[ RadSafe ] Your letter of Jan. 6

[Andree Desiree Wilson and Richard Wilson]

In this connection my paper on arsenic risk may be of 
interest. 
<http://phys4.harvard.edu/%7Ewilson/publications/ppaper807.html>"Underestimating 
Arsenic's Risk"     Regulation 24(3) pp50-53

Dick Wilson



May 11th 2001

Introduction.

The proposal by the US EPA in the year 2000 to reduce the allowable 
level of arsenic in drinking water was the first major decision of 
the EPA under the regulations which allow the use of Cost-Benefit 
analysis to make a decision.   Therefore the way in which the EPA 
make this decision is of extreme importance.  Moreover as I will 
explain below, the effect of Chronic Arsenic poisoning throughout the 
world is a major one and the decision is therefore a very  important 
one in its own right.

The paper by Burnett and Hahn must be considered as one of many ways 
in which the problem can be discussed.    They have several points of 
disagreement with the way in which the problem was discussed by both 
NAS and EPA.  In this note I discuss each of these disagreements and 
point out that some, if accepted would involve major changes in 
practices used by the EPA  and generally accepted by the American 
people, and some others are more controversial that Burnett and Hahn imply.

The arsenic problem

Arsenic has been used in commerce for thousands of years.   Recently 
it has been used in non-ferrous metals processing,  it has been 
sprayed on vineyards and other crops as arsenical pesticides before 
the more effective chlorinated hydrocarbons were discovered.   It has 
been used in medicine:  taken for a limited time at high doses, just 
below the acute toxic threshold, it cures leukemia.   It was used 
a  cure for syphilis before the superior sulphur drugs and penicillin 
were discovered.   In dilute solution it was as an over-the-counter 
cure (Fowler's solution) for mild ailments.

Arsenic has long been known to be acutely poisonous at high 
doses.   It possessed a peculiarity:   in an individual, taking doses 
just sub acute raises the acute toxic level.  This makes it the 
darling of detective story writers.  A villain can take subacute 
doses for a week, and then share a meal containing toxic arsenic 
doses with his victim.   It is recorded that the Borgias used this 
method.   But for many years it was thought that at lower doses no 
adverse effects were seen.   A few cancers were found in small groups 
of people:  Skin cancers from excessive medical use of arsenic (using 
Fowler's solution regularly over a long time)  were described by 
Hutchinson in 1888. Lung cancers from arsenic inhaled arsenical 
pesticides were found in the 1890s and from smelters in the 1920s. A 
rare liver cancer, angiosarcoma, can be caused by arsenic as easily 
as it can be caused by vinyl chloride.   But these were 
discounted,  and a threshold was assumed below which no cancers will be seen.

This view, that Chronic Health effects do not occur with ordinary 
uses of arsenic was reinforced by the standard toxicological 
predictions.  For a century public health has been considerably 
improved by tests of materials in animals before using them where 
humans can be exposed.  A chemical that is toxic to rats and mice  is 
toxic in people.  Moreover the relation is quantitative.  When the 
same fraction of the chemical is included in the human diet as the 
animal diet, the toxic effect is about the same and the probability 
of developing cancer is about the same.     Animals, particularly 
rats and mice, did not develop cancer from arsenic. Ergo, people 
don't develop cancers .   That animals do not develop cancer led 
toxicologists and almost everyone else, including myself, to believe 
that the limited data on humans was consistent with the idea that 
people who ingest moderate doses of arsenic do not develop 
cancer.   But in a paper which shocked everyone, C.J. Chen and 
collaborators from Taiwan described cancers of the bladder, kidney 
and lung attributable to ingested arsenic. This was confirmed by 
scientists from Argentine and Chile and by Dr Lu and collaborators 
from Inner Mongolia. It is apparent that, contrary to the previous 
belief of most scientists, animal data up to that time were  bad 
predictors of arsenic related human cancers.  This may be changing 
somewhat.  In a recent study Dr Ng of Brisbane, Australia persuaded 
mice to develop cancer at small doses of arsenic.  Moreover a 
metabolite of arsenic, Di Methyl Acetyl Arsenic or DMAA, 
causes  cancer - apparently both as a promoter and as an initiator.

