in this
document as Appendix C, wnicn is available on request.
b. Summary of Mixer/Applicator/Loader Exposure Estimates.
Tne total (innalational plus dermal) estimates of exposure to trifuralin
for mixer/applicator/loaders range from 103 micrograms per year for "greens"
to 2950.6-micrograms per year for sugarcane. The range for NDPA is from
1.1 X 10"* -.micrograms per year for okra, greens, peppers, and cucumbers to
32.2 X 10" micrograms per year for sugarcane. For Cy/Cg nitrosamines,
the range is 0.72 x 10--3 for okra, greens, peppers, cucumbers to
20.98 x 10 for sugar cane. Ccmmerical applicators for all crops would be
exposed yearly to a total of 1082.7 micrograms trifluralin, 11.8 x 10
micrograms NDPA and 2.36 x 10~* micrograms C-j/Co nitrosamines.
Assumptions regarding rate and frequency of application are as described by
Mittelman, (1978).
16
-------
In trie Treflan® FD 1/2/3, total (innalational plus dermal) NDPA exposure to
mixer/applicator loaders ranged frcsti 0.1S micrograms per year for cucumbers and
"greens" to 5.05 micrograms per year for sugarcane. Cormercial applicators
were exposed to 1.80 micrograms NDPA per year. These.values are approximately
two orders of magnitude nigner than the more current estimates for NDPA as seen
in Table 3. The difference primarily exists because new netnods employed in
the manufacturing of Treflan® nave reduced the NDPA contamination by a factor
of 50 from the levels present at trie time the PD 1/2/3 was issued,
Tne exposure figures as presented in Table 3 were used to calculate the
potential risks discussed in Section II. C. of tnis document. Because tne
exposure to NDPA is lower, the risk would also be expected to be lower than
that presented in tne PD 1/2/3.
3. Reentry Exposure
a. Exposure to Vapor or Particulate Matter in tne Air
NDPA and C^/Cg Nitrosamines
"Die PD 1/2/3 stated that workers reentering fields previously treated with
trifluralin could possibly be exposed to NDPA as a vapor. The PD 1/2/3 (page
77) concluded tnat tne amount of exposure to NDPA as a vapor was negligible,
partially due to rapid pnotodegradation, and tnat the risk was also
negligible. Tne Agency has found no reason to alter tnat finding, especially
considering that tne NDPA contamination in Treflan® EC currently is lower by a
factor of 50. Since this exposure to NDPA vapor is considered to be
negligible, it is reasonable to conclude tnat tne exposure to CVCg
nitrosamines as a vapor would also be negligible (Regelman, 1981b).
Similarly, Mittelman (1978) concluded tnat reentry exposure (innalational and
dermal) to NDPA fran particulate matter in the air would be highly unlikely and
if it did occur, tne amount would be insignificant. It is reasonable to
assume, therefore, that exposure to C-,/Cg nitrosamines frcm particulates
would likewise be insignificant (Regelman, 1981b).
Trifluralin
In order to determine innalational and dermal exposure to trifluralin in the
air as a vapor, the Agency used data frcm Mittelman (1978), and White et al.
(1977). Mittelman (1978) estimated the worker reentry exposure to trifluralin
resulting fran a single application of 1 lb a,i./acre for several crops. In
that estimate (Mittelman, 1978, Table 14), the air trifluralin levels were
interpolated frcm a study by White et al. (1977). For tnis document, a
regression analysis of tne wnite, et al. data was performed, and estimates were
interpolated frcm tne best-fit line (Regelman 1981b). These data am presented
in Table 4.
The number of days after treatment when each operation occurred were taken
directly frcm Mittelman's Tables 8-13 (1978) and are sunmarized by crop in
Table 5.
17
-------
*fable 3
TREFLAN EXPOSURE ESTIMATES - APPLICATOR/MIXER/LOADERS
Trifluralin NnPA®/
NDPAH/
V
Inhalation Dermal
a/
(ug/year)
cyc9 Nitrosamlnes^r
Crop
b'c/ W
Inhal. Dermal Total
Total
y
Inhalation
Dermal
Total
Soybeans
804
56.2
860.2
3.5
X
10-3
5.8
X
10-3
9.3
X
10-3
0.70
X
10-3
5.27
X
10-3
5.98
X
10-3
Cotton
689
45.6
734.6
3.2
X
10-3
4.7
X
10-3
7.9
X
10"3
0.64
X
10-3
4.27
X
10-3
4.91
X
10-3
Tomatoes
245
17.6
262.6
1.2
X
10-3
1.8
X
10-3
3.0
X
10"3
0.24
X
10"3
1.64
X
10-3
1.88
X
10-3
Cole Crcps
295
21.1
316.1
1.0
X
10-3
2.2
X
10-3
3.2
X
10-3
0.20
X
10-3
2.00
X
10-3
2.20
X
10-3
Beans
459
31.6
490.6
2.0
X
10-3
3.3
X
10-3
5.3
X
10-3
0.40
X
10"3
3.00
X
10-3
3.40
X
10-3
Trees/Vines
454
31.6
485.6
2.0
X
10-3
3.3
X
10-3
5.3
X
10-3
0.40
X
10-3
3.00
X
10-3
3.40
X
10-3
Heps
1622
115.8
1737.8
6.9
X
10-3
12.1
X
10-3
19.0
X
10"3
1.38
X
10-3
11.00
X
10"3
12.38
X
10-3
Potatoes
304
21.1
325.2
1.3
X
10-3
2.2
X
10-3
3.5
X
10"3
0.26
X
10"3
2.00
X
10-3
2.26
X
10-3
Carrots
491
35.1
526.1
2.1
X
10-3
3.7
X
10-3
5.8
X
10-3
0.42
X
10-3
3.36
X
10-3
3.78
X
10-3
Okra
100
7.0
107.0
0.4
X
10-3
0.7
X
10-3
1.1
X
10-3
0.08
X
10-3
0.64
X
10-3
0.72
X
10-3
V
Greens
96
7.0
103.0
0.4
X
10-3
0.7
X
10-3
1.1
X
10-3
0.08
X
10-3
0.64
X
10"3
0.72
X
10-3
Spanish Peanuts
372
24.6
396.6
1.7
X
10"3
2.6
X
10"3
4.3
X
10-3
0.34
X
10"3
2.36
X
10-3
2.70
X
10-3
Celery
791
56.2
847.2
3.4
X
10-3
5.8
X
10"3
9.2
X
10-3
0.68
X
10-3
5.27
X
10-3
5.95
X
10-3
Peppers
100
7.0
107.0
0.4
X
10"3
0.7
X
10"3
1.1
X
10-3
0.08
X
10-3
0.64
X
10-3
0.72
X
10"3
Mint
1277
91.3
1368.3
5.4
X
10-3
9.5
X
10"3
14.9
X
10-3
1.08
X
10-3
8.64
X
10-3
9.72
X
10"3
Dill
541
38.6
579.6
2.3
X
10-3
4.0
X
10-3
6.3
X
10"3
0.46
X
10-3
3.64
X
10"3
4.10
X
10"3
Alfalfa
491
35.1
526.1
2.1
X
10"3
3.7
X
10"3
5.8
X
10-3
0.42
X
10"3
3.36
X
10-3
4.78
X
10-3
Spring Wheat
2413
172.0
2585.0
10.2
X
10~3
17.9
X
10-3
28.1
X
10"3
2.04
X
10-3
16.27
X
10"3
18.31
X
10"3
Mustard
813
56.2
869 o 2
3.6
X
10-3
5.8
X
10-3
9.4
X
10-3
0.72
X
10"3
5.27
X
10-3
5.99
X
10-3
Saf flower
1922
136.9
2058.9
8.2
X
10-3
14.2
X
10"3
22.4
X
10-3
1.64
X
10"3
12.91
X
10"3
14.55
X
10-3
Sunflower
913
63.2
976.2
4.0
X
10-3
6.6
X
10-3
10.6
X
10"3
0.80
X
10-3
6.00
X
10-3
6.80
X
10-3
Sugar Beets
986
70.2
1056.2
4.2
X
10-3
7.3
X
10-3
11.5
X
10"3
0.84
X
10-3
6.64
X
10-3
7.44
X
10"3
Sugar Cane
2754
196.6
2950.6
11.7
X
10-3
20.5
X
10-3
32.2
X
10"3
2.34
X
10-3
18.64
X
10-3
20.98
X
10-3
Cucumbers
100
7.0
107.0
0.4
X
10-3
0.7
X
10"3
1.1
X
10-3
0.08
X
10-3
0.64
X
10"3
0.72
X
10-3
Cantaloupes
200
14.0
214.0
0.9
X
10-3
1.5
X
10"3
2.4
X
10-3
0.18
X
10-3
1.36
X
10"3
1.54
X
10-3
Watermelons
150
10.5
160.5
0.6
X
10-3
1.1
X
10-3
1.7
X
10-3
0.12
X
10-3
1.00
X
10-3
1.12
X
10-3
Dry Peas
858
59.7
917.7
3.8
X
10-3
6.2
X
10"3
10.0
X
10"3
0.76
X
10"3
5.64
X
10-3
6.40
X
10"3
English Peas
254
17.6
271.6
1.1
X
10-3
1.8
X
10"3
2.9
X
10"3
0.22
X
10"3
1.64
X
10"3
1.86
X
10-3
Field Peas
254
17.6
271.6
1.1
X
10-3
1.8
X
10~3
2.9
X
10"3
0.22
X
10-3
1.64
X
10"3
1.86
X
10"3
Commercial
1009
73.7
1082.7
4.1
X
10-3
7.7
X
10-3
00
•
X
10"3
0.82
X
10"3
7.00
X
10-3
7.82
X
10-3
Applicators
(all crcps)
-------
1
Focmms {Table 3)
a/ Derived front Mittelman (1978) Table 7. For cucumber, cantaloupe,
watermelon, dry peas, English peas, field peas, greens, hops, celery,
and dill, derived frcm addendum to Table 7.
b/ Assumes a breathing rate of 1.8 mVhr.
c/ Inhalation exposure values were estimated as follows:
Inhalation
•Exposure =
(soybeans)
hrs/year x breathing x ug/in3 in air
(application) rate (ar/hr) (supplication)
hrs/year x breathing x ug/ta3 in air
(mixing/loading) rate (rrr/hr) (mixing/loading)
- [16 hrs/year x 1.8 n»3/hr x 25.3 ug/fo3] +
[1.7 hrs/year x 1.8 m3/hr x 24.6 ug/ta3]
« 728.64 ug/year (application) +
75.28 ug/year (mixing/loading)
= 804 ug/year
d/ Dermal exposure estimates (1/) were reduced by a factor of 0.01, to
reflect the upper limit assumption of 1% denial penetration. It was
also assumed that workers during mixing/loading wear suitable protect-
ive clothing, so that no exposure during this phase of Href Ian® usage
was anticipated. Dermal exposure values were estimated^ as follows:
Trifluralin » hrs/year x ug/hr dermal x 0.01
Dermal Exposure application exposure
(soybeans)
= 16 hrs/year x 351 ug/hr x 0.01 * 56.2 ug/year
e/ Total nitrosamine levels (Mittelman, 1978) were reduced by a factor of
50, reflecting the 50-fold reduction in average nitrosamine contamina-
tion since earlier estimates (PD 1/2/3) were made (0.1 ppm vs. 5.0 ppra
= 1:50).
f/ Dermal exposure estimate figures (Mittelman, 1578) have been reduced
"" by a factor of 0.22, reflecting an upper limit dermal penetration of
22% (FD 1/2/3).
£/ Derived by dividing the Mittelman (1978) MDPA exposure estimates by 50
(see 5/ above) and then by 5, since the average C-j/Cg nitrosamine con-
tamination is now approxiaately 20% of the awrage NDPA levels (Elanco,
1980f).
h/ Since no dermal penetration data are available for C7/C9 nitrosamines,
100% dermal absorption is assumed as an upper limit (Zendzian, 1981b).
i/ Greens includes mustard greens, turnip greens, collards and kale.
19
-------
*fable 4
Estimated Trifluralin Concentrations in Air as a Vapor After Treatment—^
Days ,After Treatment
0
3
7
14
18
30
35
ug/m
0.26
0.23
0.20
0.15
0.13
0.08
0.07
DaysJVfter Treatment
ug/to
40
42
45
56
63
79
98
0.06
0.05
0.05
0.03
0.02
0.01
0.01
a/ Data were taken from Mittelman's Table 14 (1978). A regression analysis was used
approximate the best line, from which the data were then interpolated.
-------
*foble 5
Number of Days After Trifluralin Treatment When Each Operation Occurred
2nd
Incorporation
Planting
Irrigation
Tillage
Seeding
Hand
Hoeing
Insect
Scouting
Harvest
Soybeans
14
14+
40
210
Cotton
30
30
30/90
63
-
79
112
225
Beans
7
18
56
42
-
-
-
98
Tomatoes
—
-
45
42
3
35
-
105
Tree/Vine
3
7
118
134
-
-
-
210
Cole Crops
3
45
45
3
42
•
105
-------
Tne estimated air concentrations of trifluralin at each specified post-
application interval' are summarized in Table 6, using tne data fran Tables
4 and 5. For example, during tne second incorporation, 14 days after treatment
(fran Table 5) exposure of workers to trifluralin would be 0.15 ug/ta (fran
Table 4). Since tne measurements of levels of trifluralin in air becane
increasingly unreliable at or near the limit of detection, values belov-*
0.01 ug/m were assumed to be zero, and were deleted from Table 6. Estimated
hours of exposure per operation per year were taken directly fran Mittelman
(1970), and summarized in Table 7.
Finally, reentry exposure to trifluralin as a vapor in the air was competed,
and summarized by operation, in Table 8 using values frcm Table 6 and 7.
Mittelman (1978) calculated that dermal exposure would be about 12 times
innalational exposure based on the data frcm White et al. (1977). This factor
was applied to tne estimated innalation values, and summarized in tne "dermal"
column. Tne final column reflects total estimated annual reentry exposure in
micrograms per year by crop.
Theoretically, workers reentering trifluralin treated fields could inhale or
be dermally exposed to trifluralin adsorbed to soil particulate matter in the
air. Mittelman (1978) in citing soil concentration data from the White et al.
study (1977) calculated tnat the upper limit exposure would be approximately
1.28 micrograms per year for tree and vine culture. The calculation made was
an extrapolation fran tne White et al. study. The exposure was stated by
Mittelman (1978) to be insignificant. Thus, tne Agency concludes that any
possible inhalational car dermal exposure to trifluralin fran airborne soil
particles during reentry would be negligible.
b. Dermal Exposure Fran Contact with the Soil
Altnough dermal exposure to trifluralin, NDPA and the Cj/Cq nitrosamines
frcm contact with soil during reentry into treated fields may occur, it is
difficult to quantify due to tne variety of; tne field activities performed,
tne degree of physical contact with the soil by workers, the amount of exposed
body area, the duration of exposure, and tne degree of dermal penetration of
tne particular cnemical involved.
Tne PD 1/2/3 stated that reentry dermal exposure to NDPA in the soil was
theoretically possible and presented a method for arriving at an estimate.
An assumption was made at tne time that "a uniform layer of soil forms a film
an the uncovered skin 1.0 millimeter thick." This resulted in an accumulation
of soil on the skin amounting to 870 grams or approximately 2 pounds. Elanco
[30000/32:#6], as discussed later in Section III of this document, claimed tnis
was an unreasonable assumption. Although that approach was used only to
describe an upper limit of possible exposure, tne Agency agrees with Elanco
tnat tne figure is excessive and is currently using a different approacn to
calculate dermal exposure.
22
-------
Table 6^
3
Estimated Air Concentration of Trifluralin (ug/m )
2nd Hand
Incorporation Planting Irrigation Tillage Seeding Hoeing Harvest
Soybeans
.15-
.15
.06
Cotton
.08
.08
.08
.02
-
.01
-
Beans
.20
.13
.05
.03
-
-
.01
Tomatoes
-
-
.05
.05
.23
.07
-
Tree/Vine
.23
.20
-
-
-
-
-
Cole Crops
.23
.05
.05
.23
.05
~ .
a/ Values were taken frcm Table 4. Entries have been deleted if the trifluralin oonoentration was
below 0.01 ug/m .
ro
u>
-------
Table 7
Estimated Hours of Exposure per Year
2nd
Hand
Incorporation
Planting
Irrigation
Tillage
Seeding
Hoeing
Harvest
Soybeans
8
10
20
—
Cotton
13
11
2
33
—
13
-
Beans
6
13
8
20
-
-
32
Tomatoes
-
-
8
5
8
125
-
Tree/Vine
6
24
-
-
-
-
-
Cole Crops
3
"
12
6
6
16
"
ro
-------
•fable 8
Estimated Reentry Exposure to Trifluralln Vapor in the air (ug/year
2nd Hand . .
Incorporation Planting Irrigation Tillage Seeding Hoeing Harvest Inhalation - Dermal-' "total
Soybeans
2.16
2.7
2.16
_
_
_
7.02
84.24
91.3
Cotton
1.87
1.S8
0.29
1.19
-
0.23
-
S.16
61.92
67.1
Beans
2.16
3.04
0.72
1.08
-
-
0.58
7.58
90.96
98.5
Tana toes
-
-
0.72
0.48
3.31
15.75
-
20.23
242.76
263.0
Tree/Vine
2.48
8.64
-
-
-
-
-
11.12
133.44
144.6
Cole Crops
1.24
—
1.08
0.S4
2.48
1.44
—
6.78
81.36
88.1
a/ ug/m^ x hro/year x 1.8 m^/hr - ug trifluralin/year.
b/ Hittelman (1978) assured dermal exposure to be 12 tines Inhalation exposure.
io
i_n
1
-------
Dermal exposure to trifluralin would be expected to be low (Regelman, 1981b).
Exposure to tne nitrosamines snould be mucci lower, since levels in Treflan® EC
nave been reduced about 50-fold in recent years. Exposure estimates -for
trifluralin and nitrosamines have however been calculated by -the Agency
(Regelmanr 1981b and 1981c). In one study wnich appears to be typical of
Treflan® usage, West and Day (1977) analysed soils frar. fields treated at
application rates of 0.5-1.0 lb. a.i./acre. Trifluralin residues in the
surface-to-3-inch layer of top-soil ranged from 0.09-0.63 ppn (0.09 - 0.63
ug/gm soil), 26 to 176 days after treatment. NDPA was quantified at levels up
to 0.19 ug/kg soil 26 days after application of 0.75 lb a.i./acre. The
trifluralin used contained from 78-252 ppn NDPA. The NDPA figure may be
adjusted downward by a factor of about 1650=.' reflecting NDPA levels in
currently produced Treflan3 E.C* with an average of 0.1 ppn. Thus NEPA levels
could be estimated at 1.2 x 10~ ug/gm soil.
Jensen £1981) stated that approximately 1 gram of talcun powder would cover
0.082 m of skin (equivalent to 12.2 goM ). Assuming that talcum powder
and dirt have about the same densities^-', and assuming that a worker
entering a treated- field nas a total uncovered skin surface area of about
0.29 m (Mittelman, 1973), the total amount of soil on tne skin could be
about 4 grams, a figure much lower than the~870 grams estimated in the PD 1/2/3.
Soil could contain about 2.5 ug trifluralin (0.63 ug/gm x 4 gm), about
5 x 10" ug NDPA (1.2 x 10 ug/gm x 4 gm), and about 1 x 10 ug
Cj/Cq nitrosamines r assuming a ratio o£ 5:1 between UDPA and C7/Cg
nitrosamines.
For purposes of tnis exposure analysis, if we assume an average of ten separate
exposures per year, tne total annual dermal reentry exposure could be as high
as 25 ug trifluralin, 5 x 10 ug NDPA and 1 x 10 ug C7/CQ
nitrosamines (Regelman, 1981b).
It is reasonable to oonclude tnat tnesc dermal reentry exposure estimates are
upper-limit values, since pnotodegradaticn, soil metabolism and otner
degradative and metabolic pathways, as well as probable dermal penetration
below 100 percent would be expected to reduce exposure to levels below these
estimates (Regelman, 1901b).
Tne Agency does not believe, therefore, tnat trie exposure estimates for NDPA
(no greater than 5 x 10 micrograms per year) and C7/Co nitrosamines
(1 x ID-6 micrograms per year) are reliable at these extremely low levels.
Actual exposure may be orders of magnitude lower. Tne Agency must conclude,
tnen, that reentry dermal exposure to these nitrosamines from soil is likely
to be negligible (Regelman, 1981c).
