UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                      WASHINGTON, D.C. 30460
                                                        OFFICE OF
                                                 PESTICIDES ANO TOXIC SUBSTANCES
 MEMORANDUM
 SUBJECT:   Review of Data Available to the Administrator
           Concerning Formaldehyde and di(2-ethylhexyl)
           Phthalate (DEEP)
                                          *       _  /
 FROM:      John A.  Todhunter,  Ph.D. ^_^>vn /.{.  /^, .^..X.
           Assistant Administrator
             for Pesticides  and Toxic Substances (TS-788)

 TO:        Anne M.  Gorsuch
           The Adrainistrator;^-t?^JLQ'0)
                            \
 THRU:      John W.  Hernandez,
           Deputy Administrat


 Summary                     /
                            s
_At your direction I have reviewed the materials prepared by the
 Office of  Toxic Substances  in regard to formaldehyde and DEHP.   I
 have  also  reviewed the materials prepared early in 1981 by OTS
 which were appended to a letter from NRDC, dated 'September 4, 1981.
 This  memorandum is a statement of my professional conclusions and
 policy recommendations regarding the documents reviewed.   In sum-
 mary:  I find that I concur with the current recommendations
 of OTS that formaldehyde and  DEHP ought not be considered under
 section 4(f)  and that the preliminary drafts relied on by
 NRDC  in their letter to you were incomplete and flawed.

 Technical  and Policy Issues Related  to Section 4(f)

 The Toxic-Substances Control  Act (TSCA)  was enacted by Congress
 to protect human beings and the environment from those chemical
 substances and mixtures whose manufacture, processing/ distribution
 in commerce,  use,  or disposal pose unreasonable risks of  injury.
 One of the provisions of the  Act is  Section 4(f),  which states
 that:

      Upon  receipt of 1)  any test data required to be submitted
      under this Act,  or 2)  any other information available to
      the Administrator,  which indicates to the Administrator that
      there may be a reasonable basis to conclude that a chemical
      substance or mixture presents or will present a significant

-------
                              -2-

     risk of serious or widespread harm to human beings
     from cancer, gene mutations, or birth defects, the
     Administrator shall, within the 180-day period begin-
     ning on the date of the receipt of such data or informa-
     tion, initiate appropriate action under section 5, 6,
     or 7 to prevent or reduce to a sufficient extent such
     risk or publish in the FEDERAL REGISTER a finding that
     such risk is not unreasonable.  For good cause shown
     the Administrator may extend such period for an additional
     period of not more than 90 days.  The Administrator shall
     publish in the FEDERAL REGISTER notice of any such
     extension and the reasons therefor.

The purpose of section 4(f) is to identify certain chemical
substances or mixtures as high priority candidates for regulatory
action and establish a deadline for acting to regulate such
substances as are found to meet section 4(f)  criteria.  A chemical
could meet the section 4Cf) criteria when there may be a reasonable
basis to conclude that it presents a risk from one of three
effects—cancer, gene mutations or birth defects, and that the
risk of this effect is significant and either serious or wide-
spread.  None of these terms—significant, serious or widespread—
is precisely defined in TSCA.

The common technical use of these terms in the bio-medical
sciences would suggest that "significant" would address both
the probability that a risk of any magnitude could occur as
well as the magnitude of the risk; and that "serious" and "wide-
spread" would address the nature  (i.e., major vs. minor) of any
harm resulting and the size of the affected population, respectively,

For purposes of illustration, the following examples of usage
are possible:

"Significant" -

     A particular animal cancer bio-assay indicates that
     chemical compound X is a carcinogen in at least two
     species tested and all sexes in these species.  Compound
     X generally may present a risk of being a human
     carcinogen.  To be "significant" it should further be
     established that—at human exposure levels—the risk is
     probable and would be high.  A different chemical, compound
     Y, is positive in one species, one sex and at one high dose.
     It could, in theory, be a human carcinogen but it need
     not generally be considered to pose a "significant" risk
     to humans based solely on the one positive finding.
     Strong evidence for human risk should generally require
     that the carcinogenic effect be established as active in
     more than one species or at more than one dose.

-------
                         — 3 —

     If for compound X above, it were found that the
     carcinogenic effect is due to activation of a
     metabolic pathway not found in humans, then the
     risk from compound X could be deemed not "significant"
     as it would not occur in humans.

     The carcinogenic effects of compound X could also be
     found to be secondary to physiological events
     (deposition of microcallculi, hormonal changes,
     necrosis, etc.) triggered by compound X at levels
     used in bio-assays but not at levels to which humans
     are exposed.  Compound X would probably not pose a
     "significant" risk to humans.   (See footnote 1)

"Serious" -

     The concept of "serious" harm would have more utility
     in the consideration of gene mutational events or birth
     defects than in the case of cancer since malignant
     neoplasms in general are serious. Section 4(f) also
     draws a distinction between benign and malignant growths
     by use of the term "cancer" rather than "tumors".

"Widespread" -

     A teratogen could produce a minor anomaly.  IS this
     occurred at one in one hundred thousand births, it
     need not be necessarily "serious" or "widespread".
     The same outcome at one in one hundred births, however,
     would certainly be "widespread".

The determination that a chemical meets the criteria of
section 4(f) is, by statute, made by the Administrator of
EPA.

When a chemical meets the criteria for section 4(f) priority
status, the Agency must decide quickly whether to initiate
action.  The provision states that the Administrator must,
within 180 days of the Administrator's receipt of the information
(1)  These considerations as to weight of evidence to establish
human risk generally follow the recognized International Agency
for Research on Cancer (IARC) criteria (IARC, Internal Technical
Report No. 78/003) as to whether evidence is negative, limited,
or sufficient for human carcinogenicity as well as the amplifi-
cation of these criteria presented by Griesemer and Cueto
(Scientific Publication No. 27, IARC, Lyon; p. 259-281 (1930)).
The second paragraph draws from Weisburger and Williams
(in Toxicology, Casarett and Doull, Eds., 2nd edition, MacMillan,
N. Y. (1981)  as well as from a more recent paper of Squire
(Science, 214; 377-880 (1981)).

-------
                         -4-
indicating the risk, initiate appropriate action under  section  5,
6, or 7 of TSCA, or explain in the FEDERAL REGISTER why the
risk is not unreasonable.  As a purely technical matter, only
rarely would a single study provide a reasonable basis  to
conclude that a chemical presented or would present a
significant human risk.  Risk is a function of toxicity and
exposure.  Therefore, exposure evaluations (which generally
must rely on other studies than those raising toxicological
concerns) are required in order to properly quantitate  risk.
The probability that risk would indeed occur can in general,
likewise, only be assessed by review of all the available
literature on a given chemical.  Only when the diverse
pieces of information which must go into a proper risk  assess-
ment are assembled into a reasoned evaluation can there generally
be any scientific basis on which to draw reasonable conclusions
as to the nature of human risk posed by a given chemical.

In certain instances the available data on a chemical may sug-
gest that it poses a human risk but be insufficient to
reasonably support a conclusion that the chemical presents a
significant risk of serious or widespread harm [i.e., similar
to a section 4(f) determination] and/or an unreasonable risk
[a determination called for in other parts of TSCA].  The
distinction between chemicals which may pose a human risk and
those which pose either a significant and/or unreasonable
risk [applicable to section 4Cf) and/or section 5, 6, or 7,
respectively] is important from a policy standpoint.  This
distinction and the TSCA data collection authorities allow
the Agency to deal with chemicals on which the data are sug-
gestive but inconclusive by use of sections 4(a), 5(e),  and/or
8 so as to better develop a data base.  This distinction also
allows the Agency to place less priority on such substances
as may be found to pose less than significant human risk even
though the Agency may want to retain some oversight of
activities involving such a substance.

Applicability of TSCA Section 4(f);

The principle decision before the Administrator at present
is whether or not the data available on formaldehyde meet
the criteria set out in section 4(f)  of TSCA.  The technical
and policy arguments that these criteria are not met are
presented for each substance in this report in greater
detail below.  Legal conclusions as to section 4(f) are not
within my realm of expertise. In this report I have tried to
define the technical practicalities of section 4(f) and to
address the relevant policy issues.

Generally, the approach to section 4(f)  that is presented is
that, until the Administrator has reviewed the available data
and reached a conclusion as to the significance of risk,
any designation of data as having triggered section 4(f) is

-------
scientifically arbitrary.  This has tremendous practical and
scientific underpinings as, in general, no one study by
itself will be sufficient to indicate  "significant" risk
and the required determination that "there may be a reasonable
basis to conclude that a chemical substance or mixture
presents or will present a significant risk ..." should,
therefore, require a synthesis of existing and newly developed
knowledge.  Only when this synthesis is accomplished can a
reasoned conclusion, in general, be reached.

