*   UNITED STATUS ENVIRONMENTAL PROTECTION AGENCY
    *
                 •        VMSHIfJGTON DC  20460
                                                            OTFICE or
                                                       PLANNING AND MANAGEMENT
MEMORANDUM

SUBJECT: JCeponc Risk Assessment
                                                    /
FROM;     Eli/abeth Anderson, Executive Diiector >/c//'7' ( /%-T/*' '
           Interim Carcinogen Assessment Group //- '^   "'   /
^                                                  (..^
TO:        "Ec|\vm Johnson, Deputy Assistant Administrator
                  Pesticides
The Carcinogen Assessment Group has completed the risk Assessment
for Kepone.  I am herein transmitting the report of this analysis lo you,
Attachment
cc:
 A.  Aim
 \V. Talley
VR.'Dreer
* F.  Ilageman
 P.  Denney

-------
                 TUB CARCINOGEN ASSESSMENT GROUP
                         ANALYSIS OF ICE PONE
Participating Members
 Elizabeth L,. Anderson
 Robert Me Gaudily.
 Lakshmi C. Misln a
 Arnold M. Kuzmack
 Mclvin Reuber
 William Upholt
                                  R. K.Albert, MD.  Cliairman
                                   Date
                                        &

-------
Conclusion of The Keponc Risk Assessment

On the basis of animal  tests summarized in this report, there is
significant evidence that Keponc is likely to be a human carcinogen.
The impact on the induction of cancer in humans by the continued
use of Keponc in ant and roach  traps and pastes involves particularly
largo uncertainties chic to the lack of data upon which to base exposxire
extimates for the gcncial  population. In the CAG's vicnv,  the major
concern is probably tho iisk to children who accident}} inmost Ihe
Contents of traps and tubes of paste containing Kepone.  Of the various
models of assessment,  the "worst case" estimate oflifelime risk to
the  individual child is one chance in  400 of de\eloping cancer due to
accidental exposure using the assumption of repeated episodes of
ingestion and a linear non-threshold  model for cancernndur lion.

-------
                   MK'J A HO 1,1C CHARACTERISTICS
 The tissue distribution of kcpone in animals has been measured by
 several investigators (Kef.  1, 3, 4 & 7).  These studies luuc shown
 that the liver is the organ \\hcre the largest kcpone storage occurs,
 and that it is also stored in  brain,  kidney and body fat. Accumulation
 of kepone in these organs continues for longer than 150 da\s in mice
 fed 40 ppm of kepone.  Upon withdrawal of kepone from the diet, the
 half-life-«is shorter than 20 days.  Compared uilh mice, a close
 structural analogue, kcpone is less fat soluble and the hvei /fat
 storage ratio is larger lhan for  mice.   Kepone excretion is predomi-
 nantly via the liver bile and fcces route.

 in human studies of kepone factor}7  \vorkers  in Ilopeuoll,  \'a. , it has
 been found that the ratio  of fat to blood concentrations of kcpone is
 only about 7/1*. Other chlotinated  hydrocarbons are concentrated to a
 much greater degree in the  fat,  for DDT the fat/blood jatio is close
 to 300. The time it takes to equilibrate with blood after continuous
 ingc&tion  begins lias not been  measured, but based on an analogy with
 mice it may take as long as 2 years.  After  withdrawal of kcpone
 exposure  in people, the half-life in the  blood is about 4 months.
 Human liver levels have  not been measured.

 Several studies indicate that the effects of continued ingcstion are
 cumulative.  Good,  et al, (Ref.  2)  found that reproductive performance
 in mice steadily declined over a 3-month feeding period at 5 ppm.  In
 his experiments mice on the kepone diet were bred and their progeny
 were raised to maturity on the same diet.  The second litters of the
 progeny were smaller and less numerous than the first litters had
 been.  In the Allied study (Ref.  3) it was found that single closes which
 "were 0. 1  times the  L.D "30 had no effects on rats but 20 daily repetitions
"of this dose resulted in 50% mortality and  marked growth ictardation.

