One of the greatest threats to the preclinical research necessary for science-based medicine today is animal rights activism. The magnitude of the problem came to the forefront again last week with the news that animal rights terrorists tried to enter the home of a researcher at the University of California Santa Cruz (UCSC) whose research uses mice to study breast cancer and neurologic disease while she and her husband were having a birthday party for one of their children and assaulted her husband, who had gone to the front of the house to confront them. Fortunately, the license plate number of the car fleeing the scene was reported to police, leading to a raid on a house by police and the confiscation of computers and other materials. This attack appeared to be the latest crescendo in an increasing campaign of harrassment and intimidation by animal rights “activists” that has also been observed in nearby Berkeley.
This sort of threat to researchers is not a problem just in Santa Cruz and the Bay area, but in particular has been a problem in southern California as well. Just earlier this week, the University of California Los Angeles announced that it was suing several animal rights groups and individuals suspected of attacks on researchers who use animals, including UCLA Primate Freedom Project, the Animal Liberation Front (ALF), and the Animal Liberation Brigade (ALB), as well as several individuals believed to affiliate with these groups. The inciting event for this action was the second attack on the home of Edythe London, Professor of Psychiatry and Bio-behavioral Sciences and of Molecular and Medical Pharmacology at the David Geffen School of Medicine at UCLA, in early February. Her research involves the use of primate models to study nicotine addiction. In this most recent attack, an attempt was made to set her house on fire. This came on the heels of a previous attack in October, in which animal rights activists flooded her house. Prior to that, animal rights terrorists had indeed succeeded in their aim of intimidating a scientist sufficiently that he gave up animal research after a fellow researcher was targeted with a Molotov cocktail meant for her home that was mistakenly placed on the porch of an elderly neighbor. Also, we in the U.S. often forget how much more radical animal rights extremists are in the U.K., where the campaign of intimidation takes the form of death threats, intimidation of personnel of companies that supply researchers, and even in one case digging up the grave of Gladys Hammond, whose family ran a farm that raised Guinea pigs for use in medical research, and stealing her remains.
Readers may make the argument that my introduction to this discussion is unfairly inflammatory, but I have my reasons for starting this way, and I think they are good ones. First, make no mistake, the aim of the most radical of these activists is nothing short of the cessation of the use of all animals in biomedical research. Second, sooner or later, someone will be hurt or killed. As a researcher who on occasion uses mouse models of cancer myself, I state up front that I could be on the firing line just as much as the UCSC researcher or others and am justifiably disturbed when I hear spokesperson for the ALF Dr. Jerry Vlasak, for example, repeatedly advocate violence against researchers who use animals. In this article, I am not going to discuss the moral issues involved in animal research. What I am going to discuss is the seemingly scientific arguments that some opponents of animal research and animal rights activists like to invoke, arguments increasingly used in addition to the moral arguments that extremists use to justify their actions. If the arguments of opponents of animal rights research were indeed good science, then their appropriation by extremists would not allow me to do much other than bemoan the misuse of valid science as a justification for extremism. Unfortunately, such is not the case, and the bad scientific arguments used by opponents of animal research are often piled onto the extreme moral arguments that fuel actions such as those earlier this week at UCSC. Consequently, given the events of the last month or so, I thought I would take this opportunity to look at some of the common scientific indictments of animal research by its opponents.
If you examine various websites or literature from animal rights extremists looking at the issue of animal use in medical research, the forms of the scientific arguments tend, when you boil them down to the very core of their essence, to take three main forms, which are related:
- Animal research doesn’t teach us anything of value or even misleads us (i.e., it is bad science).
- Animal research does not predict human physiology or response to disease, or animals are “just too different from humans to give reliable results” (i.e., it is bad science).
- There are better ways of getting the information that do not use animals (i.e., there is better science available than using animals.)