As a policy matter, it is recommended  that risk situations
which are "significant" in the context of priority setting
via section 4(f) be clearly distinct from risk situations which
are generally not of immediate regulatory interest.  In terms
of individual lifetime cancer risks, the_various federal agencies
do not tend to regulate risks of 1 x 10   or lower aod tend to
be ambivalent about risks between 1 x  10~  - 1 x 10"°.  Certainly
(as absolute risks) these risk levels  could never be detected
any normal way and would  (using 1 x 10   as an upper bound)
represent increments of 0.03% or less  above the estimated
individual risk of persons in the,U.S."population as a whole.
(This risk is 1 in 3 or 3.3 x 10   according to the most recent
National Cancer Jnstitute figures.)  In OPTS the relative risk
range of 1 x 10~' to 1 x 10   or lower has been a low concern
range in general.   The use of a range  is appropriate since
these risks, even at the 95% upper confidence limit, are only
relative estimates for specific chemicals.  To use such risk
estimates as absolutes is inappropriate.  This point was
recently emphasized by Dr. David Rail, Director of the National
Toxicology Program, in testimony before P.ep. Gore during the
1981 NTP appropriations hearings.

FORMALDEHYDE;

Summary

Technical and policy discussion of these points is to be
found in the body of this report.

Conclusions With Respect to Formaldehyde:

I concur with the recommendations of OTS staff that formaldehyde
does not meet the criteria of section  4(f) for the following
reasons:

(A)  There is suggestive evidence that there may be human
     exposure situations—based on the integrated animal data
     base—which may not present carcinogenic risk which is of
     significance.

-------
                                  — o —
(B)   There is a limited but suggestive epidemiological base
     which supports the notion that any human problems with
     formaldehyde carcinogenicity may be of low incidence
     or undetectable.  It would not appear reasonable to
     say that a significant risk situation exists from these
     data.

(C)   Assuming that risk—of some magnitude—does exist the
     calculated relative risks tend to fall into ranges which
     normally are of from low priority to no concern such as
     those discussed in the section immediately above on the
     applicability of TSCA section 4(f).

(D)   One can, therefore (and as amplified below) conclude
     on the basis of available toxicological data, exposure
     information, and risk analysis techniques that:  (a)
     formaldehyde is a carcinogen in the rat by the inhalation
     route;  (b) its carcinogenic potential appears to vary
     significantly with species and route; (c) under certain
     exposure conditions it could present some carcinogenic
     risk to humans; and (d)  given available data the risk
     estimates suggest that certain populations may experience a
     carcinogenic risk—albeit low—due to formaldehyde
     exposure.  However, because of the nature of the
     toxicology data and the unreliability in the exposure
     data one cannot reasonably conclude, at this time, that
     formaldehyde poses a significant risk among the U.S.
     population.

Interpretation of Animal Studies with Respect to_Human Risk —
Of a variety of bioassays (Table I) done since 1954 to
assess the carcinogenic potential of formaldehyde only two
have given clear positive results attributable to formaldehyde:
a Chemical Industry Institute of Toxicology Study (CUT)
reported in preliminary form by Swenberg et al, Cancer
Research, 40, 3398  (1980)  and a study underway at the New
York University Medical Center which is an extension of an
earlier New York University study of 1979.  Both of these
have been inhalation studies in the rat.  The CUT study
used three dose levels plus a control, the New York University
study used a single high dose equivalent to the highest CUT
dose.  Only nasal carcinomas were found in either study.
Approximate mean lifetime incidences were 0% in control
rats; 0% in rats at 2.1 ppm formaldehyde; 1.9% in rats at
5.6 ppm  (not statistically significant); and 50% in rats at
14.1 ppm in the Swenberg study.  The New York University
study indicates  (by letter of August 17, 1981 from Dr. Arthur
Upton) an incidence of about 10% in rats at 14 ppm of formaldehyde,

-------
Table I:  Studies Considered in Evaluating the Carcinogenic
          Potential of Formaldehyde
Species
Rat, Q & cf,
Fischer 3_44

Rat, 0*
Sprague-
Dawley

Rat, 0?
Sorsgue-
D aw lev
!
Mouse, G*
* 0,
36<53FT

Hamster,
0?
Syrian
Rat
Rat
Rat, 0 + Q
BD
Mice (CTM,
SWR, and C3Hf)
and Rat (Wistar)
Route
inhal.

inhal .


inhal .


inhal .

inhal .
a
s . c. —
s. c.
oral
gavage
drinking
water
Dose
0, ppm
2 , 1 ppm
5 . 6 ppra
14.1 ppra
14 ppm


1'4 ppm-i-
10 oom HCI


0 ppm
2 . 1 ppm
5 . 6 ppm
14.1 ppm
10 ppm
30 ppm
formalin,
0.4%, 1 mJ
per anima]
weekly
HMT,-
40%. 2 ml
oer animal
per week
HMT,
0 . 4 g/day
HMT,
2 . 5- g/Kg/d
n
Length
24 mos.

lifetime
(19 mos.
conrolet-
ed)
lifetime-
(27 mos . ]


24 mos.

lifetime
15 mos.
until
tumor
develop-
ment
12 mos.

1 yr
2 yr
%
Incidence
(0%)
(0%)
(1.9%)
'50%)
(-10%)
"nasal
tumors

(27%)
nasaJL
tumors

- (0%)
- (0%)
- (0%)
(2%)
- (0%)
. - (0%)
(40%)
(40%)
(0%)

(0%)
(0%)
Reference
CUT study of
Swenberg et al,
nasal tumors onlv

On going NYU
study reported by
Uoton letter of
Aug. 17, 1981
NYU study of
Laskin et al


CUT study of
Swenberg et al,
nasal tumors onlv

Nettesheim, NCI -
DOE study, 1977
Watanabe e_t al,
Gann, 45 , 451
(1354) sarcomas
at injection
site only.
Watanabe e_t al,
gann , 46, 365
1955 sarcomas at
injection site.
Brendel, Drug Res.,
14_, 51 (1964)
Delia Porte et al,
Food Cosmetic Tox. ,
_6, 707 (1968)-

-------
Table I (continued)

Notes to Table:

     a - sc = subcutaneous injection

     b - HMT = hexamethylene tetramine

         which decomposes  in  aqeous solution to give  formaldehyde
         and  ammonia with  the stoichiometry:

               1 HMT €=±  4 NH,  +  6 CH_0
                               -}         £

         One gram of HMT is equivalent to 1.3 gram of free
         formaldehyde.

     £ - Delia Porte at al, Tumori, 56, 325  (1970) also report a
         3 generation study on Wistar rats with 2% KMT in drinking
         water which was negative for all three generations.  This
         level of HMT is equivalent to about 2.5 g/Kg/d of ingested
         HMT which would be 3.3 g/Kg/d of formaldehyde.

-------
                             -7-

The two studies used different strains of rat but are  in
general agreement.  The CUT study also exposed mice to
formaldehyde gas but no statistically significant increase
in tumors was observed.   CSee footnote 2.)

Other studies (some in other species and/or by other routes)
have been generally negative save for a rat study of 1954
(expanded in 1955} using a subcutaneous injection route for
0.4% formalin or up to 40% hexamethylene tetramine  (HMT, which
gives rise to free formaldehyde, 1 gram HMT =1.3 gram formaldehyde)
In this study only sarcomas at the site of injection were
found and these studies are not, therefore, generally  accepted
as being other than suggestive evidence of carcinogenecity.
Significantly, the oral administration of hexamethylene
tetramine to rats at up to 12.5 g/Kg/week was found to
produce no tumors in three different investigations of up to
two years duration (the same length as the CUT inhalation study)
conducted in 1964, 1968, and 1970.  The oral route  studies
are somewhat limited fay small sample sizes but, taking all
three studies in aggregate, there is evidence which suggests
a lack of a carcinogenic effect by this route. An important
observation is that inhalation by hamsters at similar  or
higher exposure levels than used in the CUT study  has given
a negative result.  Formaldehyde appears, therefore, to
exhibit considerable species specificity with the rat,
the most sensitive of species so far tested.

Concern that formaldehyde gas may induce tumors in  humans
should be tempered by this observation that formaldehyde
carcinogenicity appears to have a high degree of species specifi-
city and a strong dependence on route of exposure.  The rat,
an obligatory nose breather, appears to have a particular
nasal sensitivity to the deleterious effects of formaldehyde
which may not be generally shared by other animals  (man is
not an obligate nose breather)  or body sites.  Based on the
animal experimentation data, the most reasonable interpretation
is that formaldehyde has carcinogenic potential, is probably
locally acting, and is capable of both initiation and  promotion.
(2)  The statistical significance of the CUT results at 5.6
ppm in the rat  (1.9% incidence observed) and 14 ppm in the
mouse  (2% incidence observed) was determined by CUT as described
in the CUT document "Data Released to the International Agency
for Research on Cancer  (.IARC) Working Group", dated October 19,
1981, which is appended as Appendix A.