 Metabolites of kcpone have been looked for in liver and fat tissues
 but not found.  Mutagencsis studies of kepone have not been reported
 (Ref. 6).                 -

-------
               EXPERIMENTAL CARCINOGCNESIS STUDIES
 Two carcinogenesis studies arc available (Ref.  3, 5): the chronic feeding
 study in rats submitted in support of a tolerance petition for use of kepono
 on bananas, and the 25-page sujnmary of an NCI feeding study on rats and
 mice, dated Januaxy 1976.

 A.  NCI Rat and Mouse Studies

 Technical grade kepone was administered for an 80 week period to four
 groups of B6C3F1 mice and four groups of Osbomc-Mendel rats approxi-
 mately 6 \veeks of age.  The mice were killed alter 90 weeks and the  rals
 after 112 weeks.  The  four groups consist of males and females each  fed
 two doses.  The dose levels were chosen on the basis of survival and
 weight gam in a preliminary 6-week treatment panod.  As the full scale
 study progressed, it became apparent that the animals could not tolerate
 the initial doses, and the feeding level was  lowered Time-weighted avei -
 age doses for the entire experiment were calculated and are given below,
 along with Die initial doses:
       a)  Initial Dose

           Species
            Mice
            Rats

       b)  Average Dose

           Species
            Mi ce
          •  Rats
                         DOSE LEVELS,  (PPM)
   Male
Low  High
                   30
Female
    High
    ~~M
     60
   Male
Low  High
       ~~~
Female
   8
 The animals were housed in two rooms for this expcammcnt.  Kepone was
 the only compound under test in the room where iheas'ts were housed,  \
 whereas the mice were housed  in a room \\here  toxaRftene, chlordane,  and
 kepone weie concurrently under test.  Initially the re we re 50 animals in
 each of the Jour dosed groups.  Controls consisted dfimatchcd controls and
 of pooled controls.     , ,

'The survival of the animals at the end of the experiraa-nt is given below,
 v/here the  conti ols for each dosage group are combiocd:

-------
                          TERMINAL SURVIVAL

                                         Pose Group

                         Q-*             Low          High
    a)  Mice
       1  Males        G2/G8 (91%)    28/48 (58%)    25/49 (51%)
         Females     43/50 (86%)    42/50 (84%)    41/49 (84%)
     *

    b)  Rats
       ""Males        75/115(65%)   30/50(60%)    18/44 (41%)
         Females     68/110 (62%)   27/49 (55%)    18/45 (-10%)

        ^Pooled and matched controls are grouped togcthei.

 It can be seen that control mice survived the experiment in greater
 numbers than control rats,  and that survival decreased in all  groups
 as the dose increased

 Qualitatively survival correlates well (negatively) "with aAerage dose,  with
 the notable exception of female mice, where the  ratio of high to low ave-
 rage dose is  2/1 but the  percent survival is identical for the IMO  groups.
 In general rats tolerated the experiment less well than mice,  fi om the
 survival figures as well  as the clinical signs of toxicit}'.

 In lats some high dose ammals developed epistaxis  and bleeding  of the
 eyes and four animals had died or had to be sacrificed dining  the first 4 ~
 months of the study.  Females began developing  generalised ticmors at
 the fifth week in the high dose group and the 28th \veck foi  the  low dose
 group.   In all treatment  groups adverse signs  (rou^h hair coats,
 dermatitis,  anemia, tremors) occured at a moderate incidence during the
 first year and^uith a gradually increasing frequency in the second year.
 •Surviving animals at the  termination of the experiment were generally
 in very poor  physical condition.

 In mice, the -low level animals had the same appearance and behavior  as
 controls throughout the first-year, and a few low le«d males had palpable
 nodules in the abdominal area during the second  year. In the high level
 animals, females developed  generalized tremors dnnng the first week
 of the  study and these persisted until the end.  Aboii20% of the high dose
 male group were highly excitable during the  secondyzai.   Bloating or
"abdominal distenlion \\as noticed in high dose males,al week 45 and at
 high dose females at week 68.       .               •'  .  .