I tend to look at these arguments as three faces of what is in essence the same argument, specifically what I like to call an “argument from imperfection.” In other words, because animal models have many difficulties and flaws and all too often don’t predict human physiology or drug response as well as the critics think that they should, then by implication all animal research is bad science. It is an example of demanding 100% perfection or certainty, a bar that no science can ever meet and of concrete thinking typical of extremists. (Creationists and “alternative” medicine mavens are particularly fond of this sort of argument against their hated “Darwinism” or “allopathic” or “conventional” medicine–usually said with a sneer–respectively.) In its most ridiculous form, this argument takes the form of claiming that cell culture and computer models, among other modalities, can give us the same information without animals. The first reason that this argument is ridiculous is that cell culture models tend to be even less predictive of many responses than animal models for many questions and because much physiology depends upon the interaction of different cell types in their native three dimensional matrix. The second reason is that, for a computer model to be adequately predictive, it needs (1) sufficient information to input and (2) sufficient understanding of the intricacies of the physiology and biochemistry. We don’t have either. Finally, physiology requires understanding at the macroscopic level of how organs interact. Of course, these arguments are often made in less extreme forms, and I will discuss a some of these shortly. Keep in mind as I do, though, that the problem inherent in this sort of argument is that one has to look at what the alternatives to animal research are and compare their usefulness, accuracy, and reliability. If one can’t show that one’s alternative is better than animal research, then all the complaints about the imperfections of animal research don’t amount to much. It’s still the best that we have, and, as such, it’s bad science (and unethical, to boot) not to use it before trying therapies in humans. I have yet to see a compelling argument that any alternative modality predicts human response to disease and treatment well enough that we should rely on it instead of animal models.
The first of the three arguments, namely that animal research doesn’t teach us anything of value or even misleads us, is the easiest to deal with. Keep in mind that I am a cancer researcher and surgeon; consequently, my knowledge of applying animal research to cancer is stronger than for the use of animals in other fields. For example, when I see animal rights activists claim that human stem cells can be used instead of various animal models of cancer, it’s hard for me not to want to grab them by the lapels, shake them, and point out that cancer stem cells were first discovered by a researcher who observed that only a small fraction of leukemic cells in a mouse–yes, mouse!–model of leukemia could transmit the cancer from one experimental model to another. Since it is not currently possible to transmit leukemia from one person to another and would be incredibly unethical to try, the conclusion I draw from this is that whoever wrote that FAQ is ignorant of recent medical history. Indeed, Americans for Medical Progress is quite correct in pointing out how the judicious use of animal models has led to improvements in understanding and treatment of a number of diseases, such as cancer, asthma, HIV/AIDs, antibiotics, organ transplantation, and far more.
My favorite example to cite when I hear the argument that other methods besides animal research can do better than animal research in helping us understand disease (or, in its more sophisticated form, that animals may have been needed a few decades ago to discover, say, insulin, but our understanding has advanced to the point where they are no longer needed) comes from my field and my area of research interest. It also happens to come from my scientific hero, Dr. Judah Folkman, who passed away suddenly in January. It shows an area of cancer biology whose importance would have been incredibly difficult to model, appreciate, or target for therapy without mouse models of cancer. That area, of course, is tumor angiogenesis, and Dr. Folkman did his pioneering work that has now resulted in drugs like Avastin and other antiangiogenic drugs that are making it to market now and making a real impact on cancer. Dr. Folkman did it through an ingenious strategy that began from the clinical observation that sometimes tumor metastases appear shortly after the operation to remove the primary tumor.