1/ 78+252 ; 1 - 1650:1
2
2/ Handbook of Chemistry and Physics, Robert C. Weast, Ed. Chemical
Rubber Publishing Ccmpany, 54th Edition, 1973-1974.
26
-------
Table 9 suonarizes the reentry exposure to trifluralin in the air as a vapor
and in tne soil adsorbed to soil particles (Regeiman, 195155. No reentry
exposure estimates are given fear the nitrosamines because, as explained' above,
innalational or dermal exposure to nitrosamines in the air (as a vapor or
adsorbed to particulate natter) or in the soil, is likely to be negligible.
The trifluralin innalational and dermal exposure estimates (Table 9) were used
to calculate trie potential risk estimates to reentry worker presented in
_Section II, C. of this document.
C. Revised Cancer Risk Estimate.
1. Rationale for Revisions
a. NDPA
The dietary and worker cancer risk estimates found in tne H3 1/2/3 presented
estimates for NI2PA only. Trie Agency assumed at that tine that the cancer risk
from trifluralin itself was zero, since there were no data which indicated
otherwise. Wnen the' Agency received data (Elanoo, 1980d) which showed that
trifluralin itself appeared to be associated with production of tumors and
analyses wnich revealed contamination from C-,/Cg nitrosamines in addition
to NDPA (Elanoo, 1978a, 1980b, and 1981), it became necessary to recalculate
tne risk, taking tnese two additional chemicals (trifluralin and C^/Cg
nitrosamines) into account.
In the FD 1/2/3, the Agency used the "one-hit" rodel to calculate the slope
parameter or potency of NDPA as a carcinogen", using only tne lowest dose
snowing a significant response from data by Druckrey et al., 1967. Recently,
the Agency nas decided the "multistage" model provides a betterestimate for the
slope parameter; it incorporates data from all the dosed groupsboth high and
low, and is a generalization of the "one-hit" model. It enccmpasses tne "one-
nit" model as a special case (Chen and Haberman, 1981).
Tne NDPA slope parameter estimated in the FO 1/2/3 was 0.4 per ppm. It was
assumed that tne ppm dietary consumption was equivalent between humans and
animals. The Agency has reevaluated this assumption and has concluded that it
is not justified, since the calories per kilogram of food are very different in
tne diet of man as compared to laboratory animals, primarily due to tne
moisture content difference (Chen and Habeman, 1981). It is more accurate to
calculate the slope per milligram per kilogram body weight per day.
Therefore, the Agency (Chen and Haberman, 1981) recalculated the slope for i3DPA
based cn liver tumor incidence in male rats as reported by Druckery et al,
(1967) as snown in Table 10. Using the "multistage" model, the hutian
carcinogenic potency for NDPA is estimated as follows:
q*± = 0.62 x (70/0.35)1/3
= 3.6 (mg/kg body weight/day)"*1'
27
-------
Table 9
Reentry Exposure bo Trifluralin (ug/year)
Trifluralin Vapor^/^
Trifluralin
in the Soil^
Inhalation
Dermal^/
Derma l^^
Total
Soybeans
7.0
0.84
0.25
8.09
Cotton
5.2
0.62
0.25
6.07
Beans
7.6
0.91
0.25
8.76
Tanatoes
20.2
2.43
0.25
22.88
Tree/Vine
11.1
1.33
0.25
12.68
Cole Crops
6.8
0.81
0.25
7.86
a/ Regelman (1981b and 1981c).
H/ Hie dermal exposure figures were reduced by a factor of 0.01, reflecting a 1% upper limit
for dermal a isorp' ion (Zendzian, 1981a).
c/ Mittleman, 1978, stated that exposure to trifluralin particulate matter in the air would be
insignificant.
d/ Assumes 10 separate exposures per year, (Regelman, 1981b).
-------
Table 10
Liver Tumor Incidence in Male Rats Given NDP
Dree ..
(mg/kg/day) O-' 4 8 15
Response 1/30 12/14 15/16 15/15
a/ Druckery et al, 1967.
b/ A control incidence of 1/30 is used in the calculation. The frequency of spontaneous
malignant tumors was about 1 percent up to the age of 500 days.
KJ
y£>
-------
b.. Trifluralin
An estimate of carcinogenic potency of trifluralin is based on data in
Table 11 taken frcm trie chronic feeding study done by Elanco (1980d).
Using tne "multistage" model, the human carcinogenic potency for trifluralin
is estimated as follows (Cnen and Haberman, 1981):
q*x = 1.31 x 10"3 x (70/0.35)1/3
-3 -1
= 7.7 x 10 (itgAg tody weignt/day)
When this number is canpared with the slope or potency for fJBPA of 3.6, it is
evident that NDPA is the more potent carcinogen by approximately 3 orders of
magnitude.
c. C7/Cg Nitrosamines
No data exist on C7/C0 nitrosaminc oontaminants in Treflan® regarding
the potential for producing carcinogenic effects. Tne J^ency's nitrosamine
rv%1 cf"at"oc •
Xvj >3 l»g tco •
"In tne absence of acceptable oncogenic testing witn the specific N-
nitroso canpound, the Agency will assume that the contaminant is as potent
ac pi nprrinrnf^n as M—nit'Kncirr^i^t'hvlaminp M PR 42RS4 1.
Kj* C3> JL. JL » i I, Cm? iL * I -i alfe (wjh. mL Wl ¦ Y JL ii3i» I JL* I\— • |_ JL -L * -mf J •
Therefore, the Agency used data cn liver tumors produced in male rats which
nad been administered N-nitroscdietnylamine, also called diethyl N-nitrosamine
as reported by Druckrey et al., (1967) and shown in Table 12.
Using the "multistage" rrodel and data for DE21A (Druckery et al., 1967), the
carcinogenic potency for C7/Co nitrosamines is estimated as follows (Cnen
and Haberman, 19815:
q*x = 5.72 x (70/0.35)1/3
- 33 (rogAg body weight/day)
When ccmpared with tne potencies of trifluralin and NDPA {7.7 x 10 and
3.6, respectively), it can be seen tnat tne Cj/Cq nitrosamines, assumed to
be as carcinogenic as DENA, are potentially more carcinogenic. As will
be seen, however, the extent of actual exposure to tne three chemicals in
Treflan® influences the degree of risk associated with each one.
2. Dietary Risk
The Pgency calculated•tne risk of cancer from potential dietary exposure to
trifluralin, NDPA, and C7/Cg nitrosamines by using the slopes as given
above and the dietary exposure estimates given in Table 2, discussed in Section
II. B. of tnis document. The exposures and risks are presented in Table 13.
30
-------
Table 11
Combined Kidney, Urinary Bladder, and Thyroid Tumors in Male Rats Given Trifluralin^
Dose
'
(mgAg/day)
0
30
128
272
Response^
5/60
5/60
9/60
17/60
a/ Elanco, 1980d.
b/ In order to be counted as a "response", each rat had to have at least one tumor of the
kidney, urinary bladder, or thyroid.
-------
Table 12
Liver Tumor Incidence in Male Rata Given DENA in Drinking Water^
Dose
(mg/kg/day) 0 0.075 0.15 0.30
Response^ 1/30^ 1/7^ 27/30 67/67
a/ Druckery, et al, 1967.
b/ As in Table 10, the control incidence of 1/30 is used.
c/ If all tumor types were considered, the incidence would be 5/7.
-------
Table 13
a/
Dietary Risk - Trifluralin, NDPA and Ci/Cr Nitnosamines
p7"
_ NDPA
Daily Intake Lifetime
Comnodity
w
Trifluralin
telly Intake Lifetime
mg/kg diet RiskS/ mg/kg diet Risk
Asparagus
14.00
X
0-6
0.025
X
10-7
3.15
X
Barley
0.21
X
0-6
0.000
X
10-7
0.05
X
Carrots
1257.36
X
0-6
2.234
X
10"7
282.55
X
Citrus Fruit
30.10
X
0-6
0.054
X
10"7
6.76
X
Corn, grain
100.00
X
0-6
0.178
X
10-7
22.47
X
Cottonseed
10.43
X
0-6
0.018
X
10~7
2.34
X
Cucurbits,
Cantaloupe
5.20
X
0-6
0.009
X
10"7
1.17
X
Cucunber
21.24
X
0-6
0.038
X
10-7
4.77
X
Watermelon
42.90
X
0-6
0.076
X
10-7
9.64
X
Dill
1.61
X
0-6
0.003
X
10~7
0.36
X
Fruiting Vegetables,
Green Peppers
4.54
X
0-6
0.008
X
10"7
1.02
X
Tomatoes
197.46
X
0-6
0.351
X
10"7
44.37
X
Grapes/Raisins
3.87
X
0~6
0.007
X
10"7
0.87
X
Hops
0.75
X
0-6
0.001
X
10"7
0.17
X
Leafy Vegetables,
Broccoli
6.50
X
0-6
0.012
X
10~7
1.46
X
Brussel Sprouts
1.95
X
0-6
0.003
X
10~7
0.44
X
Cabbage
49.21
X
O-6
0.087
X
10"7
11.06
X
Cauliflower
3.82
X
0~6
0.007
X
10"7
0.86
X
Celery
11.43
X
0-6
0.020
X
10"7
2.57
X
Collard
7.68
X
0-6
0.014
X
10~7
1.73
X
Kale
2.88
X
0-6
0.005
X
10-7
0.65
X
Mustard Greens
5.15
X
0-6
0.009
X
10"7
1.16
X
Turnip Greens
2.89
X
0-6
0.005
X
10~7
0.65
X
Others
133.00
X
0-6
0.236
X
10-7
29.89
X
Daily Intake
mg/kg diet
CVCfl Nitrosamines
d/
Lifetime
Risk
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
0.03
0.00
2.35
0.06
0.19
0.02
0.01
0.04
0.08
0.00
0.01
0.37
0.01
0.00
0.01
0.00
0.09
0.01
0.02
0.01
0.00
0.01
0.00
0.25
10"
10"
10"
10"
10"
10"
x 10"
x 10"
x 10"
x 10-
x 10"
x 10"
x 10"
x 10-
10"
10"
10"
10"
10"
10"
10"
10"
io-
io-
0.63
0.01
56.51
1.35
4.49
0.47
10-12
10-12
10-12
10-12
10-12
10-12
0.29
0.09
2.21
0.17
0.51
0.35
0.i3
0.23
0.13
5.98
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
10-12
0.05
0.00
4.30
0.10
0.34
0.04
0.23 x 10-12
0.95 x 10-12
1.93 x 10-12
0.07 x 10-12
0.20 x 10-12
8.87 x 10-12
0.17 x 10-12
0.03 x 10-12
0.02
0.01
0.17
0.01
0.04
0.03
0.01
0.02
0.01
0.45
io-
10"
10"
10"
io-
10"
0.02 x 10-
0.07 x 10"
0.15 x 10"
0.00 x 10~
0.02 x 10"
0.68 x 10"
0.01 x 10*
0.00 x 10-
10"
10"
10"
10"
10"
10-
10"
10-
10-
10-
-------
Table 13 (continued)
a/
Dietary Risk - Trlfluralin, NDPA and C7/CB Nitrosamlnea
~W sf
Trifluralin NDPA
Daily Intake Lifetime Daily Intake Lifetime
mq/kq diet Risk®/ mg/kg diet Risk
Commodity
Daily Int
Nitrosamlnea
Lifetime
Risk
d/
LJ
Mung Beans (sprouts)
3.00
X
10"6
0.005
X
10-7
0.67
X
10~12
0.00
X
10"1
0.13
X
10"! 2
0.01
X
10-
Mustard Seed
3.00
X
10-6
0.005
X
10-7
0.67
X
10-12
0.00
X
10-1
0.13
X
10"12
0.01
X
10-
Nuts
0.79
X
10"6
0.001
X
10-7
0.18
X
10"12
0.00
X
10"1
0.04
X
10"12
0.00
X
10"
Okra
7.00
X
10~6
0.012
X
10"7
1.57
X
10~12
0.01
X
10"1
0.31
X
10-12
0.02
X
10"
Peanuts
7.06
X
10"6
0.012
X
10-7
1.59
X
10"12
0.01
X
10"1
0.32
X
10"12
0.03
X
10"
Peppermint Oil
46.20
X
10~6
0.082
X
10"7
10.38
X
10~12
0.09
X
10"1
2.08
X
10~12
0.16
X
10-
Root Crop Vegetables,
10~12
Potatoes
28.78
X
10~6
0.051
X
10"7
6.47
X
0.05
X
10"}
1.29
X
l°-}2
0.10
X
10-
Others
554.00
X
10-6
0.984
X
10~7
124.49
X
10~12
1.03
X
10"1
24.90
X
10-12
1.95
X
10-
Safflower Seeds
3.00
X
10"6
0.005
X
10~7
0.67
X
10-12
0.00
X
10-1
0.13
X
10-12
0.01
X
10-
Seed/ftxJ Vegtables,
10~12
Beans
154.22
X
10"6
0.274
X
10~7
34.66
X
0.29
X
10"1
6.93
X
10-12
0.53
X
10"
Soybeans
34.68
X
10"6
0.062
X
10~7
7.79
X
10-12
0.06
X
10"1
1.56
X
10~12
0.12
X
10-
Peas
9.52
X
10"6
0.017
X
10~7
2.14
X
10-12
0.02
X
10"1
0.43
X
10-12
0.03
X
10"
Sorghum, grain
0.06
X
10~6
0.000
X
10~7
0.01
X
10"12
0.00
X
10"1
0.00
X
10~12
0.00
X
10-
Spearmint Oil
46.20
X
10~6
0.082
X
10"7
10.38
X
10"12
0.09
X
10"1
2.08
9
10"12
0.16
X
10"
Stone Fruits
125.00
X
10"6
0.222
X
10~7
28.09
X
10~12
0.23
X
10"1
5.62
X
10-12
0.43
X
10-
Sugar, Cane & Beet
42.59
X
10-6
0.076
X
10~7
9.57
X
10~12
0.08
X
10"1
1.91
X
10"12
0.15
X
10"
Sunflower Seeds
1.95
X
10"6
0.003
X
10~7
0.44
X
10-12
0.00
X
10"1
0.09
X
10~12
0.01
X
10-
Wheat Grain and Straw
11.40
X
10~6
0.020
X
10~7
2.56
X
10"12
0.02
X
10"1
0.51
X
10~12
0.04
X
10-
TOTAIS 2.98 x 10~3 5.30 x 10~7 6.72 x 10-10 5.85 x lO"11 1.34 x lO-10 1.03 x 10"1°
a/ Data frcm E. Regelman (1981a)
B/ Trifluralin risk slope = 7.7 x 10"^/fag/kg body weight/day.
c/ NDPA risk slope = 3 * 6/mg/kg body weight/day.
d/ C7/C8 risk slope was assumed to be as large as the diethylnitree amine risk slope of 33/mg/kg body weight/day.
e/ Lifetime risk = Intake (mg/kg diet/day) x 1.5 kg diet/day x risk slope (/mg/kg body weight/day).
65 kg Body Weight
-------
a. .Trifluralin
In order to calculate tne risk to tne general population from dietary exposure
to trifluralin, tne Agency used the following formula:
Lifetime Individual Risk = slope x exposure
wnere tne slope for trifluralin is 7.7 x 10"^ (rag/kg body weight/day) 1
and wnere tne exposure is a product of
daily intake (mg/kg diet/day) x 1.5 kg diet/day
¦ T>5 kg body weight.
This assumes an average daily intake of 1.5 kilograms of food a day for tne
average 65 kilogram person. Thus, for example, tne lifetime individual risk
(R) from exposure to trifluralin from eating carrots (Table 13) is:
R = 7.7 x 10"3 x [(1257.36 x 1(T6} x 1.5 ]
"5?
= 2.234 x 1CT7
Tne risk associated witn eating carrots is tne nighest; tne risk associated
with all otrer crops is less tnan 1 x 10~ . Tne total lifetime individual
risk fran exposure to trifluralin in Treflan® in the daily diet is
5,30 x 10"'.
b» JOPA
Similarly, tne lifetime individual risk (R) from NDPA on carrots
(Table 13) in the diet is:
R = 3.6 x [(292.55 x 10"12) x 1.5 ]
"55"
= 2.35 x 10"11
Tne risks assoiated with other crops are less than 1.5 x 10~The total
lifetime individual.risk from exposure to tne NDPA in Treflan® in the daily
diet is 5.35 x 10 '
35
-------
c. C7/Cp Nitrosamines
Ttie lifetime individual risk (R) from C^/Cg nitrosamines on carrots
(Table 13) in the diet, assuming tne slope to be-as large as for DEMA,_ is:
R = 33 x [(56.51 x 10-12) x 1.5 ]
= 4.30 x 10"11
Tne risks associated with exposure to all other crops are less than
2 x 10" . The total lifetime individual risk from exposure to the C^/Cg
nitrosamines in Treflan® in the daily diet is 1.03 x 10' .
d. Comparison of Dietary Risk Estimates with Those of the
PD 1/2/3
The risk associated with exposure to NDPA as calculated in the PD 1/2/3 is
higher than that calculated in this document. The Agency calculated in the PD
1/2/3 tnat tne highest lifetime individual dietary risk associated with
nitrosamines (NDPA) was 8.7 x 10" for carrots. The analogous NDPA risk
figure in Table 13 for carrots is 2.35 x lO-"1 and for Ci/Cg nitrosamines
is 4.3 x 10"" . Tne same relationship exists for the dietary risk from all
crops listed in Table 13.
-8
The total risk for NDPA in the PD 1/2/3 was 3.3 x 10 , computed by
adding the.,risks associated with each crop. The figure in this document is
5.8 x 10 for NDPA and 1.03 x 10 for C7/g nitrosamines, a risk
lower by three and two orders of magnitude, respectively.
Three factors account for this difference. First, the Agency based exposure
to NDPA an trifluralin residue data submitted by Elanco and summarized by
Regelman (1981a), wnereas in the PD 1/2/3 the Agency based expDSure to NDPA on
trifluralin tolerances, wnich are higher than trie actual residues found by
about a factor of 5.
Tne total dietary risk from exposure to Treflan® is currently 5*3 x 10
obtained by adding risk frcro exposure to trifluralin (5.3 x 10 ) with those
associated with NDPA (5.85 x 10 ) and Cy/Cg nitrosamines
(1.03 x 10"1 ). The contribution to risk by NDPA and C-^/Cg nitrosamines
in insignificant. Tne risk is principally associated with trifluralin.
Second, current NDPA contamination in Treflan® is defined by a trifluralin to
NDPA ratio of 445,000 ppra:Q.l ppm or 4,450,000:1. (Treflan® EC contains 44.5
percent trifluralin or 445,000 ppm.) The daily intake of NDPA was calculated
by 'dividing the trifluralin daily intake by 4,450,000. The ratio used in the
PD 1/2/3 was 89,000:1, reflecting a higher NDPA contamination (445,000 ppm
trifluralin divided by 5 ppm NDPA). Therefore, because the NDPA exposure is
calculated to be much less at tne present time than previously, it follows tnat
the risk would be less as well.
36
-------
The third contributing factor to the difference between the risk associated
witn tDPA as calculated in tne PD 1/2/3 and tnat calculated in this document is
the use of tne multistage mathematical model rather than trie "one-nit" model
to estimate tne slope parameter used in trie risk assessment,
-11
Tne range for. NDPA risk (Table 13) is from less than 0.01 x 10 for barley
to 2.35 x 10" for carrots. The range for C7/C8 nitrosamines is fran
less tnan 0.01 x 10 for barley to_4.30 x 10 for carrots. Compared to
a range in the PD 1/2/3 of 0.11 x 10 to 8.68 x 10~ for sunflower and
carrots, respectively, it can be seen that the cancer risk to the general
population associated with exposure to tne nitrosamines in Treflan® is
currently significantly lower tnan was previously estimated.
Tne total lifetime individual cancer risk estimate in the PD 1/2/3 for the
general population fran exposure to Treflan® in the diet for all crops was
3.3 x 10 and was associated with the NDPA contaminant. In this document,
the total dietary cancer risk estimate is 5.3 x 10~ , principally from
exposure to the trifluralin in Treflan® (Table 13).
Thus, tne total risk associated witn dietary exposure to Treflan® as calculated
in the PD 1/2/3 has changed by approximately one order of magnitude? that is,
the cnance of getting cancer from dietary exposure to Treflan® has increased
sore what over that estimated when tne PD 1/2/3 was issued, because of the
association of trifluralin with tne production of tumors in rats as discussed.