•A significant increase was found in the incidence ofJbpnloccllular carcino-
 mns in only high dose level rats of both  sc\es  and inmice of both sexes at
 both dose levels of he pone  (P<0. 01).   The incidcncev(P<0. 05)  in the high
 dose groups were 7% and 22% for male and female io& (compaicd with
 0% in controls for both sc.xcs) and 88% and 47% for rale and female mice
                                 - 2 ~

-------
(compared with 16% for male pooled controls and 0% in females).  The
incidence  for the low dose groups of mice weie 01% for males and 52%
for females.  The incidence for low dose rats was 20% for both males
and females.  Extensive hyperplasia, fatty infiltration and degeneration
as well as occasional "neoplastic nodules" were also found in the hver
of mice and rats.
 HEPATOCBIJ ULAR CARCINOMAS IN MA T E AND FEMALE RATS
                              '1' l NC; K'C \''(5m?	
	-     Dose *               Males                 Females

      0 ppm               0/105 (0%)*             0/100 (0%K

      low dose             l/5'O (2%)               1/49(2%)

      high dose            3/44 (7%)             t  10/45 (22%)

         * Pooled control groups

  HEPATOCELLULAR CARCINOMAS IN MAI,K AMD FEMALE MICE
                       INGESTING KDPONE

1       •  Dose                      Males             Females

                 Pooled         8/49(16%):         0/40(0%)
     • 0 ppm .
                 Matched       6/19 (32%)          0/10 (0%)

       low dose                   39/48(81%)       26/50(52%)

       high dot,e                  43/49 (88%)    .  23/49 (47%)

   *The NCI report states that the hepatocellular carcinoma incidence arm
control mice of this strain is usually betwcn 7  and 10% in males  and about
l%>m females.  The 16% in these males is theicfore not regarded as usual.

The information on time until the detection of the first grossly observable
tumors is summarized below.

                      TIME TO FIRST TUMOR, DAYS

                            	Dose Group	               "
                           ••-  0 -       .Low
      a) Mice               *""
         ~Wles            87          70           62
           Females         - -          87          " 76
                                        *        "*
      b) Rats
         """fifties             -         112   '    '   108
                                         87.'         83
                                   *

                                - 3 -

-------
As expected, the time is less for each high dose group than for the cor-
responding lov/ dose group.  Ilouever the diffeience in time to first tumor
(high versus low dose groups) does not correlate well with  changes in the
average dose le\el,  since the male rats have only a four-day difference
in time to fiist tumor whereas the dose is three times higher in the high-
dose than the low-dose group.  Male mice developed tumors sooner than
females,  whereas female mice  developed tumors sooner than males.

The overall incidence of tumors is not significant  nhen benign and malig-
nant tumois  are  considered together.  When malignant tumors are con-
sidered alone,  theie  arc more malignant tumors in treated as compared
to the  contiol animals.  Malignant tumors were present in the thyroid of
low dose male rats, and there-were tumors of the reproductive system in
females of both  the low and high doses.   There was no dilfcrence in be-
nign tumors  because  both control and treated animals developed adenomas
of the  endoci me  organs  and because of the low number of control raib.

       MALIGNANT TUMORS IN MALES INGESTING KEPONEg./
                         (NCI ]JAT STUDY)

              No. Rats with          No. Hats with
            Li\er Carcinomas   Other Malignant Tumors     Total

0 ppm  «.      0/10(0%)      '       0/10(0%)   .       0/10(0%)

                                        *
low dose       1/50 (2%)              7/50 (14%)          8/50 (16%)

high dose      3/44 (7%)           .   5/44 (12%)          8/44 (18%)

         MALIGNANT TUMORS IN FEMALES INGESTING KEPONE
        '                   (NCI  RAT STUDY!!?/
                                      «
              No. Rats with          No. Rats with
    7^  ,;,    Liver Carcinomas  Other Malignant Tumors **"• - Total

Oppro        0/10(0%)       .     0/10(0%)           0/10(0%)

low dose  ,   -1/49(2%)     ,    . ,,  7/49(14%)    "      8/49(16%)

high dose     10/45  (22%)     .     9/45 (20%)          19/45 (42%)^
                      .«
   oYData concc_rmng  tumors other than the liver arezot available at
    , this time for the remaining controls in the pooled groups.

The conclusion diawri from the  NCI studies is that Kegnne is definitely
a carcinogen in  mice and rats.  There is nothing aboaftthc study (treat-
ment of annn.il.3, excess carcmogemcity in controls, jprc-mature death
due to other  causes) that would  cause any doubt aboulllic validity of the
study.

