30
November 2001
日本語訳はここをクリック（Japanese
here）
Comments on the Japan
Ministry of the Environment Report on the Test Results of Endocrine Disrupting
Effects of Nonylphenol on Fish (Draft)
Ken-ichi
Miyamoto, Akihiro Tokai, Junko Nakanishi
Research Center for Chemical Risk Management (CRM)
National Institute of Advanced Industrial Science
and Technology (AIST)
16-1 Onogawa, Tsukuba, Ibaraki 305-8569 Japan
e-mail : ken-ichi.miyamoto@aist.go.jp
INTRODUCTION
The Research Center for Chemical Risk Management (CRM), which is a
department of the National Institute of Advanced Industrial Science and
Technology (AIST), Japan, has drawn up the following comments on the Japan
Ministry of the Environment draft report on the test results of endocrine
disrupting effects of nonylphenol on fish [1]. Since the Japan Ministry of the
Environment (MoE) has not yet requested domestic public comment on the report,
the CRM has decided to first submit these comments to the OECD directly.
The CRM has concluded that the report contains major problems regarding
the exposure, effects and risk assessments and that the discussion on the risk
posed by nonylphenol (NP) in Japan is possibly misleading. Several specific
comments give further detail below.
SPECIFIC
COMMENTS
1.
Methodological problems in analyzing environmental concentration of NP
Conclusion:
The exposure assessment of NP should include an evaluation of how many sites
exhibit high concentrations and how often such high concentrations are found
at each site.
Comments:
The exposure assessment in the report of the MoE is based on an extensive
field survey conducted over two years, which consists of 2,330 concentration
values measured all over Japan. (Although the MoE report states that the
number of sites surveyed is 2,330, this is incorrect. The number 2,330 is the
number of NP concentration values.) Although this is a valuable dataset, the
analysis by the MoE is insufficient.
First,
although the report describes that Fig. 4 indicates that there were two groups
of different characteristics with roughly the 95th percentile as a demarcation
line (p.8 of the report), this is difficult to understand. Additional
statistical analysis is necessary to reach such a conclusion; however, the
report does not present any analysis other than Fig.4, which is a simple
histogram. Furthermore, the 95th percentile concentration value of 0.59
µg/L is used in the risk assessment as the highest value estimated in
general water areas in Japan without any scientific justification (p. 25).
Second, the report deals with all monitoring data as a whole. However,
this dataset is a result of surveys conducted from FY 1998 to 1999. Some sites
have six measured concentration values, which were collected at different
times. Other sites have one, two, three, or four concentration values.
Concentrations in some sites always exceeded 0.59µg/L, while only one of
two, three, or four values, two of three or six values, five of six values, or
no values exceeded 0.59µg/L at other sites. Therefore, the report does
not provide any useful information regarding how many sites exhibit the higher
concentrations, what kind of sites they are, or how often the higher
concentrations occur.
In addition, the measured concentrations of NP decrease at most sites
from FY 1998 to 1999. This may be due to voluntary activities to reduce the
use or emission of NP and/or nonylphenol ethoxylates, because the MoE
(formerly the Japan Environment Agency) had listed NP as one of 67 substances
suspected to have endocrine disrupting effects in SPEED (Strategic Programs on
Environmental Endocrine Disrupters) ’98 [2]. However, this decreasing trend
and the reasons behind it must be confirmed using recent monitoring data and
by determining the use and treatment methods of NP and nonylphenol ethoxylates.
Taking
the above facts into consideration, we have reached the conclusion that the
analysis by the MoE is inadequate for characterizing the aquatic
concentrations of NP for exposure assessment in Japan. More careful exposure
analysis will be needed using an updated dataset of the monitoring data that
includes concentrations measured in FY 2000 and later.
2.
Problems in toxicity data used in effects assessment
Conclusion:
Results of the toxicity tests conducted by the MoE contradict each other.
Comments:
The MoE conducted two long-term toxicity tests using medaka (Oryzias
latipes) in one lab. One was a partial life test, which began with
fertilized eggs at mean measured NP concentrations of 3.30, 6.08, 11.6, 23.5,
and 44.7µg/L and ended at 60 days posthatching [1]. The other was a full
life-cycle test, which started with fertilized eggs at mean measured
concentrations of 4.2, 8.2, 17.7, 51.5, and 183µg/L and was continued
until 104 days posthatching over two generations [1,3]. Table 1 shows a
summary of the results of the two toxicity tests.
According
to the MoE report [1], the test conditions for the full life-cycle test up to
60 days posthatching seem to be the same as those for the partial life test,
except for slight differences in the exposure concentrations. However, as
shown in Table 1, many endpoints represent inconsistent results. The effect on
posthatching mortality was stronger in the full life-cycle test than in the
partial life test. Conversely, the effects on growth and sex ratio were more
sensitive in the partial life test. Even though these contradictions may
influence the reliability of the toxicity tests, a result from the partial
life test was used in the risk assessment without any explanations of this
problem.
Table 1 Summary of the results of the medaka
partial life test and the medaka full life-cycle test
| Endpoint |
partial
life test |
full
life-cycle test up to 60 days posthatching |
| posthatching
mortality |
no
effects up to 44.7µg/L |
NOEC
: 8.2µg/L
LOEC : 17.7µg/L |
| growth
(total length) |
NOEC
: 23.5µg/L
LOEC : 44.7µg/L |
no
effects up to 51.5µg/L |
| growth
(body weight) |
NOEC
: 11.6µg/L
LOEC : 23.5µg/L
|
no
effects up to 51.5µg/L |
| sex
ratio determined by gonadal histology |
NOEC
: 6.08µg/L
LOEC : 11.6µg/L |
NOEC
: 17.7µg/L
LOEC : 51.5µg/L |
3.
