Sign up to receive email updates, action alerts, health tips, promotions to support our work and more from EWG. You can opt-out at any time. [Privacy]


Setting the Record Straight

EWG Response to 'A Review of the Science, Methods of Risk Communication & Policy Recommendations in Tap Water Blues'

The Messages

October 1, 1995

Setting the Record Straight: The Messages

In this section, Baker critiques what he sees as the major messages of Tap Water Blues.

Message 1. "Herbicides in drinking water are the greatest public health threat posed by pesticides today."

Baker provides no substantive criticism of this conclusion in his response. Instead, like so much else in his review, his criticism is comprised largely of his opinions of our public relations work and his claim that the risks from herbicides in drinking water are negligible. The fact that these risks are not negligible is discussed in detail below. More to the point, the EPA's recent actions against the triazines strongly support our conclusion.

On November 10, 1994, three weeks after the release of Tap Water Blues, the Environmental Protection Agency initiated a Special Review of the triazine herbicides, a process that will almost certainly lead to further restrictions on their use, and may ultimately lead to a ban. This Special Review represents the first time since 1988 that a regulatory action has been initiated against a pesticide on the basis of general risk to the public from contamination of food and water.

Message 2. "Federal drinking water standards are among the weakest of all national environmental regulations."

Baker basically agrees with our conclusion, as does the EPA. What Baker attempts to argue, nonetheless, is that drinking water standards, while weaker than the standards applied to the same herbicides in food, do not allow greater than negligible risks and that this "negligible risk has been achieved for the most part without any regulation."

The flaw with this assertion is that the risks from herbicides in drinking water are not negligible. Not only have existing regulations or lack thereof (depending on the herbicide in question) failed to achieve a negligible level of risk, they have in fact contributed to the significant risks from herbicides in drinking water faced by millions of people. As noted in Tap Water Blues and supported by the EPA (See Note 1.), the current MCLs for atrazine and simazine, and the LHA for cyanazine, permit levels of contamination from 13 to 29 times greater than negligible for each individual chemical.

The EPA defines negligible cancer risks as those that do not exceed 10-6. For example, in testimony before Congress in September 1993, EPA Administrator Browner defined negligible risk as follows:

"The reasonable certainty of no harm/negligible risk standard represents an upper bound risk of one in one million (10-6), calculated using conservative risk assessment methods."

Tap Water Blues documents, and Baker does not dispute the fact that at least 3.5 million people in the Midwest are exposed to herbicides at risk levels that are ten to one hundred times greater than negligible (10-6). The EPA supported this conclusion in the documentation supporting the initiation of the Special Review, where risks from triazines in drinking water were calculated at 1x10-5. Contrary to Baker's claim, this special review was initiated precisely because the risks from these herbicides in food and water are greater than negligible.

Baker is simply wrong when he asserts that the current cancer risks from triazine herbicides in water are no greater than would be allowed under the negligible risk standard applied to residues of these herbicides in food. Even the EPA has since recognized this fact. On October 20, 1994, EPA Administrator Carol Browner ordered the water division of the EPA to accelerate efforts to harmonize drinking water standards and bring protection offered by drinking water standards in line with those of the pesticide program (See Note 2.)

Message 3. "14.1 million people routinely drink water contaminated with five major agricultural herbicides (atrazine, cyanazine, simazine, alachlor, and metolachlor)."

As Baker clearly notes, Tap Water Blues underestimates the total population exposed to these herbicides in drinking water. In Baker's own words, "...far more than 14.1 million people undoubtedly ingest these herbicides in drinking water." Baker then dismisses this additional exposure as insignificant with the claim that "...most involve concentrations that are inconsequential with respect to human health effects." But Baker provides no data to support this conclusion.

In fact, the opposite is more than likely true. There are many areas with heavy herbicide use and poor drinking water monitoring where additional highly exposed populations certainly exist that were not identified in Tap Water Blues. Most of the state of Indiana is just one such example.

At the same time its is true that millions of additional mid-westerners beyond those identified in Tap Water Blues consume very low levels of triazine herbicides in drinking water drawn from the Great Lakes. These risks, while not "inconsequential", are at least well below levels that the federal government currently considers negligible.

