Testimony on High Production Volume Chemical Tests to the House Subcommittee on Energy and the Environment
Neal D. Barnard, M.D.
June 17, 1999
As a physician and president of the Physicians Committee for Responsible Medicine, I would like to thank Chairman Calvert and the members of the Committee for the opportunity to comment on the EPA’s High Production Volume Challenge program. We are greatly concerned about several aspects of this hurried program.
The program is based on a mistaken premise, that for a great many chemicals, basic safety information is missing. In December 1998, we conducted an analysis that casts serious doubt on this notion. Indeed, we found that a surprising amount of information is already available for many high-production volume chemicals, making further tests unnecessary and inappropriate.
We examined the test data already available on a sample of chemicals from the EPA’s HPV Challenge Program Chemical List, as updated December 9, 1998. EPA’s list summarizes available toxicity information based on the Registry of Toxic Effects of Chemical Substances (RTECS), the Hazardous Substances Data Bank (HSDB), TOXLINE, and MEDLINE. The EPA’s list suggests that gaps in knowledge are often found, gaps which were also described in the Environmental Defense Fund’s (EDF) summary report, Toxic Ignorance.
We found that the EPA, in its hurried review, simply did not check the right databases. In our evaluation, we added the Toxicology, Occupational Medicine and Environmental Series Consolidated Point Solution (TOMES CPS), which includes toxicological data drawn from a wide variety of sources, including the Chemical Hazard Response Information System (CHRIS), HazardText, Meditext, Reprotext, Integrated Risk Information System (IRIS), Reprotox, the Teratogen Information System (TERIS), Shepard’s Catalog of Teratogenic Agents, Fisher Scientific, and other Material Safety Data Sheet (MSDS) collections.
We found that data thought by the EPA to be lacking were indeed available, often conforming to or exceeding the SIDS protocol. For example, we found, in some cases, chronic toxicity tests lasting three months or longer, instead of the 14- or 28-day repeated-dose assay called for in the SIDS protocol.
In addition, human data are often available. Toxic exposures have been elaborately studied in exposed workers, in cases of accidental and deliberate overdoses, and in other settings, but these data have been and will be completely ignored in the HPV program. Indeed, if an entire schoolyard full of children or an entire neighborhood were affected by a toxin, the HPV program has no interest in these data whatsoever. It demands instead that rats and mice be used as indicators of human risk.
Here is a sampling of what our review found:
Tetraethyl lead: According to the EPA’s listing, data regarding chronic toxicity are either insufficient or are not publicly available. Following the SIDS protocol, this means that a 14- to 28-day rodent assay is in order. Clearly, however, such testing is completely unnecessary, inappropriate, and a violation of the mandate to avoid duplicative testing.
Tetraethyl lead, the antiknock agent in leaded gasoline, is a recognized carcinogen, developmental toxicant, and reproductive toxicant. Acute and chronic lead poisoning in humans is well known and thoroughly described. In addition, we identified, among others, a 20-week chronic toxicity study in rats and a 3-week exposure test in mice for immunologic toxicity. Further animal tests would be indefensible. I would be surprised indeed if anyone at the EPA honestly had any doubts as to lead’s acute and chronic dangers or felt that a 14-28 day rodent assay would be helpful.
A second example is carbon tetrachloride. There is no question whatsoever about its dangers, both to humans and animals. While, according to the EPA, chronic toxicity data are missing, we found that the effects of acute and chronic exposure in animals have been documented in several studies. For example, a 12-week rat study conducted at the University of Georgia clearly showed hepatic toxicity. More importantly, the dangers of acute and chronic carbon tetrachloride exposure have been documented in humans in several studies.
Lest these seem unusual cases, yet another example comes from turpentine. No one would suggest that further tests are necessary in order to prove that turpentine is not safe for acute or chronic exposure. Yet, believe it or not, the EPA cites a lack of data for chronic toxicity. In addition to human toxicity data, we found that 4- and 8-week rat studies had already been done.
