Connect with Us



The Physicians Committee

Growing Skepticism over Animal Tests

By Neal D. Barnard, M.D.

The August 20, 1999 issue of the journal Science reported that some common animal tests may be utterly unreliable. A strain of mice, called CD-1, has been used to test whether various chemicals have hormone-like effects but turns out to be extremely resistant to estrogen effects.1 The findings, made by Jimmy Spearow, a research geneticist at the University of California, Davis, have thrown cold water on the Environmental Protection Agency's plan to use CD-1 mice to test chemicals for hormone effects.

The report comes after other reports showing that common safety tests on animals may tell us little or nothing about risks to humans. Here is what is at issue:

When a drug or chemical enters the body, it is subjected to a range of actions that can dramatically differ from one species to the next. First, it must be absorbed. Then, it disperses into various tissues. Third, enzymes in the liver or other organs may break it apart or change it into any number of other compounds. Finally, the chemical—or whatever other chemical compounds it has become—is excreted from the body.

Common safety tests assume that animals used in tests handle chemicals in much the same way as humans do. Not so, say toxicology experts who have found that enzyme systems animals use for eliminating chemical toxins differ dramatically between species.

In general, animals eliminate chemicals much more quickly than humans do. Part of the reason is their smaller size. A smaller body weight in relation to their surface area (a 4-fold difference between human children and rats, and an 8-fold difference between humans and mice) means that, compared to humans, small animals need a much more rapid metabolism in order to maintain a stable body temperature. This same fast metabolism also tends to eliminate toxins very quickly. A chemical that is rapidly eliminated has little chance of showing its toxic effects.

The half-life (the time it takes for half of a chemical dose to be eliminated) of caffeine is only about 40 minutes in a mouse but is more than 4 hours in humans. The half-life of Valium (diazepam) in a mouse or rat is about an hour but is 20 to 50 hours in humans. And while the half-life of the barbiturate phenobarbital is 3 hours in mice, it is 50 to 150 hours in humans.2

During my residency, I was called to see a man who had overdosed on barbiturates. He had lapsed into unconsciousness. Were he to metabolize drugs as a mouse does, I might have expected him to be up and around shortly. As it was, his poisoning carried him into a coma lasting for days.

For many drugs, the elimination process depends on a series of enzymes in the liver called the P-450 enzymes. The P-450 enzymes work by attaching oxygen molecules to chemicals. Other enzymes then connect the altered chemical to a large carrier molecule that escorts it from the body. There are more than 70 varieties of P-450 enzymes. In general, these enzymes are more active in nonhuman animals than in humans. It is impossible to generalize, however, as some enzymes in animals are more sluggish than in humans.

Toxicologist Dennis V. Parke once noted that the central irony in toxicology is that pesticides are developed that are toxic to insects, rodents, or other nonhuman animals while supposedly remaining relatively safe for human use, and that scientists then use nonhuman animals as models for humans in safety tests of these very chemicals.3

Parke noted the complement of P-450 enzymes in humans "is unique and quantitatively different from that of mouse, rat, dog, and other experimental animals."3 He argued for assessing safety based on understanding how chemicals are metabolized in the human body, something that can be judged to a great degree by an analysis of their molecular structure even before they are synthesized. The Computer-Optimized Parametric Analysis for Chemical Toxicity (COMPACT) system appears to be at least as reliable as animal tests for testing the potential of chemicals to cause mutations or cancer, while using no animals at all.3

Birth Defects

In judging the ability of chemicals to cause birth defects, animal tests have repeatedly proven unreliable. A review in Environmental Health Perspectives stated, "Great species differences, largely unexplained up to now, are the rule rather than the exception."2 The main problems are marked differences in the ability of compounds to pass across the placenta to the fetus and in how quickly compounds are eliminated from the body.

In the bloodstream, various proteins bind drugs and other chemicals. If this happens to a great degree, chemicals are held in the bloodstream and are less likely to reach the fetus. The tendency of drugs to remain bound to blood proteins varies greatly and unpredictably between species.2

When I prescribed a common anti-seizure medication, valproic acid (Depakene), to a woman, I was reassured of its safety for only one reason: it had long been used in patients. Had I relied on animal tests, it would have been impossible to judge its safety. In mice and hamsters, valproic acid binds extensively to proteins in the blood. In humans, it does so to a much lesser extent, meaning that it reaches the body tissues much more easily. Will it be safe or dangerous when it gets there? Animal tests provide little reassurance.2

When a chemical reaches the fetus, differences between species continue to play a role. In humans and other primates, the fetal P-450 enzymes are able to metabolize certain chemicals. In other species, however, these enzymes do not develop until after the baby is born, for the most part.2 If the P-450 enzymes happen to make chemicals less toxic, then animal tests could overstate the degree of danger. However, in some cases these enzymes transform chemicals into more dangerous molecules, which animal tests could miss completely.

The use of animals to test the effects of chemicals rests on the assumption that the pathways a chemical takes through the animal's body are much like those in humans. Evidence now shows that this assumption proves wrong more often than not.

1. Spearow JL, Doemeny P, Sera R, Leffler R, Barkley M. Genetic variation in susceptibility to endocrine disruption by estrogen in mice. Science. 1999;285:1259-1261.
2. Nau H. Species differences in pharmacokinetics and drug teratogenesis. Env Health Perspec. 1986;70:113-129.
3. Parke DV, Ioannides C, Lewis DFV, Obrebska-Parke MJ. Current problems in the evaluation of chemical safety. Pol J Occup Med. 1990;3:15-41.


Autumn 1999 (Volume VIII, Number 4)

Autumn 1999
Volume VIII
Number 4

Good Medicine

This site does not provide medical or legal advice. This Web site is for informational purposes only.
Full Disclaimer | Privacy Policy

The Physicians Committee
5100 Wisconsin Ave., N.W., Ste.400, Washington DC, 20016
Phone: 202-686-2210     Email: