A. The Debate over Hormonally Active Chemicals
The endocrine system, a system of glands and the hormones they release, regulates the development of a fetus in the womb, sexual development and reproductive function, maturation of the brain and nervous system, and energy metabolism. Some researchers have postulated that a range of natural and synthetic chemicals in the environment could damage or disrupt human and animal endocrine systems at exposure levels much lower than what previous studies and regulatory agencies have determined to be dangerous or toxic. Proponents of this hypothesis have dubbed the implicated chemicals “endocrine disruptors.”
Many chemicals can exert toxic effects at high levels of exposure. Regulatory agencies set exposure limits for chemicals intended to protect even sensitive people from adverse effects due to chemical exposure, and few people are ever exposed to chemicals at levels above these safety limits. But proponents of additional regulatory safeguards believe that hormonally active chemicals could cause harm even at very low exposure levels. They observe that:
- The endocrine system can be affected by very small amounts of certain foreign chemicals—much less than the levels tested in traditional laboratory animal toxicity studies;
- There is evidence that some hormonally active chemicals can circumvent the normal defenses of developing organisms; and
- The environmental persistence of some of these chemicals gives them more time to do damage.
Studies in the early 1990s raised concerns over whether synthetic chemicals were causing widespread harm through endocrine disruption. Researchers in Europe published a study reporting that average human sperm counts had declined by more than 40 percent between 1938 and 1990. Other researchers reported that male alligators in a pesticide-contaminated Florida lake had abnormally small penises and reduced fertility. A breast cancer study reported that a group of women with breast cancer had higher average levels of the insecticide DDT in their bodies than a group of otherwise similar women without breast cancer. More recently, researchers have reported that some chemicals can cause changes in the size and structure of reproductive organs in laboratory animals at doses well below regulatory safety limits and near the range of typical human exposures.
However, a number of other scientists are skeptical of the extent to which endocrine disruption plays a significant role in human and wildlife health. They agree that adverse endocrine effects have been demonstrated for several synthetic chemicals in laboratory settings, and are likely to have occurred in a number of human poisoning incidents and in wildlife habitats with high contamination levels. However, these researchers question the existence and importance of health effects from the relatively low exposures to chemicals typical of the everyday environment. They raise the following objections:
- Inherent biases in human sperm count studies make them unsuitable for evaluating actual spermcount trends. Furthermore, long-term data from farm animals show no change in sperm counts over time;
- Studies of the relationship between DDT and other organochlorine chemicals and breast cancer have been inconsistent, with most studies finding no effect;
- Although some researchers have found endocrine effects in laboratory animals with very low doses of chemicals, other laboratories have not been able to duplicate these results; and,
- Even if these low-dose effects exist, it’s not clear that they should be considered harmful, because the effects are subtle, and “natural” variations in diet, stress, and other factors can cause similar effects.
Typical chemical exposures in humans are generally hundreds to thousands of times lower than exposures considered worrisome based on traditional toxicity studies. The practical importance of endocrine disruption thus depends on whether very low exposures to commonly used chemicals can cause significant harm.
B. Identifying and Studying Hormonally Active Chemicals
Aside from the body’s own hormones, chemicals with hormonal activity fall into three broad classes: (1) synthetic chemicals used in industry, agriculture, and consumer products; (2) synthetic chemicals used as pharmaceutical drugs; and (3) natural chemicals found in many foods, particularly soy.
The U.S. Environmental Protection Agency (EPA) sets regulatory safety limits for synthetic chemicals that are at least 100 times less than the highest dose found to be without adverse health effects in laboratory animals. EPA uses these conservative safety limits due to the uncertainty in whether humans are more sensitive than laboratory animals to certain toxic effects, and because humans vary in their sensitivity to toxic effects. Human exposures are substantially lower than regulatory safety limits. The key question then for assessing the risk of endocrine disruption is whether chemicals could cause adverse health effects at exposures below regulatory safety limits, and within the range of typical human exposures.
Scientists have developed a number of tests, or “assays” for hormonal activity. For example, relativepotency assays determine the ability of a given foreign chemical to mimic one or more effects of a particular hormone. These tests show that while some pharmaceutical hormones are as potent as natural hormones, industrial and consumer-product chemicals are generally hundreds to thousands of times weaker than natural hormones.