-------
                        NCI  MOUSE STUDY
100 i,  '
                 MALE MICE
   RCCM  N'PTCHEO  LOW
CONTROL  COXT^OL  DOSE
         FEMALE MICE
                                                      0s/-
                                                                25/50
                                                                 52%
                                                                            23/43
CONTROL
                     Comparison cf Incidence of Kepatoceilular  Carcinoaa
                                    '(Chlordecor.c)                   ,

-------
                                                NCI RAT STUDY
   100 -t
                        MALE RATS
FSiVIALg RATS
S   SQ'
o
c:
<
o

c
o
o
c_
u
CO
u.
O
u
o
CZ.
ID
C.
 SQ-\
    »


    I

 70-





 60-1 •





 50-





.40-





 30-
 10-
             0/105
3 I 2%
— l^ s \ 1
3™
\S^
I
i
0/100
~p%__
OL ^'^ ^g . CONTROL
1/49
27,
•^ \ \ t
1

LO'.'.' HIGH
D33H CCS£
                                Comparison of  Incidence of HepacoQellular

                                                 (Chlordecor.e)

-------
B.  Allied Chrome Ra< Study

In this study young albino rats of an unspecified strain ucro grouped into
12 groups (G male and G  female) of 40 animals per group.  They were fed
diets consisting of 0, 5,  ?0,  25,  50 and 30 ppm oflepon^, v.lmh Allied
designated as  compound  1189.  At the end of 3 and 12 months, some
of the animals from each group were sacrificed and at 2-1 months the re-
maining animals were examined.  The observations and measurements
were 1) food consumption, 2) animal weight, 3) organ \vcights, 4) hemo-
globin and hcmalocrit levels,  5) urinary sugar and protein levels,
G) histopathological examinations.  In a separate two 3ear feeding study
rats were fed  1  ppm kcponc.
    *
The animals receiving BO and 50  ppm developed muscular tremors by the
end of 2 and 3 weeks, respectively and had all died b) -1 and G months.
At 25 ppm mild  tremors  developed by 12 months.  The number of animals
alive immediately prior  to each sacrifice is shown belou.
                NUMBER OF SURVIVORS (Males/Females)
  a)  Males
  b)
   Time
 (months)

     0
     3   *
    12
    24

Females
	Time
                                   DOSE  (ppm)
                       JD_

                       40
                       38
                       31
                        7
40
39
34
 3
10

40
37
28
 4
40
37
26
 3
_50_

40
 36
  0
  0
80

40
 4
 0
 0
                                   DOSE (ppm)
0
39
31
14
5
39
34
10
10
40
37
32
9
25
40
24
3
.,„ 5(J.
39
0
0
	 80 "
-10
5
0
0
    - (months)
          0
          3
         12
         24

At the end of two years,  the maximum number which could survive  was
25 because of the planned sacrifices at 3 and 12 momlhs.  Therefore the
percent survival at the six dose levels by the end of two years was,

                       TERMINAL SURVIVAL, Percent
                                   Dose,  ppm
0
~^T
5G
5
40
10
— ro-
se
25
12
12
50
— U
0
ao
— G
0
      Males
      Females

This is a smaller survival than observed in the NCIaiiulj  at corresponding
doses (including contiols), so one would cxjxict fewer animals to live long
enough to develop cancer.           .         .           .
                                - 7 -

-------
In fact whan the  animals were examined after saci jfjcc only six of them
had gross liver lesions that were regaidcd important enough to subject
to special study  and interpretation.  The rcpoit regards the evidence
of carcmogcMiic it> as cqxu\ocal, since the opinions ol four pathologisls
differed in the interpretation of the slides, however,  the consensus was
that the lotions were carcinomas.  Furthermore, 1 1  should be noted that
carcinomas of the liver arc extremely rai c tumors in rats not treated v.'Jth
chemicals.  Lack of time does not permit i e -examination  of the lustologic
sections.
 Dose

