Rocket Science: Chapter 4: Polluting the Regulatory Process

Although a number of provisional standards for perchlorate have been issued since 1992, there is still no enforceable state or federal drinking water standard. Unfortunately, perchlorate is a prime example of the cumbersome and politically charged process of regulating chemicals in the United States today - especially when the lobbying efforts of private industry are coupled with the influence of the Air Force.

To set a drinking water standard, the EPA first derives a reference dose (RfD), the amount of a compound that is believed to pose minimal risk to all sectors of the population when consumed daily. This number is then converted into maximum contaminant level (MCL) allowed in drinking water.

The EPA issued its first provisional reference dose for perchlorate in 1992 at 0.0001 mg/kg-day, which translates into an MCL of 4 ppb based on an adult body weight. (EPA 1998; Jarabek 1998.) In 1995 the EPA reduced the uncertainty factors used in their calculations and the provisional MCL was raised to 18 ppb. (Jarabek 1998.)

In 1998 perchlorate was placed on the Contaminant Candidate List for National Primary Drinking Water Regulation. However, it was classified as a contaminant of secondary priority - one for which more data was needed before a standard could be set.

In late 1998 the EPA issued yet another revised RfD, based on new research that found effects at lower perchlorate doses. The RfD and corresponding MCL, however, were actually almost two times higher: 0.0009 mg/kg-day or 32 ppb. (EPA 1998.) This was due in part to the reduction of the uncertainty factors used in the calculation. An external peer review of the EPA's RfD held in early 1999, however, concluded that more research was needed before an official level could be set.

The EPA is now again in the process of revising the provisional RfD. A new one had been expected sometime in 2001, but the assessment has been postponed "indefinitely" because of delays in the submittal of data. (EPA 2001a.) This is far from the end of the process: The new RfD must pass through an external peer review, be converted into a goal MCL, and then an enforceable MCL, which takes health risks, treatment technology and cost into consideration.

In California, perchlorate is classified as an unregulated chemical for which monitoring is required. In 1988, the California Legislature passed a bill to establish a state drinking water standard for perchlorate by the end of 2000, but it was vetoed by former Gov. Pete Wilson. The Department of Health Services did adopt an action level of 18 ppb for perchlorate in 1997 after discovering that the chemical was contaminating groundwater all over California. DHS recommends public notification if the action level has been exceeded, and recommends drinking water source removal if the level of contamination exceeds 40 ppb.

Most California water suppliers have either taken wells that exceed 18 ppb out of service or blended them with uncontaminated water - another application of the "dilution is the solution to pollution" fallacy. Yet because the so-called action levels only triggers recommendations, water customers may not be aware that they've been drinking rocket fuel, unless they know where to look in the Consumer Confidence Reports water suppliers must now provide. If water suppliers continue serving water above the action level they are required to inform customers, but the requirement doesn't extend to informing customers that they were previously drinking contaminated water, or for how long. A state water engineer acknowledges that for most suppliers, "It [is] simply easier to shut down the wells rather than notify all the customers and deal with the public relations problem." (DC 1997.)

Misinformation is also not uncommon: A 1999 water quality report of the Arden Cordova Water Service in the Rancho Cordova area, for example, listed perchlorate as a water contaminant, but in the column stating the "potential sources of contamination," the source listed is "naturally present in the environment." (ACWS 1999.)

As of spring 2001, about 2,500 of the 16,000 public drinking water sources in California have been tested for perchlorate. The deadline for completion of statewide testing is December 2003. That data will guide the Department of Health Services in developing a Public Health Goal (PHG). Once a PHG is set, it typically takes two years before a final MCL is set. The entire state process, however, is being held up by the EPA's delay in setting a new provisional RfD.

Aerospace contractors and perchlorate makers, who stand to lose millions of dollars if a stringent standard is set for perchlorate in drinking water, have been fighting tougher standards for almost 40 years. As far back as 1962, the Manufacturing Chemists Association (now the American Chemistry Council) formed a committee on chemical propellant safety with a toxicity "task group", which included representatives from four companies in the solid propellant industry. Members of this committee participated in a Department of Defense working group called the Inter-Agency Chemical Rocket Propulsion Group (ICRPG), which was, according to the MCA memos, "the first time that a government agency has asked representatives of industry to participate in this type of committee activity." The MCA says "our active participation in the ICRPG program should be of great help in establishing safe but realistic rules and regulations without unnecessary and excessive restrictions to industrial operations." (MCA 1962, 1965.)

