An Examination of Animal Experiments
Although most of the medical research that we rely on is clinical in nature, we often see reports of various kinds of animal experiments. Such experiments are controversial, not only for their ethical aspects, but also because their results may not apply to humans. What if by relying on animal experiments, we are actually limiting our ability to improve public health?
Research on tobacco risks provided some of the strongest evidence that animal experiments can be dangerous and misleading, showing that there is no substitute for human data in searching for the causes of human disease. In the early 1960s, the tobacco lobby used all the political and scientific clout it could muster against health warnings about smoking. One piece of evidence helped their case: animal experiments did not show that inhaled smoke causes cancer. In study after study, animals forced to inhale smoke did not get cancer. As Clarence C. Little wrote in the New England Journal of Medicine, June 15, 1961, “There have been many such experiments here and abroad, and none have been able to produce carcinoma of the lung in animals.” Dr. Little worked for the Tobacco Research Committee and for Jackson Laboratory, a large-scale animal breeder. He used the results of animal experiments to argue that lung cancer is not linked to smoking tobacco. Rather, he claimed that lung cancer “is a challenge, an unsolved problem. Its etiology will probably long be an open question.” While Little’s conclusion served both of his employers, it was no help to human health. Indeed, in another editorial published at about the same time, Dr. Donald B. Effler of the Cleveland Clinic argued that animal experiments offered little support for the smoking-cancer link, and that a smoker who does not yet have a chronic cough “assumes little risk to his health.”1 The animal experiments were clearly doing more harm than good, delaying warnings about smoking.
Of course, the key evidence on tobacco came from human studies. Whether one looks at large human populations or at individual smokers, the link between tobacco smoke and cancer is inescapable, even though it was completely missed in animal inhalation experiments. So the question is, have animal experiments led us astray in other areas?
Nutrition is another area where animal experiments have raised repeated problems. While it is easy to feed vitamins, fat, or fiber to animals and to check whether their disease rates rise or fall, the relevance to humans is limited at best, due to major physiological differences between species. For example, if vitamin C helps prevent cancer, what is the impact on cancer research of the fact that rats and mice synthesize vitamin C within their bodies, unlike humans, who do not? Likewise, rats differ from humans in crucial enzyme functions. For example, rats have much higher activity of the 5-desaturase enzyme system, a part of the body’s machinery for processing fats in the diet. Because of this species difference, rats are “not an appropriate human model” for studying the effects of fats.2
Although rats have been used extensively to test the value of various iron supplements, it turns out that rats absorb iron quite differently from humans and do not give usable information. According to a report in the American Journal of Clinical Nutrition, “Our studies indicate that rodents cannot be used to assess the quantitative importance of dietary factors in human iron nutrition.”3
Research on stroke provides another example. For years, experimenters have used animal experiments to create brain damage that simulates the effects of a human stroke. They then test out various experimental drugs to see whether they reduce the damage to the brain. But a review in the journal Stroke, published by the American Heart Association in January 1990, reported that, of 25 different treatments that worked in rodents, not a single one worked in human patients. As the Stroke editorial lamented, such animal experiments were not only failing to advance science, they were actually impeding progress:
“Each time one of these potential treatments is observed to be effective based upon animal research, it propagates numerous further animal and human studies consuming enormous amounts of time and effort to prove that the observation has little or no relevance to human disease or that it may have been an artifact of the animal model itself.”4
Are animal experiments that lead researchers astray simply rare exceptions or are they typical of animal tests? Broader data come from a U.S. General Accounting Office review of the safety of all new drugs marketed in the decade 1976 to 1985. All had been animal-tested prior to approval. Of the 198 new drugs for which data were available, 102 (51.5 percent) were more dangerous than pre-market animal tests and limited human tests had indicated, so much so that they had to be relabeled or withdrawn.5
All practicing physicians remember numerous examples. Zomax was a painkiller commonly used in the early 1980s, but after it caused 14 deaths and hundreds of life-threatening allergic reactions, the drug was withdrawn. Nomifensine was intended as a new and less toxic antidepressant, and psychiatrists were bombarded with drug company advertisements pushing the new wonder drug. But many patients became ill and some died, forcing the manufacturer to pull the product. Fenfluramine and dexfenfluramine were extremely popular diet drugs taken by millions of Americans. Both drugs were withdrawn from the market in 1997 after dozens of persons taking them developed severe heart valve defects. The animal tests that were supposed to show the safety of these drugs gave absolutely no indication of heart valve damage. Every doctor’s office receives frequent notices from pharmaceutical manufacturers about unexpected drug effects which did not turn up in testing.