 0 ppm
 '5 ppm
10 ppm

25' ppm
  Dose

  0 ppm

 ' 5 ppm

 10 ppm

 25 ppm
              LESIONS OF TIIE LIVER IN MALE RATS
                                                      IJyperplasia
                                                          plus
Hype
0/8
0/8
2/1
0/5

iplasia-B'
(0%)
(0%)
0 (20%)
(0%)
LESIONS OF
Carcinomas
0/8 (0%)
0/8 (0%)
0/10 (0%)
2/5 (40%)
THE LIVER IN
Carcinomas
0/8 (0%)
0/8 (0%) -
2/10 (20%)
2/5 (40%)
FHM.ALE RATS-5/
ITyperpIasia^r

  0/18 (0%)

  1/16 (6%)

  1/15 (7%)

  0/5 (0%)
Carcinomas

 0/38 (0%)

 0/16 (0%)

 3/15 (20%)

 1/5 (20%)
Hyperplasia
    plus
Carcinomas

^  0/18 (0%)
 *,
  1/16 (6%).  ,v

  4/15 (27%)

  1/5  (20%)
      a/Two male rats had hyperplasia of the liver aftei  85 weeks.'
      - In other rats hyperplasia or carcinomas were present at
        10-1 weeks.
      b/Includes rat with hyperplasia of the hvoi only
                                - 8 -

-------
Malignant lurnoi s \\ere found predominantly in the liver in treated
animals.  There  \\eie in addition some othci malignant tumors such as
melanoma of the  brain and leukemia,  as  \\ell as subcutaneous adenomas
and adenofjln omas (a thorough histopatho]ot';ical evaluation was not
attempted by Allied).  Numbers of tumors .ne given in the tables (foi
animals that h\rd for G9 weeks,  the time of appc.aiame of the first
tumor, 01 loncuM ).  Two females ingesting 10 ppm kcpone with carci-
nomas of the h\ei also had "thrombosis  of adienal veins."

              NUMBER OF MALE RATS WITH TUMORS-/
                                                        »
                     No.  Rats with              No. Rats with
    Dose                Tumor sjy            Malignant Tumors —

     0 ppm             1/8(13%)                   1/8(13%)

     5 ppm             2/8 (25%)£/                 0/8(0%)

    10 ppm             2/10 (20%) £/                1/10(10%)

    25 ppm             2/5 (40%)                   2/5 (40%)


              NUMBER OF FEMALE RATS WITH TUMORS-/


                     No.  Rats with               No.  Rats  with
    Dose                Tumors k/            Malignant Tumors £/

    0 ppm             6/18(28%)            '      1/18(6%)

    5 ppm             9/16 (50%)                  1/16 (G%)

   10 ppm             10/15(67%)                4/15(27%)

   25 ppm             1/5 (20%)              " ^  1/5 (20%)

  a/Based on individual animal diagnoses by Allied's pathologist.
  ]>/Some lats had more than one tumor.  In such cases, the  rat
    is counted only once.
.  £/Jn this column, rats which had bemgn tumors only have been
   . excluded,  to show whether there appears to be increased
    incidence of malignancy (as \\ell as increased tuasor incidence
    in the  dcatod groups).
  
-------
 The liver/hotly weight ratio increases in both groups \vilh dose level.
 When linear regression of this ratio is clone with icspccl to dose,  it is
 found that the close at which the liver/body \\eight ratio is significantly
»(P .  05) hujhci  than controls is 15 ppm for males and 5. 3 ppm for
 females.  Therefore liver enlargement occurs at a louer dose in
 females than for males.   These concentrations are virtually the same
 at 2 years as at 3 months, indicating that no fin ther enlargement occurs
 after the first  lew months.

 The kepone lexcl in fat increases with dose.  At thiee months  the fat
 level ranges from 10  ppm in the 5 ppm dose gioup  1o 383 ppm  in the 80
 ppm  dose group.  This measurement was not done  for animals on the diet
 at longer times.
                                                                  %
 When regression of bod}r weight is done versus dose levels, it is found
 that significant depression in growth rate occui s at a concentration of
 5-10 ppm in females and 10-25 ppm in males.

 Measurement of blood cells were done at 3 month inter \als.  A tendency
 toward depression of hematocrit was obseixed at 80 and 50 ppm at 3
 months and depression of both hematocrit and hemoglobin occurred alter
 15 months in the 25 ppm group.

