Experiments on animals are predicated on the assumption that what is true for one animal species is most likely also true for others, including humans. Yet such catastrophic drug failures as the TGN 1412 incident in the United Kingdom—which nearly cost the lives of human study participants—provide a sobering reminder that animal "models" often do not correctly predict real-world consequences for people, and that continued reliance on such 19th century approaches to scientific investigation may not be to our benefit.

Consider the following:

• Improper Validation of Most Animal Tests
• Animal Welfare Considerations
• Time and Cost Considerations
• Outdated Testing Methods
• Legal Obligations
• The Opinions of Scientists

Improper Validation of Most Animal Tests

Although some animal tests in use today are as much as 80 years old, most have never been formally validated (i.e., assessed in multiple laboratories to see if they reliably give the correct answers). However, there is a great deal of scientific evidence that some of the most common animal tests may be extremely poor predictors of human effects. For example:

Eye and skin irritation tests:

• Among 281 cases of accidental human eye exposure to 14 household products, investigators with the U.S. Food and Drug Administration determined that rabbit test results correctly predicted human responses less than half the time (1)
• Failed to correctly predict human skin reactions for nearly half of 65 consumer products examined (2)
• Excessively high between-labs variability in test results and interpretation: a comparison of test results across 24 labs documented variability up to 100 percent (3).

Acute toxicity tests (often conducted using oral, inhalation and skin routes):

• May seriously underestimate human risk, as documented human sensitivity to some chemicals is as much as 2,000-times greater than in other animals (4)
• Only 65% agreement between rat and mouse test results for the same 50 chemicals (5)
• "The information obtained from conventional acute toxicity studies is of little or no value in the pharmaceutical development process" (6).

Birth defect tests (often conducted in both rats and rabbits):

• Tests in both rats and rabbits failed to detect the developmentally toxic effects of PCBs, ACE-inhibiting drugs, and other substances, and rabbits gave false negative results for toluene, tetracycline, diethylstilboestrol (DES), and other drugs (7)
• Less than 74% agreement between rat, mouse and rabbit test results for the same chemicals (8)
• Testing in a second species increases the already high rate of false positive results (9).

Cancer tests (usually conducted in both rats and mice):

• Failed to detect the hazards of asbestos, benzene, bromodichloromethane, cigarette smoke, dichlorovos, lindane, DDT, selenium sulfide, and many other substances, delaying consumer warnings and worker protection measures by decades in some cases (10)
• Only 50-70% agreement between rat and mouse test results for the same chemicals (11)
• Less than 60% agreement in the interpretation test results between labs for the same chemicals (12)
• Many biological mechanisms leading to cancer in rodents are irrelevant to humans (e.g., buildup of a2u-globulin in the kidneys of male rats, peroxisome proliferation in rodent livers, calcium phosphate-containing urinary buildup in rats) (13)
• Rodents possess cancer-prone organs for which there are no human equivalents (e.g., forestomach, Harderian gland, Zymbal's gland) (14)
• Animals are sometimes administered 100-times or more the equivalent human intake of a chemical (e.g., to consume the level of the pesticide Alar that was fed to rats and mice in one study would require eating 28,000 pounds of apples daily for 10 years) (15)
• Commonly used strains of rats and mice are highly prone to spontaneous tumor development—even "control" animals who are not administered a test chemical—which confounds the interpretation of test results (16).

As a direct consequence of shortcomings cited above, pharmaceutical regulators have reported that fully 92% of drugs that pass preclinical (animal) testing fail clinical trials, because animal studies so often "fail to predict the specific safety problem that ultimately halts development" (17).