Folkman found a mouse tumor model that mimicked this behavior and in the early 1990s did a series of pioneering experiments. In a strain Lewis lung carcinoma cells of low metastatic potential (LLC-LM), when cells are injected into C57BL/6 mice and allowed to grow subcutaneously, if the tumor is left alone, mice develop only microscopic lung metastases. These metastases do not grow and kill the mouse. If, however, the primary cancer is removed, then many large lung metastases grow rapidly. The results of the experiment above strongly implied that the primary tumor is secreting something that suppresses the growth of microscopic metastases. After this, the Folkman group did what we like to call “brute force” science, collecting mouse urine and analyzing it for tumor suppressive activity until they were able to purify a single 38 kDa peptide, which they designated angiostatin. This involved analyzing literally gallons of mouse urine. (Who said science isn’t glamorous?) Once Folkman’s group had a bunch of angiostatin on hand, it peformed the following experiment. Two groups of mice were injected with LLC-LM and the tumors allowed to grow to a certain size, after which they were surgically removed. One group was treated with angiostatin, and the control group with saline. The result was that the control group developed massive lung metastases and died, while the group treated with angiostatin had microsocopic lung metastases that never grew beyond a ball of cells. Dr. Folkman then demonstrated that it was the inhibition of angiogenesis by the angiostatin that kept these tumors in check. Ultimately, he used a similar method to discover endostatin, and later he demonstrated that endostatin could induced tumor dormancy in mice. I trust that the reader can see how these seminal preclinical observations about angiogenesis would have been virtually impossible without animal models, given that angiogenesis requires the interaction between tumor cells, cells in blood vessels, and the surrounding tissue stroma to occur.
There are numerous other examples of how animal research allows us to do things that we can’t do in humans and discover things that can’t be discovered using just cell culture or human experimentation. Angiogenesis couldn’t have been discovered through tissue culture, and computer models need the concept and measurements of a pheneomenon’s behavior to be useful. Before Folkman, although angiogenesis was suspected to be important in tumor growth, it wasn’t known to be so. My favorite other example is transgenic mice. The technology of making transgenic mice allows scientists to selectively delete a single gene (or multiple genes) and observe the effects that occur in the animal when this happens. P.Z. Myers at Pharyngula has written extensively of one class of genes whose functions in guiding vertebrate development and pattern formation were dissected largely through the use of both Drosophila models and transgenic mouse models, the homeobox genes, and indeed one of my two major areas of research interest involves the study of a diverged homeobox gene. Through these models, we have been able to study not only the structure and function of HOX genes in a manner not possible without animal models, but we’ve been able to trace the evolution of the genes all the way from worms, to fruit flies, to mice, to humans, even to the point of studying how plants and animals are related at a genetic level. These sorts of studies allow us to examine the effect of genes at the whole organism level in a manner that is not possible any other way, and, indeed, sometimes produce strikingly surprising results that lead to new avenues of research.
Another common argument of animal rights activists is that “animal testing” is a very poor means of predicting human response to drugs or other therapies. This argument is closer to the true situation than the first one, and no scientist involved in animal research would deny that there are all too often serious problems in using animal models this way. The problem is, there are even more serious problems using other methods to predict toxicity, drug response, and other parameters in human beings. The most recent example of this type of argument I’ve seen comes from, of all places, a recent issue of Skeptic, in which animal research opponents Niall Shank (a professor of philosophy who, oddly enough, authored a book critiquing intelligent design), Dr. Ray Greek (an anaesthesiologist and President of Americans for Medical Advancement, Europeans for Medical Advancement, and Japanese for Medical Advancement, all facets of a single group that appears to be totally opposed to animal research in medicine), Nathan Nobis (another philosopher), and Jean Swingle-Greek (a veterinarian) authored an article entitled Animals and Medicine: Do Animal Experiments Predict Human Responses? Unfortunately, the text is not available online, but, given that it is a fairly long article, I will summarize and excerpt judiciously to give you a flavor of their thesis and trust my cobloggers and anyone else who has read the article to keep me honest. I will also, when I deem appropriate, include statements from Dr. Greek’s organization in my discussion, given that the Skeptic article clearly flows from the same sorts of rationales found there. In essence, the article consists of arguments #2 and #3 above writ large, with a the additional tactic of setting impossible standards for animal research to meet before they would concede that it is necessary.
