5006
                         SNARL For Trichloroethylene
            Health Effects Branch, Criteria and Standards Division
                           Office of Drinking Water
                     U.S. Environmental Protection Agency
                           Washington, D.C.  20460

         The Office of Drinking Water has reviewed the current literature
         on the health effects of trichloroethylene.  Both data from
         animal tests and some studies from high level exposure in
         humans were used as basis for extrapolating to levels in
         drinking water that would result in negligible risks to the
         general human population.  When considering toxicity that
         does not include the risk of cancer, we generally use a
         child weighing 10 kg (22 pounds) and drinking one liter of
         water per day as the basis for calculations of short exposure
         (acute) toxicity and longer exposure (chronic) toxicity.
         These levels are derived using safety factors from classical
         toxicology and a logic similar to that used by the National
         Academy of Sciences in "Drinking Water and Health."  When
         considering the possible cancer risk, where it is assumed
         that there is some risk at any level of exposure, and that
         the risk increases as the lifetime exposure increases, we
         use the 70 kg (154 pounds) adult living 70 years who drinks
         two liters of water as the base, and calculate the excess
         cancer risk above the normal background according to a
         mathematical model developed by the National Academy of
         Sciences in "Drinking Water and Health," and based on animal
         tests conducted by the National Cancer Institute.

         The drinking water levels that we have calculated providing
         a margin of safety from likely toxic effects in humans
         (assuming that 100% of the exposure is from drinking water.)
         were related to the length of time that water is being
         consumed, and range from short-term emergency levels to
         long-term chronic exposure.  We have separately computed the
         potential additional cancer risk.

-------
The computed drinking water guidance levels for effects
excluding cancer risks are as follows:

    Time                           Concentration
     1  ay
    10 days __ __ 0.2  me/1 (200 ug/1)
    Chronic (long-term)            75 ug/1

The computed excess lifetime cancer risks from the NAS model
at various exposures assuming the 70 kg adult drinking two
liters of water per day for 70 years at the indicated concen-
tration are as follows:

    Concentration                  Excess Risk

     4.5 ug/1                      one in 1,000,000
    45 ug/1                        one in 100,000
    75 ug/1                        approximately two in 100,000

The development of a SNARL for trichloroethylene does not
condone its presence in drinking water, but rather provides
useful information to guide control priorities in cases
where it is found as a contaminant.  Human exposure to
contaminants in drinking water such as trichloroethylene
should be reduced to the extent feasible, to avoid the
unnecessary risks from their presence as adulterants.  The
applicable treatment technologies include aeration and
granular activated carbon.

-------
                                                    NOV 2 6 1979


                 SNARL  For  Trichloroethylene
    Health  Effects  Branch,  Criteria  and  Standards  Division
                   Office of  Drinking  Water
             D.S. Environmental  Protection Agency
                   Washington, D.C.   20460

 In  the  absence  of  a  formal drinking water standard  for
 trichloroethylene, the  Office of  Drinking Water has  estimated
 a suggested  no  adverse  response  level  (SNARL)  following the
 state-of-the-art concepts  in toxicology  for non-carcinogenic
 risk  for short  and long term exposures.  For carcinogenic
 risk  a  range of  risk estimates  are  provided for life-time
 exposures  using  a  model and  computations from  the National
 Academy of Sciences Report:  Drinking Water and Health
 (1977).  However,  SNARLS are given  on a  case-by-case basis
 in  emergency situations such as  spills  and accidents.  The
 SNARL calculations for  short-term and chronic  exposures
 ignore  the possible carcinogenic  risk that may result from
 those exposures.   In addition SNARLS usually do not  consider
 the health risk  resulting  from possible  synergistic  effect
 of  other chemicals in  drinking water, food and air.

 I.    General information and health effects

      Trichloroethylene  is used primarily as a  metal  degreasing
 agent.  It is also used, however, in dry-cleaning as a
 solvent, and in  refrigerants, and  fumigants.  Trichloroethylene
 is  slightly soluble in water.

     Trichloroethylene, like other halogenated hydrocarbons
 at high dose levels, has been reported to produce liver and
kidney ..damage ..and ..central .nervous system disturbances .in  .. .
mammals, including humans.   These effects have been  observed
as a result of short-term exposures and  the intensity of the
response was dependent upon the dosage levels.  Salvini et
al»  (Brit.  J. Med.  1971. 28::293) observed psychophysiologTcal
changes in human volunteers in a controlled inhalation study
using trichloroethylene at  as low a level as 110 ppm for two
four-hour periods.