References

1. Freeberg, F.E., Griffith, J.F., Bruce, R.D., et al. (1984). Correlation of animal test methods with human experience for household products. J. Toxicol. Cutaneous Ocul. Toxicol. 1, 53-64.
2. Robinson M.K., McFadden J.P. & Basketter D.A. (2001). Validity and ethics of the human 4-h parth test as an alternative method to assess acute skin irritation potential. Contact Derm. 45, 1-12.
3. Weil, C.S. & Scala, R.A. (1971). Study of intra- and interlaboratory variability in the results of eye and skin irritation tests. Toxicol. Appl. Pharmacol. 19, 276-360.
4. Müller, R. (1948). Vergleich der im Tierexperiment und beim Menschen tödlichen Dosen wichtiger Pharmaka. Diss. Univ. Frankfurt/Main.
5. Ekwall, B., Barile, F.A., Castano, A., et al. (1998). MEIC evaluation of acute systemic toxicity. Part VI. The prediction of human toxicity by rodent LD50 values and results from 61 in vitro methods. Altern. Lab. Anim. 26 (Suppl. 2), 617-58.
6. Chapman, K. & Robinson, S. (2007). Challenging the requirement for acute toxicity studies in the development of new medicines.  London: UK National Centre for the 3Rs.
7. Schardein, J.L. (2000). Chemically Induced Birth Defects, 3rd Ed. Rev. New York: Marcel Dekker.
8. Hurtt, M.E., Cappon, G.D. & Browning, A. (2003). Proposal for a tiered approach to developmental toxicity testing for veterinary pharmaceutical products for food-producing animals. Food Chem. Toxicol. 41, 611-19.
9. Bremer, S., Pellizzer, C., Hoffmann, S., Seidle, T. & Hartung, T. (2007). The development of new concepts for assessing reproductive toxicity applicable to large scale toxicological programmes. Curr. Pharm. Des. 13, 3047-3058.
10. Seidle, T. (2006). Chemicals and Cancer: What the Regulators Won’t Tell You. London: PETA Europe Ltd.
11. Gold, L.S. & Slone, T.H. (1993).Prediction of carcinogenicity from two versus four sex-species groups in the carcinogenic potency database. J. Toxicol. Environ. Health. 39, 143-57.
12. Gottmann, E., Kramer, S., Pfahringer, B., et al. (2001). Data quality in predictive toxicology: Reproducibility of rodent carcinogenicity experiments. Environ. Health Perspect. 109, 509-14.
13. Cohen, S.M. (2002). Bioassay bashing is bad science: Cohen’s response. Environ. Health Perspect. 110, A737.
14. Cohen, S.M. (2004). Human carcinogenic risk evaluation: an alternative approach to the two-year rodent bioassay. Toxicol. Sci. 80, 225-9.
15. American Council on Science and Health. (1997). Of Mice and Mandates: Animal Experiments, Human Cancer Risk and Regulatory Policies. New York: ACSH.
16. Haseman, J.K., Hailer, R.J. & Morris, R.W. (1998). Spontaneous neoplasm incidences in Fischer 344 rats and B6C3F1 mice in two-year carcinogenicity studies: A National Toxicology Program update. Toxicol. Pathol. 26, 428-41.
17. Food & Drug Administration. (2004). Challenge and Opportunity on the Critical Path to New Medical Products. Bethesda, MD: FDA.

Animal Welfare Considerations

Some toxicity tests consume hundreds or thousands of animals per substance examined (e.g., lifetime cancer studies consume approximately 400 rats and 400 mice; a study of birth defects and developmental toxicity consumes 1,300 rats and/or 900 rabbits; and a study of sexual fertility and reproduction generally consumes 200 litters of rodent pups--or upwards of 2,600 animals) (1). Moreover, some countries' statistics on animal use indicate that toxicity testing accounts for up to 80% of the most painful procedures to which animals are subject for all experimental purposes (e.g., death as the endpoint in acute systemic toxicity studies) (2). These concerns are exacerbated by the fact that some regulations prescribe dozens of separate animal tests to evaluate the full range of potential toxicities for a single substance (e.g., upwards of 12,000 animals may be consumed to test a single pesticide chemical according to US regulations).

References

1. Organisation for Economic Co-operation and Development. (2008). OECD Guidelines for the Testing of Chemicals. Paris: OECD.
2. Canadian Council on Animal Care. (2008). Facts & Figures – CCAC Animal Use Survey: Number of Animals Used in 2006 per Purpose of Animal Use and Category of Invasiveness. Ottawa: CCAC.

Time and Cost Considerations

Some animal tests take months or years to conduct and analyze (e.g., 4-5 years, in the case of rodent cancer studies), at a cost of hundreds of thousands—and sometimes millions—of dollars per substance examined (e.g., $2 to $4 million per two-species lifetime cancer study). The inefficiency and exorbitant costs associated with animal testing makes it impossible for regulators to adequately evaluate the potential effects of the more than 80,000 chemicals currently in commerce in the US, let alone study the effects of myriad combinations of chemicals to which humans and wildlife are exposed, at low doses, every day throughout our lives. In contrast, computer modeling techniques are lightning-fast, and many cell-based in vitro methods are amenable to "high throughput" automation using robotics—all at a much lower cost than animal tests.

References

• Charles River Laboratories. (2006). CRL Price List. Wilmington, MA: CRL.
• Corvi, R. (ECVAM), personal communication.
• Derelanko, M.J. & Hollinger, M.A. (Eds.). (2002). Handbook of Toxicology, Second Ed. Washington, DC: CRC Press.
• Endocrine Disruptor Screening and Testing Advisory Committee. EDSTAC Final Report. (1998). Washington, DC: US EPA.
• Institute for In Vitro Sciences. (2007). IIVS Price List. Gaithersburg, MD: IIVS.
• Stott, W. (Dow Chemical), personal communication.
• Webb, S. (Proctor & Gamble), personal communication.