3. Mixer/Applicator/Loader Risk
a. Calculations and Assumptions
Tne Agency calculated the risk of cancer in workers fran exposure to
trifluralin, NDPA, and C-j/Cg nitrosamines in Treflan® by using tne slopes
given in Section II.C. of tnis document and the mixer/applicator/loader
exposure estimates given in Table 3. The total lifetime individual exposure
(innalational and dermal) and risks for each crop and each chemical are
presented in Table 14.
Tb calculate tne lifetime individual risk, tne Agency used the following
formula:
|D]
Risk = slope x exposure
wnere the slopes (or patency) for trifluralin, NDPA, or C-,/Cg nitrosamines
were 7.7 x_10"^f 3.6, and 33, respectively, in units of (mgAs body
weight/day ), as described in the Dietary Risk Ejection of this document.
Tne lifetime individual exposure was calculated as follows:
Total Exposure Working Lifetime
(micrograms per year) X (40 years) X
365 days per year average
lifetime
1 Person X 1 mg
65 kg 1000
(70 years) micrograms
37
-------
Table 14
ThfKKi JBtNT RISK MflrrfiS — AppriTP-R,J|\jR/M I ^
Trifluralin
c/
d/
NDPA
C7/CR Nitrosamines"
Total b/ Total
Exposure Lifetime Exposure Lifetime
{ug/year) Risk (ug/year) Risk
Total
Exposure Lifetime
(ug/Vear) Risk
Soybeans
Cotton
Tenia toes
Cole Crcps
Bears
Trees/Vines
Hr«ng
a fryuSMg
Carrots
Qkra
V3Ub funnel iSSS
Spanish Peanu
Celery
Peppers
Mint
Dill
AJfalfa
Spring Wheat
Mustard
SaffLoher
wunLiuwcxr
Sugar Beets
Sugar Cane
Cucumbers
Carrt"*i nnppg
Watermelons
Dry Peas
English Peas
Field Peas
Cnnmprf) a]
arm 1 1 fgt frriyts
U#Xi5
(all Crcps)
860.2
1.6
X
10-7
9.3
X
10-3
8
X
10-10
5.98
X
10-3
48
X
10-10
734.6
1.4
X
10-7
7.9
X
10-3
7
X
10-1°
4.91
X
10-3
39
X
10-10
262.6
0.5
X
10-7
3.0
X
10-2
3
X
10-10
1.88
X
10-3
15
X
10-10
316.1
0.6
X
10-"
3.2
X
10-3
3
X
1Q-10
2.20
X
lO"3
17
X
10-10
490.6
0.9
X
10-7
5.3
X
lO-3
5
X
10-10
3.40
X
10-3
27
X
10-10
485.6
0.9
X
10-7
5.3
X
10-3
5
X
10-10
3.40
X
lO"3
27
X
10-10
1737.8
3.2
X
10-7
19.0
X
lO"3
16
X
10-10
12.38
X
10-3
98
X
10-10
325.2
0.6
X
icr7
3.5
X
10-3
3
X
10-10
2.26
X
10-3
18
X
10-10
526.1
1.0
X
10"7
5.8
X
10-3
5
X
io-i°
3.78
X
lO"3
30
X
10-10
107 • 0
0.2
X
10"7
1.1
X
10-3
1
X
1Q-10
0.72
X
10-3
6
X
10-10
103.0
0.2
X
icr7
1.1
X
10-3
1
X
10-10
0.72
X
10-3
6
X
10-10
396.6
0.7
X
10"7
4-3
X
lO-3
4
X
10-10
2.70
X
10-3
21
X
10-10
847.2
1.6
X
IO-7
9.2
X
lO-3
8
X
10-10
5.95
X
10-3
47
X
10-10
107.0
0.2
X
10"7
1.1
X
lO"3
1
X
10-10
0.72
X
lO"3
6
X
10-10
1368.3
2.5
X
10~7
14.9
X
10-3
13
X
10-10
9.72
X
10-3
77
X
10-10
579.6
1.1
X
10~7
6.3
X
10-3
5
X
10-10
4.10
X
10-3
33
X
10-10
526.1
iIq
X
10"7
5.8
X
10-3
5
X
10-10
4.78
X
10-3
38
X
10-10
2585.0
4.8
X
10"7
28.1
X
lO'3
24
X
10-10
18.31
X
lO"3
146
X
10-10
869.2
1.6
X
10~7
9.4
X
lO"3
8
X
1Q—10
5.99
X
lO"3
48
X
10-10
2058.9
3.8
X
10-7
22.4
X
10-3
19
X
10-10
14.55
X
10-3
X
10-10
976.2
1.8
X
10"7
10.6
X
lO"3
9
X
10-10
6.80
X
10-3
54
X
10-10
1056.2
2.0
X
icr7
11.5
X
10-3
10
X
10-10
7.44
X
10-3
59
X
10-10
2950.6
5.5
X
icr7
32.2
X
10~3
28
X
10-10
20.98
X
10-3
167
X
10-10
107.0
0.2
X
10-7
1.1
X
lO-3
1
X
10-10
0.72
X
10-3
6
X
10-10
214.0
0.4
X
10-7
2*4
X
10-3
2
X
10-10
1.54
X
10-3
12
X
10-10
160.5
0.3
X
icr7
1.7
X
lO"3
1
X
10-10
1.12
X
10-3
9
X
10-10
917.7
1.7
X
10-7
10.0
X
lO"3
9
X
10-10
6.40
X
10-3
51
X
10-10
271.6
0.5
X
10-7
2.9
X
10-3
3
X
10-10
1.86
X
lO"3
15
X
10-10
271.6
0.5
X
10-7
2.9
X
10-3
3
X
10-10
1.86
X
10-3
15
X
10-10
1082.7
2.0
X
1-
¦*4
11.8
X
lO"3
10
X
10-10
7.82
X
10-3
62
X
10-10
38
-------
Footnotes (Table 14)
a/ Exposure data from Table 3.
b/ Lifetime risks were estimated as follows:
working lifetime Total Exposure
Risk v (40 years) x (ug/vear) x 1 person x 1 ng
Slqpe average lifetime 365 days/year 65 kg 1000 ug
(70 years)
c/ Trifluralin risk slcpe = 7.7 x IQT^/wq/kq body weight/day.
6/ NDPA risk slope .= 3.6/fciq/kg body weight/day.
e/ Cj/Cq risk slope was assumed to be as large as the diethylnitrosamirie
(DENA) risk slcpe of 33/tag/kg body weight/day.
39
-------
This assumes an average working lifetime of 40 years, an average lifetime of
70 years and an average body weignt of 65 kilograms.
¦Thus, for, example, the lifetime individual risk (R) for a
mixer/applicator/loader fron exposure to trifluralin wnile treating soybeans is:
R - 7.7 x 10
860.2
"365"
40
Iff
x 1
lOT
= 1.6 x 7
-10
Tne risks for NDPA and C-j/Cg nitrosamines (Tables 14) were similarly
calculated {Regelman, 1981c), using tne appropriate slopes as above, and using
exposure data front Table 3.
The risks to mix°r/applicator/loaders fraiuexposure to trifluralin, are in tne
order of 1 x 10~ and range from 0.2 x 10" for "greens" to 5.55 x 10~
for sugar cane. Tne range for NDPA is. from 1 x 10"" for okra, "greens",
cucumbers, and watermelons to128 x 10" for sugarcane. The_range for
C-j/Cq nitrosamines is 1 x 10" for cucumbers to 167 x 10 for
sugarcane.
b. Comparison of Risk Estimates with those of tne PD 1/2/3.
The lifetime individual risk to mixer/applicator/loaders fron exposure to tne
NDPA nitrosarnine contamination in Treflan® was estimated in tne PD 1/2/3 to
range fron 0.15 x 10" for "greens" and cucumbers to 4.60 x 10" for
sugarcane. Tne estimates for nitrosamines (C^/Cg) in tnis document
(Table 14J range from 6 x 10~i0 for okra, "greens", peppers, and cucumbers,
167 x 10" for sugarcane. (The range for MDPA nitrosamines risk is
currently scmewttat lower tnan tnat for tne C^/Cg nitrosamines. It was
deemed appropriate to compare the current highest nitrosarnine range of risk
figures, C7/Cg, to those for NDPA in the PD 1/2/3.)
Tne current risk from exposure to nitrosamines is lover than that calculated
in the FD 1/2/3 due to a reduction in NDPA contamination in Treflan® frcm a
level of 5 ppn to an average of 0.10 ppm with tne introduction of new methods
enployed in Elanco's manufacturing process. Also, as discussed, a different
mathematical model was used to calculate tne slope parameters in this document.
Based on Elanco's cnronic feeding study,'the risk from trifluralin itself is
established as tne major source of tne risk associated witn exposure to
-Treflan®. Tne largest lifetime individual cancer risk estimate frcm exposure
to tne trifluralin in Treflan for mixer/applicator/loaders is 5.5 x 10~ for
sugar cane. In the PD 1/2/3, the highest risk frcm exposure tp Treflan® was
4.6 x 10" but it was due to tne levels of NDPA contamination present in
40
-------
Treflan® at that time. Similarly, in tne PD 1/2/3, tne risk to corrmercial
applicators was 1.70 x ID , wnereas currently trie hie nest risk to cawnercial
applicators is 2.0 x 10 from exposure to trie trifluralin in Treflan®. The
risks due to tne tlDPA and C-j/Cg nitrosamines do not add appreciably to tne
total risk currently associated with exposure to Treflan®; there would be no
change in tne order of magnitude. Thus, while the source of risk has changed
from NDPA to trif luralin, the risk associated with mixer/applicator/loader
exposure to Treflan® as calculated in the PD 1/2/3 has not changed
significantly; the risk is of the same order of magnitude.
In order to account for the fact that mixer/applicator/loaders also are
exposed to additional risk tnrough their diet, one can add the two risks. When
tne total dietary risk is added to the highest worker risk fog the ED 1/2/3, a
figure of 49.3 x 10" is obtained (3.3 x 10 plus 46.0 x 10 ).
Similary, when the §ame computation is done with the current risk figures, a
figure of 103 x 10~u is obtained (53 x 10 plus 55 x 10~u). This
comparison straws tnat tne risks associated -witn exposure of
mixer/epplicator/loaders to Treflan® nave increased by a factor of 2. This
is not considered to be an appreciable increase.
4. Reentry Risk
a. Nitrosamines
In tne Mr
Any risk associated witn exposure to nitrosamines in air would be
insignificant. The estimates for nitrosamine exposure in the air as a vapor
were calculated for a few crops in tne PD -1/2/3 and were found to be
negligible? the risk was also assumed to be negligible. Mittleman (1978)
stated that nitrosamine exposure in the air as particulate matter was also
insignificant and, therefore, no risk estimates were calculated in tne PD
1/2/3. Regelman (1981b) agreed tnat inhalational or dermal exposure to
nitrosamines an particulate matter in the air was negligible; therefore, no
risk estimates for exposure to nitrosamines in the air were calculated for this
document, since current exposure to nitrosamines is lower by a factor of about
50 than it was at tne time the PD 1/2/3 was issued.
In the Soil
Tne Agency had calculated in tne PD 1/2/3 dermal exposure estimates for 2©PA
in tne soil, since reentry exposure was theoretically possible, and arrived at
a figure of 0.036 micrograms per year. However, this was considered to be only
a theoretical estimate arri not an actual estimate because, as stated in the
PD 1/2/3, there is extreme variability in tne type of field activities
performed during reentry, in trie degree of physical contact witn tne soil, in
tne amount of exposed body area, in tne length of time exposed, and in the
partition coefficient of trifluralin between the soil particles and the part of
41
-------
tne numan body wnich canes in contact with the soil particles. Because of trie
possibility of a high degree of inaccuracy, and because- tne exposure was
expected to be low, if it existed at all, tne Agency determined tnat it would
not be reasonable to calculate a risk estimate for trie PD 1/2/3 tnat might be
incorrect by several orders of magnitude.
Dermal exposure to nitrosamines under reentry conditions was calculated for
tnis document to be ro greater tr.an 5 x 10 micrograms per year for NDPA and
1 x 10 micrograms per year for 0-,/Cg nitrosamines. Because exposure
estimates at tnese low levels are not reliable and because tney are negligible
amounts, tne Agency assumed that any risk associated with exposure to
nitrosamines in tne soil would also be negligible.
b. Trifluralin
Tne Agency estimated reentry exposure to trifluralin as a vapor in
Section II. B. Tne total trifluralin exposure during reentry, including
innalation and dermal contact with trifluralin vapor, and dermal contact with
trifluralin adsorbed to particles in tne soil (Table 95 was used to calculate
tne risk, to workers reentering Treflan®-treated fields. The risk estimates for
several crops are presented in Table 15.
The lifetime individual risks were calculated as described for the
mixer/applicator/loaders, using a slope of 7,7 x 10 ^ per mg/kg/body
weight/day. The risk estimates range from 1.1 x 10 for cotton to
4.2 x 10" for tomatoes.
5. Summary of Dietary and Worker Risk
Tne magnitude of risk associated wi. tfi 3 cnemical depends on its potency as a
carcinogen, as well as the extent to wnich the general public or workers are
exposed to tne chemical. If a certain chemical is highly carcinogenic as
indicated by the slope parameter, but the exposure is very low, the risk may be
acceptable if it is offset by the benefits derived fran its uses. Conversely,
if a cnemical is a mild carcinogen, but exposure is extensive, the risk may not
be acceptable if it is not offset by benefits.
For example, this document discusses risks associated with three chemicals,
eacn witn a different potency and different exposure. Trifluralin is tne least
potent of the three with a slope parameter of 7.7 x 10" , whereas
NDPA is more potent (slope = 3.6), and C-,/Cg, assumed to be as carcinogenic
as DENA, is trie most potent (slope = 33). However, tne two nitrosamines
present a lower risk to tne general population and to workers because, tnough
tneir slopes are higher, their concentration in Treflan® is lower than 1 ppn,
while trifluralin's concentration is 445,000 pprr. (44.5 percent). The exposure
to tne nitrosamines is tower than exposure to trifluralin. Thus, the potential
carcinogenic risk £rcm exposure to Treflan® is now primarily due to
trifluralin itself.
42
-------
Table 15
Trifluralin Risk Estimates - Reentry^/
Crop
Soybeans
Cotton
Beans
Ibmatoes
Trees/Vine
Cole Crops
Tbtal Trifluralin
Exposure
(micrograms per year)
8.09
9.07
8.76
22.88
12.68
7.86
Lifetime
Indlvidu^
1.5
1.1
1.6
4.2
2.4
1.5
10
10
10
10
10
10
I
a/ Exposure data fran Table 9.
-3 -1
b/ Trifluralin slope = 7.7 x 10 (mg/kg body waight/day)
9/
Risks were calculated as described for mixer/applicator/loaders in Tfcble 14.
-------
In tne ED 1/2/3, the total dietary risk associated with exposure to
Treflan® w^s 3-3 x 10~ . Tne highest mi^er/applicator/loader risk was
,4.60 x 10 ' for sugarcane and 1.70 x 10 for ccmmercial applicators. Risk
to reentry workers was highest for tree and vine crops at 2.1 x 10 , but
reentry risk was considered negligible wnen pnctodegradaticn was taken into
account. Tne risk at tnat time was due to exposure to tne NDPA carcinogen
contaminant. When the benefits of tne use of Treflan® were assessed, tne
Agency determined tnat tne benefits outweigned risks if certain recommendations
were followed (H3 1/2/35.
In tnis document (PD 4), tne highest total dietary risk associated with
expDsure to Treflan® is 5.3 x 10~ . 'Die highest mixer/aralicator/loader risk
is 5.5 x 1Q~ and tne highest reentry risk is 4.2 x 10 . These risks are
associated with exposure to the trifluralin in Treflan®. There are sore risks
associated witn tne nitrosaniines in Treflan®, but they are negligible wnen
compared to those for trifluralin as discussed above. This is due to the large
reductions wnich have already occurred in nitrosamine contamination levels.
A comparative suimary of tne risks in the ED 1/2/3 and this document is
found in Table 16.
Currently, tne dietary risk has increased by one order of magnitude, wnereas
mixer/applicator/loader risks have remained essentially the same. Hie highest
reentry risk has decreased by one order of magnitude.
Tne overall risk has not changed significantly since the PD 1/2/3 was issued.
Tne benefits, as discussed more fully in Section III. B., nave varied slightly,
but have not changed appreciably.
D. Ecological Effects
In tne PD 1/2/3, the Agency indicated tnat trifluralin was highly resistant
to leacning, was strongly adsorbed to organic matter, and did not readily run-
off frcn treated fields. Ch this basis, tne Agency determined tnat trifluralin
did not meet car exceed the risk criteria for adverse ecological effects on
aquatic oraganisms.
tower, new information and re interpretation of old data snows tnat
trifluralin could reach aquatic environments through soil runoff. Trifluralin
is persistent and nay be further long lived when sediment bound residues fail
to degrade and are slowly desorbed yielding low-level chronic exposures to
aquatic populations. The desorption or disassociation of trifluralin from this
"soil carrier" may be biologically significant (Ibuart, 1981). The very high
toxicity of trifluralin to aquatic organisms is well established (Cope, 1966;
Macek et al. 1969, 1976; Parrish, et al.1978). MAffC (Maximum Acceptable
Toxicant Concentration) levels for finfish are typically 1-5 ppb. Aquatic
44
-------
Table 16
Summary of Risks Est in® tod in the ED 1/2/3 FD 4
ED 1/2/3^/
ro 4^
Dietary (total)
3.3 x 1
-------
organisms also bioconcentrate trifluralin from 4000 - 150,000 times (Sanborn,
1974; Farrisn et al., 1978). It is not known if this bioaccumulating tendency
represents a bianagnifying problem or if any adverse effects are associated
witn a nigh trifluralin body burden. Additionally, Gouch et al. (1979) have
recently reported vertebral dysplasia in sheepsnead minnows
(Cyprinodon variegatus 3 exposed 'in early life stages to 5.5 ppb or greater
trifluralin. A 'no effect level for this ancmaly has not been reported.
The amount of trifluralin lost from treated fields in runoff as reported by
several investigators has been summarized by Waucnope (1978), A maximum loss
of 0.76 percent of tne applied treatment to cotton plots nas been reported
witn concentrations in runoff as nign as 120 ppb (water pi is suspended
sediment) and 1.6 ppb (filtered water alone) documented (Sheets, 1972). Thus
tne Agency no longer considers the transport of trifluralin residues unlikely
and tne resulting aquatic concentrations of tne herbicide insignificant.
Cn September 5, 1980, tne Agency amended tne registration for Treflan® EE to
include use on field corn, sorgnum, and barley under autnority of FIFRA
3(c)(7)(B). The Agency stated (Mountfort, 1980) tne labeling was acceptable
since Elanco had agreed in part that "When required by tne Agency, you will
submit and/or cite the data pertaining to the movement and concentrations
of trifluralin in tne environment from typical application and data on the
nazards to aquatic organisms. If the conditions of this amendment are not
compiled with, tne registration vail be subject to cancellation in
accordance witn Section 6(e) of the Act*"
Therefore, because trifluralin can be transported as bound residues in soil
runoff, nas been shown to be chronically toxic to fisn at extremely low levels,
and aquatic organisms have been snown to bioaccumulate trifluralin, tne Agency
nas determined that a field monitoring study is necessary to assess possible
adverse effects to tnese nontarget aquatic animals. The Agency still considers
tne hazard of trifluralin to terrestrial animals to be slight (Tbuart, 1981).
Tne requirement for tne field monitoring study is being proposed not because
RPAB criteria nave been met or exceeded, but because the Agency has determined
that information exists wriicn indicates that possible adverse,effects to
aquatic animals could occur in bodies of water adjacent to Treflan®-treated
fields.
46
-------
III. Analysis of Garments
After tne Notice of Determination was issued concerning tne Treflan® RPAR on
August 30, 197S, tne Agency received cements fran the Secretary of
Agriculture, tne Scientific Advisory Panel (SAP) and 17 other concerned
individuals and organizations. Many of the respondents wrote to express their
concurrence with the Agency's proposed decision: the Secretary of Agriculture
[Appendix B], Heckle [30000/32;#2], Nalewaja [30000/32:13], Jennings
[30000/32;#4], Upcnurch [30000/32:#5J, Lange [30000/32:#71, Davis
[30000/32:#14], Leggett [30000/32:#15], and Aves [30000/32:1161. Other
requested information wnich was subsequently supplied by tne i^gency: Kerapen
[30000/32:#11, Stanger [30000/32;#8], Teramura [30000/32:#91, Burr
[30000/32:110], Cert [30000/32;#111 and Baldi [30000/32:#13]. Three
ccmraenters disagrees witn several portions of the Agency's position or
rationale used to arrive at that position: Elanco [30000/32:*6], Knake
(Intersociety Consortium for Plant Protection) [30000/32:#12], and American
Cyanamld [30000/32:#17]. Since tne oomments on the PD 1/2/3 from the SAP
[Appendix A] and these last three respondents are, for the most part, lengtny
and detailed in nature, the responses are organized by topic and are discussed
below. As indicated in the following discussicn, seme aspects of the PD 1/2/3
assessment of risks and regulatory requirements nave been changed to reflect
tne recommendations of tne SAP and others. Other aspects of the PD 1/2/3
remain unchanged.