-------
                              -8-

The results of the CUT bioassay are, however,  sufficient  to
establish that formaldehyde is a potential animal carcinogen
with mode and degree of exposure quite important to the toxic
outcome.  Based on the observed differences in  response
shown by rats, mice, and hamsters, different species have
differing sensitivities.  This may be a quantitative rather
than a qualitative difference but"the negative  results of
inhalation studies with Syrian hamsters at up to 30 ppm
formaldehyde suggest that there may be qualitative differences
in response among species as well.  It cau, however, be
prudently expected that under appropriate conditions of
exposure formaldehyde gas may pose a cancer risk to the
respiratory tract of humans (i.e.: nasal and buccal cavities,
trachea, bronchial trees, and lungs, see footnote 3) .  In
the CUT bioassay (see Appendix A) , the development of
nasal carcinomas was always preceded by hyperplasia, metaplasia
and overt signs of cytotoxicity to the nasal mucosa (simply
put, the animals developed open sores in their  nasal tissues
after which tumors appeared).   In other cases,  with other
chemicals, where tumors appear to be secondary  to cytotoxicity,
no observable effect levels of exposure have been postulated
(see, for example, VJeisburger, J. H. and Williams, G. M. in
Casarett and Doull's Toxicology, 2nd ed., MacMillan, NY
(1981)).  Since formaldehyde has genotoxic properties, any
simple analysis is complicated but the findings of the CUT
research tean which supported the bioassay have bearing on
the interpretation of the results of the bioassay:

     (1)  At low exposure levels or short exposure times the
hyperplastic and metaplastic effects of formaldehyde exposure
were reversible.  This suggests that a higher exposure level or
longer exposure time may produce a qualitatively different
outcome with respect to carcinogenesis than would result from
low level or short term exposure.

     (2)  Normal endogenous tissue levels of formaldehyde
produced by metabolism range from 3-12 Ng/g tissue  (3-12 ppm).
Much of this appears to be bound up as adducts  to glutathione,
tetrahydrofolate, and proteins.  From 10%-40% of the total tissue
formaldehyde appears to exist as free formaldehyde  (0.3 Ng/g -
4.8 Ng/g).  A sustained exposure of 3 ppm formaldehyde in
air is needed to double the endogenous level of formaldehyde.
 (3)  Other body sites are theoretically possible targets but,
due to a variety of studies demonstrating rapid metabolic conversion
of formaldehyde at or near the sites of uptake  (Maloney, 6, et al;
Arch. Exp. Pharmacol., 250; 419-436  (1965); McMartin, K.E.,
Biochem. Pharmacol., 28; 645-649  (1979)) it is expected that the
respiratory tract will be the principal, possibly sole, target
organ at risk.  A high degree of  site concordance between humans
and animals has been suggested by Tomatis  (L. Tomatis et al,
Cancer Research, 38; 877 (1978)).

-------
     (3)   There appears to be an absence of cytotoxic
effects of formaldehyde at or below 1.0 ppm in air.

     (4)   Radio-isotope tracer measurements indicate that
absorption of formaldehyde gas from imspired air is essentially
complete within the nasal cavities and formaldehyde gas does
not reach the trachea, bronchi or lungs of rats at the exposures
used in the CUT study.

     (5)   Formaldehyde is a mutagen.  As such it is a potential
genetoxic carcinogen.  Aldehydes, as a chemical class, are
also promoters of carcinogenesis (Slaga, T. J., Sivak, A.;
and Boutwell, R. K.; Mechanisms of Tumor Promotion and
Cocarcinogenesis, Raven Press, NY,  1978).  Formaldehyde can,
then,potentially both initiate and promote the carcinogenic
process.   A promoting and/or cocarcinogenic effect of formalde-
hyde is supported by other studies cited in Table I, in
particular the various New York University studies.  Promotion
generally requires repeated, long term exposure to the promoter
and is reversible [(Weisburger, J. H. and Williams, G. M. ;
"Chemical Carcinogenesis" in Holland, J. F. and Frei, E.  (eds):
Cancer Medicine, 2nd ed., Lea and Febiger, Philadelphia,
1980)],  may involve damage to cellular components  (Slaga et al
(supra),  1978), and is a requisite, at some point, to the
carcinogenic process.  When taken together with the observations
reported by CIIT on the reversibility of hyperplastic and
metaplastic effects  (vide  (1) supra) and an apparent cyto-
toxicity threshold  (vide 3) supra)  the degree to which
formaldehyde promotes  its own carcinogenesis may  lead to
qualitative, as well as quantitative, differences in
response within different regions of the dose-response curve.

An argument can be made that any initiating effects of
formaldehyde would, nonetheless, produce a population of
latent malignant cells.  These cells could then produce
malignancies in response to other promoters  (endogenous or
exogenous) as the animals grew older.  In such a case the
difference in response to formaldehyde exposure at different
exposure levels would be only quantitative and not qualitative.
This argument is a classical one (See Weisburger and Williams,
1981,  supra) and has some merit as a general argument.  It
needs  to be discussed, however, in the context of formaldehyde
as it  affects the nasal mucosa.

Clearly, in the absence of detoxifying mechanisms, the
endogenous levels of formaldehyde should be potentially
capable of initiating  some cells.  The observed absence of
nasal  carcinomas in rodents as spontaneous tumors  (for
example, Dr. A. Upton communicates by letter of August 17,
1981,  that in 14 years with 1,920 rats he has not observed
a single grossly visible nasal carcinoma) indicates one of

-------
                             -10-

four things:  (a) no cells are in fact initiated by endogenous
levels of formaldehyde or repair processes at low exposure
are faster than processes leading to cell transformation;
(b) there are not endogenous promoters available to allow
expression of the transformed cells; (c)  any cells initiated
are lost before their malignant potential can be expressed
or  (d) the tumor incidence is so low as to be undetectable.

The first situation would de facto, define a practical
threshold for genotoxicity.   Formaldehyde is an endogenous
metabolite and it is possible that mammalian cells may deal
effectively with low levels of it.

The second situation could suggest that promotion by exogenous
factors  (formaldehyde could be one) was the controlling
element  in formaldehyde carcinogenesis.  This would imply a
lack of  carcinogenic effects at exposures to exogenous factors
which did not result in promotion.  The observation, by CUT,
that tumor formation could be secondary to a formaldehyde induced
cytotoxicity, hyperplasia, and metaplasia does suggest that
the promoting effects of formaldehyde may be quite important.
At  low exposure levels the hyperplastic,  metaphastic, and
cytotoxic effects observed were reversible in the CUT study.

The third situation is probably relevant to the question of
formaldehyde carcinogenesis.  Normally, the epithelia of the
nasal mucosa loses matured cells which are then daily replaced
by  underlying stem cell layers.  Thus, cells which might be
initiated by formaldehyde could be lost from the epithelia
before their malignant potential could be expressed.  This
possibility has not however, been addressed experimentally.

The fourth consideration may have some bearing but in a
variety  of studies  (CUT and those of Dr. Upton's colleagues
at  NYU)  over 2,000 rats have been observed as a historical
control.  If nasal carcinoma occurs spontaneously in rats,
its incidence would be very low.   (See footnote 4)

The observed lack or low incidence of spontaneous nasal
carcinoma in rats—given the endogenous levels of formaldehyde
present—is probably explained by a combination of the above
 (4)  One rat developing a tumor per 2,000 animals would be an
 incidence of 5 x 10   .  This is not the same as the level of
 risk to rat or, possibly, human but suggests that individual
 risk from endogenous  tissue levels of formaldehyde might be
 bounded by numbers of sinilar magnitude.

-------
                              -11-


considerations.  The net result is that there would appear
to exist some lower bound to the observable carcinogenic
effect of formaldehyde resulting from exposure to exogenous
sources.

As such, the above considerations regarding endogenous
levels of formaldehyde coupled with observations of CUT of
the reversibility of-'formaldehyde related effects at low
exposure levels suggest that low level exposures  (
-------
TABLE 13:  FREQUENCY  DISTRIBUTION OF EXPOSURE ESTIMATES IN
          FORMALDEHYDE PRL-1 TECHNICAL DOCUMENT
          These  data  are extracted from the integrated exposure
          assessment  in the PRL-1.  Only estimates  given for
          populations of a known size are considered.   High
          and  low  refer to the upper and lower bounds  of
          estimated exposure ranges.
INTERVAL  (ppm)
Number of estimated  exposed individuals

0 -
0 1
0.5
1.0
2.0
5.0
O.lb
- 0.5
--1.0
- 2.0
-5.0


2
7
3
3
1
-i*

Low
.24 x
.09 x
. O'a x
.04 x
.24 x
-
{Percentage of
(Total -Indivi-
duals Exposed
108
104
ru4-
104
106

99
0.
0.
0.
o..
-
.4a
03
01
01
55

High
2.20
4.50

5.89
3.86
1.26
x 108
x 104
-
x 104
x 106
x 106
Percentage
Individuals
97.
0.

0.
1.
0.
6a
02

03
71
56
of Total
Exposed






a/ % of total  individuals exposed

b/ includes  U.S.  population via ambient air

c/ 97% are college students who on an average are exposed to an
  estimated 8.3 ppm for very short periods and over a limited time.
  Humans cannot   tolerate formaldehyde exposures  above 5 ppm and
  so this group  is probably exposed to far less than the PRL-1 average
  estimate.
      are  mobile home residents at an average  0.4  ppm;
  40% are  U.F.  home residents at an average of  0.72  ppm
e/95% are  the  same college students as in footnote  c.

-------
                               -12-

Humans do not adapt to levels of airborne  formaldehyde over
5 ppm and levels from 2-5 ppm are generally very unpleasant.
The "high" side of the PRL-1 exposure ranges should,  then,
be viewed critically as this suggests a very sizeable group
is exposed - on a regular basis - to levels of formaldehyde
above 2 ppm.  It must be noted that for the group footnoted
as (d) in Table II, the "high" value derives from short
transients in complaint homes and is non-representative.
The appropriate averages are indicated in  the footnote.  The
exposure group footnoted as (.e) in Table II, similarly is
constituted by using a single excursion, transient as an upper
bound on the range of exposure.  The footnote (c) explains
that PRL-1 assigned these college biology  students an average
exposure of 8.3 ppm several hours a week.  Clearly, this
number is not to be relied on since humans cannot breathe
in such an environment.  The column listed under "low" is to
be taken as more representative (with the  possible exception of
the college biology students).  Table II suggests, then, that
nearly all prolonged human exposures are at levels well below
2 ppm.

Given that quantitative and possibly qualitative results of
exposure to formaldehyde appear to depend  highly on exposure
level, species, and route; that rats seem  to be particularly
sensitive to formaldehyde; and that long human experience
does not seem to indicate any pressing concerns; I must concur
with OTS' staff recommendations in this matter - that formaldehyde
ought not be considered subject to TSCA section 4{f) at .this
time.

Deficiencies in the Original OTS PRL-1 Document:

The original review of formaldehyde (PRL-1) was deficient in
a number of areas -

(1)  Exposure Estimates:  This was the major deficiency in the
original document.  In many situations the exposures given were
not obtained by analytical methods similar to that used to monitor
formaldehyde levels in the CUT study.  No attempt was made
to address the comparability of methods so it is not clear
that 1 ppm by any given method equates to  1 ppm exposure under
the CUT protocol.  This has the effect of making any subsequent
risk analysis (which must be exposure based) an "apples and
oranges" comparison.  Table III points out the various
methodologies cited in studies covered by  the PRL-1 document.

-------
TABLE III:   Number of Exposure Estimates by Type of Analytical
or Collection

Method in Formaldehvde

METHOD NUMBER OF ESTIMATES
Ac ety lace tone
Bisulfite impingers
Chromotropic acid
Chemi luminescence
Col crime try
Charcoal tubes
Solid sorbents
Draeger Tubes
FT Spectroscopy
Gas Chromatography
Ion Chromatography
MBTH Method.
Spec tropho tome try
Not Known
1
18
17
3
7
9
1
3
2
1
7
1
3
14
PRL-1


BASED ON METHOD3
as % of
1.
20.
19.
3.
8.
10.
1.
3.
2.
1.
8.
1.
3.
16.
all studies
1%
7%
5%
4%
0%
3%
1%
4%
3%
1%
0%
1%
4%
1%
aNumber of studies in data base which cite a particular method.
 Some studies cite more that one method.

-------
                              -13-


The PRL-1 review also used the highest values of  formaldehyde
found in any given monitoring study as its exposure levels  for
risk analysis.  This is not well advised since these high levels
are in most cases transients or single excursion  values  (i.e.:
not continuous) and do not, therefore, represent  the risk
picture adequately.  The tendency of the PRL-1 to use exposure
levels greater than 2 ppm when calculating exposure to some
populations is flawed as well since humans do not well tolerate
exposures to formaldehyde which are higher than around 2 ppm.
Even though some adaptation can occur, this is unknown above
5 ppm yet PRL-1 uses exposure levels above 5 ppm  as prolonged
exposures.

The poor quality of the exposure estimates is repeatedly
acknowledged in the PRL-1 and forms the basis for the recommenda-
tions contained in the earlier draft Federal Register notice
and action memorandum (see below).  As such, any  risk estimates
given in the PRL-1 must be viewed very skeptically.  It  is
not clear that such estimates would be a reasonable basis to
do other than conclude that one needed better exposure data.
This, in effect, is what OTS staff originally concluded  and
is in the earlier Federal Register notice draft of which
NRDC has a copy.

(2)  Evidence for Carcinoqenecity:  The PRL-1 concludes  that
formaldehyde is an animal carcinogen.  It downplays a number
of negative bioassays which suggest that its effects may depend
highly on species, route and site.  No attempt was made,
also, to address the question of mechanism of action or
other physiological/biochemical questions relevant to the
extrapolation from rat to human even though such  information
was available from CUT.

The discussion of human epidemiology emphasized studies which
showed marginal differences in total cancer rates  (i.e.,
the treatment of one study in which one cohort had a slight
increase in total cancer tate relative to historical controls
but a second cohort showed no such effect.  Much  was made of
the "positive" cohort but the negative cohort was downplayed).
No trend was seen toward increases in any specific cancer
(as is common for human chemical carcinogens) in  any study.
The human data do not show signs (such as correlation to a
specific tumor) usually seen for a human carcinogen and do
not, as is discussed above, generally suggest any specific
effect of formaldehyde.

-------
                              -14-

Risk Assessment;

The PRL-1 document contained a risk assessment performed by
use of the Crump multistage model and based on the high
exposure estimates presented in the PRL-1 document.  OTS
staff has also calculated a risk analysis based on median
(more representative) exposure values.  Relative risks are
calculated and expressed as 95% upper confidence limits on
lifetime risks.

Table IV presents a population distribution of relative
individual risks as calculated by OTS staff in PRL-1 using
representative  (.median or average) exposure estimates.
Risk is divided into order of magnitude ranges and the size
of the affected population falling into each range is given.
As can_be seen nearly all individual risks fall into the range
1 x 10~  - 1 x 10  .   This places them into a range in which
priority action is often not considered.  A similar situation
exists for populations of unspecified size (Table V).  Again,
if the. human risk were real, the magnitudes of individual
risks do not seem to compel a "fast track" approach.  This
seems to have been recognized in the draft FR notice prepared
in 1981 by OTS under Dr. Muir.  That notice proposed to engage
in a two year effort to collect exposure data so as to refine
the risk calculations.  This stemmed from the poor quality of
the exposure analysis in PRL-1 and the- low relative magnitude
of the estimated individual risks.

The risk calculation used by OTS included the 5.6 ppm exposure
level as a non-zero incidence data point.  CUT has since
determined this point to be statistically no different than
the 0 ppm and 2.1 ppm exposure results  (see Appendix A).
If this 5.6 ppm exposure were treated as a zero response
point, the risk estimates of Tables IV and V would shift, in
distribution, to yet lower values.

The risk estimates in PRL-1, also, do not take into account
that tumor formation in rats appears to require a long term,
high level exposure and that the human exposures are likely
to be relatively short.  Workers might see 1 ppm or less inter-
mittently over a 40 year work career assuming no change in
occupation (see footnote 5).  College students in biology
 (5)  The OSHA standard for formaldehyde is 3 ppm but OSHA
documents indicate that workplace exposures are generally
less than 1.5 ppm.  OPTS staff agree that this is an
appropriate and representative number as a general upper bound
on workplace exposures.