What Baker argues throughout his critique, however, is that these low risk populations actually justify the high risk situations because more people are at low risk than high risk, and that therefore the average risk of the two populations is negligible. This rationalization is advanced relentlessly throughout the critique, even though this average risk is totally hypothetical, and not representative of the actual risk faced by anyone.

Message 4. "More than 3.5 million people in 120 cities and towns face cancer risks more than 10 times the federal cancer risk benchmark, based on annual exposure to these herbicides."

Here Baker makes several accusations that are repeated throughout the document. Rather than repeat our rebuttal of these points throughout our reply we address them here and reference the pages in Baker's critique to which they apply. They are, first, that we fabricated a risk standard(see Baker pages 1, 4, 14, 20), second, that additive risk cannot be calculated and that therefore combined standards cannot be set (see Baker page 7), and third, that we overstate the relative risks of the herbicides compared with other sources of cancer by only focusing on the exposed population (see Baker pages 4, 11, 22, 23, 31-36).

Point One (see Baker pages 1, 4, 14, 20).

Baker wrongly accuses us of fabricating a federal cancer risk standard that is five times more restrictive than any existing federal standard.

Baker is correct that we have added the exposure to the five herbicides in our study to calculate the cancer risk faced by the population drinking herbicide contaminated water. He is also correct that we have compared this total risk to the one in one million cancer risk standard. What he fails to mention is that our method of adding cancer risks is in accord with the EPA Cancer Risk Assessment Guidelines (See Note 3) the recommendations of the National Academy of Sciences (NAS) Committee on Pesticides in the Diets of Infants and Children (See Note 4), and two previous NAS reports that explicitly examined risk assessment methods for complex chemical mixtures (See Note 6), including one that specifically studied this problem in drinking water: Drinking Water and Health: Volume 9 (See Note 5.)

Adding cancer risk as done in Tap Water Blues is widely accepted within the scientific community. As discussed previously, weakening the negligible risk standard from one in one million to five in one million depending on the number of chemicals in question (as Baker has done) is not.

What Baker cannot seem to comprehend is the overall purpose of the report, which is to point out the flaws and shortcomings of current drinking water standards and the current pesticide regulatory system. We have not fabricated a cancer risk standard, the standard is one in one million and we have used this standard. What we have done, however, is reveal the disingenuous nature of the standard when it is applied one chemical at a time. It is clear that the one in one million negligible risk standard is not a negligible risk standard at all when scores of chemicals are allowed into the water, air, and food supply each at a one in one million level of risk.

For example, 67 pesticides and metabolites have been detected in midwestern drinking water source water since 1987. At least 20 of these are carcinogens. If these 20 carcinogens are allowed in drinking water at a one in one million risk level, the overall risks from these 20 pesticides is 20 in one million, or 20 times the level that is deemed negligible. By extension of Baker's approach, a 20 in one million level of risk would be the appropriate negligible risk standard for these 20 compounds. Carrying this reasoning to its logical conclusion, the appropriate negligible risk standard for the 70,000 toxic chemicals in use today would be 70,000 in one million, or 7 in one hundred. Baker's approach has no medical or toxicological basis: simply put, a person's resistance to a carcinogen does not increase each time when he or she is exposed to an additional cancer causing chemical.

Point Two (see Baker page 7)

Baker then attacks our proposal for additive drinking water standards on the basis of their complexity. He argues that the additive risk from these toxicants is not measurable and is too difficult to calculate due to the infinite combinations of exposure levels that one might encounter in this mixture. And according to Baker, because these risk assessments are complicated, we should not attempt them, nor should we provide the public with added protection from known chemical mixtures in drinking water while scientists struggle to determine their true toxicity. EPA clearly believes that it is possible to set a drinking water standard based on total triazine exposure, and has already stated their intention to consider this approach. On October 17, 1994 the EPA rejected a petition by Ciba Corporation to weaken the current maximum contaminant level for atrazine from 3 ppb to 20 ppb. In the letter denying the petition, EPA administrator Carol Browner cited the need to examine the combined toxicity of the triazine herbicides along with their chlorinated metabolites:

"We are also considering whether to regulate the chlorotriazines (i.e. atrazine, cyanazine, simazine) as a group rather than regulating atrazine alone. The triazines have similar structure, mode of action, toxicity and degradates." (See Note 7).

