By now, a large number of consumers are aware of the hazards of the synthetic compound bisphenol-A (BPA). Effective May 11, 2016, under California state law Proposition 65, products containing BPA must possess a warning label indicating that exposure could result in female reproductive impairment. Independent research on the endocrine disrupting effects of the chemical, commonly used in plastic bottles, the lining of metal cans, and customer receipts, among other applications, has consistently demonstrated toxic effects at low dose exposures. Two recent robust studies from Denmark concur, finding deleterious effects in rats exposed to BPA at doses lower than those considered safe for human ingestion, yet not at several higher doses. Nevertheless, regulatory agencies such as the U.S. Food and Drug Administration (FDA) and the European Food Safety Authority (EFSA) conclude that BPA is safe at the levels at which it is currently in use.
Clearly, disagreement exists among academic researchers and regulators about safe levels of BPA, as well as of innumerable other chemicals. The discrepancy stems both from how data are derived to determine safe levels of exposure to known toxicants, and from whether safe levels are even derivable under traditional standards of appraisal.
When citizens inquire about the toxicity of their products, they are usually met with guarantees that hazardous substances within these items exist at levels that are too low to produce harm from routine exposures. After accidental releases of hazardous substances into our environment there is a time, either at the outset or when concentrations of the pollutant subside, that levels of the contaminant are deemed low enough to be safe for human exposure. When evidence of contamination in the municipal water system in Flint, Michigan first emerged, officials initially declared the water safe to drink. Likewise, when radiation from the Fukushima nuclear disaster crossed the Pacific and reached the West Coast of the United States, the public was met with assurances that the level of radiation was too low to do harm. “Safe levels” is a common refrain to assuage fears of chemical toxicants. Yet accumulating research, like that on endocrine disrupting chemicals (EDCs) such as BPA, reveals that the foundational principle of safe levels of chemicals at low-enough concentrations is a flimsily constructed one.
Current chemical risk assessment operates under the assumption that we can determine a lowest dose at which a compound produces negligible or no harm to human health – the Lowest Observable Adverse Effect Level (LOAEL). The presumption is that at increasingly higher doses, the substance will be increasingly harmful; at lower doses, harm will be insignificant or nonexistent. (Only a select few substances are regarded as harmful at any dose.)
Our regulatory toxicological tests are based on this supposition of positive monotonic dose-response. Monotonic refers to the slope of the dose-response curve consistently progressing in one direction and never changing sign along the way. These positive monotonic dose-response curves are commonly linear, exponential, or sigmoid. But, this expectation of monotonicity upon which we base regulation has been strongly challenged not just by the newest papers on BPA, but by an accumulating consensus. Indeed, the dose makes the poison, but in unanticipated and unpredictable ways.
Numerous substances act in non-monotonic dose-response (NMDR) manners, in which the sign (positive or negative) of the response can change throughout the measurement of the dosages. Many essential vitamins and minerals serve as examples. At too-low doses they are insufficient at providing the necessary nutritional molecules needed for functioning. At too-high doses, many can be poisonous. The desirable level of vitamin intake falls at a crucial range in the middle. Their dose-response curves, in which the response examined is nutritional benefit, would resemble an inverted U-shape. Indeed, countless nutrients possess U- or inverted U- shaped NMDR curves.
According to Dr. Pete Myers, founder and chief scientist at Environmental Health Sciences, “NMDR curves are the default expectation for endocrine disrupting compounds (EDCs).” As co-author, with the late Dr. Theo Colborn and Dianne Dumanoski, of Our Stolen Future, Dr. Myers wrote the seminal book on EDCs. EDCs are hormone mimics and as such, operate in several complex ways to trigger or suppress normal hormonal regulatory mechanisms. Consequently, they can produce negative effects at different doses, often at the very high and very low levels, rather than in-between. They act in a mode that contradicts the assumptions of low dose safety.
Dr. Myers estimates that at least 1000 EDCs are currently in use commercially, but because most chemicals in commerce have not been sufficiently tested for EDC activity, that number may be much higher. Besides BPA (and its replacement, BPS), other common EDCs of concern include: phthalates found in plastics, cosmetics, and fragrances; PCBs formerly used (and still found) in industrial applications as coolants, lubricants, and insulators; brominated flame retardant chemicals (PBDEs) in furniture and electronics; and the ubiquitous pesticides glyphosate and atrazine. Dr. Tyrone Hayes of the University of California Berkeley has conducted numerous studies on atrazine demonstrating the endocrine disrupting effects on various frog species. Perhaps the most alarming of all his findings may be the hermaphroditism and feminization of male frogs after exposure to atrazine at environmentally relevant doses – doses at or below those found routinely in rivers and streams in the United States.