Species differences are one reason for this problem. Another is the effect of stress.6 The laboratory environment is one of constant and inherent stress. Animals start out being shipped as freight and end up in the chronic confinement of a laboratory. They are unable to move freely, unable to get away from their own wastes, and, at intervals, are taken from their cages for instrumentation, blood tests, surgery, weighing, or whatever else is on the laboratory schedule. These effects are routine for the laboratory staff, but can be terrifying for animals. When animals are stressed, their immune function, hormone levels, cancer rates, and susceptibility to viral and bacterial infections all change. Stressed animals frequently exhibit illnesses of various kinds, leaving experimenters to sort out which symptoms are caused by the drugs being tested and which are caused by lab conditions or other unknown factors. The result can be like a jigsaw puzzle with 50 extra pieces and no way of telling which are the irrelevant ones.
Animal experiments sometimes focus researchers’ attention and resources on areas that are distant from where clinicians might like them. Cross-species transplant experiments are one example. Chronic hepatitis, a significant cause of liver dysfunction, is a serious disease killing about 6,000 Americans every year. However, several existing treatments do have a measure of success. Approximately 20 percent of patients clear the hepatitis virus after a human liver transplant. Other promising therapies involve alpha-interferon and hepatitis B immune globin, and they need to be refined and studied in the clinical setting.
Some experimenters, however, embarked on a very different plan: a series of baboon-to-human liver transplants, an experiment that committed millions of dollars and many lives—both human and animal—with little likelihood of near-term success. Two patients received baboon livers, and both promptly died. Neither was permitted to have a human liver transplant, due to organ shortages and the fear that a new human liver might become re-infected with hepatitis. In 33 previous transplants of other organs from baboons or chimpanzees into humans, none of the patients survived as long as the very first recipient of an animal organ, who died just nine months after the 1963 experiment.
A similar lesson came from the tragedy of Baby Fae in Loma Linda, California. The unfortunate infant was born with a malformation of the heart. She was given a baboon’s heart and died 20 days later. A better treatment, the Norwood procedure, which repairs the infant’s heart, was denied to the infant, and the surgeon never even looked for a human donor.
In order to stop the body from rejecting the animal organ, powerful drugs are used to suppress the immune system. The result is the loss of the body’s normal defenses against viruses, bacteria, and cancer. The first baboon liver recipient died of a massive brain yeast infection. The second died with a massive abdominal infection.
In some cases, animal experiments consume substantial amounts of money and other resources, while other research needs go begging. For example, there have been numerous experiments seeking to cause and to prevent cancer in rats, while there have been few dietary intervention studies seeking to reduce cancer in women or men. Likewise, diabetes prevention in rats has been thoroughly studied, while researchers seeking to study means of preventing diabetes in children are struggling for funding. Human intervention studies lose out when large portions of the research funding pie are consumed by animal experiments.
In the name of birth defects research, experimenters have sewn kittens’ eyelids shut, leaving them sutured for an entire year, and reared other kittens in the dark for their entire lives.7 The experimenters wanted to show that depriving cats of normal vision would prevent typical brain development, in spite of the fact that it is already known that changes in vision or other senses can affect the brain. Other experimenters have been paid to give daily cocaine injections to pregnant rats to examine the effect of prenatal cocaine exposure on spatial learning.8 Experimenters have even received grants to decapitate neonatal opossums to compare their gonadal development to that of other marsupial species.9
Meanwhile, funding for the large-scale epidemiological research, clinical trials, and birth defects monitoring programs that hold the promise of elucidating the causes of congenital malformations is not forthcoming.
