Philosophy of AR > Animal Testing - Index > Anti-Vivisection Index

Vomiting ferrets and vomiting shrews -- a scientific debate

Animal Aid has consistently criticized the use of ferrets and shrews as "models" for the study of vomiting in humans. Following are two specific examples, both of which took place at the St George's Hospital Medical School, London.

Vomiting ferrets (published in 2002):

"In an experiment funded by the Pfizer drug company, 18 ferrets were injected with loperamide, an opiate that induces vomiting. Some of them were pre-treated with an anti-emetic. Retching (up to 85 times), vomiting (up to 10 times), gagging, mouth scratching, intense licking and 'wet-dog shaking' episodes were counted for the following two hours. All the animals were then killed and their brains analyzed, in an attempt to locate the 'vomiting centre'. The authors suggest their findings may have relevance 'in the clinic', though exactly what relevance ferret vomiting and retching has for people is not explained".

Vomiting shrews (published in 2005):

"Researchers at St George's Hospital Medical School, London, chose the house musk shrew as a model for studying nausea and vomiting in humans (a known side effect in people undergoing chemotherapy). It was observed, by chance, that some of the shrews spontaneously ate their own vomit. This observation persuaded the researchers to study vomiting in shrews, unrelated to their chemotherapy-in-humans experiments. Three groups of 16 shrews (specially bred at St George's) were forced to vomit, either through the effects of motion sickness, or else by being administered nicotine. The animals were subsequently killed with carbon dioxide, and then their necks were broken. This was done in order to study and weigh the stomach contents of each shrew. The researchers concluded that 'This curious behaviour observed under laboratory conditions, if replicated in the wild, may have significant ecological consequences for shrews'.

In December 2005, Animal Aid was contacted by a research scientist, who tried to justify the use of such animals in medical research. Andre Menache MRCVS replied on behalf of Animal Aid as scientific consultant. In deference to the request for anonymity, Animal Aid has not published any identifying details of the research scientist.

Research Scientist:

Ferrets, unlike rats, vomit and retch and so are the most likely animal to test anti-emetics on. No computer model of this effect is even in the process of being developed due to its unbelievable complexity. This is important for the clinic to understand how to moderate the severe effects of chemotherapy as well as some anaesthetics. As of your article (2002), small animal imaging of the brain had not been developed enough to use. Novel imaging techniques can make some terminal animal experiments unnecessary in the not-to-distant future.

Animal Aid:

Thank you for your message to Animal Aid concerning the testing of anti-emetics in ferrets. I would like to respond to your comments, as one scientist to another, as follows:

1. As a veterinary surgeon, I am acutely aware of the concept of "species differences" within the animal kingdom in general, and within domestic animal species in particular. In treating domestic animals in a clinical setting, I need to be aware of inter-species sensitivity with respect to various anti-emetics, as well as intra-species (age, breed, etc) sensitivity, with respect to the selection and dose of an appropriate anti-emetic. I would assume that the same principle would apply to human clinical situations as well.

I therefore find it difficult to see how ferrets could predict optimal doses and anti-emetic drug of choice for humans of different ages, who may be experiencing a wide range of uniquely human physiological or pathological conditions. Animal models are typically poor predictors of human response (see refs). One can try to look for animal-human correlates in retrospective studies - by which time the result in humans would be known - so why bother with the animal experiment in the first place?

2. We live today in an era of evidence-based medicine. I therefore challenge you to point me to a systematic review in the scientific literature of anti-emetic ferret studies that clearly demonstrate human clinical benefit.