While synthetic endocrine disruptors are the most commonly discussed examples of chemicals that exhibit NMDR patterns of toxicity, they are not the only substances that do. Heavy metals such as lead, cadmium, selenium, arsenic, and manganese show NMDR patterns as well. In fact, even though the presumption of monotonicity pervades all of risk assessment, it is unclear whether even the majority of compounds actually do act in that simplistic manner. What is clear, as Pete Myers states, is that “by ignoring NMDR curves, risk assessment as currently practiced is deeply flawed and unquestionably allows people to be exposed to harmful chemicals at dangerous doses.”
One of the major flaws lies in the methods of chemical toxicity testing. Most toxicity tests utilize a maximum of three doses as reference points. As we know from basic algebra, plotting three points cannot possibly lead to an accurate estimation of any curve. In order to determine the level at which negative health effects might emerge, says Myers, “You need to have tested an extraordinarily wide range of doses and have, preferably, at least five doses across that range.” He adds, “You can’t say anything about the absence of (NMDR) with just three doses.” Thus, with such a small set of reference points, many substances could appear to follow monotonic dose-response with the attendant fall-back assumption of safety a very low levels of exposure. But, untested low doses could actually be the most harmful.
Further complicating determinations of safe levels of chemicals, dose-response curves are specific to precise endpoints. Endpoints are the biological outcomes – such as cancer, reproductive toxicity, or neurological impairment for which toxicologists test. Even if all of the possible endpoints could be or were tested for each chemical (which they are not), each chemical may follow a different curve for each endpoint assessed. For example, arsenic acts monotonically for cancer risk, but inflammatory markers in the umbilical cord of pregnant women are lowest at intermediate levels of arsenic exposure, demonstrating a NMDR curve for that endpoint. Hence, the same chemical may be both safe and unsafe at the same exact level of exposure, depending upon which health effect one examines.
Another issue with establishing safe levels of any single chemical through traditional toxicity measurements stems from the fact that cumulative exposures are not accounted for, nor are aggregate exposures. Chemicals in combination may act synergistically. Roundup herbicide, for instance, causes cell cycle dysfunction (which can lead to cancer) and apoptosis (programmed cell death) in certain product formulations which contain different “inert” (yet toxic) ingredients. These toxic effects are either not produced or produced to a much lesser degree from glyphosate (the “active” ingredient) alone.
Additionally, the time of exposure within the lifetime of an organism can determine whether or not the chemical produces toxic effects and at what dose. Early development and puberty/adolescence are critical stages of life (“windows of vulnerability”) at which exposure to toxic substances may generate greater harm than at other life stages. Lead exposure in children, particularly during embryonic, fetal, and postnatal periods, produces neurological deficits that do not occur in equivalent adult exposures. By overlooking additional complexities such as these in deriving safe levels, chemical testing protocols as they stand are greatly in need of repair to adequately reduce health risks.
In the face of such evidence that our notion of “safe levels” of toxicants is outdated, why are such antiquated modes of risk analysis still utilized to determine regulations? “Because too much money is at stake” says Dr. Myers. “Using procedures capable of detecting NMDR curves would be likely to require lowering a large number of reference doses so much that the chemical would be required to be removed from the market.” The removal of so many chemicals would more reliably ensure safety, but would impede commercial and industrial profits.
Given the inadequacy of the current risk assessment paradigm, changes are warranted to better protect public health. Tony Tweedale, founder of RISK (Rebutting Industry Science with Knowledge) Consultancy, suggests that studies must “test for the effects of real world doses” and “test the whole dose response curve,” rather than simply a few high dose points. He also advises drawing from the thousands of peer-reviewed academic studies for policy decision-making, because “tens of thousands more experimental and supporting etiologic and epidemiologic papers (are) being tragically ignored.”
Chemical regulations based on current unsound testing practices cannot possibly be considered adequate. In fact, in 2014, the National Academies of Sciences (NAS) offered updates to the EPA’s traditional risk analysis methods to better address NMDR and other deficiencies in chemical risk assessment. Among their proposals is augmentation of risk evaluations to include “statistical considerations, uncertainty analysis, life stage or susceptibility issues, and modes of action.” The EPA has yet to act on these recommendations.
Because of the faulty paradigm under which current risk assessment and regulation proceed, one cannot confidently dismiss the contribution of the innumerable commercially utilized chemicals toward human diseases and negative health outcomes. As such, assertions by the FDA and EFSA about the safety of BPA or other toxicants at current levels should be taken with a note of skepticism. Cautions such as those now abundant in California should be heeded.
A society that values human health and safety over commercial growth would acknowledge the tremendous defects and scientific uncertainty implicit in our current paradigm of assessing chemical toxicity. We cannot even begin to approach a valid judgment of “safe levels” within the context of the more than 85,000 chemicals currently in commerce (of which only a small percentage have been tested for safety even under current protocols). Chemical regulation based upon the precautionary principle would not only be relevant under such conditions of uncertainty, it would be most the prudent option for the benefit of public health.