-------
       TABLE IV:  DISTRIBUTION OF RELATIVE INDIVIDUAL LIFETIME
                  RISKS AS DERIVED FROM PRL-1 FOR FORMALDEHYDE
Range of Risk


10~6 - 10~4 (0.003)b
-d -3
10 ' - 10 (0.15)
-1 -~>
Population Sizea


2.14 x 108
5
1. 52 x 10
% of Total
A

99.9

0.071
4 !
110 - 10 ~ (1.5) : 3.13 x 10 • j 0.015
-2
10 (3)

0


B
i
95.0

4.1

0.9 \


A:  Total of all identified populations

3:  Excluding ambient air and water exposures
aThis table applies only to exposed populations for which
 a size was estimated in PRL-1

'°NCI data indicate that individuals of the U.S. population have
 an average individual lifetime risk of contracting cancer of
 3.3 x 10~-> the number in parenthesis is the % increment
 above this average represented by a risk in the center of each
 indicated range save for the last range in which case the %
 increment is that due to the boundary risk.

-------
TABLE V:  DISTRIBUTION OF GROUPS WITH UNSPECIFIED POPULATION
          SIZE BY RELATIVE RISK INTERVALS
          Both maximum likelihood and 95% upper confidence
          estimates of relative individual lifetime risk are
          shown.  For comparison, according to recent National
          Cancer Institute figures the absolute risk of contracting
          cancer over a lifetime is approximately 1/3 or 3.3 x  10
          For each risk interval the percentage increase in lifetime
          risk  (assuming the risk estimates are absolutes which
          they are not) is presented.
Interval
range
<10~5
10~5 - 1CT4
10~4 - 10~3
10~3 - 10~2
% incremental
0.003%
0.03%
0.3%
3%
Number of size
Undetermined groups
MLEb j ULOCC
Maximum likelihood ppper Limit of Confidence
estimate
6
•>
2
0
0
7
3
ld
a/ % increment above NCI determined lifetime risk of 3.3 x 10

b Maximum likelihood estimate
                                                            _ 1
c/ Upper limit of 95% confidence


 /Represents one group with a 20 fold range in estimated exposures

-------
                              -15-

labs are, similarly, a large group  (1.2 million individuals)
which the PRL-1 assumes are exposed over 70 years.  In fact,
exposure of nearly all of these students would be generally
on the order of one year or less.

Another factor not considered in the PRL-1 risk estimates is
that the animal evidence would suggest that inhalation of
formaldehyde by humans should produce tumors  (if any are
seen) primarily of the oral, pharyngeal and nasal mucosa
(see footnote 3).  These are well defined sites for which
the limited epidemiological evidence does not suggest an
association with formaldehyde exposure under what are poorly
documented but probably representative conditions.

In summary, the PRL-1 does serve to identify some possible
areas of concern.  From a technical standpoint it would appear
to be premature to "fast-track" any regulatory actions since
we have estimates only on the upper bound of any risks and
these do not, when reasonable assumptions are made, appear
to be significant.  The draft FR notice  (discussed below)
which was prepared by OTS on the basis of PRL-1 review
acknowledged as much by specifying no regulatory actions per  se.
The proposed response was, rather, to conduct monitoring surveys
of the actual exposure conditions so as to better quantitate
the risk.  This de facto recognizes that the PRL-1 numbers
were highly unreliable and that a proper risk assessment had
yet to be done.  PRL-1 does not seem to sustain a reasonable
conclusion that formaldehyde presents a significant risk of
cancer to humans.

Deficiencies in the draft Federal Register notice on formaldehyde;

In 1981 OTS drafted a proposed Federal Register notice
on formaldehyde for the signature of the Administrator.

This notice presented a logical sequence and I have analyzed
the rationale employed as follows:

     (A)  Formaldehyde is an animal carcinogen;

          therefore,

     (B)  it is a potential human carcinogen.

     (C)  Formaldehyde is a high volume chemical;

          therefore,

     (D)  it presents wide spread exposure potential.

     (E)  Some individual exposures are estimated to be high/-

          therefore,

-------
                              -16-


(F)  some exposures are serious.

     Given (B),   (D) and (F), then

     (G)  Formaldehyde is a potential human carcinogen that
          presents serious or widespread exposures;

          therefore,

     (H)  formaldehyde meets the section 4(f) criteria for
          "significant" risk of "serious" or "widespread" harm.

There are a number of flaws in this line of reasoning.  Some
are explicitly acknowledged in the draft FR notice but
others are more subtle and need to be addressed first:

(1)  The meaning of the term "significant" in the draft FR
notice is not consistent with the common technical meaning
as discussed above.  No case is made for other than formalde-
hyde being a potential or possible human carcinogen.  The
degree to which this may or may not be an actuality (i.e.,
the significance of this risk in terms of probability of
existence and magnitude if it exists—as discussed under
Technical Issues above) is not addressed.  Therefore the
draft FR notice uses the words "significant" and "potential"
synonymously.

(2)  The assumption that a high production volume equates
with a high or "widespread" exposure potential is very weak.
In general, the degree of correlation between production
volume and exposure will depend on uses and handling methods.
A blanket assumption, as used in the FR draft notice, is
not justifiable in the absence of supporting evidence.

(3)  The conclusion that some individual exposures are "serious1
is premised on faulty assumptions as to exposure levels as
previously discussed.

The plain language of the statute calls for initiation of
regulatory action under sections-5, 6 or 7 as the response
to be made under section 4(f).  The proposed gathering of
exposure data—the response given by the FR draft—appears
to be an acknowledgement that the significance of any risk
from formaldehyde has not been reasonably evaluated since,
as a result of gathering this exposure data, it is further
proposed to evaluate risk to see if further action is
appropriate.   The premises and conclusions in the FR draft
appear to be faulty.

-------
DEHP;

No detailed analysis is presented of the DEKP situation since
work culminating in the PRL-1 and draft Federal Register
documents was based on incorrect exposure assumptions.  The
policy analysis given in the section on formaldehyde applies to
the DEHP case.

For purposes of discussion and comparison, the "risk estimates  in
the PRL-1 document are given and then discussed briefly:

Table:  PRL-1 Risk Estimates
Group
hemodialysis patients
(45,000)
hemophiliacs
(19,000)
U.S. population via food
70 yr Individaul
Risks
1 X 10~3
1 X 10~4
7.5 X 10"7
Potential Cases
Over 70 years
45
1.9
188
   (250 million)

U.S. population via drinking           -
  water  (250 million)           6 X 10                   150

 Occupational  (as in PRL-1
  originally)                          4
  Manufacturing (335)           1 X 10~4                 0.03
  Processing  (40,000)           1 X 10~                    4
 Occupational  (as revised by
OTS)
Manu
Processing (40,000)           1 X 10                  0.04
  Manufacturing  (335)           1 X 10 i               0.004
                                      ~°
a    Assumes 95% upper confidence limit is an absolute risk,
     it is in fact relative and an upper bound on any relative
     risk present.

b    Revised on the basis of exposure data presented at the 1980
     pthalates conference.


It must be pointed out that the first two groups in the Table are
not affected by TSCA since their exposure is a result of contact
with medical devices under the purview of FDA.  These are also groups

-------
which are at high risk from their diseases and the benefits
of these medical devices should probably far outweigh the risks
indicated.  In any case, they cannot be factored into any
analysis of TSCA applicability.

Estimates for the dietary risks to the U.S. population as a
whole fall below the range in which regulatory agencies do
not take any actions.  The exposures were estimated as described
in PRL-1 and are of good quality, in my opinion.

The drinking water estimate is the average number contained in
PRL-1.  I believe it to be representative and, again, the risks
are below the range in which we would normally not act.

The original Federal Register document was driven largely by
the occupational risks given in the table.  The risks begin to
push up into the range in which we would normally consider some
action  (regulatory or non-regulatory)  but are, nonetheless, lower
than what might push a priority designation.  As a result of
the 1980 phthalates conference, OTS staff realized that the
exposure estimates which had been used to generate these risks
were completely incorrect.  In essence they postulated a DEHP
concentration (as vapor in air) higher than the vapor pressure
of DEHP could generate.

The exposures and associated, relative risks were recalculated and
are given in the PRL-1 Risk Estimates Table as revised occupational
risks.  With the revision of the occupational exposure numbers
reason to take any action with respect to DEHP disappeared.

There has not, since the detection of the errors in occupational
exposures, been any reasonable basis upon which DEHP could be-
considered to pose significant risk to humans - given current
use patterns - and I concur with OTS staff that DEHP cannot be
recommended to you as a candidate for section 4(f) designation.

As a final comment on DEHP, there are enough anomalies in the
result of the NCI bioassay that I remain unpersuaded that DEHP
should be considered a human carcinogen.

-------
                Appendix A
                 DATA RELEASED TO

THE INTERNATIONAL AGENCY FOR RESEARCH ON CANCER (IARC!