A. Garments Relating to Risk
1. Worker Exposure-
a. Exposure Estimate of N-nitrosodipropylamine (NDPA)
Knake [30000/32:#12] asked about the concentration which would result if one
to two pints of trifluralin containing less tnan one part per million
nitrosamine contaminant (NDPA) were mixed witn a million pounds of soil.
Additionally, he stated tnat if tne Agency was concerned about applicator
exposure, a few trips to the field to see actual use practices mignt nave
helped,
Sgency scientists calculated tnat if tne product in question was Treflan®,
tne resultant concentration of the nitrosamine contaminant would be
0.5 to 1 x 10~ microgram/gram soil or 0.5 to 1 part per trillion, assuming
a uniform distribution of nitrosamine tftrougnout soil. This fact, however, has
no impact on estimates of worker exposure during application, since tne Agency
estimates were based cn residues of NDPA in air and dust samples collected
during actual field experiments (Day et al., 1978). Concentrations of tne
nitrosamine NDPA in dust collected behird a spray rig applying Treflan® ranged
from 0.01 to 0.07 microgram/gram. or 10,000 to 70,000 parts per trillion.
Knake's point regarding field trips is well taken. Curing the trifluralin
47
-------
review, field trips by Agency scientists and other staff members were
undertaken as were consultations and discussions with applicators, USDA
representatives, University representatives and registrants^ The information
obtained was applied to tne Agency's estimates of risk outlined in the
H) 1/2/3. Tnus, the above cement [30000/32:#12] does not alter the worker'
exposure estimates calculated by tne Agency and presented in tne FD 1/2/3.
b. M ixer/Applicator/Loader Exposure to NBPA
1) Inhalation
Elanco 130000/32:#6] stated tnat tne EPA estimate of exposure to NDPA via
inhalation of dust was excessive by about 13.5 times the median value measured
by the company under field conditions (Day et al., 1978).
Elanco in a subsequent submission {Elanoo, 1980c) objected to toe Agency's use
of tne mean of four vacuum sweeper samples (96 x 10 ug tEPA/m ) in its
applicator/teixer/loader inhalation exposure estimates.— Instead, Elanco
suggested that data from tne dust pansi' should haVe been inclined as
WO 1 1 Tn "i +* 1 FI arifTt " f i 177dH«-TR7Df" f-a 4"VT1Tl "fvTll IT" At'hOT"
wtrX-L* JJ1 auUiL1\Jl f i uiicli ucrlHTl J OaUt LITvJli Iwla QuIJcX
experiments— {data wnicn were totally excluded as invalid in the Agency's
evaluation), including them in a Log-Probit statistical treatment using all tne
data fran trie vacuum sweeper and dust pan samples.
As a result, Elanoo estimated that dust-NDPA levels were actually
7.3 x 107 ug/n , about 13.5 times lower than the Agency's figure of
96 x 10 ° ug/m . Tne Agency's decision to use only the valid vacuum
sweeper data^? rather tnan tne dust pan samples and tne results of invalid
samples—7 was based cn the following rationale (Begelman, 1981b).
a) In their original submission (C&y et al., 1978) Elanoo reported that tney
collected large dust samples in their field nonitoring study to provide a basis
for determining the concentration of trifluralin and NDPA cn particulate matter
in tne air breathed by mixer/applicator/loaders. The particulate samplers
collected large samples of dust with the concentrations of the chemicals being
higher on tne dust fran the vacuum sweeper (VC) than from tne dust pans (DP).
3/ Day et al., 1978, Experiments # DMS 7-3VC, CMS 7-4VC, CMS 7-5 VC, and CMS
7-6VC.
4/ Day et al., 1978, Experiments # DMS7-3 DPI, DMS 7-3 DP2, CMS 7-4 DPI, DMS
7-4 DP2, DMS 7-5 DP, and CMS 7-6 EP.
5/ Day et al., 1978, Experiment # DMS 7-7 DPI, DMS 7-7 W 2, CMS 7-7 VC 1,
and DMS 7-7 VC 2.
48
-------
Tney stated this was due most probably to trie smaller sized particles (0.5u -
lOu) collected by tne VC. They stated tnat theoretically only tne respirable
particulates snould reach the lungs, but larger dust particles could be
deposited in tne nasal passages. Elanco tnen used tne chemical content of tne
VC dust samples (presumably more representative of respirable particles) to
calculate tne "...maximum exposure to NDPA and trifluralin via innalable
particles." Trie results were set out in Table IV-15 of tneir submission, with
exposure estimates given for botn respirable and total particulates. Tne
Agency agreed tnat tnis approacn was valid and tnat using the data frcm tne VC
samples would give a reasonable estimate of typical inhalational exposure. The
Agency used tnese figures in tne exposure estimate set forth in tne FD 1/2/3.
Elanco in tneir additional comments en tne FD 1/2/3 submitted September 4,
1980, (-ElanoOr 1980c) proposed tnat the data on the chemical content of tne .
dust pan samples snould also be included. However, in tneir original
submission, tney cnose not to incorporate these data into their inhalational
exposure estimate because of tne reasons described above.
b) The larger particle DP samples contained lower levels of trifluralin (and
NDPA and C^/Cg nitrosamines) tnan the VC samples, presumably due to tneir
lower surface-to-volume ratio. This presumption is supported by a comparison
of tne means of tne two data sets, in which the VC samples appear to contain
nearly three times tne average..concentration of NDPA as the DP samples
{96 x 10 vs. 36 x 10~ ug/m ). Since the particle-size distributions
were apparently quite different in the two groups, it vrauld seem inappropriate
to average all reported data. Instead, the VC samples (which were more
representative of respirable particles) were used exclusively by the Agency and
initially by Elanco,to estimate maximum exposure to innalable particles. The
average of 96 x 10"" micrograms per cubic meter was based on four
measurements frcm experiments in which TreflarP application ard incorporation
operations were carried out simultaneously. Tne Pgency used this approach in
the ED 1/2/3 and still deems it valid.
c) Elanco*s rationale for using tne log-probit technique to evaluate tnis
group of data is unclear frcm their rebuttal cements. The Agency does not
agree that Elanco has proven that the VC data are "skewed", or that, a
log/probit analysis is justified.
d) We do rot understand tne justification for "fuzzing-over" data (i.e.
including data in tne "count" of the number of samples, but not including tne
specific'values in the correlation computation). It is the Agency's
understanding that tnese "fuzzed-over" data—' were collected'under
application/incorporation conditions quite different from those of the otner
data; i.e., applications and incorporation of Treflan® were carried out
separately. Thus, those measurements do not reflect total particulate phase
exposure during simultaneous application and incorporation (Day et al., 1978,
page 5).
49
-------
e) Trie Agency could nave taken, tne conservative approach that the highest
measured VC sample (296 x 10 ug HDPA/m ) was more representative than
tne mean of that group of data. The Agency's estimates would than have been
three times higher tnan actually computed.
f) Finally, there were only four valid VC data points. Considering this
limited amount of data, tne Agency.determined tnat averaging them was the best
possible approach.
2) Dermal
Elanco also suggested [30000/32:#61 that the EPA estimate of dermal exposure
based on a single positive result in a cloth charcoal adsorber sample was
excessive-and that alternative estimates submitted by them reduce this dermal
exposure to one-sixth that calculated by EPA.
The EPA estimate of dermal exposure was based on an.unsatisfactory data base
since experiments designed to measure potential dermal exposure.were
unsuccessful. Attempts were made by Cay et al. (1978) to determine dermal
exposure to.the applicator by analysis of cotton gauze arm pads, cotton shirts,
and ootton gloves worn by the applicator. These studies, however, were not
'useful in predicting dermal exposure due to the extremely poor recoveries of
NDPA and trifluralin from samples innoculated in the field, contamination of
field "blanks", and the inability of the cloth material to prevent dissipation
of NDPA and trifluralin during the work day {Mittelman, 1978).
In recognition of these difficulties, Etey et al. attempted to simulate dermal
exposure by constructing an "adsorber" consisting of a layer of charcoal held
fast around 2 glass jar by a tigntly bound cotton cloth. Only one study using
the cnarcoal-cloth adsorber revealed positive levels of NDPA in tne charcoal.
The Agency considered this to be a valid measurement and based the estimate of
dermal exposure in the ED 1/2/3 cn this value.
Elanco (1980c) stated that the negative results for NDPA measurements frcm the
experiments using the cotton gauze arm pads, cotton shirts, and cotton gloves
can be used. They calculated tne maximum levels of NDPA that could have
been tnere by using the limit of sensitivity of their method of detection.
Using these levels, the average dermal exposure is 0.0138 ug/nr, a figure one
sixth that of the EPA's figure of 0.083 uq/ftz.
Tne Agency maintains tnat tne studies using cotton gauze arm pads, cotton
snirts, and ootton gloves are not useful in predicting dermal exposure and tnat
tne measurement derived frcm the study using tne cnarcoal-clotn adsorber is the
appropriate figure to use. The Pgency agrees that the resultant exposure
estimate is clearly biased toward conservatism, but in the absence of adequate
data, no alternative action was deemed appropriate {Mittelman, 1980). The
50
-------
Agency assumes tnat dermal exposure can occur at least up to tne one level
measured in the field study (Day et al., 1978). Currently, however, tne
average NDPA contamination in Treflan is approximately 1/50 of tne level
present wnen tne FD 1/2/3 was issued. Tne revised exposure estimate outlined
in Section XI.B. of tnis document takes tnis into account.
c. Reentry Field worker exposure
Elanco 130000/32:#6] asserted that there is no evidence nor is it reasonable
to assume tnat field workers who reenter trifluralin-treated fields are
exposed to NCPA? that EPA innalational exposure estimates were as much as 100
times too nigh. Elanco also criticized EPA's use of laboratory results instead
of field studies indicating negligible exposure.
Elanco (19B0c) furtner explained tnat tney used results frcm a study by Wnite
et al. (1977) which measured trifluralin volatilization loss under field '
conditions. NDPA was not measured because it would have been below tne level
of detection. Elanoo calculated the theoretical levels of NDPA that would have
been present by using a trifluralin to NDPA ratio of 39000:1, a ratio which
they criticized EPA for using in tne dietary exposure estimate presented in the
PD 1/2/3. (Elanco1s
-------
Table 17
HDPA Xnnalstion Exposure (ug/year)A/
Crop Mittelroan Elanco Batio
Beans 3.0 0.024 125:1
Itomatoes 5.0 0.060 74:1
Tree and Vines 15.0 0.145 103:1
Cole Crops 6.0 0.077 78:1
1/ Elanco, 1980c.
52
-------
Trie Agency might nave used sucn an approach if no data were available to allow
tne derivation of a more appropriate model for tne estimation of
inhalationalexposure. Trie innalation exposure estimates reported -in tne PD
1/2/3 were based an actual field measurements (Day et al., 1978) in conjunction
witn a laboratory model system (Oliver, 1978 5 as described by Mittelman (1978,
Addendum 5). Since tnis approacn was based on actual tOPA measurements from
botn field and laboratory experiments, it was considered to be a better model
tnan tr.at used by EUanco in calculating tneoretical levels of NDPA.
Tne Agency affirms that its field/laboratory model affords a more realistic
estimate. Botn metnods (Elanco's and tne Agency's) arrive at estimates of •
exposure for innalation wnicn do not pose a significant carcinogenic risk. The
risk calculated ¦ from tne Agency's exposure figures was outweighed by tne
benefits of trifluralin use, as tne risk vould be if calculated frcm Elanco's
estimates. Currently, tne risk would be even less tnan that calculated by tne
Agency for tne PD 1/2/3 primarily since NDPA contamination is currently much
less tnan it was previously estimated to be, as discussed in section II. B. 3.
of tnis document. The Agency concluded in tne PD 1/2/3 that inhalational
exposure and risk under reentry conditions were negligible, and presently
reaffirms tnat conclusion. If Elanco's model were used to calculate NDPA
exposure, it too would be negligible; there would be no cnange, tnerefore, in
tne Agency's regulatory position if either model were adopted.
The "worst-case" EPA estimate of dermal exposure was also criticized by Elanco
[30000/32:#6] as being exaggerated, sines it suggested that a worker would
collect nearly two pounds of soil on his'exposed skin.
Tne EPA estimate of dermal exposure to reentry workers based on contact witn
NDPA contaminated dust was presented for illustrative purposes to indicate that
dermal NDPA exposure during reentry is theoretically possible. Since no
reentry dermal exposure data exist, it was considered appropriate to determine
tne "worst-case" situation, i.e.,the worker is literally covered--with dust.
The Agency used this approach enly to describe an upper limit to possible
exposure.
The EPA dermal exposure estimate was based on tne assumptions tnat a worker
entering trie field has a total uncovered skin surface area of 2,900 cm and
tnat a uniform layer of sgil forms a 1.0 ran thick film on the uncovered skin
and that tnere are 3 x 10 grams of soil per cubic meter. Using tnese
assumptions, the quantity of soil on tne exposed skin would be 0.87 kg or
almost 2 lbs.
Currently, however, tne Agency is using a different method to estimate the
amount of dust a worker could collect on his skin (Jensen, 1981). This
approacn was used in the revision of tne exposure estimates discussed in
Section II. B. 3. of this document. It was estimated that a worker could
collect approximately four grams of dust en his exposed skin, rather than
pounds as originally calculated in the PD 1/2/3.
53
-------
As discussed in Section II. B. 3. of this document, the exposure to NDPA at
current contamination levels in Treflan© was calculated to be 5 x 10
micrograms per year, a level judged to be negligible. No risk was calculated
for tnis exposure, since tne risk would also be assumed to be negligible;
pnotcdegradaticn, soil metabolism, other degradative and metabolic pathways,
and dermal penetration of 22 percent would result in even lower levels of
exposure than that calculated.
2. Dietary Exposure
Elanco stated tnat '"There is no evidence, or is it reasonable to assume, tnat
there is any dietary exposure to NDPA as a result of the use of Treflan"
130000/32:#6].
In its PD 1/2/3 the Agency acknowledged that West and Day (1977) could not
detect NDPA at a test sensitivity of 0.1 to 0.2 ppb in crops grown in
agricultural land treated for successive years with Treflan® EC containing as
much as 450 ppm NDPA.
However, tne Agency reviewed laboratory studies which indicate that
nitrosamines can be taken up into plants grown in C-NDPA treated soils over
at least 49 days !Dean-Paymona and Alexander, 1976? Berard, 1977; Berard and
Rainey, 1977), From levels reported in these studies and those of West and Day
(1977) the Agency concludes that it is reasonable to assume that HCPA could be
found in various Treflan® treated commodities at or above tne test sensitivity
of 0.1 ppb (Donoso, 1980). Because of this the Agency felt it was prudent to
develop tne dietary exposure case described in tne PD 1/2/3 on trifluralin.
Elanoo further stated that it considered two of the Agency's assumptions used
in the dietary exposure estimate to be unsound;
o That trifluralin residues up to and including the established tolerance
limit are "probable."
o Tnat NDPA residues in crops will fcc in the sane ratio to trifluralin as
that observed in the fonnulaticn. (Elanoo, however, used tne same ratio
of 89,000si when they calculated inhalation exposure to HDPA as
shown in section III. A. 1. c. of tnis document discussing field worker
exposure.)
Elanco furtner stated tnat the FDft and others have analyzed over 20,000
samples covering more tnan 27 crops in which no detectable trifluralin residues
were found at a detection limit of 0.01 ppm.
54
-------
Tnougn many such surveys nave failed to detect trifiuralin residues in crop
samples, FDR lias found trifiuralin residues in canmodity samples on a number of
occasions in its surveillance and compliance programs, Tnis was indicated in
trie trifiuralin PD 1/2/3. Trie Agency, however, did not oonsider these data
sufficient to be used in its dietary exposure estimate because the number
of positive findings were few and ill-defined; nine samples were identified as
"raw agricultural products", miscellaneous items", arid "fisn and marine
animals".
Tne Agency developed a "worst case" exposure in tne PD 1/2/3 based on
infnrfifif m ^ r5or*r*pnt* an?1 rvf rrnn arrPAfifl t* i cyr\ nn
XI i LtwiM4 I'G * ivl i sJi ft I XerdTi i Cti. CfQ> Wvv J alb JL Oi iCm X* < i»WL>i * Wdlwi I Wi I
trifiuralin tolerances, data cn food factors, and a then-established
trifiuralin to NDPA ratio of 89,000:1 in formulated Treflan®. Using the
oSS»>mip>iiilOfi lJjqL lu W3S ^X}SSXDX6 iyJricL1!!L l-TTXlXUTBXXn OOOXQ 06 |3EnE*SSr*l- 3>w IwS
tolerance levels, tne Agency calculated in the ED 1/2/3 trie "Potential Dietary
Exposure to NDPA" and determined, using tne above published information, tnat
the total probable dietary exposure to IIDPA would be about 1.92 x 10 mg/kg
body weight/day. The i^gency stated in the PD 1/2/3 that these residues may in
fact be much lower. The Agency's dietary carcinogenic risk estimate based on
tnose exposure figures represented the "worst case".
Elanco offered alternative estimates of -dietary exposure to IJDPA
associated with use of trifiuralin [30000/32:#63. One such estimate was
further explained in a letter from Elanco (Day, 1980). The estimates were
extrapolated fran study results of trifiuralin residues in crops performed by
Elanco. Trifiuralin residues used were either the highest level reported or the
sensitivity of analytical methods for trifiuralin that were used. It was shown
that the total dietary risk based on those NDPA exposure estimates was
8.05 x 10 , based on exposure to Treflan® containing 5 ppm !®PA.
Because of tne acquisition of new data on the carcinogenicity of trifiuralin
itself and because a new exposure estimate was needed, tne Agency decided to
base tne new exposure estimate cn actual residue data sufcmitted over the years
by Elanco on tne several crops expected to oontain trifiuralin and nitrosamine
residues (Regelman, 1981a). It was determined tnat because Treflan® had been
shown to contain two chemicals (trifiuralin and nitrosamines, including NDPA
and Cj/Cq) snown to be associated with tumor production in laboratory
aril tnal c ^ 1 #¦""! I'vii wwq •% aQ vv*¦? ttc* ¥• 4^ 4 mss+'Q itw»va *"¦»T nl it
aaiinicLLor iu vouiu Eje note appropriate xjd arrive 3Z an esuiiuoue uure cj.oseiy
approximating "real v/orlc" conditions. Hie Agency determined that it would be
reasonable to use these residue studies for the exposure estimate as they
comprised a large data base which was not highly variable.
It was assumed tne data were representative of treated crops generally. A
discussion of dietary exposure to trifiuralin and nitrosamines was presented in
section II. B. 1, of tnis document. Tne total dietary risk calculated by the
Agency (Section II. C. 2 for NDPA was 5.85 x 10"* , less than that calculated
by Elanco (Day, 1980), because the Agency's estimate is based on 0.1 ppm NDPA
in Treflan® instead of 5 ppm. The total dietary risk calculated for
trifiuralin by the Agency was 5.3 x 10~ , the highest per crop being
2.2 x 10" for carrots. Tne Agency's calculations for C7/Cg nitrosamine
55
-------
—1 0
total dietary risk was 1.03 x 10 and does not contribute significantly to
tne overall risk. - These estimates, because tney are based on actual
trifluralin residue data, reflect rnre closely tne actual concentrations of tne
tnree cnemicals in tne diet tnan a "worst case" (theoretical maximum) estimate
would.
3. Tbxicology
a. Cancer Risk Assessment
1. Elanco's-Carments
Clanco (30000/32:#6] stated: "Tne cancer risk assessment was determined by
tne use of questionable methods Elanoo cannot accept trie model used by EPA
to extrapolate frcm results in published animal studies. The model is
apparently intended to be a generalization of the familiar one-hit exponential
model. However, it can be demonstrated tnat tne particular form of tne model
used by EPA is invalid wnen tested against actual data. It can also be shown
tnat tne metnods of calculation based on tne unique form of tne model lead to.
erroneous slope estimates. Since linear risk extrapolation relies on sucn
estimates, all risk assessments made by EPA must be considered invalid.
Elanco recognizes tnat tne entire subject of risk assessment is in a state of
flux, and tnat objections may legitimately be raised to all approaches.
However, tne extrapolation procedures used previously by EPA in tne decision
document supporting tne denial of tne petition to suspend tne registration of
Treflan® (Federal Register, August 8, 1977) come far closer to
acceptability tnan tne method used in the present case."
Elanco submitted additional canments (Elanco, 19E0c) explaining more fully tne
reasoning benind tneir position. Tney had specific suggestions for calculating
tne risk assessment using a model which tney claimed would fit the NDPA data
better arri give a more precise estimate of risk.
The Agency's Carcinogen Assessment Group (CAG) reviewed Elanco's canments
(Chen and Haberman, 19815. Tne CAG stated that Elanco's main criticism was
tnat tne Agency's "one-nit" model, P =1 - exp(-bdtm), determined by using tne
Druckrey relationship, (t^c)n x d - constant, may not adequately represent
tne real dose-response curve at time points other tnan t^p. Based on tne
cumulative tumor incidence in the highest dose group (out of nine dose groups)
from tne study by Druckrey et al. (1967), Elanco demonstrated that tne model
constructed by the Agency did not adequately fit tne DENA data because tne
model overestimated tne cancer incidence at early ages and underestimated the
cancer incidence at later ages (e.g., tne observed incidence was 100 percent
at 760 days after correcting for mortality while tne Agency's model predicted
only 65 percent).
56
-------
While Blanco1 s point is correct, it is not directly relevant to the "Agency's
risk assessment for HDPA (or DENA) because the Agency made no attempt to
estimate tne time course of tumor formation using its "one-nit" model and/or
estimate tne lifetime risk at nigh doses as Elanco hac done. Instead, trie
Agency used Druckrey1 s potency 'parameter, k = d(tcjg)n, solely to determine
tne potency parameter b wnen t = 1 (lifetime). The slope of tne "one-nit"
model, b, was intended to be used only for estimating lifetime risk at low
doses. Tne Agency realizes that determination of tne slope, b, wnicn is
intended for use at low doses, may be inaccurate for estimating risk at high
doses, and • tnat this procedure may incorrectly equate the value of n in
Druckrey*s fomulation vritn tne value of m in tne "one-hit" model. However,
tne procedure was used because of the Agency's desire to estimate the slope
for a large number of N-nitroso compounds frcro Druckrey's data, which
are available only in tne form of tabulated values of k and n. This procedure
appeared to Ids trie only method one could use to obtain a solution.
However, even trough this assumption may be suspect, at least in the
situations wnere it can be independently verified, its use does not appear to
nave introduced mucn error. For DEHA and NDPA, where both tumor incidence data
and Druckrey's parameters exist, tne Agency's procedure used in tne PD 1/2/3
does not differ appreciably from the multistage procedure wnicn the Agency
currently uses.
Tne carcinogenic_cotency of M3PA using tne Druckrey index is estimated as
4.38 (rag/kg/day) which is close to 3.62 as estimated by tne multistage
model. It snould be noted that tne NDPA slope (potency) estimated by tne
Agency in the PD 1/2/3 of 0.4 (ppm)~ differs from 4.38 because it was
assumed that ppm dietary consumption is equivalent between humans and animals.
Tne Agency currently finds that this is not justified since tne calories/kg of
food is very different in tne diet of man compared to laboratory animals,
primarily due to tne moisture content difference. Tne current procedure used
is as follows: the ppm dose in animals is converted to mg/kg/day; this is tnen
converted to rag/kg/day for humans to obtain tne actual equivalent dose for
humans.
Tne carcinogenic potency of DENA using tne Druckrey index is estimated as
43.46 (mgAg/day) wnicn is again very close to 33.45 as estimated by the
multistage model for C7/CQ nitrosamines.
Tnese models (tne "one-hit" and "multistage") are considered to be relevant
only at lew doses and may not fit at nign doses, as the following example using
tne multistage rrodel snows. Without converting to human equivalent.dose, tne
carcinogenic potency of DENA for animals is q^* = 5.72 (mg/kg/day)"" , using
tne multistage model. Altncugn the lifetime cancer risk at low doses is
approximated by P(d) = 1 - exp(.-5.72 x d), (where d = dose), this formula is
not applicable at nigh doses. For instance, one can not use this formula to
calculate the risk at d = 0.15 nig/kg/day as Elanco had done. If the formula
57
-------
were used at d = 0.15, the risk would be only 0.58 instead of being 27/30 or
0.90 as observed. Tne risk predicted by trie multistage model (when trie fall
roodel is used) at d = 0.15 is 0.89 whicn closely agrees with the observed
incidence of 0.90.
Therefore, the'fact that CAG's procedure ("one-hit" model) does not fit well
with Druckcry's time-to-tumor data in the highest tested dose group does not
necessarily imply tnat the method is invalid at low doses.
The dietary and worker risks associated with exposure to NDPA were
recalculated using the multistage model and are described'in Section II. C,
along with the calculations for trifluralin arid C^/Cg nitrosamines. ¦
2) Merican Cyanamid's Ccnments
American Cyanamid, the registrant for Prowl®, an alternative herbicide for
sane trifluralin uses, Garmented verbally in April, 1980, to the Agency tnat
the assumptions used in arriving at the carcinogenic risk estimate for the N-
nitroso contaminant of Prowl® were not the same as those used for the N-nitroso
contaminant of trifluralin. They requested that the HDPA/trifluralin
carcinogenic risk be recalculated,using the same assumptions. Itie previous
inhalation estimates in the PD 1/2/3 were based upon an assumed breathing rate
for workers of 1.2 cubic meters per hour. This assumption has been changed to
1.8 cubic meters per hour to correspond to the breathing rate assumed in more
current risk estimates calculated by the Agency for the types of work performed
by mixer/applicator/loaders and field workers.
Tne new estimates for inhalational exposure were obtained by multiplying the
previous exposure estimates in the H) 1/2/3 by 1.5. This was done because a
breathing rate of 1.8 cubic meters per hour represents a 50 percent increase
over tne previous assumption of 1.2 cu m/hr. This was discussed in
section II.B. of this document.
The risk to workers was then calculated and is presented in Section II. C.
of tnis document. The estimates of the lifetime individual risk for
m ixer/appl^ cator/loaders., for NDPA presented in the PD 1/2/3 ranged from
0.15 x 10. to 4.6 x 10-'. The new estimates range from 1 x 10 to
28 x 10"1 . The new estimates are two orders of magnitude lower, principally
because of decreased NDPA contamination levels currently in Treflan®.
b. NDPA Oncogenic and Mutagenic Risk
The Agency asked the SAP to corrment on the determination that lew risks fran
EMA and gene effects are associated with the NDPA contaminant in Treflan®. The
SAP (Appendix A, October 15, 1979) agreed with the Agency cn this issue and
agreed with tne Agency's rationale for attributing the oncogenic risk of the
trifluralin product to its HDPA contaminant at that time. The SAP also deemed
58
-------
as reasonable tne Agency's assertion of lower risk at 1 pprr, NDPA contamination
tnan at 5 ppm, but stated that tnis assumption would have -to be confirmed by
tne ongoing oncogenic study witn tne current Treflan® product containing levels
of less tnan 1 ppm MOP A.
Tne results of tnis- study were not available at tnat time. As discussed
previously in tnis document, tnis study demonstrated that trifluralin itself,
essentially free of HDFA contamination, is associated witn tne production of
tumors in laboratory animals wtien administered in rsign doses.
The Agency reaffirms its position as discussed in tne PD 1/2/3 tnat tne
mutagenic risks for DNA/gene effects from NDPA currently associated witn
trifluralin products would be low. However, the Agency and the SAP are also in
agreement tnat tnis must be verified by properly designed maimialian studies,
not currently available. Tne registrants will be required to conduct studies
to assess tnese potential toxic effects,
c. Spindle Effects
Oi tne basis of publisned studies, the Agency concluded in tne PD 1/2/3 tnat
trifuralin interferes witn tne spindle apparatus in dividing plant arc! animal
cells and would tnus nave tne potential to cause abnormal segregation of
cnranosemes. Elanco 130000/32:16] asserted tnat "trifuralin is not a spindle
poison in mammalian species...." and "... definitive studies snow tnat
trifuralin is very selective in its interaction witn tne microtubular protein
of plant cells and tnat tne ocmpound does not react witn mammalian microtubular
protein." Tne Agency nad expressed concern about trifluralin's potential to
exert a "spindle effect" in mammalian cells because of its reported mode of
action in plants. Elanoo sufcmitted a review article (Hess, 1979a) and tnree
studies (Hess and Bayer, 1974; Hess and Bayer, 1977; and Hess, 1979b) on
trifluralin's mode of action. Cn the basis of these studies, Elanco maintained
tnat "trifluralin does not disrupt cell division processes or any spindle
process when tested in vivo in animal cells."
Tne Agency disagrees witn Elanco's position regarding the lack of trifluralin
induced disturbances in tne formation or tne function of tne cell division
spindle in manraals (Mauer, 1980). Although studies are available delineating
tne probable mode of action of trifluralin in plant cells (consistent witn its
efficacy as a herbicide), insufficient or inconclusive evidence exists to
exclude a oomparable effect on the mitotic spindle {microtubular assembly) in
mammals; in fact, there are animals studies suggesting such an interaction
(Bobinson and Herzog, 1977).
Furthermore, as an evolutionary conservative cell organelle, the apparatus for
cell division does not differ significantly between plants and animals," hence
similar spindle effects mignt be expected to occur in mammals exposed to
trifluralin.
59
-------
Regarding trie first part of Blanco's objection ("spindle poison"), tne Agency
nas .already presented evidence for disruption of tne cellular spindle apparatus
in both (nontarget) plants arid animals (FD 1/2/3). Altnougn the studies were
not well documented (e.g., as to levels of the tlPDA contaminant), demonstration
of trifluralin's ability to interfere with the formation or function of tne
microtubular elements of trie cell division spindle was clearly illustrated at
ratner low "dosages" in three species of plants (at concentrations of 0.1 tnru
1.6 ppb), as well as in two species of salamanders (10~ to 10 rt).
Further, in an extended series of studies in Drosophila, tne continuous
feeding of larvae at levels of 0.01 and 0.10 percent of a trifluralin
preparation oontaining 177 ppra NDFA nas consistently'given evidence of
nondisjunction as discussed in the PD 1/2/3 and subsequently by Murnik (1979).
A repeat study by tne same investigator witn "pure" trifluralin (i.e., no
detectable NES?A) was negative, however. Thus, altnougn tnese limited animal
studies may be considered inconclusive, they do suggest that high
concentrations of trifluralin products (with or without stated levels of tJDPA)
nave tne capacity to disrupt the mitotic spindle in dividing animal cells, and
thereby nave the potential to cause abnormal segregation of chratosanes
(nondisjunction) (.Mauer, 1980).
As to tne second Elanco contention ("selective action on plant microtubular
protein"), three components can be identified: (1) binding of trifluralin to
tubulin protein sub-units, wnicn normally polymerize into microtubules to
constitute the cyto-arcnitectural elements of the cell division spindle (among
other cellular processes); (2) assumed chemically-induced prevention (or
inhibition) of sucn microtubular assembly and/or function; and (3) disruption
of the oell division process and other spindle processes _in vivo in animal
cells (e.g., orientation).
In a series of published studies in plant cells (cotton, wheat, oorn, etc.),
trifluralin has been shown to bind specifically to plant tubulin,. thus
preventing tne appearance of microtubular elements of the cell division spindle
(Hess, 1979b; Hess and Bayer, 1974; Hess and Bayer, 1977; Bartels and Hilton,
1973 and 1974). In contrast, the same investigators reported no binding to pig
brain tubulin jxs vitro nor any inhibition of polymerization into microtubules
in tnese preparations at tne aqueous saturated concentration (1 £ M). Further,
at supersaturated concentrations (1 to 10 u M) trifluralin apparently nad no
effect en tne cell division of normal or transformed (XSFL-2) monolayer cell
cultures of sheep ling, nor on cleavage of fertilized sea urchin eggs (Hess and
Bayer, 1977).
The node of action of trifluralin on tubulin and microtubular assembly has
been compared to that of tne well-known spindle poiscn, colchicine. However,
tnere are differences in tne pharmacological responses between tnese chemicals
wnich nave led sate investigators to suggest essential chemical differences
exist between plant and animal tubulin. For example, at least 1000 times as
much colcnicine as trifluralin was required to completely prevent microtubular
formation in tne alga, Chlamydomonas (2.5 - 3.75 nf-5 vs. 3.75 - 5.0 u M;
comparable no effect concentration levels were 0.25 rati and 0.1 u
60
-------
respectively). Similarly, mitotic disruption in wneat root tip cells and
cultured blood lily endosperm cells occurred witn i.O and 0.3 u M trifluralin,
respectively, canpared to 2.5 and 0.1 mil colchicine (Hess, 1979b). In
contrast, 0.1 u M oolcnicine appears to be sufficient to disrupt cell division
in' He La cells (Hess, 1979b), as well to bind to pig brain tubulin preparations
as per L. Wilson (SAP, October 9-10, 1979), whereas as noted above,
supersaturated aqueous as per concentrations of trifluralin (up to 10 u M)
appear to be inactive in certain animal cell systems. It nas also been
suggested tnat different binding sites for colchicine and trifluralin are
involved wnicn mignt also contribute to the presumed differences between plant
and animal tubulin, and thus tne differential responses -(Hess and Bayer, 1976).
Trifluralin's inability to affect animal microtubules, however, may only be
apparent, since, in the studies cited above, tne chemical was generally tested
only to' tne limit of its aqueous solubility. Other recent animal studies have
reported effects on microtubular formation at aqueous (< 1 u as well as at
supersaturated concentrations (greater than 10 u M). For example, microtubule
oral tend formation was delayed by C.2 u M trifluralin in tne large ciliate,
Stentor coeruleus (Banerjee et al., 1975), wnile polymerization of purified
pig brain tubulin into microtubules was prevented by concentrations of about
10 u M and above (Robinson and Herzog, 1977).
Hence, tne animal studies cited in the Elanoo rebuttal are insufficient and/or
inconclusive to exclude possible interference witn tne mammalian cell division
spindle in vivo (Mauer, 1980). In terms of tne magnitude of any risk from
spindle effect, tne Agency nas concluded the evidence is inadequate, sines, (i)
t* hfly*o a y-Q FV\ t j»a 1 "i 4 al 1 ar\ ftp c-e «i n af 4 /mv*m norm 4 ra 1 t •% "i % 4 V
uiitrLt: die ISJ VaXlCJ in V1VQ lucsIkiacU. Xa.n SuuultrS r SuuaLlC OIT yStluliiai f cuiG ilxj 1l
is not clear wnether it is trifluralin itself, a metabolite, or a contaminant
tnat is active. Tne SAP agreed with tne Agency's proposed position requiring
additional testing to clarify these uncertainties.
d. Spindle Effects Threshold
The SAP catmented, "A threshold (does) exist(s).for spindle effects frcm
compounds such as colchicine." (The inference is that this is true for otner
spindle inhibitors such as trifluralin.)
Tne SAP (Appendix A, November 30, 1979) based its opinion on the kinetics of
the affinity for the same tubulin binding site of colchicine and bencrnyl,
another cnemical capable of interfering with tne cell division spindle, and the
resulting inhibition of polymerization of tubulin to microtubules. The Panel's
opinion was strengthened by studies with manmalian cells it cited (referenced
in the SAP submission as: Cox and Puck, 1969; Wilson, Anderscxi, and Chen, 1976)
using the otner well-known anti-nitotic agents, colcemid and the vinca
alkaloids, as well as by a mouse micronucleus test with benanyl (referenced in
tne SAP submission as Seller, 1977), all suggesting a threshold for spindle
effects.
61
-------
The Agency disagrees with the opinion expressed by the SAP (Mauer, 19805. A
threshold for spindle effects and nondisjunction cannot be unequivocally
demonstrated by the studies cited for eitner tne antimitotic alkaloids, nor. for
benanyl. Tne _in vitro studies (Cox and Rick, 1969; Wilson, Anderson, and Chen,
19%) were corriucted witn ceil lines (CHO, EHB) displaying inconstant
cnroraoscme numbers; unsynenronized cell populations were treated; and no
control cell clones were isolated for" comparison•to the treated isolates. The
dose range between tne concentration of agent needed to block movement of
single chranoscmes and that needed to cause complete mitotic arrest is
extremely narrow (e.g., in tne case of oolcemid, 0.015 ug/ml and 0.03 ug/ml}.
Any reference to a mechanism involving what relative percentage of tubulin must
be bound to account for selective inhibition of polymerization leading to "non-
disjunction" is tenuous. ¦ Only 3 percent of cellular tubulin bound appears to
be critical to block tne entire spindle thus leading to polyploidy. Cellular
tubulin is involved as well in cellular processes other than the mitotic
spindle. Tnere is no suggestion for thresholds involved in these processes.
The in vivo study cited (Seiler, 1977) was actually performed with methyl
benzimidazole carbamate (MHZ), the principle (active) metabolite of benanyl. .
The author reported no effect (no micronuclei) of MBC given orally to mice at a
(single) dose of 50 wq/kq corresponding to a blood level of 8 ug /ml (the
solubility limit of the canpound). Furtner, the inference of a threshold for
spindle effects (microtubular) was drawn by the investigator from ais _in vitro
experiments with isolated pig brain preparations, in which 5 x 10 M
(approximately 10 ug /fnl) was tne minimal effect level (25% inhibition of
tubulin association), while 10 f! (12 ug/ml) gave maximum effect (100%
inhibition), Finally, the micronucleated erythrocytes which are oounted as the
end-effect in tnis test are not exclusively the product of spindle innibition;
micronuclei also occur following treatment with chromosome breakers. No effect
levels one fifth to one tenth those for micronuclei have been reported with
benomyl/MBC for reproductive effects in rats at 7.5 rog/kg, only a portion of
which may be attributable to tne chemical's effects on tubulin. - Hence, the
circulating threshold may conceivably be <8 ug /nl for these effects.
Tnere are no comparable data available on the effects of trifluralin which
could be related to a threshold. Although benanyl (MBC) and colchicine may
nave tne same tubulin binding site, evidence is available that trifluralin does
not snare tne same site (Hess and Bayer, 1976).
B. Garments Relating to Benefits
Elanco [30000/32:#6] alleged that EPA greatly underestimated the economic
impact from the cancellation of Treflan®; that tne economic loss would be
significantly greater if current information relative to the planted acreage
and tne price of the agricultural ccnracdities were used.
62
-------
1. Base Planted Acres
Trie EPA analysis in tne PD 1/2/3 used data wnich represented the most
current information available on acreage at tnat time. Normally tne Agency
uses an average for the most recent tnree to five years; generally, tnis method
is a good indicator of a "typical" production year after a hypothetical
cancellation. To project acreage into tne future is imprecise and is not
normally undertaken in an analysis of this sort (Gaede, 1980a).
The 1974-1976 average number of planted acres was used as the analytical base
for all trifluralin use sites other than cotton. A 1971-1976 average was used
for cotton since the unusually low acreage for 1975 would have biased the three
year average downward excessively.
The Agency has obtained more current data and finds that changes have occurred.
(Gaede, 1980b). The original EPA/USDA estimate of the percentage of planted
soybean acres treated with trifluralin was approximately 10 percentage
points lower tnan more, current 1979 estimates; in addition, the EPA/USDA
estimate of the percentage of planted cotton acres treated with trifluralin was
approximately 10 percentage points higher than more current estimates.
Deviations from tne estimate are expected frcra an analysis that was not
intended to represent any ore specific year. Thus, unless drastic structural
changes occur in U.S. agriculture, the estimate will probably be
representative. Such changes could not be anticipated at the time of the
analysis nor can they now be projected precisely (Gaede, 1980a).
2. Base Cannodity Price
Tne Elanco contention of a downward bias in agricultural commodity prices is
inaccurate. The same logic used in the estimating procedure for base acres was
also used for crzimodity prices. The results of tne analysis were stated in
constant dollars (absent the influences of inflation, etc.) that corresponded
to the years used for acreage data as stated in footnote 17 of tne PD 1/2/3
(Gaede, 1980a).
Tne Agency -concludes that tne Elanco contention of bias associated with both
base Harvested acres and commodity prices is inaccurate. The EPA/USDA economic
assessment was undertaken with tne best available information provided by
experts in relevant professions.¦ This assessment is deemed appropriate as it
now stands.