 The urinary excretion of sugar and protein were  made at 3 month inter-
 nals.  Kepone  dosing had no effect on sugar but did cause protein to be
 excreted-m the urine.  The effect got progressively larger and occurred
 sooner as the dose increased.  Kidney tissue was examined especially
 to provide e\idcncc  for this proteinurea.  Females sho\\ed a trend toward
 increasing incidence and severity of chronic 1 idncj lesions with increasing
 dose. Renal disease in treated males was not as noticeable because un-
 treated rats also had renal disease.

 In the second study in which rats ingested 1 ppm  Kupone,  2 of 1 0 (20%)
 male rats and  2 of 13  (15%) female rats had malignant tum'ors, compared
 to 0 of 17 female controls and 0 of 14 male controls.                   v

 In summary, the Allied study does indicate that kepone is carcuiogenic
 in rats,  par lieular ly for liver,  and could also be for tumors developing
 in other organs, even though the overall mor lality  w^s high.  These
 findings are similar to those of the NCI study.
                       «  \                                  «
 C. Comment

 Carcinoma^ of  the liver were induced by Kepone at  hSgh and low dose
 levels in mitu and at higher dose levels in rats   Carcinomas of the
 types icpoiled  in  this study have the morphologic c Ixau-actei istics of
 milgnanl tumois.  do invade and mctastasi/e to the  lioigs and/or other
 organs and giuw iollowmg transplantation to other  aunmnls of the same
 strain.  11 should  be noted that  the incidence ol lung snetaslases is  not


                                - 10 -

-------
reliable unless, among other things,  serial sections arc done, which
was not the case in these studies.  JLikcwise,  multiple sections are
necessary to demonstrate mvasiveness of carcinomas.

The development of carcinomas of the liver in treaiod rats is ex-
tremely impoi (ant.  Carcinomas of the liver in rats arc preceded by
the development ol hyperplasia and hyperplastic nodules,  i.e. benign
lesions.  The in st hypcrlasia of the livei was noted in albino male
rats  after 83 weeks and the first carcinomas at the end  of 101 weeks.
If the ammalt. had survived for longer peiiods of tune,  these lesions
would have continued to progress.   Control rals did not develop car-
cinomas ol the  liver in these  studies.  Fmthermore, carcinomas of the
liver are extremely rare in any strain of untreated rat.

D.  Conclusions

Male and lemale BOC3F1 mice ingesting eithci the low or high dose
levels of kepone m the diet develop carcinomas of the liver.  Carcino-
mas  of the liver were  foxmd in Osborne Mendel male and female rats
ingesting the highe: dose level of kepone.  Albino male and female
rats  ingesting high doses of kepone also had cai cinomas of the liver.
Such carcinomas of the liver  not only have Ihc morphologic characteris-
tics of malignant tumors  bui they also kill the host animal and are
capable of invading and metastasizing to the lungs and of growing on
transplantation 10 animals of  the same strain.  The overall incidence
of all malignant tumors was increased 111 both Osborne Mendel and
albino male  and female rats.
                                          \
The  conclusion drawn  from these studies is that kepone  is definitely a
carcinogen in animals.  There is nothing about the suuhcs (treatment
of animals,  excessive carcinomas in controls, premature death due to
other causes) that would cause any doubt about the validity of these
studies.                                                <
                                 - 11 - '

-------
                      H1SK EXTRAPOLATION
A risk extrapolation has been performed for the possible eai cinogenic
effects of allowing existing slocks of Kc-ponc to be used up. Only a lincai
extrapolation was possible with the limited cxposxire (In la available,  and
even that is parametric.  Use of non-linear extrapolation mo 1 hods would
require estimates not only of average exposures but also of the distribu-
tion around the average, this is a particular problem in the picsent case,
where accidental mgoslion by children may be a significant poi tion of the
total human exposure.