The number of people in the world affected by arsenic is large. The 
N.Y. Times on April 7th 2001 referred to a small study in Chile. In 
that "small" study cohort, 500 people got cancer in excess of the 
background, a larger number than the 450 more than background that 
developed cancer from the atomic bomb explosions in Hiroshima and 
Nagasaki. In Bangladesh 30 million people are exposed to arsenic at 
levels higher than EPA presently permit (>50 ppb in the water). I 
have been quoted many times, without ever being contested, that the 
Bangladesh catastrophe makes Chernobyl look like a Sunday School 
picnic. Some thousands have already died from secondary effects of 
the skin lesions and there are estimates by responsible that a 
million people will die eventually.   The crucial question is : "What 
does that  imply for Americans exposed to arsenic at present 
levels?   As often happens the whole world is watching us as we 
discuss what to do.

About 10% of Chileans or Taiwanese who drank water containing 500 ppb 
of arsenic developed cancer - mostly lung cancer.  A fundamental 
issue is how one estimates a risk at lower levels from the generally 
accepted measurements at higher exposure levels.     On this issue 
there is a strong disagreement so that it is important to understand 
the two extreme positions.   On the one hand there are scientists and 
physicians who think that linear dose-response is most likely, and on 
the other hand,  some scientists believe that there exists a 
threshold below which no effect will be seen.   This disagreement 
permeates all studies of environmental cancer risk.    50 years 
ago,  Sir Richard Doll and Dr Armitage discussed a mult-stage theory 
of cancer, and how the theory might describe both naturally occurring 
cancers and cancers caused by an external agent.   They noted that 
most cancers caused by an external agent cannot be distinguished from 
those that occur naturally and may be fundamentally 
indistinguishable.   It was then reasonable to assume that the 
external agent affected one stage in the cancer development in the 
same way as the natural processes do.    25 years ago a  group of 
scientists including one of Doll's collaborators,   Sir Richard 
Peto,  pointed out that the argument is very general.    If an agent 
increases the probability of any step in the cancer formation process 
in the same way as natural processes,   then almost any 
biological   dose-response relationship becomes linear at low 
doses.     This formed the basis for the EPA assumption of low dose 
linearity as a default for all carcinogens.   It is the default 
linearity argument that I described in the pages of Regulation 18 
months ago.  If one uses the default dose-response and starts from 
the data from Taiwan or Chile where there is a risk of 10% (mostly a 
lung cancer risk) for people who drink water containing 500 ppb of 
arsenic, the default assumption leads to the 1% risk at presently 
allowable levels, 50 ppb, that NAS in 1999 reported as 
possible.    This is a higher risk than Burnett and Hahn take as an 
upper limit.   Moreover it is highly probable, as Burnett and Hahn 
point out, that there are other cancers and other chronic effects due 
to arsenic.  While it is becoming popular to attribute any 
unexplained ailment to such exposures as arsenic in the cured wood in 
one's fence,   some of these attributions are probably correct.  A 
doubling - or even quadrupling - of the 1% figure may not then be unreasonable.

Some toxicologists have suggested  that the dose of arsenic to the 
relevant organ does not increase linearly with ingested dose. This 
remains unproven and is certainly not a general situation for 
carcinogens.   For many carcinogens, the number of DNA adducts is 
almost  proportional to ingested dose over 5 orders of magnitude - 
showing unequivocally that the pollutant in question reached the cell 
- even when some toxicologists were arguing that it does not.   We 
are even unsure of the relevant organ for the induction of internal 
cancers. We think we understand why an inhaled material can cause 
havoc in the lung. But how does ingestion of arsenic produce lung 
cancer? Even after the Taiwan data were generally accepted, this part 
was questioned because it seemed so unlikely that ingested material 
would affect the lung in an important way. But the data from Chile 
seem unequivocal. Some biologists and toxicologists insist on their 
belief that there is a threshold below which there is no effect, or 
at least non linearity. Even worse than the fact that we were told, 
until recently, that arsenic does not cause cancer in animals and 
therefore is unlikely to cause cancer in people, is the fact 
that  recent studies show that a metabolite of arsenic, 
DiMethylArsenic Acid (DMAA) does cause bladder cancer in 
rodents.   Therefore it eems to me that if  these toxicologists and 
biologists base their conclusion upon animal data, they  have the 
burden of explaining why they were wrong for nearly a 
century.    Indeed between  the major arsenic conferences of the 
Society for Environmental     Geochemistry and Health (SEGH) in San 
Diego in 1997 and 2000 there seemed to me to be a change in attitude 
of most particpants - they no longer believed that animal toxicology 
was useful for the present discussion.