3. Relative Importance of Prowl® as a Treflan® Alternative
Tne American Cyanamid Company [30000/32:#17], proposed to "...bring the
Agency's attention to errors in the benefit analysis,...". American Cyanamid,
tne registrant for Prowl®, alleged that less than catiplete consideration was
given to alternative products and the EPA and USDA were perhaps not aware that
Prowl53 is an effective alternative for trifluralin. Prowl® arrf other
63
-------
dinitroanilines were considered by SPA and USDA as effective weed control
alternatives for cotton and soybeans. Comparative soybean and cotton yields
for otner dinitroanilines (Prowl®, flucnloralin, dinitramine, and profluralin)
were estimated to provide yields tnat ranged from 95 to 100 percent of tnose
achieved witn trifluralin.
American Cyanamid stated further tnat trie preliminary benefit analysis was in
error arri outdated. It is true that changes in either the herbicide or
agricultural markets would cause tne basis for the analysis to imprecisely
describe current conditions. However, the Agency maintains that the analysis
is still relevant, since minor deviations are normal for an analysis that was
never intended to represent any one specific year. Hie most important issue at
hand is whether the relative economic importance of trifluralin, as represented
in tne August 1978 analysis, is still valid (Gaede, 1980b).
The econonic analysis was presented in terms of a "typical" production year
for a time period of three to five years following a hypothetical cancellation,
talcing into account the long-term uncertainty associated with herbicide
markets. The Agency was cognizant of limits to the production capacity of
other dinitroanilines 2s well as the uncertainty of wnen they would becaite more
market competitive with trifluralin.
American Cyanamid indicated that a new production facility would be at full
capacity in 1980. In a subsequent cannunication American Cyanamid stated that
the production facility is now fully operational (American Cyanamid Company,
1981). This statement implies that trifluralin's market shares.for soybeans
and cotton can currently be challenged. While there is seme increase in
Prowl'-d's market shares -on both cotton and soybeans, the penetration of these
markets is still in the early stages and significant inroads into trifluralin's
market snare have not occurred (Gaede, 1981). Since Prowl® is priced lower
than trifluralin and since the registrant is offering price rebates to farmers,
market penetraticr. strategies are- at least in tne initial stages- Whether
there will be higti levels of price and other forms of market competition in the
future is unknown.
American Cyanamid stated that the Agency's market share estimates of otner
dinitroanilines were in error. Hie 10 percent market share estimated by
USDA/EPA for other dinitroanilines an soybeans was an accurate estimate of use
in 1979 (Gaede, 1980b),* though the market Share improved in 1980, the previous
estimate is probably still reasonable for a "typical" production year.
For tne cotton market, the Agency projected an approximate market share of 8.4
percent, a figure wnicn is below the current situation. The ramifications of
tne low estimate upen the relative economic importance of trifluralin will be
snown to be minor in Section III. B. 4.
American O^anamid [30000/32:#17] also indicated they had pending registrations
at that time for Prowl® on potatoes, beans, peas, dry beans, peanuts, and
sunflowers. Subsequently registrations were granted for use on potatoes
(November 4, 1980), sunflowers (April 9, 1981), and peanuts (April 9, 1981),
64
-------
Regarding potatoes, 72,000 acres were treated with trifluralin in 1978,
amounting to 5.3 percent of the U.S. potato acreage. Benefits were estimated
as $387,000 in 1978. Since 5,3 percent is a fairly small percentage,
trifluralin would be considered a minor use herbicide for this site. For
Prowl®, no data were readily available tq assess the current market share.
Even if substantial market inroads were achieved by Prowl®, tne overall cnange
in the total benefits of trifluralin use would not be of major significance
(Gaede, 1901).
Regarding peanuts and sunflowers, 301,000 and 650,000 acres were respectively
treated with trifluralin in 1978, amounting to 19.6 percent and 65.0 percent of
tne total U.S. peanut and sunflower planted acreage. It would be reasonable to
expect that the Prowl® market penetration for these two sites (peanuts are!
sunflowers) would be low, since the product is newly registered (April 9,
1981). Hence, tne current overall benefits of trifluralin use vould not be
expected to cnange significantly because of tnese two new registrations. Thus,
tnougn Prowl0 nas tiiree newly registered use sites* the early stages of market
development cannot be expected to nave mucn significance upon tne overall
current benefits of trifluralin use (Gaede, 1931).
Hie primary contention of American Cyanamid is tnat tne relative eoonattic
importance of trifluralin has diminished to such a degree that the Agency
snould reconsider its regulatory stance. The following paragraphs will provide
a' qualitative rather than a quantitative assessment of tne relative econanic
importance of trifluralin both in current and future herbicide markets.
4. The Relative Economic Importance of Trifluralin in the Current
Soybean/Cotton Herbicide Market
Soybeans - The EFA/USDA estimate of tne percentages of planted soybean acres
treated witn trifluralin has remained relatively stable. However, the number of
trifluralin acre treatments has increased by more tnan 50 percent relative to
tne August 1, 1978, EPA/USDA "typical" year estimates because of the following
factors:
a. The natter of soybean planted acres has increased sharply from an
average of 52.3 million in the 1974-76 analytical base period to over a
total of 70 million in 1979 and 1980 as well.
b. Tne proportion of soybean acres treated with any herbicide has been
increasing ewer time.
c. Trifluralin's acre treatments of soybeans has expanded since
the EPA/USDA analysis was completed.
For all otner dinitroaniline herbicides used on soybeans, a 10 percent market
snare was estimated by EPA and USDA in 1978. Relative to the 1980 soybean
herbicide market, tnis estimate is somewhat low but appears reasonable for a
"typical" production year approacn (Gaede, 1980b and Gaede, 1981).
65
-------
In absolute terms, trie current economic importance of trifluralin would be
accentuated since tne number of acre treatments has rapidly escalated relative
to tne "typical" production year scenario developed in tne EPA/USDA analysis.
Cotton - For cotton, tne EPA/USEft original estimate of tne percent of
planted acres treated with trifluralin is approximately 10 percentage points
nigner tnan more current estimates. However, the current number of acres
treated witn trifluralin is nearly identical to tne EPA/USDA original estimate
of 3.31 million acres (Gaede, 1980b).
Trifluralin's percentage snare of trie ootton nerbicide market nas remained
relatively stable over tne past few years, wnile the market snare of other
dinitroanilines has exceeded tne 8.4 percent estimated by EPA and USDA.
Trifluralin is still by far tne leading herbicide used an cotton (over 50
percent of tne planted acres) and greatly exceeds tne usage of all otner
dinitroanilines.
The relative market snare and economic importance of trifluralin may have
diminished sligntly when compared to tne "typical" year scenario developed in
the EPA/USDA analysis (Gaede, 1980b and 1981). However, the total current
economic importance of trifluralin to the cotter, industry, as indicated by the
total number of acre treatments, continues to greatly exceed the otner
dinitroanilines, though sane are as efficaceous.
5. The Relative Economic Importance of Trifluralin in tne Future
Soybean/Cotton Herbicide Market
In 1978 tne future uncertainty of herbicide markets limited tne EPA/USDA
economic analysis to a "typical" year within three to five years after a
hypotnetical cancellation. Since future herbicide markets are still nignly
uncertain, forecasting is not highly precise as estimates are extended further
into tne future.
However, it would be reasonable to assume (Gaede, 1980b) that trifluralin's
future market snares and economic importance to soybeans, cotton, and otner
crops will diminish to same extent because of the following factors;
a. Alternative weed controls that are cost-effective substitutes will
eventually enter the market (e.g., tne potential registration of Prowl© for use
cn beans and peas). It is possible that the market entry of viable alternatives
may occur quite rapidly since industry research as well as submissions for
registration of other dinitroanilines may have been temporarily kept in
abeyance until the EPA regulatory stance on nitrosamines was better understood.
66
-------
b. Production capacity £or other dinitroanilines (e.g., Prowl®} will probably
increase, a necessary condition to support a serious market penetration effort ¦
by sucn mecnanisms as competitive pricing and rebates. However, given the
currently nigh interest costs for investment capital and inventories, large
increases in prcducticn capacity and market penetration by ocmpetitors may be
delayed.
c. In tne near future it would also be reasonable to assume that Elanco would
attempt to protect its market snares by developing various market strategies
(e.g., trifluralin price reductions). Sucn actions oould conceivably improve
tne overall economic benefits of herbicide use by providing farmers with lower
cost weed controls, wnicn may in turn contribute to slightly lower food prices
for consumers. It is also possible that Elanco could maintain its market
snares for Treflan® (trifluralin) for an extended period of time because of
such factors as econarties of size in production, patent protection until
September, 1985, and strong repurchase intentions of trifluralin users.
Sharp declines in Treflan's® market shares are hence difficult to imagine in
tne near future (Gaede, 1900b).
In summary, tne joint EPA./USDA- economic analysis completed in August 1978 is
deemed adequate to serve as a benchmark for evaluating herbicide market changes
or other deviations free a "typical" agricultural production year. The
qualitative magnitudes and directions of change in economic impacts of a
trifluralin cancellation that have occurred since tne EPA/USDA analysis was
completed in August 1978 do not justify any change in the Agency's risk/benefit
analysis or in its regulatory stance.
C. Comments delating to Testing Requirements/Regulatory Options
1. Benzimidazole Metabolites of Trifluralin
Tne Agency was concerned about tne exposure of applicators to benzimidazoles
(possible metabolites of trifluralin) arri therefore proposed in the FD 1/2/3
to require tnat the registrants perform mutagenicity testing on tne
ben2imidazole metabolites of trifluralin. Nelson et al., (1977) reported tele-
presence of benzimidazole derivatives in the etnylacetate extract of a
trifluralin/tnairmalian hepatic microsome incubation mixture. The SAP, in its
preliminary public review of tne trifluralin RPAR (SAP, Sept. 20, 1979),
indicated uncertainty as to why the Agency required tnis testing, since the
benzimidazoles nad not been detected in excretion products of animals during
in vivo testing.
Elanco [30000/32:16] also commented tnat studies on the benzimidazole
metabolites would not be relevant since tney had only been' detected in in
vitro microsomal studies and not in metabolism studies conducted with
•trifluralin in maimalian species (goats, rats, dogs). Hence, in Elanco's view,
additional mutagenic studies related to trifluralin metabolites are also
considered unnecessary.
67
-------
Trie Agency has reassessed tne _in vivo and _in vitro studies of trifluralin
metabolism 'Mauer, 1900). A surar.ary of tne vivo studies follows: Golab et
al. (1969) fcund that about 90 percent of trie radioactivity in tne feces and
urine of a C-trifluralin fed goat was concentrated in polar substances
located at tne origin of TLC (thin layer cnrortatography) plates. These polar
materials were not identified. Since no attempt was made to isolate or
identify benzimidazole products in this study, tney could have been present.
Qrcnerson and Anderson (1966) reported that about one to three percent of
radioactivity in tne feces of rats fed radiolabeled trifluralin could be
extracted with methanol. Thin layer chromatography of this metrtanolic extract
revealed that the material was associated with polar areas. Thus it is
possible, tnougn not established, that the polar material found in the in vivo
studies of Qraerscn and Anderson (1966) and Golab et al. (1969) contains the
benzimidazole precursors 2,3-dihydroxy-2-ethyl-7-nitro-l-propyl-5-
(trifluoranethyl) benzimidazoline and/or its dealkylated analog, and 2-etnyl-7-
nitro-l-propyl-5-(trifluDrometnyl) benzimidazole 3-oxide and/or its dcalkylated
analog. However, benzimidazole products were not actively looked for in any of
tne in vivo studies and none were reported. Thus, tne available evidence seems
to indicate that benzimidazole precursors, but not benzimidazole products
themselves, may have been present in trifluralin metabolic products found in
urine and feces of rats and a goat (Mittelman, 1980).
The Agency has also reassessed the pnotodegredation and soil degradation
studies. The pnotodegraaaticn study by Leitis and Crosby (1974) revealed that
several benzimidazole products and precursors were produced after irradiation
of trifluralin in a methanol solution. Soderquist et al. (1975) Sound
benzimidazole products and precursors after irradiating trifluralin vapor in a
photoreactor for to up 12 days. These investigators also collected air samples
for seven days after the application of trifluralin to field soil surfaces.
Ch tne third day they found concentrations of benzimidazole products at the
level of detection, 0.5 ng/m . In contrast, about 13.1 ng/m of-"the
precursors were reported cn the same day. Since benzimidazole products are
pnotodegradation products of these precursors, it is highly unlikely tnat any
detectable level would be found durirg application. Therefore exposure of
applicators to benzimidazole products would likely be very low. Assuming a
concentration of 0.50 ng/m total maximum inhalational exposure to applicators
•would be about 7.5 ng/person/day (assuming a respiratory rate of 1.3 cubic
meters and an eight hour work day).
Therefore, as indicated in tne FD 1/2/3 and discussed above, degradation
products of trifluralin including seme substituted benzimidazoles have been
reported to occur in treated soil, as well as under ultra-violet decomposition
conditions. There have also been reports of their in vitro occurrence as
metabolites of treated rumen microbial cultures (Golab et al., 1969), and as
indicated, in liver microsomal preparations from normal and phenobarbital-
induced rats (Nelson et al., 1977).
68
-------
However, vmen tested in the standard bacterial mutagenicity (£mes) assays, the
trifluralin metabolites, including trie benzimidazoles previously identified in
trie in vitro microsomal assay, have proved to be negative in a preliminary
report (Kelson, 1978); such tests have previously detected other benzimidazole^
(Seller, 1972), (Cne measurement in the Nelson, 1978, study was positive, but
in a telephone conversation with Dr. tfelson in September, 1980, ne indicated it
•was an erroneous result tnat was not reproducible in subsequent testing.)
In sunnary, compared to their unequivocal "identification in environmental fate
and _in vitro studies, the presence of benzimidazole derivatives in mammalian
metabolism studies is uncertain. In addition to the major metabolic pathways
amply demonstrated for trifluralin in tne martmalian studies cited in tne
ED 1/2/3, that is, by dealkylaticn and reduction, a large number of polar
derivatives with chemical characteristics ccraparable to benzimidazoles were
formed, but not identified (Ehmerscn et al., 19665. Each one of these
represented a minute quantity, and tne totality probably represent a minor
pathway in the species studied. 'However, since seme of these metabolites could
represent the same benzimidazole products previously identified in vitro, the
Agency nas raised tne issue.
Cti tne other hand, unless more definitive, compelling evidence becomes
available, tne Agency is not prepared to pursue the requirement for additional
testing in this regard (Mauer, 1980). Therefore registrants will not be
required to perform mutagenicity testing on the benzimidazole metabolites of
trifluralin at this time. Tnis represents a change in the Agency's proposed
testing requirements outlined in the PD 1/2/3.
2. Reproduction and Teratology
In the PD 1/2/3 the Agency identified areas needing further testing based on an
evaluation of reproduction and teratology studies previously submitted by
Elanco. Wnile admitting tne shortcomings'of experimental design" wnicn led
Agency reviewers to judge the studies unacceptable for regulatory requirements,
Elanco [30000/32:#6] contends tne studies snowed trifluralin had no adverse
effects cn developmental or reproductive processes of animals wnen administered
at large daily dosages (up to 2000 ppra in tne reproduction study; up to
1000 mg/kg in tne teratology study), and suggests additional testing may not
be necessary.
In its review of tne trifluralin RFAR, tne SAP was concerned tnat the
reproduction test did not satisfy tne testing requirement wnicn was part of its
reccmttendod battery to estimate potential risks to man from spindle effects.
Hence, the Panel recommended a multigeneration reproductive study modified to
include a dominant lethal test.
A newer rcproBuction study (Elanco, 1977) was submitted to the Agency with the
intended purpose of answering seme of tnese questions posed by the SAP in its
review of tne trifluralin PD 1/2/3. Frcra the findings reported in tnis study
69
-------
Bianco suggests that tre lack of impairment or modification of reproductive
performance in tne rat following administration of high dietary levels of
technical trifluralin (up to 0.2 percent = 2000 ppm) "...further supports the
position that trifluralin does not affect spindle formation in a mammalian
system in wruch there are rapidly developing tissues."
The "newer" study submitted by Elanco is actually a modified one generation rat
teratology study performed in 1975, and was designed to satisfy special
criteria for Japanese regulatory requirements. Since it is rot a multi-
gene ration reproduction study, it" does not fulfill the SAP's reccranendation
(Mauer, 1980). Furthermore, it would not fulfill, tne sensitivity criterion
(power of test) the Agency believes is required for assessing spindle effects,
even if modified by inclusion of a dominant lethal test. Finally, at no time
were cytogenetic studies performed, specifically, analysis of fetal or maternal
tissues for numerical chrorrosomal aberrations.
The type of testing (a multigeneration reproductive study) which would fulfill
the Agency's requirement and whicn was recommended in the PD 1/2/3 is outlined
in the Federal Register published on August 22, 1978 (43 FR 37336).
3. Mutagenicity, Including Heritable Spindle Effects
Tne Agency proposed in the H3 1/2/3 that registrants would need to perform
additional testing of trifluralin for mutagenicity, including heritable spindle
effects. Garments included tne following.
a. EN A/Gene Effects
Elanco [30000/32:16] commented that ample mutagenicity studies have already
bean conducted showing trifluralin is negative for any mutagenic effect, and is
not a spindle poison.
Elanco asserts that tne Agency's requirement for additional mutagenicity
studies is unneccessary since sufficient studies have already been performed,
all of whicn have snown negative results, and hence demonstrate that
trifluralin (toes not interact with genetic material to induce mutagenic effects.
Ihe Agency disagrees witn Elanco. Although tne Agency has judged that the
risks of direct DUA/gene (and chranosanal) effects are low, and mainly
associated with tne nitrosodipropylarame (NDPA) contaminant of trifluralin
products, further tests are required including studies to assess transport to
the manmalian gonad (Mauer, 1980). In its review of the Agency's RPAR on
Treflan® (SAP, October 9-10, 1979) tne SAP agreed with the Agency's
position, and suggested these low risks be "verified through properly designed
studies en tne prcduct as it is currently produced witn levels of less than
1 ppm NDPA." (Appendix A, October 15, 1979).
70
-------
As reviewed in the PP 1/2/3, both the tecnriical grade of trifluralin and
various formulations [Treflan®, containing know, (up to 177 ppm) or unspecified
amounts of tne principal contaminant, NDPA] nave been reported" as negative for
direct EtJA/gene-cnrcrroscme effects in studies involving: bacteria (gene
mutation and D"1A damage tests); fungi (mitotic recombination); Drosopnila (gene
mutation and chromosome breakage);¦and roaixialian cells in culture (unscheduled
syntnesis). Subsequent microbial studies performed in an Agency
laboratory, however, nave revealed positive results in tne Pmes Assay as well
as in the yeast mitotic recombination test, but at concentrations 2.5 to 9.5
times tne highest concentration in previously reported studies (Chen, 1979).
Two preparations were tested in the Agency study; . A technical grade of
trifluralin containing less tnan 3.0 ppm NDPA; and an NDPA-bolstered technical
preparation, to wnicn 100 ppm of tne contaminant had been deliberately added.
At tne dosages used (2400 to %00 micrograms) botn tne technical grade chemical
and the nitroso-contammated preparation induced'framesnift mutations in
Salmonella (7sr.es) strains, but only in tne absence of metabolic-activation.
Wnen tested in tne presence of a metabolic activating system, only tne
trifluralin preparation with added HDPA was active in tne yeast test.
In contrast, I3DPA tested by itself has consistently snown positive genetic
activity in most of tne same _in vitro tests which nave been performed with
trifluralin preparations, but only in the presence of metabolic activation,
fence tne Agency had concluded tnat the NDPA contaminant of trifluralin
preparations represented a potential risk of direct DNA/gene effects to humans
(Hauer, I960). However, the risk was considered low because of the negative
results of trifluralin preparations in tne majority of the tests conducted to
date, as well as tne low potential for human exposure to tnese products.
On tne other hand, information on maimalian in vivo testing is lacking and tne
Agency recorraended in tne PD 1/2/3 tnat further testing for botn gene mutation
and cnrcmosomal breakage be performed in order to better evaluate risk for
tnese genetic effects in mammalian systems. The Agency,also recommended
studies to -assess the potential for trifluralin/NDPA to reach tne mammalian
germinal tissue in a metabolically active form (Mauer, 1980).