     Long-term exposures of mice to trichloroethylene produced
carcinogenic effects in both male and female animals (National
Cancer Institute, 1976).  In addition to the carcinogenic
effect,  trichloroethylene has been reported to be mutagenic
in microorganisms,  transforms cultured mammalian cells to
carcinogenic cells, and binds with tissue macromolecules,
thus supporting the carcinogenic potential of trichloroethylene,

-------
      There  has  been some  controversy  over  the  current evidence
 linking  trichloroethylene and  carcinogenicty in  animal
 studies.  Although  the  NCI bioassay was  positive, others
 have  argued that  the effects may  have been due to contaminants
 (epichlorohydrin  and epoxybutane)  in  the tested  trichloroethylene,
 NCI has  agreed  to retest.   The NAS in its  1979 report,
 however,  recognizing the  issue, accepted the NCI result and
 computed  a  risk value based upon  carcinogenic  potential.

      Recent studies on  the metabolism and  elimination of
 trichloroethylene in rats  and  human volunteers reveal that
 the metabolites of  trichloroethylene,  namely trichloroethanol
 and trichloroacetic acid,  are  not.  immediately  eliminated
 from  the  body.  Trichloroethanol was  found to  have a half-
 life  of  12  hours  in human  volunteers.  This would mean that
 repeated  daily  exposure to trichloroethylene via drinking
 water would result  in some accumulation  of trichloroethanol
 in the body.  Moreover, the metabolite trichloroacetic acid
 has been  reported to bind  to plasma proteins.  This property
 of trichloroacetic  acid may result in interaction with drugs
 and chemicals having similar properties, thereby resulting
 in toxic  effects.   (Ertle  £t^al. Arch. Toxicol.  29, 171-188,
 1972.)

 II.   SNARL  Development

      Trichloroethylene is  a carcinogen in  mice,  and also
 causes non-carcinogenic bioeffects.   One-day,  10-day and
 chronic SNARL values  based on  non-carcinogenic bioeffects
 are computed incorporating appropriate factors of safety.
Estimates of concentrations projected  to increase the lifetime
 cancer risk by  one  in 100,000  and  one  in a  1,000,000 are
 also  provided using  the NAS model..  The non-carcinogenic
 SNARL recommendations are  made considering the child and
other possibly  sensitive members of the population.

     Using a study where human volunteers were exposed via
 inhalation to 110 ppm (590 mg/m > of  trichloroethylene for
an 8-hour period where psychophysiological symptoms were
observed, a one-day  SNARL  value of 2 mg/1  could be calculated
for the child.

-------
      (590 n>8/m3)(8 m3/day)(0.30)       x J_ =  2.02 mg/1
      (1 I/day)(100 uncertainty factor)   7

 where:    1/7 = child/adult body weight ratio
           0.30 = absorption factor
           1 I/day = child daily water consumption
           100 = uncertainty factor via 10 factor because
           a human experiment was used and 10 factor
           because data did not specify the no observed
           adverse effect level

      To calculate a SNARL for 10 days metabolic and pharmacokinetics
 data are required.   Since that data is not available a
 conservative method would be to divide the one-day SNARL of
 2 mg/1 by 10 whereby the 10-day SNARL value would become
 approximatley 200 ug/1.

      Since the one-day and 10-day SNARL values  are determined
 for emergencies  and spills for a short period of time,  it
 should be assumed that drinking water would be  the primary
 or sole source of human intake of trichloroethylene.   This
 is in opposition to that for a chronic SNARL where a lesser
 contribution from drinking water may be appropriate.
 Therefore,  a relative  source contribution factor has  not
 been incorporated into the suggested one-day and 10-day
 SNARL values of  2 mg/1 and 0.2 mg/1,  respectively.

      The  NAS (1979) has  computed a  one-day  SNARL of  105 mg/1
 and 15  mg/1  for  the  seven-day  SNARL.   Their  calculations
 were  based upon  the  observation of  intoxication  of  adults
 and the application  of uncertainty  factors.   Our calculations,
 however, were  based upon psychophysiological  parameters  and
 extrapolated to  the child with the  appropriate uncertainty
 factors.