Let N = the number of people exposed to Kcponc through use of the
existing  stocks, y_ =  the number of years necessary to u^e up  the stocks
(•ourjcntly estimated at '1 and,  in any case, small compaird to a human
lifetime), and p = the pi oportion of the Keponc stocks that vill eventually
be consumed by" humans  via the various  possible loutcs of exposure.
Total stocks are currently estimated at  537.-8 Ib (sub]cct to levision).
Then the average daily exposui e to Kepone, expressed as an equivalent
ppm in the diet, is

               __      537.8 Ib x 106 x p
               2. 205 Tb/kg x N x 1. 5 kg/day x 3G5y days

             = 4.455 x 105 p/Ny

To derive the  slope of the dose-response curve, the  NCI data for liver
tumors for male mice fed 20 ppm (the case that yields the highest
result) and the formulation niade by Dr. Todd Thorslund of OPP has
been used.  (The lowest  slope for mice derived from the NCI  study is
Jower by a factor of 5.  Values for rats are lower by factors of 9 to 77. )

Under the one-hit theoiy  (he  pi obability of a individual developing a tumor
due to an exposure to a dose  x  of a agent may be expressed as

                     H(x) = 1 - e~Bx

where "B" is a unknown  parameter to be estimated from a bioassay where
the  observed vaiiables aie the control and exposure  incidence of tumors.
JvCt R be the incidence in »i exposure group and  0 the conliol incidence.
Under the usual assumption that whatever causes the tumor in the controls
is independent of the action of the agent,  we have that

            '  ' R =  O -I  (1 - O) R(x)  = O  I (l-O)U - c ~B>V)

Solving for B, we have that

               B =  ( In(l-O)  - In(l-R)] /x

In the case of Ki-ponc, estimates of O and R,  when X is 23 ppm in the diet
of a  mouse ovei its entire life  span arc O = 0.10 and 11 - 0. 80  respectively
BO that

-------
               B = [ln(l -O) ~ In(l-R)] /x = 0. 085

 Since O and R are  estimator! 13 as also estimated.

 A question can be  laised as to whether the  appropriate measme of
 increased risk is R(\) 01  (1-O)R(\).   Since the difference is small in
 this case, R(x) will be used for convenience.

 Some may prefer to use an upper confidence  limit for 13.  This is not
 done here because a best  estimate of the risk is bem£ attempted and
 because the estimate  of B is alread}' based on the  worst-case data
 from the NCI stud}''.   If a  95% confidence interval wheie chosen,  the
 results below Avould be  about 50% higher,

 Using tins as an estimate of the response for humans,  \\e find the
 lifetime risk of the a\erage exposed  individual to be

              U, 005 x 4. 455 x 10 5 p/Ny = 3. 79 x 10 4 p/N;>
 If _N people are exposed, v,e obtain the number of cases cau^od per
 year by multiplying this quantity by N/70 and the iiumbei of rases
 in y_ years by multiplying  b;> Ny/70.  Thus, the total numbci of cases
 pro]ccted as a result ofuising up the existing stocks of Kepone is

                                  540 p

 Obviously,  an estimated of jp is needed to arrive at a final estimate.

 For example,  if it is estimated that 1-/10 of 1% of the total stocks would
 ultimately -re a ch people (p - 0. 001), then about half a case of cancer
 would be projected to result.

 This calculation also assumes that the effect of exposure to Kepone would
 be the same if the same dose  \\ere taken in a single accidental ingestion
 or spr&ad  out ovei a period of several yeais.  While tins assumption is
 consistent with the one -hit thoc-ry, it may not be true in fact.  Anatysis
.of the experimental record for various carcinogens is needed to develop
 better methodologies for dealing with  this situation.

 Tins calculation does not  take, into account the special circumstance that
 exposure .of simll childi en to Kepone  may be greater than exposure  of
 other age groups.  'I his could affect the  results in two \\ays.  First,
 small children ma\  be moio or less susceptible than othri  groups.
 Se< on
-------
 jcasonablc worst case.   The following is illustrative and con easily be
 varied to be moic or less conservative.  It is also oriented lo  the
 "accessible" formulations   Accidental cxposvncs from the "inaccessible"
 ant traps would piesumablc be much less.

 Assume that a child cats a 3-07, tube of ant paste  every G months for
 three years.  Such paste contains 0.125% Kcponc.  Assuming the child's
 daily food consumption to be about 0. 75 kg, his average  dietary exposure

 to I^eponc would be about 0. 8 ppm for that period of lime.  Using the
 above methodology for calculating risk,  his probability of conti acting
 cancer as a result of exposure to Keponc would be about  0. 25%.  This
 means that 400 individuals \\ould have to be exposed to this high level
 fpr there to be an exported one additional case of cancer.