In a continuation of tne review of tne trifluralin RPAR, tne SAP has reiterated
its Hay, 1978, recommendations for a "core" battery of tests selected frcm tne
mutagenicity guidelines proposed by the Agency (43 FR 37336, August 22, 1970)
for tne mutagenicity screening of all pesticides (Appendix A, November 30,
1979). Tne SAP recommended tnat tne test battery consist of an Ames Assay,
(with a dose-response course, if possible), a point mutation assay in mammalian
cell cultures (using mouse or hamster cells), and an in vivo (manmalian)
cytogenetics assay. In addition, the SAP reccnmended modification of the multi-
generation reproduction study to include a daninant lethal test (wnich would
also be required for eacn pesticide, in addition to the core battery).
Finally, it recommended an oncogenicity study in tne overall evaluation of
mutagenicity.
71
-------
Using tnis testing scheme, tne Panel concluded that a suitable microbiological
test has been performed for trifluralin (Appendix A, November 30, 1979).
•bvevcr, sines adequate information frcm mammalian tests was lacking, the Panel
recommended: a gene mutation test in marmalian cells (using Chinese namster
ovary, mouse lympncma cells, or"another validated mammalian cell system), and
a cytogenetics assay in mice or rats, examining bone marrow cells, -
spermatogonia and lymphocytes. If botn these assays are negative, the SAP
would be satisfied that no significant mutagenic risk is posed by trifluralin,
and no further testing should be required.
Thus, tne SAP and the Agency are in basic agreement with respect to tne
requirement for additional testing to assess direct DNA/gene and chromosomal
effects in mamr.alian test systems. However, the Agency differs with the SAP on
several minor counts (Mauer, 1980).
Fin;t, the Agency is not entirely convinced about the adequacy of the
microbial testing of formulated trifluralin preparations. As detailed in the
PD 1/2/3, results of tests with the formulations were considered inconclusive
sinoe these tests were performed without metabolic activation. In addition, as
detailed above, subsequent tests in an Agency laboratory have generated
positive results (albeit "weak," in terms of the higher dosages used) in
comparable microbial assays using the tecnnical cnemical. It is essential that
this apparent contradication be resolved by further testing in microbial
systems.
Second, wnile the Agency is satisfied with data indicating that both
trifluralin with nc detectable NDPA (<1.0 ppm) as well as a Treflan®
formulation reported to contain 177 ppm of tne contaminant did not induce sex-
linked recessive lethal mutations or direct chromosomal damage (dominant
letnals, breakage) in Drosophila testing, no comparable germinal tests in
mammals are available. Hence tne Agency requires performance of an adequate
dominant lethal test ("adequate" in the sense of an appropriate number of
experimental units per dosage group) in at least one species (mouse or rat)
according to standard protocols [43 FR 37336], in addition to the in vitro and
satiatic mammalian tests. Tnis should be separate fran a modified multi-
generation study, as tne SAP recommended.
Lastly, tne Agency cannot agree with tne SAP's minimal standard which
suggests that if their recommended "core" battery is negative "no significant
mutagenic risk is posed by trifluralin and no further tests should be required."
It is the Agency's opinion that appropriate evaluation of trifluralin/NDPA
products for potential mutagenic risk to man requires at least two elements
(Mauer, 1980): (i) assessment for the presence of active compound in the
mammalian gonad [43 FB 37336, Sec, 163.85-1]; and (ii) results frcm appropriate
germinal testing [43 ER 37336, Sec. 163,84-1j.
72
-------
b. Spindle Effects Testing
Trie SAP commented that: "Tne best methods for predicting adverse nealtn
effects in man frcm spindle poisons are tne multigeneration reproductive test
and the dominant letrial test."
In its written review of tne trifluralin PD 1/2/3, .trie SAP agreed witn the
Agency's position that additional testing for spindle effects be required,
especially in mammalian systems (Appendix A, October 15, 1979). The SAP
subsequently recommended a sequence of tests wnicn tne Panel believes will be
capable of estimating risk to man of spindle effects (Appendix A, November 30,
1979).
Specifically, tne Panel recanmended tne following testing scheme, consistent
with its previous recommendations on tne proposed mutagenicity guidelines
(43 FR 37336): a multigeneration reproductive test, modified to provide a
statistically significant assay for dominant letnal mutations; if the
multigeneration and dominant letnal tests are negative, tnen a cytogenetics
assay should be conducted on the sperm and bone marrow cells of an adequate
number of tne tested animals used in these or other studies. In the Panel's
opinion negative results in these tests would indicate tnat "no significant
risk to man exists frcm spindle innibition by trifluralin in normal
agricultural use."
In addition to in vivo tests, tne Panel also suggested research funding for the
development of in vitro assays for spindle effects of pesticides, contaminants
and breakdown products (such as an examination of metapnase arrest in mammalian
cells), as well as approacnes to evaluate the predictability of _in vitro
tubulin binding of spindle poisons.
Using tnis testing scheme, the Panel concluded that trifluralin had been
inadequately tested for potential spindle effects.
The Agency differs with tne SAP in tnat tne Agency does not consider the in
vivo tests recommended by tne Panel (multigeneration reproductive, doninant
letnal and cytogenetics) sufficiently sensitive to adequately assess risk frcm
spindle innibitors (Mauer, 1980). The Agency does agree, however, witn tne
desirability for trie development of sensitive _in .vivo assays for spindle
effects, as well as the funding of 2T650c*irct"i on tne predictive value of tubulin
binding. 'Tne Agency cannot suggest at tnis time an alternative set of tests
sinoe none of tne assay systems currently proposed for use in screening
programs can unequivocally detect chemicals producing spindle effects
significantly relevant to man, such as aneuploidy by nondisjunction or any
otner mecnanisms. Tne following is a portion of a summary statement by OPP
staff scientists delineating tne position of the Agency in tnis regard:
(Hill, 1980)
73
-------
"The Scientific Advisory Panel (SAP) sufcmitted to trie Office of Pesticide
Programs (OPP) its evaluation of chemicals which may interfere witn tne
cell-division spindle (benanyl, tniophanate-methyl and trifluralin) on
November 30, 1979 (Appendix A). Tne SAP is willing to accept data from a
few test systems which are presently available"(reproduction and dominant
letnal tests plus in vivo cytogenetics) to evaluate safety in regard to
spindle effects. Using this scneme the Panel concluded that more testing
would be required for trifluralin and tnat safety had been demonstrated for
benanyl and thiophanate-methyl.
"Co tne one hand, C3PP agrees with tne Panel that nothing in the review of
all tne information available on these chemicals# including subchronic
toxicity and reproduction studies, indicated significant adverse effects
frart -this toxicity. Ch the other hand, OPP does not think these systems
nave tne capability to assess fully the degree of hazard from spindle
inhibition.
"Basically, QPP staff is of the opinion that there is a lack of information
bearing cn the sensitivity ard adequacy of various tests to evaluate risks
frail spindle innibition and, more generally, inhibition of microtubular
protein polymerization."
"In addition to in vivo tests, tne Panel also suggested research funding
for the development"of" in vitro assays for spindle effects of pesticides,
contaminants and breakdown products (such as examining metaphase arrest in
maranalian cells), as well as approaches to evaluate the predictability of
in vitro tubulin binding of spindle poisons."
"The OPP staff agrees with the SAP as to the need for further investigation
on tne usefulness of various systems for evaluating risks. To this end,
tne .Agency plans to identify outside scientists wno can nelp,-us delineate a
meaningful research program to fill scne of the information gaps and to
sunmarize possible means of risk assessment.
"The problem of assessing risks is heightened when one considers on a
broader scale tne developnent of scnemes for chemicals which inhibit
microtubular polymerization. In this case, one is concerned not only with
spindle effects but also all other effects produced by chemicals acting via
tne nolecular mecnanism of microtubular innibition. Since these inhibitors
affect all microtubule-related processes, including cell shape, cell
movement, intracellular molecular movement, and cellular secretion, many
functions may be affected by tnese 'cellular poisons.' Given these
effects, we must decide what are tne appropriate toxioological endpoints of
concern, now to evaluate then and, lastly, how to evaluate risks. Because
of tne magnitude of tne problem we have decided to start with spindle
effects and to enlarge tne scope of problems over time."
74
-------
Tnerefore, the Agency will not require furtrier tests to assess risks frcm
spindle effects at tnis time (Mauer, 1980). When appropriate testing is
developed, the registrants will be required to perform them to assess effects
due to inhibition of spindle fiber formation or function.
4. Labeling Requirements
In the memorandum of October 15, 1979 (Appendix A), the SAP reccranended that
tne Agency adopt Option 3 as detailed in the FD 1/2/3. This option called for
tne cancellation of all registrations for products containing trifluralin
unless registrants modified labeling of their products to reflect less than
1 ppm N-nitroso-dipropylamine {NDPA) contamination.
Elanoo has stated that "the proposal to require registrants of trifluralin
herbicides to include in tne label a statement concerning the level of the NDPA
contaminant will not serve to provide meaningful information to users, and
therefore snould rot be adopted. If, however, it is adopted, Elanco urges EPA
to postpone implementation until EPA establisnes a general policy on
nitrosamines in pesticides. Tne nitrosamine issue is much broader than NDPA in
trifluralin. EPA should delay action until development of a policy that would
apply uniformly to all registrants of pesticides containing nitrosamines...."
[30000/32:#6].
Tne Agency has reconsidered its proposal to require a statement of the level of
NDPA contamination in trifluralin on the product label and has determined that
any requirements of this type vail riot be proposed until Agency policy on tnis
subject is finalized. The Agency therefore withdraws its requirement for
trifluralin product label statements on tne level of NDPA contamination at this
time.
Notwithstanding tne foregoing, the Agency vail continue to require that the
confidential statement of formula for trifluralin- containing products be
amended to include tne level of N-nitroso contamination under the inert
ingredients statement.
The Agency initially proposed that this statement should read as follows:
N-nitroso~di-n-propylamine (NDPA)....<1 ppm.
Elanco has provided information on total N-nitrosamine content in ttdxtc tJhsn
600 lots of Treflan® EC; the total N-nitrosamine contamination ranged frcm
<0.01 ppm to 0.96 ppm (Elanco, 1980b).
In view of tne risk and exposure analyses discussed in this document which show
if the total N-nitrosamine content is kept below 1 ppm, the risk fran e_xposurc
to the nitrosamine contamination in trifluralin products is negligible, the
Agency has determined that in order to assure the public is protected frcm
exposure to nitrosamines, technical trifluralin products may contain no greater
than 0.5 ppm total N-nitrosamine contamination.
75
-------
Trie Agency will require that the statement in the Confidential Statement of
Formula for all technical trifluralin products to be placed under the inert
ingredients section shall read as follows:
o
Ibtal N-nitrosamine contamination .... no greater than 0.5 ppm.
During the formulation of end-use products, sane formulators heat the
crystalline technical trifluralin to 70°C over several hours in order to
liquify it. Because it has been shown by Probst (1981) that NDPA can be
generated during heating at 70°C, the limit of total N-nitrosamine
contamination in formulated products will be set as a function of the amount of
trifluralin in the end-use product with allowance for seme nitrosamine
generation.
For example, assuming a 0.5 ppm limit in technical product, a 50% formulated
product, with no generation of nitrosamines, v.ould be expected to have no
greater than 0.25 ppm total N-nitrosamine contamination (0.5 * 2). However,
the Agency will allow for a two-fold increase in nitrosamines. Therefore, in a
50% formulated product, the allowable limit would be 0.5 ppm (0.5 ppm in
technical trifluralin x 50% trifluralin in the formulated product x 2 to allow
for possible nitrosamine generation).
Similarly, a 5% formulated product might be expected, wtih no nitrosamine
generation, to nave no more than 0.25 ppm total N-nitrosamine contamination
(0.5 r 20). However, to allow for twice what would be found in a straight
forward dilution, tnat allowable limit would be 0.05 ppm (0.5 ppm x 5% x 2).
All allowable limits for total N-nitrosamine contamination should be calculated
in this same way; i.e., by multiplying 0.5 ppm by the % trifluralin in the
formulated product and tnen multiplying by 2 to allow for possible generation
of more nitrosamines during- formulation.
Elanco [30000/32:#6] stated tnat "Once the EPA has establisned a limit on
tne level of NDPA contaminant in herbicides containing trifluralin, the
requirement must uniformly apply to all registrants of such products." Elanco
further stated:
"lYie level of NDPA impurity tnat EPA proposes to establish for trifluralin
must apply to all trifluralin registrants and products. There is no
justification for EPA to apply a more lenient standard for a producer less
capable than the current registrant-producer. The position doc orient
suggests, however, that higner NDPA contamination levels may be permitted
for new trifluralin pesticide products for wnich a level of 1 ppm or less
'cannot be met without substantially increased'costs.
76
-------
"Establishment of a nigner ND?A limit for a new pesticide on this basis
would be improper and unjustified. First, Elanco incurred extraordinary
costs in acnieving consistent NDPA levels of below 1 ppm in trifluralin and
it vould be arbitrary and unreasonable for EPA to allow other producers to
avoid similar costs to acnieve similar levels.
"Second, it is ;/ell established tnat the benefits to be taken into account
in determining whether a pesticide causes unreasonable adverse effects on
the environment include adverse economic implications on agricultural
producers and consumers, but do not include production cost impacts on
pesticide producers alone. [See FIFRA Section 2(bb)? EDF v. EPA, 489 F.2d
1247, 1254 (D.C. Cir. 1973)].
"Finally, in view of tne substantial benefits of trifluralin and the
.conclusion of EPA tnat a NDPA impurity level of less than 1 ppn is
necessary to bring trifluralin into risk-benefit balance, it is
inconceivable that a producer of a proposed new pesticide with no
established benefits could snow tnat a higher NDPA limit would be
appropriate for its pesticide."
Although tne Agency does not agree with all of tne points in tne Elanco
argument, it does agree with the Elanco contention tnat the standard for
maximum U-nitrcsamine rontamination in all trifluralin containing products
srould be the same. As indicated in its ED 1/2/3, the Agency determined that
use of trifluralin contaminated witn greater than 1 ppm total N-nitrosamine
would be unreasonable because the risks would be increased unnecessarily witn
no offsetting increase in benefits. Therefore, any future manufacturer of
trifluralin containing products will have to acnieve a level of no greater than
0.5 ppm of N-nitrosamine contamination in tneir technical trifluralin products
in order too be registered. For formulated products, the level all
manufacturers will nave to acnieve will be calculated on a percentage basis
with a built-in factor of two as described above.
77
-------
IV, Conclusions and Requirements
A comparison of trie risks due to exposure to TIDPA as set out in tne PD 1/2/3
witn tne maximum risks due to exposure to trifiuralin, NDPA, and C7/Cg
nitrosamines calculated in tnis document shows that tne risks to the general
population and workers exposed to Treflan® nave not changed significantly. Tne
total dietary risk nas • increased cane order of magnitude from 3.3 x 10" to
5.3 x 10" , Tne risk to mixer/applicator/loaders has remained essentially
tne same and is in tne order of 1 x 10~ . Tne risk -to reentry workers was in
tne order of 1 x 10 due to exposure to NDPA (without taking
pnotodegradation into account) and now is in tne order of 1 x 10 due to ¦
exposure to trifiuralin. Thus, the overall risks associated with exposure to
Treflan® have not cnanged appreciably. Tne source of the risk has cnanged,
but the degree of the risk has not.
Similarly, tne benefits derived from the use of trifluralin-containing products
have remained relatively constant. At the time of the FD 1/2/3, the economic
impact, if trifluralin-containing products had been cancelled, would have been
a $300 million loss. There are alternatives for seme trifiuralin uses as there
were then, but trifiuralin is still the major herbicide in use for cotton and
soybeans, tne two most prevalent uses. The qualitative changes in economic
impact that nave occurred since the EFA/USDA analysis was completed in August,
1978, do not justify any change in the estimate of benefits of trifiuralin use.
Therefore, since there is not significant change in risk estimates nor any
significant cnange in benefits,the position presented in the ro 1/2/3 remains
uncnanged. Based on tne best available information, the risks of using
trifluralin-containing products are outweighed by the benefits of its use if
tne registrants meet certain requirements proposed by the Agency.
After reviewing cortments from the Secretary of Agriculture, the Scientific
Advisory Panel, and otners wno commented on the Agency's findings and
recommendations concerning trifiuralin as set forth in Fbsition Doc orient 1/2/3,
tne Agency has decided to implement the proposed regulatory action (Option 3)
as set forth on pages 137-140 of the PD 1/2/3 with a few modifications.
A comparison of Cption 3 and the regulatory action proposed by this document
(FD 4) can be seen in Table 18.
The regulatory action wnicn tne Agency will implement is as follows:
The Agency will issue a section 6(b)Ll) notice of intent to cancel all
tecnnical trifiuralin registrations -J unless registrants amend the terms
and conditions of registration to limit the total N-nitrosamine content in
technical trifiuralin products to a level not to exceed 0.5 ppm.
6/ Sane registrants have received state registrations for trifiuralin to
meet special local needs under the authority of 24(c) of FIFRA. These
registrations are federal registrations governed by FIFRA, They are subject to
tne 6(b)(1) notice of intent to conditionally cancel all trifiuralin
registrations.
78
-------
Table 18
A Comparison of Proposed Regulatory Actions
Proposal
1. Limit N-nltrosamine
contamination
PD 1/2/3
NDPA < 1 ppm
2. Labeling modification
3. Quality control
of N-ni trosainine
contamination
4. Mutagenicity
testing
5. Testing for
spindle inhibition
6. Reproductive
effects testing
Required
Required
Required, including
benziinidazole
Required
Required
PD 4
•total N-nitrosamine in the technical
product, no greater than 0.5 ppn.
"total N-nitrosamine in formulated
products to be calculated on a
percentage basis including a
multiplication factor of 2 to allow fo
any generation of nitrosamines during
formulation.
Require only modification of
Confidential Statement of Formula
Required
Required, excluding
benz imida zoles
Required, but appropriate
protocols not yet defined
Required
7. Teratology testing
Required
Required
8. Field monitoring study Not Required
for ecological effects
Required
-------
Trie Agency will issue a section 6 (B){1) notice of intent to cancel all
formulated trifluralin-containing registrations— unless registrants amend
the terns and conditions of registration to limit the total N-nitrosamine
content in formulated products to a level not to exceed tnat level which has
been calculated on a percentage basis, assuming an upper unit of 0.5 ppn in the
technical trifluralin as starting material, including a multiplication factor
of 2 to allow for any generation of nitrosamines during formulation.—
In regard to those who are currently operating under State-approved trifluralin
registration^ with Federal trifluralin registration applications pending final
EPA. decisiorw and those who may have submitted applications, for new
trifluralin registrations, their applications will be denied unless those
applications are amended to satisfy the above requirement. Once these
applications are amended to ccmply with this requirement, the registrations
will continue to be reviewed by the J*gency to assess whetheir all registration
requirements, have been satisfied.
The registrants will be required to cotiply with tnis regulatory action 30 days
after notification.
A. Amendment to the Confidential Statement of Formula
Tne Agency will not require that product labeling be modified at this time.
However, tne amendment to tne confidential statement of formula shall appear in
that statement under the inert ingredients statement and for technical
trifluralin registrations snail read as follows:
Total N-nitrosamine contamination... no greater than 0.5 ppm.
For formulated products containing trifluralin, the statement snail read as
follows:
Tbtal N-nitrosamine contamination....no greater than (number to be inserted
must be calculated as follows: 0.5 ppm total N-nitrosamine allowed in
technical trifluralin x XI trifluralin in formulated product x 2 to allow
for possible nitrosamine generation)
These limits are consistent with the FD .1/2/3 where it was determined tnat the
benefits of using trifluralin-containing products outweighed the risks if the
level of HDPA was kept below 1 ppm.
7/ Once registrants have amended tne terms and conditions of registration
to canply with the maximum total N-nitrosamine contamination requirement, it
will be unlawful (under section 12(a)(1)(C) and (E) of FIFRA) to sell or
distribute trifluralin products wnose registrations nave been so amended if
tney are contaminated with total N-nitrosamine at those levels specified above.
8/ under 40 CFR 162.17, all state-registered pesticide products, unless
governed by section 24(c) of FIFRA, must be registered under FIFRA. Pending a
final EPA registration decision, however, state registrants may continue to
sell or distribute the pesticide product if solely within intrastate
commerce.-
80
-------
In order to maintain consistency in the area of regulation of nitrosamine
contamination in trifluralin products, the Agency has determined that all
manufacturers of products containing trifluralin will be required to acnieve
these levels for tneir tecnnical and formulated products in order to fee
registered.