First, it should be pointed out that the authors concede that “fundamental biological discoveries in the past three centuries were made by studying animals” and that “animal studies continue to be of important scientific value in the context of basic biological and biomedical research.” Of course, given that they make this concession, it’s odd that the authors can then argue that animal research is so useless in understanding human physiology and designing drug therapies, and they can’t resist disingenuously adding later something that makes me wonder if they truly understand clinical research:
Animal models claimed to be predictive of human responses are widely used in drug testing, environmental toxicology and disease research. Animals, in the case of predictive models, are clearly used as substitutes for human subjects. Unless researchers believed that such models were causally analogous to humans in relevant respects, there would be no rational basis for their use as predictive models.
Yes, and no. Yes, we consider animal models to be somewhat predictive of human responses, but, no, we do not use animals as a substitute for human subjects. If that were the case and animals were viewed as substitutes for humans in testing drugs, then there would be no need for such extensive clinical trials after animal studies were completed. Rather, animal studies should be best viewed as the first test of a new drug or treatment on a whole-organism level in order to look for unexpected, toxic, or other effects that might not be apparent in cell culture. In other words, animal tests are a screening process, not a substitute for human studies. They are also a convenient tool that allows us to test hypotheses that we cannot test in humans, either for reasons of practicality or ethics. We can certainly argue about how good a tool or screening test animal studies can be, but it is disingenuous and incorrect to argue so strongly that animals are meant to be “substitutes” for human subjects. In fact, the argument above is consistent with, albeit a weaker statement of, part of the manifesto of Dr. Greek’s organization, as stated on its website:
Americans For Medical Advancement (AFMA)/ Europeans For Medical Advancement (EFMA)/ Japanese For Medical Advancement (JFMA) is a mainstream science-based research and educational institute dedicated to improving policy and decision-making regarding the use of the animal model in biomedical research.
AFMA/EFMA/JFMA opposes animal-modeled research as a modality for seeking cures and treatments for human disease based on overwhelming scientific evidence that findings from animal models cannot be reliably extrapolated to humans. We seek to demonstrate, through rigorous research and analysis, that the reliance on animal-modeled research, as well as other pseudoscientific endeavors, harms rather than helps humans, and prolongs human suffering by inhibiting medical progress.
If this isn’t dogmatic, I don’t know what is, particularly the implied claim that all animal-modeled research is “pseudoscientific.” Continuing in the Skeptic article, the authors then go on to list a number of references that supposedly show how horrible a tool animal studies are in terms of predicting human response. They tend to pick the worst examples and put even examples that are not so bad in the worst possible light, arguing that animal studies are so inaccurate that they cause more harm than good. That is a tenuous argument based on claims that false negatives and false positives generated by some animal research cause more harm than good by “misleading” researchers and that they “directly” harm patients by allowing harmful treatments to be tried on humans because they supposedly were found safe in animals, so much so that the whole enterprise should be scrapped as hopelessly irredeemable. The latter claim is such a mischaracterization of clinical drug development that it makes me wonder if the authors honestly believe that drugs are approved for general use after only animal trials. Apparently, they have never heard of a phase I clinical trial, where the goal is to give human volunteers increasing amounts of a new drug until toxicities are observed. Animal studies serve, until more predictive studies are found, as a rough guide for (1) dosage and (2) possible toxicities. Moreover, a logical consequence of another part of their complaint, namely that compounds that were found to be harmful in animals and therefore were not taken to clinical trials, is that we should test in humans therapies whose development stopped because of animal testing. After all, if animals and humans are so different that the animal results must be viewed as meaningless, then how else would we identify potentially life saving drugs that were missed? Again, if we as medical scientists actually operated under the straw man assumption that the authors attack (that animal models are such accurate predictors of human response), then we wouldn’t bother to do phase I clinical trials at all! We could just test a new drug in animal models and then go straight to phase II trials of efficacy in humans. Indeed, Niall et al are not only attacking a straw man argument here but dubiously conflating toxicity testing and the recommendations made from it. Health policymakers, many of them unfortunately physicians, often err on the side of extreme caution when an animal experiment using even doses much higher than any human could ever encounter, shows that compound A causes cancer in mice, even as scientists point out the disconnect.