I would like to refer you to two documents published by Animal Aid, one of which is a critique of GM animal research, written by Dr Jarrod Bailey, a geneticist - and the other, a critique of dog models of cardiovascular research, written by Dr J Pippin, a US cardiologist. Both critiques are viewable on our website: 


eg. Clin Pharmacol 1962;3:665-72
Zbinden, G (1991) Predictive value of animal studies in toxicology. Regul. Tox. Pharm. 14: 167-177
CMR Workshop ˇ Animal Toxicity Studies: Their Relevance for Man Quay 1990 p 49-56 and p57-67
Spriet-Pourra, C and Auriche, M (Eds) 1994 SCRIP Reports PJB, New York Garratini, S (1985) Toxic effects of chemicals: difficulties in extrapolating data from animals to man. Annu. Rev. Toxicol. Pharmacol. 16: 1-29
Zbinden, G (1993) Regul. Toxicol. Pharma col. 17: 85-94
Calabrese (1984) Suitability of animal models for predictive toxicology: Drug Metab Rev 15: 505-523
Oser, BL (1981) J. Toxicol. Environ. Health 8: 521-642
Calabrese, EJ (1987) Principles of Animal Extrapolation. Wiley, New York
Olson, H., Betton, G., Stritar, J., and Robinson, D. (1998). The predictivity of the toxicity of pharmaceuticals in humans from animal data-An interim assessment. Toxicol. Lett. 102-103, 535-538
Regulatory Toxicology and Pharmacology 2000;32:56-67
Drug Metabolism and Drug Interactions 2000;16:143-155

Dr Ralph Heywood, former director of Huntingdon Research Centre, said,
the best guess for the correlation of adverse reactions in man and animal
toxicity data is somewhere between 5 and 25%."

Research Scientist:

Thank you for your thoughtful and intelligent reply. I don't think that there is a responsible scientist that would disagree that allometry is an inexact science and that animals are not always a good predictor of human physiological response. As I am sure you are aware, there are some models that are very good (i.e. cardiac remodeling response in rabbits) and some that are abysmal (i.e. hepatotoxicity in rats). Conscious dogs in ventricular fibrillation research have the advantage of normal sinus arrhythmia without intervention but, as with all models (even computer) there is disagreement among scientists on the meaning of the data, which extends to humans, as well. This research has, however, lead to some important discoveries that are not possible with human subjects unless discovered by providential intervention.

I am sure you would agree that before we can ethically give human subjects any drug, there must be some realizable and safe response in a physiologically active model. I have been and continue to work on developing mathematical models of physiological systems as well as have thousands of biomathematicians that regularly publish in the Journal of Applied Physiology and, given the toxicity of some of the powerful compounds in development, no intelligent and caring person would want their sister or mother to take a drug for breast cancer that was only tested on a relatively simple computer model or in such a simple environment as a petrie dish. I point you to the decades of research done by Arthur Winfree, Physiome, LLC and Johns Hopkins University on a complex cardiac computer model that failed to provide information that was useful to drug safety. I cannot wait for the day when we can look at the human genome and design a drug that will work unconditionally with no toxicity; however, the genome is proving to be much more complicated than the hopeful but unknowing public anticipated and genomics, which is now grown to panteomics, is really just an infant in this whole drug discovery process.

While you make some very good points on some of the animal model issues, I have found that data, which when correctly designed for and collected, does not support the lone use of either in silico or animal modeling. In any case, there is absolutely no chance that the FDA would ever allow a First-In-Human Study without first testing both efficacy and safety in the best animal model. The only alternative would be unthinkable to all but the most severely misanthropic- to define a human subpopulation that should be experimented on and use them as disposable commodities. Even drugs that appear safe kill the animals used in research and are dropped on the spot. It is not worth pursuing a compound that will destroy your company when there are a string of backups that were developed in parallel to take its place.