                  WORKING GROUP

                 OCTOBER 19. 1«8i
 Chemical  Industry Institute of Toxicology (CUT)
                P. 0.  Box  12137
          Research Triangle Park,  NC  27709

-------
                                   FORMALDEHYDE



3.1   Carcinogenic!ty studies in animals



  (b)  Inhalation and/or intratracheal administration



  Mouse



       Groups of 120 male and 120 female B6C3F1 mice, 6 weeks of age, were exposed



  to 0, 2.0, 5.6 or 14.3 ppm (0, 2.46, 6.89, 17.59 ng/m3) formaldehyde (> 97.52 pure)



  vapor by whole-body exposure for six hours per day on 5 days per week, for up



  to 24 months.   A 90 day sub-chronic study utili'zed to. set the exposure levels



  demonstrated necrosis of the nasal  mucosa and death after 2 weeks exposure to 40



  ppm.   The number of mice involved in scheduled sacrifices and unscheduled deaths



  is shown in Table 1.   All tissues from control and high exposure mice and multiple



  sections of nasal cavity and all gross lesions from all low and mid-exposure



  mice, as well  as all  tissues from unscheduled deaths in these groups were



  evaluated histopathologically.







       Concentration related effects  of formaldehyde exposure on mouse survival



  were not apparent, however, substantial  mortality did occur in- all  groups of



  male mice.  This was  primarily attributed to  fighting and ascending urinary



  tract infections.  Exposure of mice to 14.3 ppm formaldehyde vapor  resulted in



  diminished body weight.  Using life-table analysis, a number of lesions  in  the



  na-sal cavity wer^ significantly increased (p  < 0.0167).  These lesions  included



  epithelial dysplasia  at 5.6 and 14.3 ppm in males and females;  squamous meta-



  plasia at 5.6 and 14.3 ppm in males and  14.3  ppm in females; purulent or  sero-



  purulent rhinitis at 14.3 ppm in males and females; and olfactory epithelial



  atrophy in 14.3 ppm female mice.  Squamous cell  carcinomas  occurred in  the  nasal



  cavities of 2 male mice,  but no females.   Neither life-table analysis  nor the



  Fischer exact test revealed statistical  significance (Kerns  et al.,  1982a,b).

-------
Table 1.   Number of formaldehyde-exposed mice involved in  scheduled



                 sacrifices  and unscheduled  deaths
Exposure
(ppm)
0
0
2.0
2.0
5.6
5.6
11.3
14.3
Sex

Male
Female
Hale
Female
Male
Female
Male
Female

G mo
10
10
10
10
10
10
10
10
Schedul
12 mo
10
10
10
10
10
10
10
10
ed Sacri
10 mo
0
20
1
20
0
20
0
19
Mce
24 mo
21
31
22
26
19
11
17
20

27 mo
0
16
0
12
0
11
0
9
Unscheduled
0-24 mo
70
30
77
34
01
19
02
34
Deaths
24-27 mo
0
4
0
0
0
9
0
11
Total

120
121
120
120
120
120
119
121

-------
Rats
     Groups of 120 male and  120 female  Fischer  344  rats,  7 weeks of age  were
exposed to 0, 2.0, 5.6 or 14.3 ppm  (0,  2.46,  6.89 or  17.59 ng/m3) formaldehyde
(> 97.5% purity) vapor by whole-body exposure for six  hours-per day on 5 days
per week, for up to 24 months.  A 90 day sub-chronic  study utilized to set  the
exposure levels demonstrated weight loss and mucosal  erosions of the nasal
turbinata in rats exposed to 12.7 ppm and mortality and ulceration of  the nasal
turbinates and trachea after 2 weeks of exposure to 40 ppm.  The number  of  rats
involved in scheduled sacrifices and unscheduled deaths is shown in Table 2.
All tissues from control and high exposure rats and multiple sections of nasal
cavity and all gross lesions from all low and mid-exposure rats, as well  as all
tissues from unscheduled deaths in  these groups, were evaluated histopathologically,

     Rat survival was adversely affected by exposure to 14.3 ppm.  A summary of
neoplastic lesions in the nasal cavity is shown in Table  3.  Life-table  analysis
of thes'e data revealed significant  increases (p < 0.0167) in squamous cell
carcinomas in both male and female rats exposed to 14.3 ppm formaldehyde vapor.
Significant increases in other neoplasms were not detected in formaldehyde
exposed rats.  A variety of other lesions were significantly increased in
formaldehyde exposed rats.  These included epithelial  dysplasia at 2, 5.6 and
14.3 ppm in males and females; squamous metaplasia at 2,  5.6, and 14.3 ppm in
males and females; goblet cell hyperplasia at 2, 5.6,  and 14.3 ppm in males  and
at 2 ppm and 14.3 ppm in females;  hyperkeratosis at 14.3 ppm in males and
females; purulent or seropurulent rhinitis at 2, 5.6,  and 14.3 ppm in males  and
at 5.6 ppm and 14.3 ppm in females;  squamous atypia  at 14.3 ppm in males  and
females; olfactory epithelial atrophy at 14.3 ppm in males and females; respira-
tory epithelial  hyperplasia at 14.3 ppm in males;  goblet cell  metaplasia  of  the
olfactory epithelium at 14.3 ppm in males and females; and squamous  epithelial

-------
hyperplasia at 14.3 ppm in males.  The incidence of trachea! epithelial dysplasia



and metaplasia was different from control animals at 14.3 ppm  in  both males and



'females.  The incidence of squamous metaplasia in the nasal cavity  in rats from



the 2.0 and 5.6 ppm exposure groups regressed during the three month postexpcsure



period (Swenberg et al., 1980; Kerns et a!., 1982a,b).

-------
Table 2.   Number of formaldehyde-exposed rats involved in scheduled



                 sacrifices and unscheduled deaths
Exposure
(ppm)
0
0
2.0
2.0
5.6
5.6
11.3
11.3
Sex

Male
Female
Ma I e
Female
Ma I e
Female
Male
Fewale
Scheduled Sacrifice
6 mo
10
10
10
10
10
10
10
10
12 mo
10
10
10
10
10
10
10
10
10 mo
20
20
20
20
20
20
20
19
21 mo
51
17
50
11
11
11
13
11
27 mo
10
9
10
10
10
9
5
0
30 mo
6
1
1
1
3
5
0
0
Unscheduled Deaths
0-21 mo
6
13
10
16
19
19
57
57
21-30 mo
1
7
6
9
7
6
4
0
Total

120
120
120
120
120
120
119
120

-------
                                  Table 3.   Summary of neoplastic lesions in the nasal



                                cavity of Fischer-344 rats exposed to formaldehyde vapor.
Exposure group (ppm)
Sex
No. of nasal cavities examined
Squamous eel) carcinoma
Nasal carcinoma
Undifferentiated carcinoma
or sarcoma
Carcinosarcoma
Os teochondroma
Polyploid adenoma
0
M
110
0
0
0

0
1
1

F
111
0
0
0

0
0
0
2.0
M
110
0
0
0

0
0
4

F
no
0
0
0

0
0
1
5.6
M
119
1
0
0

0
0
6

F
116
1
0
0

0
0
0
11.3
M
117
5la
lb
2b

1
0
4

F
115
52a
1
0

0
0
1
Significantly increased over 0, 2.0, and 5.6 ppm (adjusted Cox-Tarone p < 0.0167).



 One animal also had a squamous cell carcinottia

-------
3.2  Other relevant biological data
  (a)  Experimental systems
  Toxic effects
       Acute cell degeneration, necrosis and inflammation, were evident in the
  nasal cavities of rats exposed to 15 ppm formaldehyde vapor for 1 to 9 days
  (6 hrs/day) (Swenberg et al., 1982).  Initial lesions were most severe on the
  tips of the rnaxillo and nasoturbinates.  Acute degeneration and sloughing of
  the respiratory epithelium, with edema and congestion- were evident at the end
  of 1 day's exposure.  This was followed fay ulceration,  necrosis and an influx
  of inflammatory exudate at days 3-9.  Early squamous metaplasia was detected
  over the naso- and maxi11oturbinates, median septum and lateral  wall  after as
  little as 5 days of formaldehyde exposure.   Examination of turbinates from rats
  exposed 5 days and allowed to recover for 48 hours demonstrated considerable
  regeneration.   Areas which were frequently ulcerated,  such as the lateral wall,
  had single thin strap cells covering areas  normally occupied by 5 or more cuboidal
  epithelial cells.  In contrast to these changes in the  respiratory epithelium,
  mild serous rhinitis was the principal lesion in rsgions of olfactory epithelium.
  Mild degenerative and inflammatory changes  were also evident in the nasopharynx.