Thirty years later, as accumulating evidence of contamination and potential health effects raised the spectre of tight drinking water standards, industry formed the Perchlorate Study Group (PSG), consisting of Aerojet, Alliant Techsystems, American Pacific/Western Electrochemical Company, Atlantic Research Corporation, Kerr-McGee Chemical Corporation, Lockheed Martin, Thiokol Propulsion Group, and United Technologies Chemical Systems. In 1992, in cooperation with the Air Force, the PSG began a high-stakes campaign to block or weaken proposed standards.

In 1992 the EPA issued its first provisional RfD for perchlorate, 4 ppb. The calculations used a conservative uncertainty factor of 1,000. The PSG realized that the easiest way to raise the RfD was to reduce the uncertainty factor; it could easily get the RfD raised if the data deficiency value was reduced. The group paid an unknown amount for a more thorough literature review to be conducted and submitted enough additional data to the EPA for the agency to reassess the RfD. In 1995 the EPA reduced the uncertainty factors to 300, and issued a new provisional RfD of 18 ppb.

For the PSG and the Air Force, however, this was still not high enough. In 1996 the PSG and the Air Force hired a private firm named Toxicology Excellence for Risk Assessment, or TERA, to derive an "analogous" reference dose. In 1997, TERA sponsored a peer review of its own RfD which concluded that there still wasn't enough toxicology information on perchlorate to do a credible risk analysis.

Soon after that, the review panel outlined a series of studies to be conducted by the Air Force and the PSG. According to a 1997 Air Force memo, the goal of the studies was clear: "[C]omplete the necessary studies to fill in the missing data gaps in order to see if the provisional reference dose can be raised." (Rogers 1997.) The reference dose subsequently recommended by Aerojet, McDonnell-Douglas and other PSG members was 20 to 100 times greater than the EPA's provisional reference dose. (Aerojet/McDonnell-Douglas 1997.)

A June 1997 Defense Contract Management Agency (an agency of the Department of Defense) situation report on the perchlorate contamination near Lockheed Martin's former site in Redlands, explained the high stakes for the U.S. military:

"Significant Political, Legislative, Military or Diplomatic Impacts: There may be far reaching ramifications when the public learns of this situation. Adverse media attention and congressional interest similar to/greater than Aerojet Rancho Cordova situation may occur. The government will have to deal with legal liability issues. . . . This discovery is likely to increase Lockheed Martin's environmental tort litigation case load and their environmental cost claims against the government. . . . Future procurement programs could be adversely affected due to increased environmental related costs." (DCMDW, 1997.)

The report acknowledges that people most at risk "are those with existing thyroid gland problems, pregnant women and children," but urges: "This information has not been released to the public by environmental and public health regulators. Please keep this information close hold. Failure to do so will adversely affect . . . proactive environmental efforts." Just what those proactive efforts might be are not identified.

In 1998, after Native Americans living along the Colorado River raised concerns about the accumulation of perchlorate in their irrigated cash crops, the Defense Department financed an EPA study on the uptake of perchlorate by lettuce. Lettuce was chosen in part because 80 percent of the nation's winter lettuce crop is irrigated by the lower Colorado River. The Defense Department at first declined to give the scientists conducting the research permission to share their results with EPA scientists. After activists complained about the secrecy, preliminary results were released in a 1999 conference that showed relatively high levels of perchlorate uptake in young lettuce plants, but low levels of perchlorate in mature plants. But the full results have still not been released, and the Air Force has denied EWG's repeated requests for the data.

The Air Force has also declined to release information in response to a Freedom of Information Act inquiry by a law firm regarding military research on the health effects of perchlorate, the extent and causes of contamination, and studies which involved giving perchlorate to human subjects. The Air Force responded that this information was "fully exempt from disclosure" until the EPA peer review had been completed because this would expose the AF's "deliberative process." (RLG 2001.)The denial has been appealed.

In August 2000, Lockheed Martin launched a $1.75 million study at Loma Linda University near San Bernardino, Calif., in which "volunteers" were paid $1,000 to take a daily dose of perchlorate for 6 months. The doses ranged from 0.5 mg to 3 mg of perchlorate a day; the highest amount corresponds to a dose 83 times higher than California's current action level. Three other studies where "healthy male volunteers" are being fed perchlorate-laced drinking water were also underway at that time. Two of these, being conducted at Harvard and Oregon State universities, are being funded by the PSG, and the other, being conducted by a scientist in Germany, is apparently being sponsored by the Air Force. (TERA 2001.)

In November 2000, when EWG broke the news of the Southern California tests, Loma Linda researchers acknowledged only that they were sponsored by Lockheed Martin. However, in its latest SEC filing, Lockheed discloses: "We also are coordinating with the U.S. Air Force, which is conducting preliminary studies of the potential health effects of exposure to perchlorates in connection with several sites across the country." (SEC 2001.)

Almost all of the recent toxicology research on perchlorate has been conducted or funded (or both) by the Air Force, the PSG, or a specific company. The EPA's 1998 report offering a revised RfD describes a number of occasions when it analyzed test data submitted by the Air Force or the PSG and found that significant results had been ignored.

For example, the EPA criticized a PSG-sponsored study by Argus Laboratory which overlooked the significance of a perchlorate-induced increase in the size of one brain region by almost 30 percent and of unexpectedly high increases in motor activity. The report stated that EPA disagreed with the argument "that these effects are 'not suggestive of a nuerotoxic effect' because of an 'unknown biological significance.' The EPA considers increase in the size of any brain region to be a potentially adverse effect." (EPA 1998.) Moreover, the EPA noted that "Argus Laboratory and the sponsor (PSG) have failed to respond adequately to the request for an explanation." (EPA 1998.) Similarly, when the EPA analyzed thyroid data provided by the Air Force, they found significant physiological changes at lower doses than had been reported and said the Air Force "did not provide a reason for discounting the significance" of changes at lower doses. (EPA 1998.)

Although children are at far more risk for the effects of thyroid hormone disruption than adults, federal and state regulators have consistently ignored children in their calculations. Children not only drink more water relative to their body weight, but their developing brains and bodies are more susceptible to hormone disruption. Studies on rats, guinea pigs and rabbits have consistently found perchlorate-induced effects in the thyroids of pups to be greater than the effects on those of the mothers. (EPA 1998.)

When it comes to the standard assumptions of body weight and water consumption used to convert the RfD to an MCL, however, the EPA uses an adult body weight of 154 pounds (70 kg) and drinking water consumption of two liters of water a day. (Jarabek 1998; EPA 1999b) These faulty assumptions were used in all of the EPA's MCL calculations (1992, 1995 and 1998), as well in the DHS' derivation of the current action level. (EPA 1998; CADHS 2001.) DHS has indicated, however, that the California drinking water standard now in development will take children's health into account, and have stated that it is concerned about "the ability of perchlorate to interfere with the production of hormones by the thyroid gland, and the need for thyroid hormones for normal prenatal and postnatal development." (EPA 2001a; CADHS 2001.)

Although all children consume significantly more water than adults relative to their body weight, this is especially true for the almost 40 percent of infants who, in their first four months of life, drink formula made with tap water. (EWG 1999.) These bottle-fed infants consume more than seven times as much water as adults relative to their weight. If just this one error is accounted for, both the EPA's MCL and the California action level would be 7.5 times lower than their present values.

Despite weighty evidence to the contrary, the EPA apparently doesn't consider significant changes in thyroid hormone levels to be a problem for developing children. The EPA ignored critical data showing effects at concentrations ten times lower than the study on which the agency based the most recent provisional RfD.

In its 1998 report, the EPA emphasizes how devastating and permanent developmental effects are "caused by a lack of thyroid hormones" rather than by tumor development or thyroid structure change. The agency notes that "the earliest biological effect, changes in thyroid and pituitary hormones, is the precursor lesion for both the potential carcinogenic and neurodevelopmental effects." However, this seems to have been forgotten when it came to developing an RfD: The principal study used by the EPA to derive the 1998 RfD was one which showed changes in thyroid structure at perchlorate concentrations of 0.1 mg/kg/day. Yet, this level of perchlorate is ten times higher than was shown to affect thyroid hormones. (EPA 1998.) The explanation given by the EPA is that it was unclear what "degree of change to designate as adverse."