How one addresses the two approaches to discussing risks at low dose, 
the macro approach of default linarity and the microapproach of the 
toxicologist is a core problem that underlies all EPA 
regulations.   I would like to see NAS and EPA study should address 
the default linearity approach and the  difference from  the 
toxicological approach, and do so in all generality so that  we may 
learn from the behavior of other agents and substances, and also 
learn how to apply the insights gained for the future.  Meanwhile 
Burnett and Hahn take a probability distribution of health 
effects.  I would take a bimodal distribution corresponding to the 
difference in perspective, which changes with time even in the 
absence of new data.

Does one, and should one discount lives lost in the future?

Burnett and Hahn argue, as Raiffa and Weinstein did in a seminal 
paper that one should discount lives saved in the future with the 
same discount rate as used in direct economic analyses.  Others have 
disagreed.  But it is abundantly clear that American society as a 
whole has not agreed with discounting.   With a 7% discount rate on 
lives lost in the future,  little money should be spent on waste 
disposal.  Indeed with even a 0.1% discount rate, the US is clearly 
spending much too much on consideration of disposal of high level 
nuclear waste.   To choose to discount lives is therefore a policy 
decision which  would create profound precedents and should not be 
done lightly.

What does compliance mean, and is there an alternate way of meeting 
the legitimate requirements?

If it is cheap the easiest way of meeting any proposed new arsenic 
regulation will be to modify a new water supply so that at all times 
the concentration of arsenic in all the water is below the permitted 
level.  But that may not be necessary for meeting any public health 
goal based on chronic, as distinct from acute, effects.   What is 
important is that the total amount of arsenic ingested over a long 
period - perhaps 5 years- be below the permitted level.  For example 
the water supply of a large community might use mixing and filtration 
to meet the average required level.  But a small water system in a 
community of 1,000 to 5,000 people may have problems.  They should 
clearly be allowed on or more options.  One might be to have supply 
for drinking water separate from that used for such activities as 
washing.  Another might be to allow residents to use bottled water 
for drinking and cooking.   One could have a 2 step regulatory 
system.  A mandatory 50 ppb standard, and an advisory 10 ppb standard 
for those water systems where there is effective participation in the 
decision by all water users.  The key to compliance is that all 
residents in the water system be informed of the discussions on the 
calculated effects on health and be able, through ordinary democratic 
procedures, in the decision of what to do.    At the present time I 
do not find that these possibilities are discussed by Burnett and 
Kahn, by  the EPA,  or by the vocal representative of the affected 
water systems.