Animal experiments are among the slowest methods of determining health risks, and sometimes delay appropriate public health measures. Alar, the chemical growth-regulator used by apple growers, was shown to be a likely carcinogen by computer analysis of its chemical structure. Yet, it was allowed to remain on the market while animal tests were run. The animal tests were not much help in settling the question: some of the animals got cancer while others did not, and the results are still a matter of muddled dispute. Since typical cancer tests in rodents require two years of chemical exposure for each animal, followed by more years for examinations of organ samples, no results are available for anywhere from five to ten years. And because animal tests often yield ambiguous results (test results in rats differ from mice in about 30 percent of cases), they often lead only to more tests seeking to solve the ambiguities. In Alar’s case, 25 years after computer analysis showed its cancer-causing potential, animal tests were still being run. Animal tests served mainly to buy time for the manufacturer, delaying Alar’s removal from the market.
If our goal is a better understanding of human illness, there is no substitute for a major investment in human population studies and clinical research. Large population studies showed the links between tobacco and cancer, between cholesterol and heart disease, and between high blood pressure and stroke. They revealed how the AIDS virus is transmitted, and how to prevent the disease.
Human clinical research showed that pancreatic damage was the cause of diabetes long before Banting and Best ran their well-known dog experiments, and has led to innumerable advances in every field of medicine. An editorial in Stroke stated that “the answers to many of our questions regarding the underlying pathophysiology and treatment of stroke do not lie with continued attempts to model the human situation perfectly in animals, but rather with the development of techniques to enable the study of more basic metabolism, pathophysiology, and anatomical imaging detail in living humans.”4 Computerized tomography (CT), positron emission tomography (PET), magnetic resonance imaging (MRI), and other methods elicit valuable data in the course of treating patients. Vital information can easily and ethically be elicited from human subjects in the course of treatment or in large epidemiological studies, and that such data are far more useful than studies on animals.
The value of epidemiological studies and human clinical trials is clearly demonstrated in birth defects research. While the incidence of most birth defects is on the rise,10 human-based research has revealed ways to prevent certain major defects. Observational studies demonstrated the link between folic acid deficiency and neural tube defects,11 and led to clinical intervention trials confirming that folic acid supplementation can prevent up to 70 percent of spina bifida and other neural tube defects.12,13 Discovery of fetal alcohol syndrome (FAS) also came from human observational studies, explaining many instances of facial abnormalities, growth retardation, and cerebral involvement.
A stunning recent success of human research trials is the identification of magnesium sulfate as a prophylaxis against mental retardation and cerebral palsy in very low-birth-weight babies. Researchers observed that brain hemorrhages are much less likely to occur in infants whose mothers had preeclampsia. It was not the preeclampsia that protected the babies; it was the magnesium sulfate that is commonly used to treat it. Following this lead, researchers looked at how often cerebral palsy and mental retardation occurred in very low-birth-weight infants in Georgia. They found that magnesium sulfate protects against both conditions. In cases of very low-birth-weight infants, maternal magnesium sulfate therapy could potentially prevent 63 percent of cerebral palsy cases and 49 percent of mental retardation.14
Other methods are showing their power, too. When the National Cancer Institute (NCI) scrapped its mouse tests for screening potential new anti-cancer treatments because of concerns that results in mice do not apply to humans, NCI began using human cancer cells, taken from cancer patients during routine surgery and kept alive in standardized cell cultures. The cellular method is cheap, fast, and, unlike animal experiments, it involves the right species.
Cellular tests, aided by computer analyses, can predict whether chemicals are likely carcinogens. They do so in a matter of days, rather than years as required by animal tests. The Ames test, for example, is a long-established method that demonstrates whether chemicals can cause the chromosomal damage that is the first step in cancer. The Ames test, using salmonella bacteria, is inexpensive and highly accurate. The test does not work for all types of test chemicals, however, suggesting that what is needed is a battery of rapid cellular tests that provide better results than any single test alone. Many such tests already exist, and others are under development.