Outdated Testing Methods

Between the time that most commonly used toxicity tests were conceived and today, there has been a revolution in biology and biotechnology. Advances in tissue engineering and robotics have given birth to rapid "high throughput" in vitro (cell culture) systems, while emerging technologies such as bioinformatics, genomics, proteomics, metabonomics, systems biology, and in silico (computer-based) systems offer still more potential alternatives to animal use. In June 2007, the US National Academy of Sciences called for a major paradigm shift in toxicology that would "rely less heavily on animal studies and instead focus on in vitro methods that evaluate chemicals' effects on biological processes using cells, cell lines, or cellular components, preferably of human origin. The new approach would generate more-relevant data to evaluate risks people face, expand the number of chemicals that could be scrutinized, and reduce the time, money, and animals involved in testing" (1).

References

1. http://www8.nationalacademies.org/onpinews/newsitem.aspx?RecordID=11970

Legal Obligations

As public opposition towards animal testing has grown, animal use has been broadly prohibited where alternative methods are "reasonably and practicably available" (e.g., EU Directive 86/609 (1), as well as legislation in the US States of California (2), New Jersey (3), and New York (4). Animal testing bans may also be sector-specific, as in the case of the 7^th Amendment to the EU Cosmetics Directive (5), which currently bans the marketing of any formulated cosmetic products that have been animal tested, and will soon culminate in an EU-wide marketing ban of cosmetic products whose ingredients have been animal tested following 2009 and 2013 cut-off points.

References

1. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:31986L0609:EN:NOT
2. http://www.hsus.org/web-files/PDF/ARI/ca_sb2082.pdf
3. http://www.hsus.org/web-files/PDF/ARI/nj_a909.pdf
4. http://www.hsus.org/web-files/PDF/ARI/ny_a07402.pdf
5. http://ec.europa.eu/enterprise/cosmetics/doc/200315/200315_en.pdf

The Opinions of Scientists

"Even if a chemical is found to be nontoxic in animal studies, the safety of the chemical cannot be assured."

– Dr. Barbara Shane, US National Toxicology Program (1)

"Currently available animal models, used for evaluating potential therapies prior to human clinical trials, have limited predictive value in many disease states."

– US Food & Drug Administration (2)

"The problem is we don't know what the findings really mean."

– Dr. Robert Maronpot, US National Institute of Environmental Health Sciences (3)

"[E]ven if the LD₅₀ could be measured exactly and reproducibly, the knowledge of its precise numerical value would barely be of practical importance, because an extrapolation from experimental animals to humans is hardly possible."

– Dr. D Lorke, Bayer AG, Germany (4)

"[R]egulators have chosen animal tests to forecast human cancer risks. To this end, animal data are filtered through a series of preconceived assumptions that are presumed to overcome a host of human/animal differences in biology, exposure and statistics – differences that in reality are insurmountable."

– Dr. Gio Batta Gori (5)

"In the present state of the art, making quantitative assessments of human risk from animal experiments has little scientific merit."

– Statisticians Drs. David Freedman and Hans Zeisel (6)

"Animal studies of lead, mercury, and PCB's each underestimated the levels of exposures that cause effects in human by 100 to 100,000-fold. Regulatory decisions that rely largely on toxicity testing in genetically similar animals under controlled laboratory conditions will continue to fail to reflect threats to the capacities and complexity of the human brain as well as important gene-environment interactions."

– Physicians for Social Responsibility (7)

"The one or two or three hundred millions of dollars a year that we're now spending on routine animal tests are almost all worthless from the point of view of standard setting... [I]t is simply not possible with all the animals in the world to go through new chemicals in the blind way that we have at the present time, and reach credible conclusions about the hazards to human health. We are at an impasse. It is one that has deep scientific roots, and we had better do something about it."

– Nobel Laureate Dr. Joshua Lederberg (8)

References

1. Shane, B.S. (1989). Human reproductive hazards. Environ. Sci. Technol. 30, 1193.
2. Food & Drug Administration. (2004). Challenge and Opportunity on the Critical Path to New Medical Products. Bethesda, MD: FDA.
3. Maronpot, R. cited in Brinkley, J. Many say lab-animal tests fail to measure human risk. The New York Times, A-1 (23 May 1993).
4. Lorke, D. (1983). A new approach to practical acute toxicity testing. Arch. Toxicol. 54, 275-87.
5. Gori, G.B. (2001). The costly illusion of regulating unknowable risks. Regul. Toxicol. Pharmacol. 34, 205-12.
6. Freedman, D.A. & Zeisel, H. (1988). From mouse-to-man: the quantitative assessment of cancer risks. Stat. Sci. 3, 3-56.
7. Physicians for Social Responsibility. (2000). In Harms Way: Toxic Threats to Child Development. Boston: PSR.
8. Lederberg, J. (1981). A challenge for toxicologists. Chem. Engin. News. 1, 5.