Also on the EPA’s list is butane, for which testing for acute toxicity to fish, daphnids, and algae, and terrestrial toxicity was noted by EPA to be lacking. However, we found detailed studies showing that butane’s solubility is negligible, it has little likelihood of bioconcentration, and it is rapidly volatile from most surface soils. Attempting to test it in animals is a ludicrous proposition.
Here are our overall findings: Of the 70 chemicals cited in the EDF’s report as lacking basic test data, we took a sample of 35. Of these, 22 remained in the HPV program. The remainder had been eliminated for various reasons. Of these 22 chemicals that were supposedly missing basic test data, we found test results easily satisfying the SIDS test battery for 13. In other words, of this sample of chemicals thought to be missing test data, more than half were not missing these data at all.
The EPA now acknowledges that its review was flawed and that more data are available than initially believed, but despite our bringing this problem to the Agency’s attention six months ago, it has shown no interest in modifying its headlong rush into testing. Turpentine, carbon tetrachloride, cyclonite rat poison, and the other chemicals I mentioned are still on the HPV list.
Some at EPA have defended keeping these chemicals on the list, saying that we need to know their effects on the environment. This is nonsense. The tests the EPA is calling for, in the case of turpentine, carbon tetrachloride, tetraethyl lead, and cyclonite rat poison, among many others, are not environmental tests at all. The ecotoxicity tests are already finished for these chemicals. For each of these chemicals, the tests EPA is calling for are 14- to 28-day rodent poisoning studies that are intended to provide a crude estimate of the effects of ingestion. As we have seen, these tests have already been done.
The EPA aims to throw the burden on industry to bring available test data forward, and then, if any tests appear to be missing, it will require industry to contract with laboratories to run all the missing tests. This check-the-box mentality misses the fact that many of these tests are utterly irrelevant and a waste time, money and animal’s lives. For example, if turpentine, rat poison, or leaded gasoline had actually not already been used in 14- or 28-day rodent tests, I doubt that anyone in this room would feel better if industry were to force-feed these chemicals to small animals. We already know, from many different lines of data, that these chemicals are highly toxic.
The real issue is, are we exposed to these chemicals and, if so, what can we do to stop these exposures? If current laws do not provide adequate protection, then our priority should be to seek a modification of the laws, not to demand further use of crude tests.
The EPA may hold that test results will give it the ammunition it needs to go after polluters. The HPV program may, in fact, do the opposite. The HPV tests are simple and crude. Because animal tests are very often inaccurate, the program may in fact be used to exonerate toxic chemicals, as I will describe in more detail momentarily. The EPA’s challenge is not to again show that these chemicals are toxic to rats or mice, but to show that humans are, in fact, exposed to them. The HPV program ignores human exposures, and demands testing, whether humans could potentially be exposed or not.
It must be remembered that the HPV program’s goal is not to perform a thorough toxicity battery on any chemical. Rather it is to determine whether crude screening information exists. In this situation, animal tests are often so inaccurate that they do more harm than good. The inability of animals to replicate human exposure results leaves plenty of room for interpretation and disagreement. Tobacco presents a tragic example.
The link between cigarette smoking and lung cancer is abundantly clear from human studies. However, animal tests muddied the waters. Tobacco tar painted on animals’ skin led to tumor production, while animals forced to inhale tobacco smoke in one study after another simply failed to develop bronchogenic carcinoma. The tobacco industry was able to argue that since humans inhale tobacco smoke, rather than paint it on their skin, it did not appear to be a dangerous substance. Far from helping public health programs, these tests worked to the tobacco industry’s advantage. In the New England Journal of Medicine of June 15, 1961, Clarence Little pointed out, “There have been many such experiments here and abroad, and none have been able to produce carcinoma of the lung in animals.” “Lung cancer,” he wrote, remained “a challenge, an unsolved problem,” much to the benefit of his employer, the Tobacco Industry Research Committee.
