By Kristie Stoick, Physicians Committee for Responsible Medicine
Risk assessment, it has been said, is more an art than a science. True, in a general sense, one uses a standard equation: HAZARD X EXPOSURE = RISK. Arriving at an answer traditionally requires the use of prescribed methods of measuring hazard and exposure. The regulator requests, and receives, a check-box list of safety tests to fulfil the hazard side of the equation, multiplies this hazard by expected exposures, and voila! Risk has been assessed. Perhaps inspired by the sentiment expressed in our opening sentence, however, risk assessors—whether they be of the corporate, government, or NGO perspective—are beginning to look less and less to "check the box" toxicology assessments and more towards testing strategies that often require less testing, but are still adequately protective of public health and the environment. The art, if you will, is getting equal consideration.
As early as 1984, the U.S. National Academies reported on strategies to assess the hazards of chemicals for the National Toxicology Program (NTP), a program of the National Institute for Environmental Health Sciences. (1) The Environmental Protection Agency has used "testing strategies" at least since the inception of the High Production Volume Chemical Challenge Program (HPV) in 1999 and continued with other programs such as the Voluntary Children's Chemical Exposure Program (VCCEP) and the Endocrine Disruptor Screening Program (EDSP).
Investigators began to pursue these approaches for practical reasons: A standard checklist of acute and chronic toxicity tests for one chemical can cost into the millions of dollars and take at least 3-4 years to conduct. However, as animal-protection groups become more involved in chemical-testing policy and society becomes less willing to tolerate in vivo testing, a broader array of companies are seeking testing strategies that avoid specific tests or reduce the numbers of animals used per test.
Initially, testing strategies were typically of a global "priority-setting" nature and included some sort of assessment of human exposure, with the degree of exposure dictating which chemicals should be assessed further and what types of tests an assessment would entail. For example, in the HPV program, chemicals that met strict production parameters that resulted in a very low potential for any human exposure (Closed-System Intermediates) were not tested for repeated-dose and reproductive toxicity endpoints. The REACH legislation, enacted in the European Union in 2007, bases the number and type of toxicity tests required for registration of a chemical on the production volume of the chemical, rather than actual human exposure.
Other strategies use basic chemical characteristics that would obviously preclude testing: explosive chemicals cannot be tested for practical reasons; the insolubility of a chemical leads one to avoid aquatic testing; the corrosive nature of a chemical would render many effects uninterpretable in a chronic study. As commonsense as some strategies may seem, it is only with prodding by animal-protection scientists that the chemical industry and regulatory bodies have begun to accept them on a wider scale.
Regulators can now think more and more about the toxicity characteristics of a chemical and how those characteristics can be used to avoid (or trigger) further in vivo testing. For example, much attention has been paid to multi-generation reproduction studies. Until now, most regulators have taken for granted that reproductive toxicity studies must be extended to two and three generations; several different data analyses have begun to question this maxim and propose creative tiered approaches to the assessment of reproductive toxicology. (2, 3, 4) In general, a regulator might use the results from initial, short-term studies, such as in vitro binding, genotoxicity or cell transformation assays, acute toxicity tests, or even sub-chronic toxicity tests to determine whether conducting full long-term, chronic, or carcinogenicity studies should be recommended. Indicators could be built into sub-chronic tests to make decisions regarding the potential for effects on the neural, reproductive, or immune systems. In most cases, these "triggers" would not be hit, and longer-term animal intensive testing would not be conducted.
Casting off a check-list approach can be an inspiring process. My colleagues and I had one of these "A-ha" moments while reviewing a test plan submitted to the EPA by Dow Chemical as part of the HPV Program. Dow was planning to conduct a combined reproductive/developmental toxicity study by the dermal route of exposure for a mixture of four chemicals termed Commercial Hydroxyethylpiperazine (CHEP). Having recently hosted a workshop on in vitro percutaneous absorption methods, I suggested that Dow test-drive one of them and see whether CHEP was even absorbed through the skin. Dow scientists were so excited by the idea that they took it in a different direction and used a (Quantitative) Structure Activity Relationship [(Q)SAR] model to determine the absorption potential of CHEP. The decision tree below shows how Dow used the EPA DERMWINN model.
To check the applicability of the model, experimental and modelled absorption data for ten analogous amines were compared; the ratio of experimental to modelled data was of acceptable concordance (values ranged from 0.1 to 10.5) and the absorption of CHEP was modelled. The components were predicted to minimally absorb through the skin, with estimated average total absorption at 1.99 mg/kg/day. When considered together with other factors such as production and use, protective equipment, and available data on piperazine (the component predicted to penetrate in the largest amount), this estimation supported the conclusion not to conduct a stand-alone dermal reproductive and developmental toxicity study.
While this example simply uses the risk assessment equation "low exposure = low risk," the fact remains that creative thinking and a willingness to think outside the (check)box saved over 600 animals in this case alone.
With the acceptance and continued use of in vitro and in silico approaches to measure percutaneous absorption, investigators and regulators should consider how this approach can avoid dermal systemic testing, including for acute, sensitizing, sub-chronic, and chronic endpoints.
The excitement surrounding these new risk assessment paradigms is palpable. Creative thinking, weight-of-evidence approaches, and plain common sense can go a long way towards eliminating or reducing the numbers of animals killed in chemical testing programs—which is science we can all support.
©2007 Kristie Stoick