[CPEO-BIF] TCE Toxicity

[Lenny Siegel <lsiegel@cpeo.org>]

[The following is text of the written testimony submitted by Boston 
University Professor Dr. David Ozonoff to the New York Assembly hearings 
on TCE and vapor intrusion. I found particularly interesting his 
explanation (item #1) of the relationship between the drinking water 
standard and the air exposure standard. - LS]


November 10, 2004

Jeff O’Donnell
Committee Assistant
New York State Assembly
Program & Counsel Staff
Capitol/ Room 520
Albany, NY  12248

Dear Sir

This testimony relates to your Committee’s interest in the public health 
aspects of exposure to tricholoroethene (TCE) via vapor intrusion into 
indoor environments from contaminated groundwater. I have had a long 
interest in the health effects of the chlorinated ethylenes, TCE and its 
very close relative, PCE (tetrachloroethylene), and have authored 
numerous peer-reviewed epidemiological studies on these chemicals. I 
have attached my Curriculum Vitae for your reference as to my standing 
and qualifications to give my opinions on this important matter.

TCE has been implicated in at least four kinds of adverse health 
effects: neurotoxic effects (effects on the central nervous system); 
cancer; birth defects; and autoimmune disease (scleroderma, lupus 
erythematosus and mixed connective tissue disease). For historical 
reasons and force of circumstance much of the study of the effects of 
TCE on human beings has taken place in the occupational environment, 
raising the question as to what effects might be expected, if any, at 
the substantially lower levels normally encountered in an indoor air 
environment from vapor intrusion. While this is not an easy question to 
answer, I have substantial concerns about effects even at these levels. 
This is founded on two main arguments.

1. We have been studying the effects of PCE in drinking water for almost 
15 years and are able to see substantial increased cancer risks in these 
circumstances, which are orders of magnitude lower than occupational 
exposures. Drinking water exposures within homes come both from 
ingestion and air stripping and dermal absorption. The latter are 
roughly of the same order of magnitude as ingestion. The current MCL for 
drinking water is 5 micrograms per liter. Thus allowing for a doubling 
of dose from air/dermal and a consumption of 2 liters/day (a likely 
overestimate) this is a dose of 20 micrograms per day. Since the average 
adult breathes approximately one cubic meter/hour, this corresponds 
(roughly) to an indoor air exposure of 1 microgram/m3. The MCL is an old 
standard based on outdated cancer risk estimates. The proposed level of 
5 micrograms/m3 is thus not consistent with the current (now fairly old) 
water standard.

However there is reason to believe the 10^-6 risk level is considerably 
lower than previously thought. To be health protective one normally uses 
the most conservative estimates. I call your attention to work by Cronin 
et al.   In this physiologically-based pharmacokinetic model ("PBPK 
model") account is taken of the actual physiological processes of 
absorption, distribution and metabolism in animals and humans. Mice and 
humans, for example, are built on the same "organization chart" of 
blood, brain, digestive system, lungs, etc., but the extent to which the 
different "boxes" in the chart absorb, distribute, exchange and 
metabolize TCE might vary from species to species. The object of the 
PBPK exercise is a better extrapolation of a representative response 
from mouse to human by substituting the correct species-specific values 
for things like absorption coefficients, rate constants, and other 
values ("parameters") that determine the details of how TCE is handled 
in an organism.

A major difficulty with PBPK modeling is the paucity of information on 
two things: the correct biologically effective endpoint to use (is it 
the peak concentration of the metabolite TCA?, cumulative TCA?, the 
metabolite DCVC?, some DNA-adduct?, the metabolite trichloroethanol?); 
and accurate determination of the parameters that describe important 
physiological processes, like the rate of absorption or excretion in 
different species and between different organs and tissues in the same 
species. Considerable uncertainty in the correct parameter estimates can 
lead to very large differences in estimates of biologically effective 
dose, and hence of dose-response modeling. The way parameter 
uncertainties are currently handled is to incorporate a distribution of 
values in the model, and then determine the impact on the dose estimate 
(so-called Monte Carlo modeling). There is a wide range of legitimate 
estimates using PBPK models when coupled with the linearized multistage 
model used by NYDOH. Cronin et al., for example, have estimates as low 
as .02 micrograms/m3 TCE as the 1 in 1,000,000 risk in air.

I note that the choice of a linearized multistage model, as used by 
NYDOH, is not the only possible choice. Choosing another biologically 
plausible model for the dose-response function can also result in a 
large variation in estimated risks. These variations for estimates of 
the risk of TCE in drinking water were investigated by Cothern, et al. 
Four different functional forms were used, including the one chosen by 
NYDOH (the multistage model), and the estimated risks compared. Cothern 
et al. note there are no biologically based criteria for choosing one 
model over another. The results are a dramatic example of the effects of 
model choice. It turns out that the NYDOH choice was one of the most 
permissive (in the public health protection sense) compared to the least 
permissive, the Weibull model.  The difference in estimated risks among 
the models was almost a factor of 10,000, i.e., the Weibull model 
predicted risks from TCE in drinking water to be 10,000 times higher 
than the risks from the most permissive model.

2. My second line of argument relates to the kinds of health effects one 
might expect to result from extremely tiny exposures. To summarize a 
longer argument, the effects are just those where some kind of 
biological amplification of damage occurs. The classic example is 
cancer, where a tiny alteration in the genetic program of the cell (the 
DNA code) that makes a cell into a cancer cell is reproduced each time 
the cell divides. Thus the original damage is biologically reproduced 
along with it and the offending tiny amount of chemical need no longer 
be present. This is essentially the reason we believe there is some 
cancer risk at every level of exposure. The dose response relation 
discussed above tries to estimate the size of that risk at various 
exposures.

However, there are other biological systems where such intrinsic 
amplification might be expected, among them the immune system (one need 
only think of bee stings and the dramatic, sometimes fatal effect of 
tiny exposures); the nervous system (where tiny signals are amplified 
into large responses); and human reproduction (where a entire organism 
comes from a single fertilized egg).  Thus the health effects seen in 
occupational environments are plausibly also present, although at a 
lesser frequency, at much lower exposure levels as well.

There has been a huge volume of scientific work done on this 
environmentally prevalent contaminant and these comments only highlight 
two considerations I though pertinent to the question before your 
committee. I am attaching a table I did a few years ago of epidemiologic 
studies of cancer and TCE/PCE. I hope you find it useful.

Please do not hesitate to get in touch with me if you have further 
questions. I regret that my schedule does not permit me to appear before 
your Committee in person.

Sincerely yours





David Ozonoff, MD, MPH
Professor of Environmental Health
Boston University School of Public Health


--
Lenny Siegel
Director, Center for Public Environmental Oversight
c/o PSC, 278-A Hope St., Mountain View, CA 94041
Voice: 650/961-8918 or 650/969-1545
Fax: 650/961-8918
http://www.cpeo.org


_______________________________________________
Brownfields mailing list
Brownfields@list.cpeo.org
http://www.cpeo.org/mailman/listinfo/brownfields