While the complexities that Baker cites do exist, they do not negate the reality of the risk that people face from drinking mixtures of these chemicals. In fact, it is precisely because scientists cannot accurately assess the risk from chemical mixtures in drinking water that we need to look toward strategies that reduce exposure of at the source, rather than blithely dismissing the risks as negligible, when in truth we do not know the toxicity of these mixtures.


It is true that the EPA has been loath to adopt the practice of adding cancer risk from related exposures (i.e. toxicants in the same media, or from the same source), even in obvious cases like the triazine herbicides where the toxicity of the pesticides is so nearly identical, and where they occur together in drinking water. This has begun to change, however. The Special Review of the triazine herbicides (atrazine, cyanazine, and simazine) will assess the combined risk and benefits of these herbicides, a first in the history of the EPA. The standard that will serve as a benchmark for negligible risk is the one in one million cancer risk standard.

Point Three (see Baker pages 4, 11, 22, 23, 31-36).

Finally, Baker voices one of his major concerns, which he calls the "paradox of inequitable distribution of negligible risks." In essence, Baker argues that it is legitimate for a subset of the population to be exposed to high risks if another part of the population is exposed to lower risks, and if enough low risk people can be averaged with the high risk population so as to make the total population risk seem insignificant. A comparison between Baker's use of his data and our use of Bakers data in Tap Water Blues is instructive on this difference in approaches.

Baker's own data for Ohio, Illinois, and Iowa -- which understate risk because they only look at one herbicide, atrazine, and do not quantify metabolites -- indicate that over five million individuals drink atrazine contaminated water that presents a greater than a one in one million risk. Baker argues, however, that the risk for these people is actually overstated, and that the appropriate way to characterize the risk to these five million exposed individuals is to average the risks they face with the risks faced by the additional 18 million people in these states who drink water with little or no contamination by these compounds. In essence, Baker argues that we can reduce the risks faced by the citizens of Columbus, if we simply average their risks with the citizens of Cleveland who drink water with little or no triazine contamination.

If widely adopted as policy, Baker's reasoning would justify exposing millions of people to unacceptably high levels of weed killers in tap water (or any other toxicant) any time one could identify an equally large number of people somewhere who are not. Baker is playing with numbers and ignoring lives. In contrast, the aim of the risk assessments in Tap Water Blues is to accurately characterize the risks faced by exposed individuals so that policies may be devised to protect them.

Message 5. "Across the Corn Belt, we estimate that at least 241,000 individuals are supplied by surface water sources where average annual herbicide levels exceed federal standards."

Here Baker argues that drinking water standards are strong, that they provide adequate protection from triazine herbicides in drinking water, that we ignored the strengths of these standards in our report, and that even people drinking water that is out of compliance with federal standards are unlikely to face any health risks.

Baker attempts to downplay the risk of the triazines by citing an outdated (1990) characterization of the toxicity of atrazine that makes no mention of the cancer risks that the EPA currently regards as the critical toxic effect of atrazine and the triazines as a class. This reference does not reflect current scientific thinking on the toxicity of the triazines, nor does it capture the EPA's concern over the inadequacies of current drinking water standards for the triazines.

On October 17, 1994, EPA Administrator Carol Browner pointed out several weaknesses of current drinking water standards (beyond the use of the safety factor methodology) in her denial of a Ciba petition to weaken the MCL for atrazine from 3 ppb to 20 ppb:

It is, for instance, possible that the Agency may have underestimated the risk from atrazine exposure in drinking water. EPA is assessing whether it should include the chlorodegradates of atrazine in its risk assessment. EPA based the MCL on the toxicity of atrazine alone, and this approach may have underestimated the actual risk posed by the use of atrazine. In other program areas the Agency has included the chlorodegradates when calculating the dietary risks of atrazine and other triazine compounds.