C. Risks from Chemicals with Hormonal Activity
Scores of laboratory animal studies have confirmed that high doses of hormonally active environmental chemicals—that is, doses much greater than everyday environmental exposure levels—can cause a range of adverse effects. There is also limited evidence in humans of such effects due to chemical accidents and use of some pharmaceutical drugs. For example, diethylstilbestrol (DES), a synthetic estrogen roughly as potent as estradiol, was given to several million pregnant women between 1947 and 1971 in the mistaken belief that it reduced the risk of miscarriage. DES caused high rates of infertility in daughters and increased rates of undescended or abnormal testes in sons of DES-treated women.
Although these studies show hormonally active chemicals can harm a developing fetus, it is not clear if these results can be generalized to endocrine disruption by exposures to chemicals at the low levels found in the everyday environment. Both dose and potency determine toxicity, and the doses in these studies were far greater than exposures to chemicals at the low levels found in the everyday envrionment. Furthermore, DES is thousands of times more potent in its hormonal effects when compared to hormonally active environmental chemicals. The practical risk of endocrine disruption for human and wildlife health instead centers on the potential for adverse effects at doses well below regulatory safety limits, including exposures that could be encountered in everyday life.
Several researchers have studied low-dose effects of hormonally active chemicals, but with mixed results. For example, some researchers have found that the industrial chemical bisphenol A (BPA) increases mouse prostate gland weight at doses well below the lowest dose at which BPA had previously been found to have a physiological effect, and near the estimated “worst-case” human exposure level. However, other researchers were unable to duplicate these results. Similar discrepancies have occurred in laboratory animal tests of other chemicals.
Because of the controversy regarding low-dose effects, EPA and the National Institute of Environmental Health Sciences convened an independent expert panel to re-analyze data from 49 low-dose studies. The expert panel concluded that endocrine effects have been demonstrated for a number of chemicals at doses below their previously determined no-effect levels. Only BPA had any effects at doses near the range of human exposure. However, because the BPA effects could not be duplicated by some laboratories, the panel concluded that it is not clear whether the apparent low-dose effects of BPA represent a general property of the chemical. Because of the subtlety of the effects, the panel also concluded that it is not clear whether low-dose effects should be considered “adverse,” or merely biological changes that can’t be assigned a value label such as “bad” or “good.” For example, natural factors such as variation in diet, stress, and hormone levels during pregnancy appear to have as much influence on study results as low doses of the chemicals being tested.
Laboratory studies provide information on the types of chemical effects that are plausible at given dose levels. The next step is to go out into the world and see if there is evidence for actual endocrine disruption or other harmful chemical effects in humans or wildlife at typical chemical exposure levels. Epidemiological studies attempt to determine if particular health outcomes, such as cancer, endocrine disruption, or asthma, are associated with particular risk factors, such as diet, genetics, smoking, or exposure to environmental pollution. Although epidemiological studies are an important part of risk assessment, they are not as definitive as laboratory studies, because the subjects are not randomly assigned to “treatment” and “control” groups, introducing the potential for bias in study results.
Neurological Effects. Laboratory animal studies have found that exposure to a group of chemicals called polychlorinated biphenyls (PCBs) in the womb can cause later learning and behavioral disorders. Although the biochemical mechanisms for these effects are unknown, some researchers have suggested endocrine disruption, as well as alterations in neurotransmitters (the chemical signaling mechanism in the brain and nervous system) as possible factors.
Five studies have assessed whether humans are adversely affected by exposure to PCBs and a number of other persistent chemicals at the relatively low levels encountered in the everyday environment. Some of these studies have found that children who had higher PCB exposures in the womb performed more poorly on tests of intellectual and neurological development. Where effects were observed, children in the top 5 to 20 percent of PCB exposure generally performed up to several percent worse than less-exposed children on one or more neurological tests.