Problem in the selection of assessment endpoints for ecological risk
assessment
Conclusion:
The MoE selected vitellogenin induction and testis-ova production as endpoints
for ecological risk assessment, which, as far as we know, have never been used
in any ecological risk assessment. We believe they are not ecologically
relevant endpoints for the risk assessment of NP in Japan.
Comment:
The selection of assessment endpoints for ecological risk assessment should be
at the population, community, or ecosystem level, but not at the individual
level except for threatened or endangered species [4]. However, the MoE used
vitellogenin induction and testis-ova appearance, which are not considered to
have a direct influence at the population level or higher, to estimate the
predicted no-effect concentration (PNEC).
Kang
et al. [5] indicated that the development of testis-ova in male medaka did not
have direct effects on reproductive ability. They demonstrated that male
medaka exposed to 29.3, 55.7, 116, or 227ng/L of 17ß-estradiol exhibited
normal spermatogenesis in spite of the development of testis-ova. This
resulted in unaffected fecundity and fertility [5].
In
general, ecological risk assessment should be based on the endpoint of
population or of higher ecological hierarchy. If this is difficult,
alternatively, the organism-level endpoint that is directly related to
population dynamics, such as survival, growth, and reproduction, should be
applied. Therefore, the endpoints for the ecological risk assessment of NP in
Japan should be based not on biochemical or pathological effects such as
vitellogenin induction or testis-ova appearance but on ecologically relevant
endpoints.
4.
Others
Conclusion:
The citation of the Canadian risk assessment report [6] in the MoE report
easily leads its readers to misunderstand the original intention of the
Canadian report.
Comments:
The MoE report describes the Canadian risk
assessment report as follows (p.23).
“In its ecological risk assessment, it says that
the predicted no-effect concentration in the most conservative approach is
0.17µg/L obtained by dividing an acute toxicity 96hLC50 value of 17µg/L
in winter flounder by an assessment factor of 100, the predicted no-effect
concentration in the conservative approach is 0.39µg/L obtained by
dividing the largest no-observed-effect concentration (NOEC) value of 3.9µg/L
of chronic toxicity in mysid shrimp by an assessment factor of 10, and the
predicted no-effect concentration for endocrine disrupting activity is 1µg/L
obtained by dividing a threshold value of 10µg/L of the induction of
vitellogenin in male rainbow trout plasma by an assessment factor of 10. Then
it compares to the respective predicted no-effect concentrations and to the
predicted environmental concentrations. As a result, it says concentrations in
river water, factory effluents, and effluents from wastewater treatment
facilities exceed predicted no-effect concentrations in some cases.”
In fact, the ecological risk assessment in Canada was performed on
three tiers. At the final tier, the risk was evaluated by comparing the
estimated exposure value (concentration) with the estimated no-effect value
(concentration) of 1µg/L, which is the most important estimated
no-effect concentration used in the Canadian risk assessment. However, this
concentration was estimated based not on the vitellogenin induction in male
rainbow trout as described in the MoE report, but on the extrapolated
concentration using a species sensitivity distribution. The endpoint of
vitellogenin induction in rainbow trout was used to compare the chronic
toxicity of NP with the endocrine effects. Furthermore, the Canadian report
points out that using the endpoint of the whole-organism vitellogenin
induction in rainbow trout is hyperconservative.
The above citation in the MoE report indicates that there may be some
intent to justify the use of the induction of vitellogenin and testis-ova as
the endpoint of the ecological risk assessment of NP in Japan in an inadequate
manner.
We sincerely recognize the importance of the work carried out by the
MoE on risk assessment of NP. Regarding the prudent avoidance of unnecessary
misunderstanding of the risk posed by NP, we urge the MoE to verify or
calibrate the scientific data and findings as soon as possible by consulting
third party scientific groups from the viewpoint of not only a scientific
perspective but also a regulatory perspective. We hope the discussion in OECD
will work toward this end.
References
[1] Environmental Health Department, Ministry of
the Environment, Government of Japan. (2001) Report on the test results of
endocrine disrupting effects of nonylphenol on fish (Draft), August
http://www.oecd.org/pdf/M00019000/M00019987.pdf
http://www.oecd.org/pdf/M00019000/M00019988.pdf
http://www.oecd.org/pdf/M00019000/M00019989.pdf
[2] Japan Environment Agency. (1998) Strategic
Programs on Environmental Endocrine Disruptors ’98 (in Japanese)
[3] Yokota, H., Seki, M., Maeda, M., Oshima, Y.,
and Tadokoro, H. (2001) Life-cycl toxicity of 4-noniyphenol to medaka (Oryzias
latipes). Environ. Toxicol. Chem. 20.
2552-2560
[4] Suter, G.W., Efroymson, R.A., Sample, B.E., and
Jones, D.S. (2000) Ecological risk assessment for contaminated sites. CRC
Press, Boca Raton, Florida, USA, p.37
[5] Kang, I.J., Yokota, H., Oshima, Y., Tsuruda,
Y., Yamaguchi, T., Maeda, M., Imada, N., Tadokoro, H., and Honjo, T. (2001)
Effect of 17ß-estradiol on the reproduction of Japanese medaka (Oryzias
latipes). Chemosphere, in press
[6] Environment Canada, Health Canada (2001)
Priority substances list assessment report, Nonylphenol and its ethoxylates
http://www.ec.gc.ca/cceb1/eng/final/NPEs.pdf