 Another approach \\ould be to estimate the  numbers of UnIdron likely
 to bo accidental]) exposed lo Kepone.  Data in the Pesticide* Upisodc
 Report System sho\\s  an a\erage of 12. 5 cases per year in the  home or
 with children in vhich Kepone might ha\e been accidental!)  ingested
 foi those >ears for \\lmh substantial data is available.  If on)\ 1% of
'the actual cases \\ere reported (probably a pessimistic assumption),
 then there would be  about 1200 such cases per year or about 3, 750 over
 three years.

 If these  cases  are assumed to involve an average of haK an ant trap, or
 4 gm, then each child \\ould eat 5 mg of Kepone or about G. 7 ppm in his
 diet for  that da>.

 Male mice exposed to tins dietary level for a lifetime \\oulcl be projected
 to have an incremental cancer incidence of about  575* (Since the dose
 is very close to the  experimental doses,  this resultnould not vary much,
 whatever extrapolation method is used. )  Since the erposuios \\ould only
 last for  1/70x305 of the individuals' lifetimes, the nuk to each individual
..would be approximately 0. 57/(70x3G5) =  2. 2 x 10  • Again,  tins calcula-
 tion neglects the lo-u competing risk early in  life.  YSlh these assump-
 tions, there would be  3750 x 2. 2 x 10~^ = 0. 083 case*.  Tins means,
 roughly  speaking, that there \\ould be a probability rffl/12 that one
 case  of cancer would lesult from this exposure.
                                    •
 Summary

 Because of the lack of fn m exposure data,  three  diffxionl assumptions
 weic made." Fust,  the highest rir.lv to a single indivtlual horn the
 worst reasonably pofc^ibJc exposure; was estimated ly> assuming that a
 child-would cat six 3-o/,  lubes of ant paste oxer 3 yens.   'Ihis  would
 result in a cancel ru.k of 1/100,  exli apolalmg fiomilie animal data.
 The numbci of ehildien exposed at this level,  if any,is likely to be
 vciy  small.
                                     •

                                 - 3 -

-------
Second, using data from the Pesticide JJpisodc Reporting S\slem as a
slai ling point,  it \sas <\<~,< \nncd that 3, 750 chikhcn might each cat about
half an ant trap containing Kcponc.  This  would result an c> pee led 0. 003
cases of cancer or a piobabihly of 1/12 of one case of cancer.

Finally,  an attempt was made to estimate parametncall} the effects of
all routes of exposure, including vaporr/alioni spilling onto kitchen
surfaces, etc.  in addition  to accidental ingestion by childrc-n.  Using a
linear no-threshold extrapolation,  the resulting number of cases of
cancer would be-540 multiplied  by the proportion of the Kcponc which
readies humans.

In addition to thr- uncertainties inherent in any extrapolation fiom
animals to humans and fi om high to low doses,  additional uncertainties
Eire introduced in this  ease because  the exposuie is likely to be concen-
trated on children and is 111 ely  to be the  result of single accidental
ingeslions  rather than conhnuous in the  diet.

-------
                             Refer cncos


1.  llubcr,  J. J.  "Same Physiological Effects of the Insecticide Kcpone
   in the Laboratory Mouse",  Toxicology and Applied Pharacology 7
   51G-524 (1965)          .                                      ~

2.  Good, F. E. , G. W. \Vare, D. F. Miller  "Effects of Insecticides on
   Reproduction in the L/aboratory House:  I. Kepone", J. Economic
   Entomology 58-751-757

3.  United Fruit Company, EPA Pesticide Petition No.  OCOH19, Section
   C

4.  Curley,  A.  "Koponn Meeting at State Health Dcpa? Imcnl, Jlichmond,
   Va.  September] 5,  1975" EPA memorandum SeptemlxM  10,  1973.
            ^
5.  National Cancer Institute,  "Report on CarcmogcnrMS Bioassay of
   Technical Grade  Chiordeconc (Kepone)" January 197G

6.  Environmental Mulagen Information Center, Oak JJidgc Tenn.
   (July 1976)

7.  II. B. Matthews,  National Institute of Environmental Health Sciences.
   (July, 19 76)

-------