Another oddity is that the authors couch much of their arguments in evolutionary terms. Indeed, they more or less claim that evolutionary considerations are why one absolutely can’t extrapolate responses in one animal to those of humans. Consistent with this article, the website of Dr. Greek’s group, there is a bold claim rejecting the contention that we should, as most scientists would advocate, apply the “three Rs” (“Reduce, Refine, and Replace”) to animal models:
- Applying knowledge gained from animals to humans harms humans most of the time (see A Critical Look at Animal Experimentation for many examples)
- Intractable differences between species mean that animals cannot ‘predict’ how the human body will respond to a disease or a drug. Their use violates the most fundamental principle of biology: evolution. Therefore the ‘animal model’ paradigm should be rejected as unscientific.
Science already has a wealth of superior (not ‘alternative’!) human-based methods at its disposal. They are responsible for the medical care we enjoy today and are the only way to prevent, cure and treat human illness – yet many are starved of funds while animal experimentation is highly funded. The animal experiment lobby maintains that animal experimentation is an expensive business – it is. But it is not just costing society enormous sums of money, it is costing us far more in terms of human health.
The authors use these same arguments in a weaker form in their article for Skeptic, but not once do they show a concrete example illustrating how and why one of these methods is so “superior” to animal experiments–and by how much. One would think, if these methods were so obviously superior and already available, that it would be child’s play for the authors to spell out one or two specific examples in detail, with references from the peer-reviewed literature and concrete numbers, given how much verbiage they devoted to the deficiencies of animal testing. (More on that later,) As for evolutionary considerations, contrary to the fallacious arguments made by Niall et al, appeals to evolution do not tell us that animal models are useless, only whether one model is more likely to be useful than another. There’s a reason why there are no good mouse models of HIV in rodents, while there are useful, albeit imperfect, primate models (some imperfections of which our always-intrepid critics of animal research point out gleefully in the article). Some physiological mechanisms are more highly conserved across species than others, and some animal tumor models, such as lung carcinogenesis due to carcinogens from cigarette smoke, actually recapitulate the genetic changes observed in human tumors pretty well, suggesting common, evolutionarily conserved mechanisms of carcinogenesis, as do some rodent models of human colon cancer, where chemoprevention trials in special strains of mice and rats actually line up fairly well with human trials. Although Niall et al correctly point out that modern medicine is moving towards personalized treatments based on gene expression profiles, we are not there yet.
In some cases the authors also appear to exaggerate just how bad the situation is, even when it comes to references that they themselves cite. One example is a recent BMJ systematic review that–big surprise–found that animal models were good predictors for some conditions and poor predictors for others and was in fact more of a criticism of the quality of published animal studies than the utility of animal studies. It concluded:
Systematic reviews can provide insights into the limitations of animal models. For example, the animal models for stroke, where there was agreement with the results from clinical trials, seemed more representative of the condition in humans than the animal models for head injury, where there were differences in the results. In stroke, the time from the occlusive event to the start of treatment was similar in animal and human studies. In head injury, treatment was given within five minutes of injury in the animal models but up to eight hours after injury in the clinical trials. None of the animal experiments used models that mimic the complex situations that usually follow traumatic head injury. Comorbidities are clearly relevant in stroke, which occurs in older people with hypertension and diabetes but also in people with head injuries, often accompanied by other injuries and by hypotension and hypothermia. Comorbidities were examined in the stroke models but not in the head injury models.
That there is a gap between clinical research and clinical practice is well established. Our work highlights another gap—specifically the lack of communication between those involved in animal research and clinical trialists. Systematic reviews of animal experiments could promote closer collaboration between the research communities and encourage an iterative approach to improving the relevance of animal models to clinical trial design. When models do not represent the clinical context they could be adapted accordingly. Furthermore, as is the case for human research, systematic reviews could help identify and improve deficiencies in the conduct and reporting of animal research.