      The NAS chose to  work with uncontrolled  case histories
where trichloroethylene  was  accidentally  ingested.  The
 study which  the Office of  Drinking Water  chose to evaluate
and extrapolate, while being an inhalation study, was conducted
under controlled conditions.

     A  longer exposure SNARL for trichloroethylene, can be
calculated using a study by Kimmerle and Eben entitled
"Metabolism, Excretion and Toxicology of Trichloroethylene
after Inhalation."  This study evaluated the subacute exposure
to trichloroethylene via inhalation in adult rats for some

-------
 14 weeks following exposure to 55 ppm (300 mg/tn3),  five days
 a week.  Indicies of toxicity include hematological investigation,
 liver and renal function tests, blood glucose and organ/body
 weight ratios.  Liver weights were shown to be elevated
 while the other test values were not different from controls.
 The elevated liver weights could be interpreted to  be the
 result of hydropic changes or fatty accumulation.  The no-
 observed-effect level was not identified.

      Using the method of Olsen and Gehring (1976) whereby
 the lung-whole body ratios for humans (adults) and  rats
 (adults) are assumed to be roughly equivalent, the  total
 dose of trichloroethylene to the child can be determined  and
 a  longer term SNARL can be calculated to be approximately 75
 ug/1 when the principal source of trichloroethylene is
 assumed to be from drinking water.

      (300 mg/m3)    8 m3/day (5)(1)(0.30)     =   73.5 ug/1
      (1 I/day)              (7)(7) (1000)

 Where:

 55 opm (5.46)  = 300 mg/m  minimum effect level
 8m           = according to Olsen/Gehring
 5/7           = fraction converting  from 5  to 7-day exposure
 1/7          = child/adult  body  weight ratio
 0.30          = absorption  rate
 1  I/day       = child consumption per  day
 1000          = uncertainty  factor due to animal  study
                where minimal  effect was reported

      In cases where  other sources of exposure  are prevalent
 and,  for example, drinking water  Ls assumed to account  for
 a  portion of the total exposure,  say 20%, of  the  trichloroethylene
 intake, then the SNARL value would become 15 ug/1.  By-and-
 large, however, the  75 ug/1  SNARL would be assumed  to be
 appropriate under normal  circumstances in the absence of
 other major sources of TCE.

     A chronic SNARL  approximately equivalent to the SNARL
of 75 ug/1 can be justified on the basis that  (1) long-term
exposure to low doses of  trichloroethylene  probably does not

-------
bioaccumulate much more over a  lifetime than in 3-6 months,
and  (2)  the  SNARL was calculated for the child and not the
adult  thus providing a somewhat larger safety margin.

     Since trichloroethylene is considered a carcinogen, at
least  for mice, and using the risk estimates generated by
the  National Academy of Sciences (NAS), it is possible to
identify that range of trichloroethylene concentrations that
would  increase the risk of one excess cancer per 10  or
10   people exposed over a lifetime.  From the NAS model it
is estimated that consuming 2 I/day over a lifetime having a
trichloroethylene concentration of 4.5 ug/1 or 45 ug/1 would
increase the risk by one excess cancer/million exposed .or
one  excess cancer/100,000 exposed, respectively.  This is
the  range of risks where many EPA regulatory values for
other  carcinogens have been.

     These risk extrapolations were based on an assumption
that there is no threshold level for carcinogens.  The
state-of-the-art at the present time is such that no experimental
tools  can accurately define the absolute numbers of excess
cancer deaths attributable to trichloroethylene in drinking
water.  Due to the biological variability and a number of
assumptions required, each of the risk estimating procedures
lead to a different value.  There is wide variation among
these  estimates and also in their interpretation.  For this
reason we report the results of the NAS risk computations,
which  is a conservative approach, as a range of values in
the one in one hundred thousand to one in one million incremental
risk (risk above background) for a carcinogen.  The NAS risk
estimates are based on the multistage model concept.  "At
low dose the multistage model is often mathematically equivalent
to the linear or single hit model.  Therefore, its use for
extrapolation is consistent with the conservative linear
risk estimation.  If the precise mechanism of carcinogenesis
is represented by a threshold or log-normal dose response
relationship, the multistage model may considerably over
estimate the risk at low dose levels.  However, this possibility
cannot be reasonably quantified."  (NAS-1979)

-------