       The mouse was similar, but less severe than the rat in its  acute response
  to formaldehyde toxicity.  Five Hays exposure to 15 pptr formaldehyde  vapor  caused
  degeneration, focal necrosis and inflammation to the naso- and maxilloturbinates
  and to the lateral wall, but minimum toxicity to areas  lined by  squamous  and
  olfactory epithelium.  By comparing these-acute data with results  from the
  6 month interim sacrifice (Kerns et al.,  1982a,b}, it is clearly evident  that
  adaptive changes have occurred.  The extent and severity of formaldehyde-induced
  toxicity diminishes with time.  This may  be due to changes in respiratory

-------
physiology, as well as alterations at the cellular level, ie. squamous metaplasia,



epithelial hyperplasia and increased detoxification pathways.







Effects of formaldehyde exposure on cell turnover



     A prominent response to cell loss associated with toxicity is compensatory



cell replication.  Surviving cells undergo division in order to replace dead



cells.  Exposure of rats to formaldehyde vapor for 3 days (5 hrs/day) resulted in



increased cell replication in rats at 5 and 15 ppm and mice at 15 pom.  No



increase was detected in rats exposed to 0.5 or 2 ppm or mice exposed to 0.5,



2 or 6 pom (Swenberg et a!., 1982).







Studies on the distribution and metabolism of formaldehyde in the nasal  cavity


                                                                      14
     Biochemical investigations on the absorption and distribution of   C-



formaldehyde have demonstrated that H-CO is primarily absorbed in the upper



respiratory system (Heck et a!., 1982).  Following a 6-hr exposure, the amount


    14
of   CH-0 absorbed appears to be directly proportional to the airborne concen-



tration.  This result is consistent with the high water solubility of Ci-uC.

                                           i A

It  is  important as well that the amount of   CH^Q absorbed d.id not appear to



vary following pre-exposure.  Hence, these findings, which are based on single



exposures, may also be relevant  to the chronic toxicity of Cr^O.







     The  extensive distribution of radioactivity to other tissues indicates that



absorbed   CH?0 or its metabolites are rapidly removed by the mucosal  blood



supply.   It is,  however, unlikely that this radioactivity is primarily due to



  CH-0.   Rapid metabolism in the blood and tissues would probably reduce the



concentration of free CH^O to extremely low levels.

-------
     The disposition and pharmacokinetic studies  indicate  that  inhaled  QUO  is
extensively metabolized.  It is likely that  folic acid plays  an  important, perhaps
a pre-eminent, role in  this incorporation.   However,  the possibility  that    CrLO
may form covaient adducts in vivo, or that " CO- may  itself be  incorporated
via carboxylation reactions cannot be excluded.  The metabolism  of  inhaled
14
  CH-0 appears to be similar to that which occurs following other routes of
adnii m'stration.
     In order to localize this absorption within the nasal cavity, naive or
                                                14
pretreated rats and mice were exposed to 15 ppm * C-formaldehyde for 6 hours
and prepared for whole-body autoradiography.  Formaldehyde-associated   C
was heavily deposited in the anterior nasal cavity of rats and mice.   The
amount of radioactivity correlated well with the distribution of lesions in
similarly exposed animals.   That is, activity was greatest in regions of res-
piratory epithelium over the maxillo- and nasoturbinates and the lateral  wall.
An exception was noted for the ventral portion of the nasal cavity lined with
squamous epithelium.  Radioactivity was heavily deposited  in this area, however,
minimal toxicity occurred.   This is most likely accounted  for by the relative
insensitivity of squamous epithelium to formaldehyde intoxication and supports
the concept that induction of squamous metaplasia is a host defense mechanism
to'formaldehyde toxicity.  A portion of the radioactivity  located in the  ventral
squamous regions may be the result of muco-ciliary flow and gravity,  rather than
direct exposure to formaldehyde vapor.  As  such, the radioactivity may  represent
covalently bound material rather than reactive formaldehyde.   The extent  of
formaldehyde-associated radioactivity deposited in regions of olfactory mucosa
of posterior sections of the nasal  cavity was much less,  with radioactivity
primarily confined to the nasopharynx.  No- differences were apparent  in for-
maldehyde distribution between naive rats and mice.   When animals that  had  been

-------
exposed to 15 ppm of non-radioactive formaldehyde for 9 days prior to exposure


   14
to   C-formaldehyde were compared to the naive rats .



mice had less radioactivity (Swenberg et al., 1S82).
   14
to   C-formaldehyde were compared to the naive rats and mice, the pretreatad
     This decrease in radioactivity is associated with a decrease in minute



volume in mice exposed to formaldehyde (Sarrow et al., 1982).  If the minute



volumes for rats and mice exposed to 15 ppm formaldehyde are used to calculate



the amount of formaldehyde inspired, and this amount -is normalized to the



surface area of the nasal cavity (Gross et al., 1982], the "dose" of formaldehyde

                                                     2
available for absorption is 0.154 and 0.075 ug/min/cm  in rats and mice,



respectively (Barrow et al., 1982; Swenfaerg et al., 1982).  Thus, the mouse nasal



mucosa is exposed to half the amount of formaldehyde than the rat nasal  mucosa



is.  This dose correlates well with tumor data, where the incidence of nasal



carcinoma is similar in rats exposed to 5.5 ppm and mice exposed to 14.3 ppm



of formaldehyde vapor (Kerns et al., 19825).






Mutagem'city and other short-term tests



     The effects of formaldehyde were evaluated in the C3H/10T4 Cl 8 call



transformation system.  Treatment of calls with 0.1 - 2.5 yg/ml  of formaldehyde



alone did not result in significant rates of transformation.   If formaldehyde



treatment was followed by continuous treatment with 0.1 ug/ml of the tumor



promoter 12-0-tetradecanoyl phorbol-13-acetat2 (TPA), transformed foci  were



produced.  Methanol and formic acid lacked significant transforming activity



under either treatment regimen.  The results suggest that formaldehyde  is



an initiating agent for C3H/10TH Cl  8 transformation (Ragan and Boreiko, 1981;



Boreiko and Ragan, 1982; Soreiko et al., 1981).

-------
     Formaldehyde has been shown to cause DNA-protein crosslinks in mouse L1210



cells (Ross et al., 1981; Ross and Shipley, 1980) and V79 cells (Swenberg at al.,



1982).   The DNA-protein crosslinks were repaired within 24 hours in both cell



types.

-------
                            REFERENCES
     Barrow, C. S., Sceinhagen, W. H., and Chang, J. C. F.  (1982).
Formaldehyde sensory irritation.  Proceedings of the Third C I IT Conference
on Toxicology: Formaldehyde Toxicity, Hemisphere Publishing Corp., New
York, in press.

     Soreiko, C.  J., Couch, 0. 3., and Swenberg,  J. A. (1581).  Mutagenic
and carcinogenic effects of formaldehyde.  Proceed ings of che AU1/3NL
Symposium on the Genotoxic Effects of Airborne Agents. Brookhaven National
Laboratories, Upton, New York, in press.

     Boreiko, C.  J., and Ragan, D. L. (1922).  Formaldehyde effects  in
the C3H/10TT cell transformation assay.   Proceedings of the Third CUT
Conference on Toxicology: Formaldehyde Toxic!tv,  Hemispnere Publishing
Corp., New York,  in press.

     Gross, E. A.,  Swenberg, J. A., Fields, S., and Popp, J. A. (1982).
Comparative morphometry of the nasal  cavity of rats and mice.  J.  Anat.,
in press.

     Heck, H., Chin, T. Y., and Schmitz, M. C. (1982).  Distribution of
    ]forma Idehyde in rats after inhalation exposure.  Proceedings  of
the Third CUT Conference on Toxicology: Formaldehyde Toxicitv, Hemisphere
Publishing Corp., New York, in press.

     Kerns, W. 0.,  Oonofrio, 0. J., and  Pavkov, K. L. ('982a).  The
chronic effects of formaldehyde vapor inhalation in rats  and mice.
A preliminary report.  Proceedings of the Third CUT Conference on
Toxicology: Formaldehyde Toxicity, Hemisphere Publishing  Corp., New
York, in press.

     Kerns, W. 0.,  Pavkov, K.  L., Donofrio, 0. J., Cornell,  M. M.,
Gralla,  E. J., and Swenberg, J. A. (1°82b).  Inhalation careinogenicity of
chronic formaldehyde exposure in rats and mice, in preparation.

     Ragan, 0. L.,  and Boreiko, C. J. (1982).  Initiation of C3H/10TJ
cell transformation by formaldehyde.   Cancer Lett. 13, 325*331•

     Ross, W. E., and Shipley, N. (1980).  Relationship between ONA
damage and survival in formaldehyde-treated mouse cells.   Mutat. Res. 79,
277-283.

     Ross, W. E., McMillan, D. R., and Ross, C. F. (1981).  Comparison
of ONA damage by methyImelamines and  formaldehyde.  J. Nat 1 . Cane. Inst. 67,
217-221.