In other words, according to this article animal models of stroke, for example, are actually pretty good at predicting human response to ischemic brain injury. At the risk of injecting anecdotal evidence, I’ll also point out that research using mouse tumor models that I did 10 years ago and that colleagues with whom I worked until 1999 did shortly after I left their laboratory showing that combining antiangiogenic therapy with either radiation or chemotherapy produces a synergistic antitumor effect has been validated in humans. Indeed, that is now how antiangiogenic therapies are generally used, as combination therapy. I could also point out a number of other reviews that find that animal models can be useful and predictive for cancer chemoprevention treatments, antibiotics, and neuroscience. In rabbits, for instance, evidence for a preclinical effect of drugs against fungal infections can be translated into humans if toxicity and pharmacology can be optimized in humans. In cancer, the use of mouse tumor models is more problematic in that single tumor models tend to have a fairly poor predictive value of human response. However, results are much better when a panel of in vitro and in vivo tumor models are used.
The most pervasive problematic aspect of the Skeptic article (and, indeed, all arguments of this type advanced by opponents of animal research) is the utopian “impossible dream” fallacy coupled with differential standards, in which they demand standards of accuracy of animal research that are impractical or even unobtainable for nearly any disease model, while failing to make the case that the substitutes proposed will yield results at least as accurate and useful or preferably more so. This is very much like the tactics of advocates of “alternative medicine,” who are prone to demanding a utopian standard of perfection for what they perceive to be “conventional” therapies while asserting without good evidence that their therapies are better or expecting us to accept their speculations that they are. The authors of this Skeptic article do in essence the same thing, detailing the problems with animal models while making overblown claims for potential replacements. For example, here is part of their section What Will We Use If We Don’t Experiment On Animals?:
There are two answers to this question.
1. If a test or research method does not accomplish the purpose for which it is used, it should be abandoned. If testing drugs for liver toxicity in animals does not predict liver toxicity in humans, then the test is a waste of resources and, moreover, can be dangerous since the results cannot be counted on to reflect the human condition. So, even if no other testing and research modalities existed, the animal model should still be abandoned. By analogy, there is no cure for AIDS but (hopefully) we would not treat AIDS patients with trephination—drilling holes in the skull—even if trephination happened to be the only available procedure.
This argument sounds seductive on its surface but greatly exaggerates the lack of utility of animal testing. Trephination is indeed utterly useless. Even the worst description of animal testing as made by these authors does not render it the equivalent of trephination, and their comparing it to trephination is clearly a huge exaggeration. The authors then propose a second answer:
2. In the case of human medicine, there are myriad research and testing modalities that are scientifically productive. Anything that is humanbased is, ipso facto, going to be more reliable than anything animal-based. Examples include human embryonic stem cell research; epidemiological studies of patterns of human disease and their associations with environmental causes; in vitro research using human cells and tissues; the use of gene chips or microarrays to study patterns of gene expression in humans; clinical research; autopsies; mathematical and computer modeling; post-marketing drug surveillance; basic scientific research in the fields of biology, physics and chemistry; and technology-based research methods such as those using positron emission tomography, functional magnetic resonance imaging, and others; these are viable means for discovering truths about human disease and drugs.