We are also considering whether to regulate the chlorotriazines (i.e. atrazine, cyanazine, simazine) as a group rather than regulating atrazine alone. The triazines have similar structure, mode of action, toxicity and degradates. We do not account for the potential additive increase in cancer risk due to exposure to the components of the chlorotriazines mixture, and therefore may be understating risks when regulating the contaminants individually (See Note 8.) In contrast to Baker's view, the current EPA position is that atrazine and the triazine herbicides are so toxic and widely present in food and water that they warrant special regulatory review based on the cancer risk that they present to the general public and pesticide applicators. The EPA is conducting this special review using the quantitative cancer risk assessment methods employed in Tap Water Blues, not the safety factor methodology advocated by Baker.

We clearly acknowledge the existence of safety margins in the methodology used by the EPA's Office of Drinking Water. Tap Water Blues, however, makes three fundamental criticisms of this methodology:

  • The safety factor method is inappropriate for these carcinogens and is not used to regulate carcinogens by any other regulatory division of the EPA;
  • The safety factor approach applied to the triazines has been rejected by the Office of Pesticide Programs which has lead regulatory authority for pesticides at the EPA, and;
  • The methodology produces weak health standards.

These standards:

  • Are far higher (up to 29 times higher for just one herbicide, cyanazine) than allowed from these same herbicides in food;
  • Are particularly inappropriate because most exposure occurs via water;
  • Do not take into account exposure to multiple chemicals with similar toxicological properties, and;
  • Result in MCLs, or legal levels of contamination that are not based on the most sensitive endpoint (cancer) observed in animal studies, but instead on a health endpoint that is in many ways irrelevant to the toxicity and potency of the triazine herbicides.(See Note 9.)

Messages 6 and 7
"In the Corn Belt, an estimated 65,000 infants drink these herbicides from birth via infant formula reconstituted with contaminated tap water."
"Earlier studies have shown that children are particularly susceptible these chemicals."

Baker's principal contention here is that these statements are meaningless and designed to frighten the public.

Baker is wrong on both counts. These statements are neither meaningless nor designed to frighten the public. (Interestingly, one of them -- "Earlier studies..." -- was not even in Tap Water Blues, but was taken from a press advisory.) They are designed instead to educate the public on the reality of herbicide contaminated drinking water, and the potential risks that this contamination presents to children.

As our study shows, infants and young children are exposed to a disproportionate amount of lifetime exposure simply because they drink more water per unit of bodyweight. And as Baker acknowledges in his critique,

"... children are probably more susceptible to all chemicals than adults are."

Meanwhile, two years after the National Academy of Sciences recommended that pesticide regulatory standards be strengthened explicitly to protect children, not a single food or drinking water standard for pesticides has been changed to meet this goal. As it stands today, neither drinking water nor food standards for pesticides make any specific accommodation for the sensitivity or higher levels of exposure faced by infants and children.

In attempting to support his claim that it is safe to feed infants and young children herbicide contaminated tap water, Baker makes a very revealing analogy between atrazine and aspirin. Baker goes to great lengths here and elsewhere (page 38) in his critique to compare the acute toxicity of aspirin with the acute toxicity of atrazine, although no scientist that we are aware of considers the acute toxicity of atrazine or aspirin a serious issue. On these grounds alone the comparison is completely irrelevant. More seriously, however, it reveals an ignorance or lack of concern about the potential health risks that children face when they consume chemicals that are known to be toxic but are thought to be safe for children based on the results of studies in sexually mature animals and their safe use in human adults.

As it turns out, the relevant health risk for children from aspirin is not acute toxicity, but Reyes syndrome, a condition that kills over half of the children it afflicts. Notably, the link between aspirin and this fatal condition in children was not uncovered by routine animal testing, but was discovered only after children died from consuming aspirin at levels that were previously thought to be safe. Today aspirin in not recommended at any dose for children with a fever.

The lesson that aspirin and Reyes syndrome teaches us is not Baker's claim that all toxic substances are safe for children if only the doses are small enough. Instead, the lessons are twofold; first, that some substances are not safe for children at any dose, even when they are safe for adults; and second, that in some cases no amount of animal testing will reveal these differences, as was the case with aspirin.

That is why, in 1993, the Committee on Pesticides In The Diets of Infants and Children of the National Academy of Sciences concluded that, " Profound differences exist between children and adults...these differences can affect the toxicity of pesticides in infants and children," and that "in the absence of data to the contrary, there should be a presumption of greater toxicity to infants and children." (See Note 10.)