-------
         UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                     WASHINGTON, D.C. 20460
                                                       OFFICE OF
                                                PESTICIDES AND TOXIC SUBSTANCES
tMEMORANDUM

SUBJECT:  OTS Formaldehyde Workplan

FROM:     John A. Todhunter, Ph.D.
          Assistant Administrator  'v'
            for Pesticides and Toxic Substances  (TS-788)

TO:       The Administrator  (A-100)
As a follow-up to the analysis and report memo on  formaldehyde
I recently prepared at your request,  I have attached a copy of
the Office of Toxic 'Substances formaldehyde workplan which"I
recently approved.

Basically, it is proposed to place into action most of the
items contained in the draft Federal  Register document which was
prepared last spring  (1981) when there was some confusion as to
whether or not formaldehyde was a TSCA section 4(f) chemical.
At that time it was believed in OTS that regulations were
inappropriate and unsupportable.  OTS was faced with the
possibility that formaldehyde might set off section 4(f) and,
to be ready in case it did, drafted the referenced Federal
Register document.  Due to OTS' position at that time vis a
vis regulation of formaldehyde, the Federal Register document
draft posed no regulatory options but, instead, proposed actions
of the sort included  in the attached  workplan.

Allow me to point out that item I, Evaluation of hazard, is
underway and part A thereof is essentially completed with
part B nearing completion.  Parts II  and III would develop
over a 24-30 month period.  They would be supported by part IV.

As was pointed out in my analysis memo, OTS does not recommend
that formaldehyde be given priority status - relative to other
OTS concerns - under  section 4(f) of  TSCA.  This only indicates
that OTS does not hold the carcinogenic risk to be of pressing
concern.  There could be other, non section 4(f), health
concerns (given formaldehyde's irritant properties) and, also,
a small possibility that there could  be some, as yet unidentified,
small subpopulation whose exposure might raise a question of
carcinogenic risk we might have interest in pursuing.  For these
reasons, even though OTS and I agree  that we cannot recommend
section 4(f)  status for this chemical, OTS and I agree that a
workplan would be prudent so as to maintain an Agency presence in
this matter.

-------
             PLAN FOR THE FORMALDEHYDE INVESTIGATION

                             OUTLINE


  I.     Evaluation of hazard

        A.   Analysis of available animal data to evaluate its
            relevance for assessing carcinogenicity in humans

        B.   Analysis of existing and planned epidemiologic
            studies and identify the need for additional studies

 II.     Evaluation of exposure

        A.   Update of the exposure assessment

III.     Evaluation of risk

        A.   Comparative analysis of various risk models

        B.   Time-to-tumor analysis

        C.   Calculation of risks based on the updated exposure
            assessment

 IV.     Establishment of mechanisms for exchange of information

        A.   Establish interagency workgroup

        B.   Publish Federal Register notice requesting
            information

        C.   Contact and meet with industry and labor groups

  V.     Outside peer review of hazard, exposure, and risk
        evaluations

-------
                                — 2 —
             PLAN FOR THE FORMALDEHYDE INVESTIGATION
                     DESCRIPTION OF PROJECTS

I.    Evaluation Of Hazard  (HERD)

     A.   Analyze available animal data relating to formaldehyde's
     carcinogenicity and evaluate its relevance for assessing
     carcinogenicity in humans.  HERD will evaluate data from
     bioassays, mutagenicity studies, and research conducted to
     clarify the mechanism of action of formaldehyde.  HERD will
     evaluate the strength of the evidence of formaldehyde's
     carcinogenicity, identify likely mechanisms of action, and
     assess the validity of extrapolating from formaldehyde's
     demonstrated carcinogenicity in animals at high doses to its
     action in humans at concentrations typical of human
     exposures.

     B.   Analyze existing and planned epidemiology studies to
     evaluate their anticipated usefulness in identifying
     formaldehyde's human carcinogenicity and identify and
     evaluate the need for additional epidemiology studies.  HERD
     will evaluate studies planned by the National Cancer
     Institute to determine the role they will play in predicting
     formaldehyde's human carcinogenic potential.  They will
     identify and evaluate the value of sponsoring additional
     epidemiology studies.

     Product:  Report describing findings.
     Time for completion:  8 months.

II.   Evaluation of Exposure  (EED)

     A.   Update and refine the exposure assessment based on
     available data and identify categories that should be
     dropped from further consideration.  EED will evaluate
     existing data on population sizes and exposure levels not
     included in the PRL-1 for accuracy and statistical
     representativeness, using predictive fate modeling or
     professional judgment.  EED will estimate release levels and
     exposure levels for categories for which monitoring data are
     not available.  Based on these analyses, EED will identify
     categories that result in minimal or insignificant exposure
     and will make recommendations of categories that need not be
     considered further.  The updated exposure assessment will go
     as  far as possible without conducting new monitoring
     studies.  Decisions on any further exposure work will be
     made after evaluation of risk.  It is anticipated that
     further work may be limited to incorporating information
     received from contacts and the interagency workgroup
     described in IV.

     Product: Exposure assessment.
     Time for completion: 5 months
     Extramural funds: less than $25,000.

-------
                                — 3 —
III.  Evaluation of Risk  (HERD)

     A.   Perform a comparative analysis of a number of risk
     models using the CUT data, based on hypothetical exposure
     scenarios.  This analysis will show the range of risks that
     various extrapolation models predict.  It will help OTS
     evaluate the relevance of different models for predicting
     the risk resulting from exposure to formaldehyde by
     considering the best fit to the CUT data and data on
     biological mechanisms of action.  It also will assist OTS in
     identifying the exposures that are associated with risks of
     concern.  Models to be used include the linear, one-hit,
     multistage, multihit,  probit, logit, and Weibull.

     B.   Analyze the final CUT time-to-appearance data using
     time-to-tumor extrapolation techniques.  This analysis will
     identify whether these techniques improve the accuracy of
     the risk prediction described in III.A. and will indicate
     the utility of and resources involved in routinely
     conducting time-to-tumor analyses.  Techniques to be used
     include the Hartley-Sielken General Product model and Crumps
     time-to-tumor model.  The analyses can be started only after
     the final data from the. CUT study are available.

     C.   Recalculate formaldehyde risks based on the updated
     exposure assessment and the final CUT data, using
     appropriate model(s) selected from the analyses performed in
     III.A. and III.B.

     Product:  Findings of the comparison of risk models and
     time-to-tumor analyses and updated risk assessment.
     Time for Completion:  8 months.


IV.   Establishment of Mechanisms for Exchange of Information (AD)

     A.   Establish an interagency workgroup to share information
     and to serve as a conduit to receive the results of research
     conducted by agencies, industry, and scientific groups as it
     becomes available.   The workgroup will assure that
     information related to the hazard, exposure, and release of
     formaldehyde from products is shared by the agencies to
     avoid overlapping or duplicative efforts.

     B.   Publish a notice in the Federal Register requesting
     information relating to hazard and exposure.  The notice
     will describe EPA's interest in obtaining information and
     will request comment and contact with interested parties.
     The Federal Register notice will alert academic and
     international groups not directly involved with the affected
     industries of our desire for their input and comments.

-------
                               -4-
     C.  Identify and contact industry and labor groups.  OTS
     will hold a series of meetings with these groups to obtain
     information about hazard concerns, current exposure levels,
     production processes, and other technical information.
     These contacts will be maintained throughout the course of
     the investigation to assure their continued input.

V.   Outside Peer Review of Hazard, Exposure, and Risk
     Evaluations

     A.  Each of the new evaluations developed in sections  I,  II,
     and III will undergo internal and external scientific  peer
     review in accordance with Agency guidelines.

-------
               SCHEDULE (Starting January 4, 1982)


Elapsed Time:  5 months:

o     Updated exposure assessment completed.

o     Interagency group established.

o     Federal Register Notice published.

o     Decision to eliminate certain exposures  from  further
      consideration, based on the exposure assessment made.

Elasped Time:  8 months:

o     Hazard evaluation completed.

o     Risk assessment completed.

o     Identification of relevant industry and  labor groups, and
      meetings begun.

o     Decisions made, such as the extent of OTS'1 continuing
      formaldehyde efforts, based on the conclusions of the
      hazard evaluation; whether-to eliminate  additional
      exposures from further consideration, based on the risk
      assessment; whether or not to conduct additional hazard  or
      exposure studies.

Elapsed Time:  11 months:

o     Report on progress and information from  interagency group
      and public contacts.

Elapsed Time:  14 months:

o     Peer reviewed reports on new information  available on
      formaldehyde's hazard, exposure, and risk.

o     Disposition meeting to decide the direction of the
      investigation and further research needs, based on the
      findings to date.

-------