The lesson is simple: If our goal is to gather preliminary information about chemical risks, we should not underestimate what we have in hand now, nor should we assume that further animal tests will more clearly indict dangerous chemicals. Indeed, they may well do the opposite. The HPV program could discourage regulators from taking action on chemicals that are suspected hazards based on worker exposures or other clinical and epidemiologic observations.
The HPV program is rushed and poorly designed, making a comprehensive data review impossible. The evaluation of high-production chemicals should instead be put on a schedule that allows a careful and deliberate analysis. As it is, the EPA plans not to review test plans, not to look at test results, and, indeed, not to do anything whatsoever based on the results of the program.
When additional testing is desirable, the EPA must require tests that are more modern, more accurate, and more humane than those currently in the HPV protocol. So far, it has side-stepped that responsibility. Particular priority should be given to human cellular tests. The tremendous advantage of testing in what we might call “the right species” is obvious when we consider the pathetic performance of animal tests.
Recently, Dr. Bjorn Ekwall and colleagues in the Multicenter Evaluation of In-Vitro Cytotoxicity (MEIC) trial based in Uppsala, Sweden, showed that the rat and mouse lethal dose (LD50) tests, which are required under the HPV program, are only about 65% accurate in predicting human risk. (rat LD50 R2 = 0.61, mouse LD50 R2 = 0.65). The rodent tests showed varying degrees of inaccuracy for 50 test chemicals, and were wildly inaccurate predicting human lethal doses for nine chemicals. For one of these, digoxin, the problem was clear: Rodents have less Na/K ATPase enzyme activity, compared to humans, rendering rodent tests uninterpretable. (The remaining eight chemicals included nicotine, propoxyphene hydrochloride, mercury chloride, thioridazine hydrochloride, warfarin, chloroform, sodium oxalate, and atropine sulphate.)
In fact, in many cases, rat LD50s are not even good predictors of mouse LD50s. The MEIC review found that rat tests grossly underpredicted results of mouse tests for several chemicals. An animal LD50 test could exonerate a chemical or make it appear safer than it is. It should not be forgotten that rodent LD50 tests have never been formally validated.
The LD50 is a World-War-I-era test that is a grossly inaccurate measure of human risk. It is to medical investigation what a divining rod is to geological investigation.
Human cell tests are far better predictors of human toxicity. The MEIC program tested 61 different in-vitro assays for their correlations with human lethal blood concentrations, and compared the results with LD50 tests on rats and mice. It tested 50 chemicals which varied widely in their chemical properties, using 29 independent laboratories. Unlike the 65% accuracy of rat and mouse tests, a combination of three human cell tests predicted the toxicity of chemicals with 77% precision, and the addition of a fourth test improves the battery’s accuracy to approximately 80%. Statistical adjustments can be made depending on a test chemical’s ability to pass through the blood-brain barrier. Moreover, human cell tests may also provide information that is useful in judging ecotoxicity.
When testing is required, it is vital that the EPA call for the most accurate methods available, not the archaic and extraordinarily cruel animal LD50 test.
If validation of these testing methods for interagency use is desired, this process could be completed in a matter of months. Given the time, money, and animals’ lives that will be lost in this program, this more cautious approach is clearly desirable. While the EPA rightly states that test data should be accepted internationally, it has, in fact, taken a key role in setting international standards. An international effort to eliminate the cruel LD50 test, which kills half the animals used, was actually scuttled by the EPA, which wanted to retain this archaic test in the HPV program.
In summary, the EPA recognizes that its analysis of testing data, which was the basis for the hurried approach taken in this program, was seriously flawed. The required tests, particularly the LD50, are so crude that they are likely to exonerate toxic chemicals, rather than indict them. In the process, hundreds of thousands of animals will be killed, without any effort to discern whether humans are actually exposed to the chemicals in question or not. It is time to step back, and take a more sensible and effective approach to public health.