Here’s where the authors’ double standard towards animal research starts to become obvious. I note that in the several pages of prose leading up to this statement, the authors list and reference many, many problems with animal research. They provide citations for these problems and, when possible, provide numbers for the poor predictive value of some animal models of some diseases. Yet, when they describe their preferred models and tests, suddenly there is not a single reference to tell us how predictive these other methods are, not a single concrete number to show that these tests, or combinations thereof, can predict human response better than the animal models that the authors would like to see replaced by them. Instead, we are told the authors’ preferred methods are “scientifically productive.” No kidding! I wouldn’t argue with the contention that they can be scientifically productive; what I’d argue with is whether the authors’ preferred methods can presently predict human responses better than our current tools that include animal research. If they can’t, then what the authors are proposing is that we throw out a useful (albeit flawed) tool (animal studies) in favor of either unproven techniques with a lot of promise that has not yet been realized (genomics and proteomics, functional MRI), known methods that study human response after the fact, given that they are done after a drug is in release (post-release surveillance, epidemiology), or methods that are clearly no better and probably worse at prediction than animal models (cell culture and computer models), all of which are also being used in addition to animals anyway. I would be more than happy to replace my animal tumor models with a combination of the above–when it can be shown that the above tests do an equivalent or better job of modeling human physiology, biochemistry, and pharmacology. The authors fail to show that they can, preferring argument by assertion rather than by evidence. Indeed, I’m not sure that they can ever show that some aspects of whole-organ physiology can ever be modeled that way, but I’ll never know because they didn’t try. Yet such complexities do not disturb the authors of this article:
We agree that life processes are interdependent; for example, the liver influences the heart, which in turn influences the brain, which in turn influences the kidneys and so on. Thus, the response of an isolated heart cell to a medication does not confirm that the intact human heart will respond as predicted by the isolated heart cell. The liver may metabolize a drug to a new chemical that is toxic to the heart whereas the original chemical was not toxic.
We also concede that cell cultures, computer modeling, in vitro research etc., cannot replace the living intact system of a human being. But while animal models may be intact systems, are they intact systems in ways that are causally relevant to human intact systems? Shifting the focus from genes, cells and tissues to intact animal systems does not evade the long reach of our concerns about causal disanalogy.
This, and their comment about how anything “human-based is, ipso facto, going to be more reliable than anything animal-based” also betrays their agenda. (It also makes me wonder if they think that a study of cultured human cells would be superior in figuring out the effects of hemorrhagic shock on cardiac pumping capacity than the use of a well-designed animal model. Hint: It would be highly unlikely to.) They seem to think that scientists use animal models because we don’t realize that it would be better to do research whenever possible on human cells and human tissue. The reason we do not is that most of the time it is neither practical nor ethical to do such research on humans, particularly research involving what we most want to know about these days, macroscopic responses to genetic and biochemical signals and changes.
Finally, let me show you an example of what I referred to earlier as the “impossible dream” fallacy, namely the impossible standard to which the authors demand that animal research be held:
We realize that our claims are controversial, but our arguments are straightforward. If our arguments are unsound, they should be easy to refute. Here is how:
1. Explain why animals, when used as predictive models for the study of human disease and to test drugs, are not used as CAMs ["causal analogical models."] (Remember, we fully accept that animal studies can yield fruitful insights in the context of basic biological research. If you want to know about rat biology, you must study rats. The issue here is whether you can study humans in ways that are predictively efficacious by studying rats).
2. Show that animal models, when used as CAMs, are successful far more often than not. This can be accomplished by comparing the results of drug toxicity studies in animals with studies in humans or by comparing the results of induced diseases in animals with the same disease in humans.
The first one is more a game of semantics than any substantive criticism in that animals don’t have to be “causal analogical models” to be useful to biological research or predictive enough of human responses to remain useful. It is in essence defining the use of animals in such a way that one could never demonstrate sufficient accuracy to satisfy the authors, namely as a true predictive surrogate for human beings. The second one, however, is the “Bingo!” of double standards. The authors demand that animal models must be successful predictors “far more often than not,” a standard to which they do not even try to hold their preferred methodologies or even claim that they can achieve. Remember, epidemiology is prone to all sorts of confounding factors that make incorrect conclusions very easy, as we discuss quite frequently on this blog. Do they honestly claim that epidemiological studies are accurate “far more likely than not”? Post-marketing surveillance is not predictive; it is after the fact. Genomic approaches, although promising, are very expensive (although they are decreasing in cost). They are also computationally intense, require many human tissue samples to produce statistically valid patterns that must then be validated as predictors, and require very sophisticated statistics and rigorous methodology, without which they are extremely prone to false positive results, something that occurred time and time again early on in such studies. Finally, computer models, as sophisticated as they are becoming, are still too unreliable to trust as a primary modality for preclinical drug testing.