The EPA is on a slippery slope: Assuming that some degree of change in thyroid hormone levels might not do much harm, but admitting that we don't really know. The EPA's solution: Ignore the hormone data entirely and hope that no one would notice. If the RfD were based on changes in hormone levels rather than on changes in thyroid structure, it would be ten times lower than the current value.

These two errors in EPA's calculations - ignoring the effects on children and ignoring critical data - have resulted in a provisional MCL which is only nominally below the perchlorate concentration likely to have an effect on a child's thyroid hormone levels. For an 11.5-lb. bottle-fed infant drinking 1.1 liters of water a day, water with a perchlorate concentration of only 40 ppb would likely have an effect on the child's thyroid hormone levels which are critical for proper development. It is possible that lower levels of perchlorate may also have an effect since no lower doses of perchlorate were tested. The EPA's provisional reference dose of 32 ppb is clearly not low enough to protect the health of infants and children.

Uncertainty factors are used to make standards conservative in order to account for more susceptible portions of the population, differences between the physiology of study animals and humans, data deficiencies and other unknowns. In 1998, the EPA reduced the uncertainty factor used in deriving its RfD to 100, radically underestimating the uncertainty surrounding perchlorate and failing to provide sufficient protection for sensitive human populations. Just some of the unknowns surrounding perchlorate's effects: whether it is being concentrated in breast milk; whether rats are more or less sensitive to thyroid hormone disruption than humans; whether the studies that have been done can detect subtle nuerodevelopmental effects stemming from perchlorate exposure; whether perchlorate is concentrated in the tissues of food crops grown with tainted water.

Considering the large percentage of the population who are already hypothyroid, the EPA's RfD also fails to adequately protect this sensitive subpopulation. The EPA typically builds in an extra tenfold margin of safety to account for people who might be more susceptible to the harmful effects of a certain compound than others. In the case of perchlorate, however, the safety margin was only threefold. This is a serious error in light of the fact that hypothyroidism is becoming increasingly common in America. A recent study found that, on top of the 10 million Americans already diagnosed with a thyroid problem, an additional 13 million (or almost 10 percent of the population in total) may have an undiagnosed thyroid condition. Furthermore, 90 percent of these people are likely to be hypothyroid rather than hyperthyroid. (Canaris et al. 2000.)

The RfD also overlooks the fact that perchlorate is only one of a number of anti-thyroid compounds to which we are exposed during our everyday lives, including pesticides, dioxins and PCBs. (EPA 1997.) Furthermore, recent research has shown that mixtures of pesticides and nitrates in drinking water have synergistic affects, and concentrations of these chemicals at the same order of magnitude as current MCLs have been found to have significant effects on thyroid hormone levels. (Porter et al 1999.) It is likely that perchlorate would compound these effects, yet no such consideration is addressed by the EPA or figured into the RfD calculations. Moreover, a reviewer of the EPA 1998 provisional RfD noted that the EPA has not adequately addressed the subtle adverse effects of altered thyroid hormone levels in fetuses and children, nor have they addressed the possibility that perchlorate blocks the uptake of iodine into milk and that perchlorate itself may be concentrated in breast milk. (EPA 1999c.)

Several internal EPA reviewers argued for an uncertainty factor of at least 300, which would decrease the current provisional RfD by a factor of three. (EPA, 1998.) In light of the major scientific uncertainties, the significant proportion of the population who would be highly susceptible to perchlorate's effects, and the many other anti-thyroid compounds we are already being exposed to, EWG believes that a UF of 1,000 is the minimum necessary to safeguard public health.

EWG's analysis shows that the MCL should be 7.5 to more than 2,000 times lower than the standards proposed by the EPA. While the EPA has been citing a proposed value of 32 ppb, its RfD and MCL calculations neglect many considerations critical to public health.

Following the EPA's guidelines of RfD and MCL derivation, but using assumptions, appropriate to protect children, EWG's calculations show that the MCL for perchlorate should be between 0.04 and 4.3 ppb. The highest value in this range (4.3 ppb) is the result of changing only the EPA's "standard assumptions" of adult body weight and drinking water consumption, to those of bottle-fed infants. The lower values in the range also take into consideration other key isues, such as the thyroid hormone data the EPA ignored and a more realistic estimation of the uncertainties.