Trifluralin registrants will be required to certify tnat tnis level is an upper
limit in accordance with section 163.61-6 of tne Guidelines for Registration of
Bssticides in tne United States, Subpart D, Chemistry Requirements, as proposed
on July 10, 1978 [43 ER 29709-29710]. A person wro distributes, sells, offers
for sale, Holds for sale, snips, delivers for shipment, or receives and (having
so received) delivers or offers to deliver a pesticide product, the chemical
composition of wnich differs frcra tne amended cnemical.composition statement,
will be in violation of FIFRA section 12(a)(1)(C) and subject to sanctions
under section 13 and 14 of FIFRA.
The registrants of trifluralin roust also advise the Agency of tne quality
control procedures tney will institute to assure the Agency tnat the level of
total tl-nitrcsanine stated in the confidential statement of formula is not
exceeded. In addition, registrants must maintain accurate quality control
records on tnese products and make such records available to the Agency on
demand.
Tnese requirements apply to all current registrants of trifluralin-contaming
products and any present or future applicants for registration of trifluralin-
containing products.
B. Testing Requirements
Tne Agency considers the data on tne mutagenic potential (including UNA, gene
and chromoscmal effects) of trifluralin to be inadequate for a precise
determination of risk. Additionally, the Agency considers the data on
reproduction, teratology, arri ecological effects to be inadequate.
FIFRA Section 3(c)(2)(3) states, in part, tnat:
"(i) If the Administrator determines tnat additional data are required to
maintain in e-f F&ct. an pyi <^t" ina nf" p* rx^t"i pinp ,
*2*1 I «ik it * ¦ J*. Ap W • Vr/ii JL W ,£¦> Jh> £3 W «Ja>^wr • 4 Jtm G W dLr ¦ £¦» f W » .V—¦
Administrator shall notify all existing registrants of the pesticide to
wnicn tne determination relates and provide a list of such registrants to
any interested person."
1. Mutagenicity Testing Requirements
Though tne Agency will not require testing on benzimidazoles, registrants of
fwv'vH 11 cs es 4 4 ft/**! ¥"v 4 "1 i tva 1 4 n 1 1 vtvii i t foe# ~"**¦4^11 iv a 1 i n sst"!/""
couict.s ¦Lnxn'j v.xrx xl jluir 2jlxn v»ii,x dg to v*sst»» wxrxx.ai.ixr5xxn one
provide the Agency with other data concerning the mutagenic potential of tnis
compound. Tne registrants will be required to perform further microbialtests
and a gene mutation test in manmalian cell cultures,as well as a dominant
letnal test, an in vivo cytogeneticstest and/or a micronucleus test. Wnen tne
81
-------
Agency identifies,witn the help of outside scientists, appropriate tests to
assess risks from spindle inhibitors, the registrants willbe required to
perform tnese for trifluralin also. In addition, appropriate evaluation of
trifluralin products with iJDPA for potential mutagenic risk to man requires an
assessment far tne presence of active compound in the mammalian gonad as well
as appropriate germinal testing.
2. Other Testing Requirements
Tne Agency will also require the registrants to perform a new multigeneration
reproduction test, as well as teratology tests to satisfy data gaps existing
for those criteria.
Finally, the Agency has determined that a field monitoring study is necessary
to assess possible adverse effects to aquatic organises -in the area of
TrefIan®-treated fields.
A Notice will be sent, subsequent to the issuance of this position document,
to the registrants of trifluralin-containing products informing tnem of the
rpnn i r^p-Tru^nt" "i nn fhficp fpcfc anH Hpcirri hi nn in f-hp
protocols they must follow and the actions they must take to canply with this
itotice.
Tne Agency would like to note tnat there had been a moratorium on the
publication of all decision documents involving pronouncements on products
containing nitrosamine contaminants. The figency lifted tne moratorium for this
ED 4 because it was unlikely that any changes would be made with respect to tne
basic decision on trifluralin. Removal of TrefIan® products from the
moratorium was necessary to expedite publication of this PD 4 to conclude the
issue and to remove any impediment the moratorium may have had on the ccmpany1 s
marketing and development program.
82
-------
Bibliography of Comments
American Cyanaraid Canpany, from Garbariro, J., April 2, 1900.,
[30000/321 # 17].
Aves, A., November 21, 1979, American Soybean Association,
130000/32: #16].
Baldi, A., November 15, 1979, Ace to Agriculture Chemicals Corp.,
[30000/32: #13]
Burr, J., November 6, 1979, Extension. Service, Oregon State Ihiversity,
(30000/32: #10].
Dart, E., November 5, 1979, Moses lake, Washington, 130000/32: #11].
Davis, D. E., October 25, 1979, Auburn University, [30000/32: #14].
Elanco, November 12, 1979, Division of Eli Lilly and Canpany,
fror, G.W. Probst [30000/32: #6].
Flaitm, B. R., Letter, October 18, 1979, U.S. Dept. of Agriculture.
Fowler, H. W., Jr., Memorandum, October 15, 1979, FIFBA Scientific Advisory
Panel (SAP).
Fowler, II. W., Jr., Memorandum, November 30, 1979, FIFBA Scientific Advisory
Panel (SAP).
Jennings, V. M., Ctetober 31, 1979, Cooperative Extension Service, Iowa State
University, [30000/32:.. #14].
Kemper., H., September 27, 1979, Cooperative Extension, University of
California, [30000/32: #1].
Knake, E., November 5, 1979, lYie Intersociety Consortium for Plant
Protecticn, [30000/32: #12].
Lange, A. H., November 5, 1979, Cooperative Extension, University of
California, [30000/32: #7].
Leggett, H., December 13, 1979, National Cotton Council of America,-
[30000/32: #15],
Merkle, G., October 16, 1979, Tfexas A & M University, [30000/32: #2].
Nalewaja, J. D., October 31, 1979, North Central Weed Control Conference,
Inc., {30000/32: *3],
\
-------
Stanger, C., November 6, 1979, Oregon State University, [30000/32: #0],
Teramura, K., November 6, 1979, .Malheur County Onion Growers Association,
[30000/32; #9].
a
Upcnurch R.P., October 31, 1979, University of Arizona, [30000/32: -5].
-------
References
American Cyanamid Canpany , Letter from J.J Garbarino to Marcia Williams
(SPRD), dated May 23, 1901.
Banerjee, S, J.K. Kelleher, and L. Marquilis, 1975. Trie Herbicide
Trifluralin is Active Against Microtubule-based Oral Morphogenesis in
Stentor Coeruleus, Cytobios, 12:171-178.
Bartels, P.G. and J.L. Hilton, 1973. Comparison of Trifluralin, Oryzalin,
Pronanide, Propham, and Colchicine Treatments on Microtubules,
Pesticide Biochemistry and Physiology 3:462.
Bartels, P.G. and J.L. Hilton, 1974. Comparison of Herbicides and
Colcnicine Treatments on Microtubules. Abstracts Weed Sci. Soc. Amer.
#205.
Barton, A., 1901 Risk Assessment for Treflan. Memo to Carol Langley,
SPRD dated August 3, 1981.
14
Berard, D.E., 1977. Translocation and Metabolism of C N-
riitrosodipropylamine in Soybean Plants. Eli Lilly and Co., unpublisned
data, [proprietary!
Berard, D.F. and D.P. Rainey, 1977. Absorption of 14C-M-nitroso-
dipropylaraine by Soybean Plants. Agricultural Biocnemistry Lilly
Researcn Laboratories, Division of Eli Lilly and Co., Greenfield,
In., 46140.
Binkert, F., and W. Schmid 1S77. Pre-implantation Embryos of Cninese
Hamster. I. Incidence of Karyotype Ancmalies in 226 Control
Qnbroyos. Mut. Pes. 46:63.
Chen, J. 1979, CBIB Virology Unit Report, Product Chemistry Guidelines
Test: Microbial Bioassay of Pesticidal Cnemicals.
Chen, C., W. and B.H. Haberman, 1981. Memo to Marcia Williams (SPRD)
entitled "Carcinogenic Potency for Trifluralin Including N-nitroso-
dipropylaraine (NDPA) and Diethylnitrosartine (DEXlA)" dated July 9, 1981.
Cnernozenski, 1960. Cancer Res. 28:992
Cope, O.B. 1966. Contamination of freshwater ecosystems by pesticides. J.
Appl. Ecol. 3: 33-44.
Ccucn, J.A., J.T. Winstead, D. J. Handen, and L.R. Goodman, 1979. Vertebral
Dysplasia in Young Fish Exposed to the Herbicide Trifluralin. J.
Fisn Diseases, 2:35-42.
Day, e.W. Jr., D.G. Saunders, and J.W. Mosier. 19~S. Special Treflan
field monitoring studies II. Elanco submission #15 (confidential)
February, 1978.
-------
Day, E.W., 1980. Letter to J. Donoso dated Feb, 25, 1980.
Dean-Raymond, D. and M. Alexander 1976. Plant Uptake and Leacning of
Dimetnylnitrosamine. Mature. 262: 394-395
Donoso, J., 1980. EFB Response to External Canments on the Trifluralin
PD 1/2/3. Mono of 4/28/80 to H. Williams, SPRD.
Druckery, H., R. Preussman, S. IvanKovic, and D.'Scnmaftl. 1967.
Organotropism and Carcinogenic Activities of 65 Different N-nitroso
Compounds in BD Rats. Zeitschrift fur Krebsforscrtung 69:103-210.
(Translated for EPA by Literature Research Co., Annandale, Va).
Elanco, 1977, A Study of trie Effects of Trifluralin on Pregnant Rats and
Tneir Progeny, R-1265, by Worth, H.M., J.L. Emmersari, and D.M.
Morton, Toxicological Division, Lilly Researcn laboratories.
Elanco, 1978. Progress Report cn tne Reduction of the Nitrosamine
Impurity in Trifluralin Production. II. Elanco submission #26
[confidential] - December, 1978.
Elanco, 1960a. Personal C&Tmunication on August 18, 1980 from Ralph Hill
" to Carol Langley (SPRD) Re Results of Hitrosamine Assays on Technical
Trifluralin.
Elanco, 1980b. Letter dated September 3, 1900, from Ralph Hill to Carol
Langley (SPRD) Re Results of Nitrosamine Assays on TREFLAN EC
[confidential].
Elanco, ISCOc. Comments: Trifluralin Position Document, submitted by
Jonn Murphy to Carol Langley (SPRD) on September 4, 1980__
[confidential].
Elanco, 1980d. Tne Chronic Toxicity of Compound 36352 (Trifluralin) given
as a canponent of the diet to the B6C3F1 mouse for 24 Months - Studies
M-9067 and M-9077; and the Chronic Toxicity of Compound 36352
(Trifluralin) given as a Component of the Diet to Fischer 344 Rats for
Two Years - Studies R-87 and R-97, Toxicological Division, Lilly
research Laboratories [confidential], September, 1980.
Elanco, 1980e. "Trifluralin Worst Case Risk," submitted by John Murphy to
Carol Langley (SPRD) on September 22, 1980, [confidential],
Elanco, 19Q0f. Position Statement on New Trifluralin Studies dated,
September 23, 1980.
Elanco, 1981. letter frcm Ralph M. Hill to Carol E. Langley, (SPRD)
regarding trifluralin impurities [confidential], dated April 30, 1981.
Emmerson, J.L. and R.C. Anderson. 1966. Metabolism of Trifluralin in trie
Rat and Cog. Tbx. and Appl. Pnarm. 9; 84-97.
-------
Epstein, S.S., E. Arnold, J. Andrea, W. Bass, and Y. Bishop, 1972.
Detection of Chemical Mutagens by the Dominant Lethal Assay in tne
Mouse. Toxicol. Appl. Pharmacol. 23:288.
FDA, 1981, Pesticide Besidue Data for 1978, 1979, and 1980 for Trifluralin
submitted by Ellis Gunderson to Tar. Miller (SPRD), September 29, 1980.
Fowler, H.W., Jr., et al., September 20, 1979, Transcripts of Proceedings,
FIFRA Scientific Advisory Panel (SAP), Twenty-Sixth Meeting.
Fowler, El.w., Jr. et al., October 9-10, 1979, Transcript of Proceedings,
FIFBA Scientific Advisory Panel (SAP), Twenty-Seventh Meeting.
Gaede, H.W., 1900a. EAB Response to External Comments on tne Trifluralin
rn 1/2/3. Memo of 1/14/80 to M. Williams, SPRD.
Gaede, H.H., 1980b. Review of tne 4/2/80 Comments by toerican Cyanamid on
tne Preliminary Benefit Analysis of Trifluralin. EAB Memo to M.
Williams, SPRD, May 8, 1980.
Gaede, H.V.'., 1981. Memo from Carol E. .Lang ley (SPRD) to the File through
Harry Gaede on "Benefits of Trifluralin Use," May 29, 1981.
Golab, T., R.J. Herberg, E.W. Day, A.P. Rawn, F.J. Holzer and G.W. Pretest.
1969. Fate of Carbon-14 Trifluralin in Artificial Rumen Fluid and
Ruminant Animals. J. Agric. Food Cnem. 17(3): 576-5B0.
Haberman, B.H., 1931. Review of Trifluralin (Treflan) Chronic Rat and
Mouse Studies for Pathology data. Memo of 6/8/81 to Cinao Chen,
Carcinogen Assessment Group, dated June 8, 1981.
Hess, F.D. and D. Bayer, 1974. The Effect of Trifluralin orutne
Ultrastructure of Dividing Cells on the Root Meristem of Cotton
(Gossypium Hirsutum L, 'Acala 4-42'). J. Cell Sci. 15:429-441.
Hess, F.D., and Bayer, 1976, "Interaction of Trifluralin and Microtubule
Protein," Abstract #161, Weed Scientific Society of America.
Hess, F.D. and D. Bayer, 1977. Binding of the Herbicide Trifluralin to
Chlamydcronas Flagellar Tubulin. J. Cell Sci. 24:351-360.
Hess, F.D., 1979a, Mode of Action of the Herbicide Trifluralin, Exhibit A
of tne Elanco Submission, f30000/32;#6].
Hess, F.D., 1979b, Tne Influence of the Herbicide Trifluralin, cn Flagellar
Regeneration in Chlamydomonas. -Exp. Cell Res. 119:99-109.
Hook, E.W., and Porter, eds., 1977. Population Cytogenetics Studies in
Humans.
Jensen, J. 1S81. Quantitative Exposure Estimates for Chlordane in Soil
Samples in Frayser (Memphis), Tennessee. -Memorandum to Anne Barton,
HED Science Policy Staff. January 8, 1981.
B1
-------
Kasza, L., 19S1. Interpretaticn of trie role of Urinary Calculi in tne
Development of- Neoplasms in tne Urogenital System in Rat Experiment.
Memo to Haner K. Hall, SPRD, dated March 16, 1981.
Langicy C.E., 1981. Memorandum to E. Regelman (HED) regarding tne review of
Trifluralin Residue, dated March 12, 1981.
Leitis, E., and D.G. Crosby. 1974. Pnotodeccmposition of trifluralin.
J. Agric. Food Chem. _22(5): 842-848.
Macek, K.J.t C. Hutcninson and O.B. Cope, 1969. Itae Effects of Ttemperature
on Bluegills and Rainbow Trout to Selected Pesticides. Bull. Environ.
Contain. Toxiicol. 4(3): 174-183.
Macek, K.J., M.A. Lindberg, S. Sauter, X.S. Buxton, and P.A. Costa, 1976.
Toxicity of Four Pesticides to Hater Fleas and Fathead Minnows. U.S.
EPA, Ocol. Res. Series, Mo. EPA-60Q/3-76-099, 68 pages.
Mittelman, A., 1978. I. Environmental Fate of N-nitroso-n-dipropylamine
(HDPA) and Trifluralin. II. ^Jorker Exposure Estimate for N-nitroso-n-
dipropylamine (NOPA) and Trifluralin. HED, OPP, EPA.
Mittelman, A., 1980, EFB Response to External Ccnraents on tne Trifluralin
ED 1/2/3. Hero to 2/29/80 to M. Willaims, SPED. '
Mountfort, R., 1980, ?nendments to add Field Corn, Sorghum, and Barley,
Treflan EC. letter to Ralpn Hill, Elanco Products Canpany, dated
September 5, 1900.
Murnii;, M.R., July 19, 1979, Letter and Report to Steve Berkcwitz, SPRD.
National Institute of Environmental Healtn Services, August,'1979, Env.
Healtn Perspect, V 31.
Nelson, J.O., 1978. Unpublished data in letter to P.C. Kearney, USDA.
September 28, 1978.
Nelson, J.O., 1900, Personal communication from Dr. Nelson to Carol
Langley, SPRD, September 9, 1980.
Nelson, J.O., P.C. Kearney, J.R. Plinroer, and R.E. Menzer. 1977.
Metabolism of Trifluralin, Profluralin, and Flucnloralin by Rat Liver
Microsomes. Pest. Biochem. and Pnysiol. _7: 73:82.
Oliver, J.E., 1978. USDA, unpublisned data.
Parker, 1974. Genetics , 78:163
Parrisn, P.R., E.E. Eyar, J.M. Enos, and W.G. Wilson, 1978. U.S. EPA, Gulf
Breeze, Fl. EPA-600/3-78-010, 53 pages
-------
Piko, M. and Bcmsel-HeLmreicn, 0., 1960. Triploid Rat Embryos and Other
Chromosomal Deviants after Colchicine Treatment and Polyspermy.
Mature, 186:737.
Plageraarin, 1970, JIJCI 45:509.
Probst, G.VJ. August 6, 1981, "Reduction of Nitrosamine Impurities in
Pesticide Formulations," "Lilly Research Laboratories, A Division of
Eli Lilly and Company, Indianapolis, Indiana, 46285, U.S.A.
Regelraan, E. 1981a. Metro to Carol E. Lang ley (SPRD) entitle "Revision of
Dietary Exposure-Treflan HD 4", dated May 6, 1981.
Regelman, E. 1981b. Memo to Carol E. Langley (SPRD) entitled "Revision of
Treflart PD 4" dated flay 26, 1981.
Regelman, E. 1981c. Compilation of Dietary and Worker Exposure and Risk
Estimates sufcmitted to Carol E. Langley CSPRD) July 28, 1981.
Robinson, D.G. and N. Herzog, 1977 "Structure, Syntnesis, and Orientation
of Microfibrils. A Survey of the Action of Microtubule Inhibitors on
Microtubules and Microfibril Orientation in .Oocystis solitaria."
Cytobiologie, European Journal of Cell Biology, 15, 463-474.
Sanborn, J.R., 1974. Tne Fate of Select Pesticides in the Aquatic
Environment. U.S. EPA, Ecol. Res. Series, No. EPA-660/3-74-025, B3
pages.
Russell, L.B., 1976 in tollaender, ed. Chemical Mutagens: Principles
and Methods for Their Detection. 4:55.
Seller, J.P., 1972. Mutagenicity of Benzimidazole and Benzimidazole
Derivatives. (1) Foreward and Reverse Mutations in Salmonella
typniiturium Caused by Benzimidazole and Sane of Its Derivatives.
Mutation Res. 15:273-276.
Soderquist, C.J., D.G. Crosby, K.w. Moilanen, J.M. Seiber, and J.E.
Woodrow. 1975. Occurence of Trifluralin and Its Pnotoproducts in
Air. J. Sgri. Food Chem. 23(2): 304-309.
Taylor, J.W. 1965. J. Cell Biol, 25:145. '
Tiepolo, 1967. Cytogenetic, 25:51.
Traut, H. and w. Scheid, 1974, "The Induction of Aneuploidy by Colcemid Fed
to Drosopmla Melanogaster Females." Hut. Res,, 23:179.
Waucnope, R.S., 1978. The Pesticide Content of Surface Water Draining from
Agriculture Fields, A Review,, J. Environm. Qjal. 7(4): 459-472.
West, S.D. and E.U. Day Jr., 1977. Residues of N-nitro-sodipropylamina
and Trifluralin in Crops frcra Fields Treated with Treflan. Eli Lilly
and Co., unpublished data. [Proprietary],
-------
Wnits, A.W. Jr., L.A. Harper, R.A. Leonard, and J.W. Turnbull. 1977.
Trifluralin Volatilization Losses, from a Soybean Field. J. Ehviron.
Qual. 6(1):105-110.
Wilson, L., Transcript of SAP Meeting, October 9 & IC, 197S.
Sendzian, R.P. Memo to Carol E. Langley (SPRD) entitled "Dermal Absorption
of Trifluralin," dated May 26, 1981a.
Zendzian, R.P. Mono to Carol E» Langley (SPRD) entitled "Trifluralin:
Furtner Comments on Exposure and Absorption", dated July 27, 1981.
I
------- |