The authors give another indication of where they’re coming from near the very end of the article. First they crow about how none of their critics have been able to meet their intentionally impossible standards, while neglecting to recognize that they can’t meet their own standards, and then they bring out a dread accusation:
Thus far, we have not been able to find such data contradicting our arguments; more importantly, none of our critics have been able to present this data either. One hypothesis that explains this is that there are no such data. Either no one has compiled it, or it simply doesn’t exist. We suspect that these are hypotheses worthy of further research. Until such data can be found, analyzed, and interpreted we must tentatively conclude that the use of allegedly predictive animal models remains in vogue not for scientific reasons but for non-scientific reasons. Those who have an interest in social policy being guided by science should demand that good science prevail and, thus, that society turn its attention to more fruitful methods of biomedical research.
Actually, we are trying to use methods that use fewer or no animals. There isn’t a biomedical researcher alive that I’m aware of who doesn’t wish that there were another way to get the answers we seek. Animal research is expensive, messy, labor- and time-intensive, and falling under increasingly onerous federal regulations. As for the claim that animal models remain in vogue for non-scientific reasons, that’s about as good a case of the pot calling the kettle black as I’ve seen in a long time. Here’s why. Most responsible animal researchers actually do subscribe to the “3 Rs” mentioned earlier. It’s a reasonable strategy for minimizing the use of animals now while still using them to study diseases where they have relevance, all with the the long term goal of eventually rendering animal models unnecessary, particularly if we regulatory bodies approve only high quality studies. Unfortunately, that reasonable strategy is not enough for the authors of this article, however, at least not for Dr. Greek as evidenced on his group’s website. He considers animal research to be so rotten and useless that it should be stopped now, even though he can’t show that there are other methods, whether they are “alternatives” or mainstream, that can do the job even as well as (according to the article) the poor standard animal research presently shows. He comes to that conclusion through straw man arguments about how animals are actually used and defining the bar for success so high that few imaginable modalities could reach it while excusing his “alternatives” from the same bar. His vision, if it came to pass, would be a disaster for biomedical research. His arguments, as embodied in the Skeptic article, are the real pseudoscience.
When it comes down to it, opposition to animal research is almost always far more ideological than scientific in nature. I purposely did not go into the moral issues involved with animal research because such issues depend very much upon the background and moral framework from which each of us proceeds. Animal rights extremists, such as those described at the beginning of this post, tend to emphasize their moral objection to such research above all, or how they view animals as being equal to humans and deserving of the same rights. That is how they justify their campaigns of harrassment and even violence against medical researchers. Indeed, some of them argue that even in the hypothetical case in which animal research would definitely result in a cure for AIDS or cancer, they would still oppose it. Other opponents of animal research, such as authors of the Skeptic article, differ from the extremists in that they are not violent, will acknowledge when pushed that animal research has produced useful scientific discoveries, and do not make arguments that animals are morally equal to humans or that we should therefore not eat them, experiment on them, or keep them as pets. However, in their seeming reasonableness, their opposition to animal research is not, their claims otherwise notwithstanding, based any more on science than that of the animal rights extremists. That they would apply such a double standard to animal research compared to their favored non-animal research modalities, coupled with their absolutist language and cherry picking of studies, show this quite conclusively. Just remember, whenever you hear seemingly “scientific” arguments against animal research that emphasize how bad and inaccurate it is, ask for concrete examples from the peer-reviewed literature that show that non-animal modalities are consistently equal to or better than animal experiments to answer the question being asked. You’ll be hard-pressed to find them.