Late 1996 brought two long-awaited breakthroughs. First, a new study showed that safety tests using human cells are more accurate than animal tests. Second, a new company began offering methods for developing new drugs with no animals at all.
In the Multicenter Evaluation of In Vitro Cytotoxicity tests (MEIC), researchers from the U.S., Europe, Japan, and other countries tried 68 different test-tube methods to predict the toxicity of 50 different chemicals, such as aspirin, digoxin, diazepam (Valium), nicotine, malathion, and lindane. The effects of the chemicals in humans were already known from poison control centers. The study’s goal was to see how well the cellular tests matched actual human experience and to compare them with data previously reported for animal tests.
Rat LD50 tests—lethal dose tests that measure the dose of a chemical that kills 50 percent of the animals given it—were only 59 percent accurate, and the mouse tests were about 70 percent accurate. But the average human cell test was 77 percent accurate. Accuracy was boosted to 80 percent when results from three different human cell tests were combined.
With personnel formerly of Glaxo Wellcome, SmithKline Beecham, and Shire Pharmaceuticals, Pharmagene Laboratories, based in Royston, England, became the first company to conduct new drug development and testing using human tissues and sophisticated computer technologies exclusively. With tools from molecular biology, biochemistry, and analytical pharmacology, Pharmagene conducts extensive studies of human genes and investigates how drugs affect the actions of these genes or the proteins they make. While some have used animal tissues for this purpose, Pharmagene scientists believe that the discovery process is much more efficient with human tissues.
Many examples exist of animal experiments wasting valuable time and research funds, and yielding results that misled researchers. Happily, research techniques that do not involve animals are the keys that can open the door to the prevention and treatment of human disease.
1. Effler DB. One surgeon’s attitude toward cigarette smoking. Surg Gynecol, Obst 1960;111:232-3.
2. Siguel EN. Cancerostatic effect of vegetarian diets. Nutr Cancer 1983;4:285-91.
3. Reddy MB, Cook JD. Assesment of dietary determinants of nonheme-iron absorption in humans and rats. Am J Clin Nutr 1991;54:723-8.
4. Weibers DO, Adams HP, Whisnant JP. Animal models of stroke: are they relevant to human disease? Stroke 1990;21:1-3.
5. U.S. General Accounting Office. FDA Drug Review: post approval risks 1976-85. April 1990.
6. Barnard N, Hou S. Inherent stress: the tough life in lab routine. Lab Animal 1988;17:21-7.
7. Sur M, Frost DO, Hockfield S. Expression of a surface-associated antigen on y-cells in the cat lateral geniculate nucleus is regulated by visual experience. J Neurosci 1988;8(3):874-82.
8. Levin ED, Seidler FJ. Sex-related spatial learning differences after prenatal cocaine exposure in the young adult rat. Neurotoxicology 1993;14(1):23-8.
9. Fadem BH, Tesoriero JV, Whang M. Early differentiation of the gonads in the gray short-tailed opossum (monodelphis domestica). Biol Neonate 1992;61:131-6.
10. Edmonds LD, James LM. Temporal trends in the birth prevalence of selected congenital malformations in the Birth Defects Monitoring Program/Commission on Professional and Hospital Activities, 1979-1989. Teratology 1993;48:647-9.
11. Blatter BM, van der Star M, Roeleveld M. Review of neural tube defects: risk factors in parental occupation and the environment. Enviro Health Persp 1994;102:140-5.
12. Czeizel AE, Dudás I. Prevention of the first occurrence of neural-tube defects by preconceptional vitamin supplementation. N Eng J Med;327:1832-5.
13. Centers for Disease Control and Prevention. Neural Tube Defects and Folic Acid Brochure.
14. Schendel DE, Berg CJ, Yeargin-Allsopp M, Boyle CA, Decoufle P. Prenatal magnesium sulfate exposure and the risk for cerebral palsy or mental retardation among very low-birth-weight children aged 3 to 5 years. JAMA 1996;276:1805-10.