Costs of meeting requirements

The costs of any action to meet a regulation are far more uncertain, 
and often much lower,  than opponents of the regulation tend to 
claim.   The most well known example is the cost of reducing the 
occupational exposure to vinyl chloride in the polyvinylchloride 
(PVC) industry.   Estimates of fixing the industry were high.  The 
exposures were reduced by sealing the equipment to stop fugitive 
emissions.  But when this was done there was a net saving of 
material, and actually a saving of money!    Industry tend to 
estimate the cost of applying any rule tomorrow and removing arsenic 
from the water tomorrow (and maybe having the equipment sent by 
FEDEX!).   An important feature of the proposed arsenic rule is the 
date of 2006, rather than 2001,  for enforcement.   This allows time 
for innovative approaches.   For example, the Los Angeles water 
district uses the Los Angeles Aqueduct which brings arsenic from Lake 
Mono and Lake Crawley in the eastern Sierras.   The water contains 
arsenic at an average level of 23 ppb.  But careful management, 
filtering and mixing from different sources brings this down over the 
last 10 years to less than 2 ppb in most locations.  Exact costs are 
not available,  but they seem not to be large since the process was 
done slowly.     Increasingly developers in western states are 
depending upon wells to supply their wtaer instead of the surface 
waters that were almost universal in the past.  These often contain 
arsenic.    But there exists an organization, the U.S. Geological 
Survey that can help guide drillers where to place new wells.  At a 
special USGS conference in February 2001, it appears that USGS have 
only recently been fully consulted. It may be that now that U.S.G.S. 
have been consulted, we are likely to find out that it will be 
possible to drill wells into aquifers with  free of arsenic at a cost 
the cost which is quite inexpensive compared to the cost or removing 
arsenic from existing water supplies.   These arguments suggest that 
the cost of meeting the new standard will be considerably lower than 
that proposed by the water industry and lower than that proposed by EPA.

Consistency with past and future regulations.


Since the arsenic lung cancer risk even at 500 ppb is approximately a 
doubling of the risk that a heavy cigarette smoker has, the risk at 
50 ppb, by this argument, is only a 10% increase.   On the one hand 
this emphasizes the magnitude of the problem cigarette smoking poses 
for public health, it also suggests that arsenic at the 50 ppb level 
might be considered insignificant.   But this argument has never 
suceeded in the public discussions of pollutant levels.

Emerson once said that excessive consistency is the hobgoblin of 
small minds. Nonetheless a regulatory agency such as the E.P.A. 
should have a very clear reason for any lack of consistency.  These 
reasons  should be stated to the people.  Unfortunately over the 30 
years of EPA existene that has not been one of its virtues.     I 
will discuss a few possible implications here.   If one assumes that 
exposures of people to arsenic at high doses is at least as 
significant as exposures of mice to thrichloroethylene, and use the 
same EPA procedures, one finds that the regulatory level should be 5 
ppt - an obviously unattainable goal.    If even the cost-benefit 
procedure used by EPA in the year 2000 for regulating arsenic had 
been applied to regulation of trichloroethyelene or chromates, that 
regulation would have been less severe than at present.  If Burnett 
and Hahn's procedures had been adopted there would have been no 
regulation at all.  What then do we make of the public opinions that 
are implicit in the  interesting movies "A Civil Action" and "Erin 
Brokovich"?    It seems to me that these, as representations of 
public opinion, cannot be ignored.    Painful though it may be to 
revisit past decisions, the country is ill-served if one does  not 
learn from them. Of course it was not as expensive to regulate TCE as 
it will be to regulate arsenic, and even less expensive to continue 
to regulate TCE. Society may therefore wish to retain the existing 
regulations,  but that should be done after careful examination 
containing the new perspective we hope that the regulators and the 
public will learn from arsenic.

I noted earlier that if lives are discounted at anything approaching 
the same rate as money, hazardous waste regulations and in particular 
high level nuclear waste regulations must be revisited.    But the 
comparison of arsenic with the proposed EPA regulatory procedures for 
nuclear waste in Yucca Mountain show a much more serious problem of 
consistency even when no discounting is sued.   We are pulling 
arsenic out of the ground into the environment and forgetting about 
it. Yet unlike chlorohydrocarbons which break down chemically, and 
nuclear wastes which eventually decay, arsenic remains carcinogenic 
for ever. The proposed Yucca mountain regulations apply only after an 
unlikely malfunction. But they are sufficiently restrictive that if 
they were applied to arsenic in the year 2001, most of the drinking 
water, and most of the agricultural land in the U.S.A. being in 
non-compliance even with no accident. This brings to the fore three 
possible explanations. Either (i) the proposed Yucca mountain 
regulations are excessively restrictive or (ii) the E.P.A. should 
propose that arsenic above ground should be treated as a toxic waste 
or (iii) the E.P.A. should come up with a clear scientific reason why 
the comparison is invalid or even a modicum of consistency is 
contrary to the public interest.