The results from these studies are inconsistent regarding the type and timing of observed health effects. For example, the Lake Michigan study found declines in test performance in school-age children while the North Carolina study did not, even though the PCB exposures were similar. Effects sometimes also appeared and then disappeared, or vice versa, among the same group of children assessed at different ages. Where effects were observed, they were relatively subtle—only a few percent in most cases. One of the studies also found that the association of higher PCB exposure with lower test performance disappeared for children who were breastfed. Thus the extent to which low-level PCB exposure has permanent negative effects on children’s development remains unclear, but appears at worst to be relatively small.
Another factor to consider in assessing current risk from PCBs is that the children in these studies were born roughly 8 to 20 years ago. PCB exposure, as well as exposure to other persistent chemicals in the environment, has declined substantially during the last 20 years, and continues to decrease. Thus, whatever the effects of these chemicals on children born in the past, current effects are now lower and will likely continue to decline.
Male Reproductive Health. In 1992, researchers in Denmark published an analysis concluding that human sperm counts had declined by more than 40 percent between 1938 and 1990. However, other researchers have argued that the samples of men who elect to donate semen are never representative of the general population, and that the degree of bias can vary in different directions at different times and in different places. For example, a study in Australia found that average sperm counts varied by more than a factor of two among five separate groups of sperm donors recruited by the same doctors, at the same hospital, using the same recruitment methods.
Farm animals provide a check on human sperm-count studies, both because humans and farm animals are likely exposed to most of the same chemicals present in the environment, and because studies in farm animals presumably avoid the potential for selection bias inherent in human studies with voluntary sperm donors. Researchers have found that sperm counts in farm animals have been constant during the last 70 years. Evidence for human sperm-count declines thus appears to be relatively weak.
Female Reproductive Health. Increased lifetime exposure to estrogen increases the risk of developing breast cancer. Because of the link between estrogen and breast cancer, some researchers have proposed that estrogenic chemicals in the environment might increase the risk of developing the disease. Although a few studies have found an association between exposure to persistent chemicals such as DDT and increased risk of breast cancer, most have not. Furthermore, the chemicals in question–DDT, dieldrin, and hexachlorobenzene–are exceedingly weak estrogens. As a result, a link between these chemicals and breast cancer appears to be biologically implausible. It thus appears that the potential for everyday exposures to estrogenic foreign chemicals to increase the risk of breast cancer has not been properly evaluated.
Wildlife Studies. There are a few cases in which wildlife health effects have been linked specifically to the mechanism of endocrine disruption due to environmental contamination. Examples include the following:
- In a mollusk species called the dogwhelk, the marine fungicide tributyltin (TBT) can cause females to develop masculine characteristics, such as a penis, at exposure to concentrations as low as one-billionth of a gram of per liter (one part per trillion). Some European populations of dogwhelk began to recover after regulations reduced TBT use. Nevertheless, TBT concentrations are still high enough in some regions to endanger sensitive mollusk species.
- A high rate of hermaphroditism has been found in some freshwater fish in England that live just downstream of some sewage treatment works (STW) wastewater discharges. In most cases, the causative agents appear to be natural estradiol excreted by women, and ethinyl estradiol, a synthetic estrogen in birth control pills, both of which are not completely removed by STWs. In a few other cases, estrogenic chemicals called alkylphenols, released from industrial plants, are responsible for the effects.
- Many Great Lakes birds accumulated organochlorine chemicals in their bodies by eating contaminated fish in the 1960s and 1970s. These chemicals were likely the cause of high rates of eggshell thinning, deformities, and mortality in chicks. Although some Great Lakes bird species have made dramatic recoveries due to reductions in organochlorine chemicals during the last 20 years, other species continue to decline in the most contaminated locations.
Endocrine disruption has been unequivocally demonstrated in humans and animals at relatively high doses of chemicals—many times greater than typical human or animal exposures to environmental contaminants. Endocrine disruption has also been demonstrated in a few aquatic species due to low exposures to a few chemicals. However, the evidence for adverse hormonal effects from low-level chemical exposures in humans is much weaker. Subtle neurological effects may have occurred in some children due to PCB exposure in the womb, though the evidence is inconsistent. There does not appear to be credible evidence for hormonally active chemicals causing increases in breast cancer risk, or declines in sperm counts. Overall, the evidence suggests it is unlikely that adverse health effects due to endocrine disruption have occurred in humans from exposures to small amounts of foreign chemicals in the environment.