I can assure you that my fellow scientists, statisticians, technicians and managers work diligently every single day to come up with the best method of drug discovery and development. It is certainly not in anyone's best interest, scientists, managers, marketers or financiers to use methods that waste time and money with less than optimal guarantee of success. Failures waste time, money and, sometimes, companies. Computer models would reduce the drug discovery process to months instead of decades if they worked. Exploration and discovery would be limited only by computing power and physical limitations of robotics. Can you imagine what a drug company could do with 15 years of patent-protected marketing instead of 5 years!? To this end, I submit that if better models come along, it is in everyone' s interest, yours, mine and Pharma, to get the right answer in the most efficient manner possible. When viable alternatives to animal modeling arise, they are being used at every available opportunity but for now, in the time and world we live in, we, meaning caring and thoughtful humans are limited by the knowledge we have gained up to this point in our brief history. I applaud your work to develop alternatives to animal models and wish you luck. You will need it as well as I. We are not on opposite sides.

Animal Aid:

Thank you for your reply. One of the pitfalls in these sorts of communications is to talk in generalities rather than address specific issues. I will therefore try to cover as many specific points as possible, in reply to some of your statements. I would appreciate it if you could reciprocate, particularly with respect to the point I made about systematic reviews in critically assessing animal studies and clinical trials.

When you talk about "very good" animal models (your example "cardiac remodeling response in rabbits"), on what do you base this assessment? Is it based on systematic reviews? Can you point me to a specific result in veterinary or human clinical medicine?

2. You mention the use of conscious dogs in ventricular fibrillation research as having led to some important discoveries. This is a fairly general statement, which is unhelpful in the context of our current dialogue. In contrast, you made no comment about the specific critique of dog models of cardiovascular research, written by Dr J Pippin, to which I referred you.

3. Yes, I do agree with your statement that before we can ethically give human subjects any drug, there must be some realizable and safe response in a physiologically active model. But which active model are you referring to? The active dog model or the active cat model, each of which is species-specific? The human male or the human female, each of which is gender-specific (with respect to heart drugs, for example)? The human adult or the human child, each of which is age-specific?

4. And yes, I agree with your statement that no intelligent and caring person would trust potentially toxic drugs on simple computer models or petri dishes. In reply to this, I would ask that you read the brief document entitled Lethal business ( ), which illustrates the lack of scientific methodology in some of our current toxicity assessments. It also covers some of the modern methods of toxic risk assessment, which clearly go far beyond simple computer models and petri dishes.

5. You mention the point that regulatory authorities would not allow a First-In-Human Study without first testing both efficacy and safety in the best animal model. Sadly, you are technically correct. I use the word "sadly" because - in my view - these regulatory requirements are based on 50-year-old laws and 150-year-old science. Modern science teaches us that the best animal model (the chimpanzee, with whom we share approximately 98% of our DNA) is immune to AIDS, hepatitis B and common malaria, and that's the best model we've got... If you have the time and inclination, I strongly urge you to read Dr Jarrod Bailey's just-published critique of non-human primates in medical research (attached).

I firmly believe that animal experimentation is on the way out -- not only because it is cruel -- but because it represents outdated science. Those of us who lived 150 years ago could be forgiven for believing that a beating dog heart is no different to a human heart. Today, there is no longer any excuse. We now know that animals and people are very different from each other, and that we as human individuals are quite unique in our response to medical drugs. The word "similar" is no longer a scientifically robust term to use when peoples lives are at stake. We live in an era of evidence-based medicine. There is a growing demand within the scientific research community for transparency and accountability.

It is regarded by many in the scientific community as indefensible to conduct more research while existing research has yet to be evaluated. This principle already applies to clinical research and is monitored through organizations such as the Cochrane Collaboration but the same principle should apply to animal studies while they form a part of preclinical science. In a recent letter to the Lancet it was noted that â...academia as a whole has still not grasped that it is unscientific and unethical to embark on new research without first analyzing systematically what can be learned from existing research; or that it is impossible for consumers of research results to assess what contribution a new study has made unless its results have been set systematically in the context of an up-to-date review of the totality of the relevant evidence."

It follows that a moratorium on animal research would be scientifically defensible while a large-scale programme of systematic reviews is conducted to determine the value of existing animal studies (ref.

Research Scientist:

Thank you Andre for your quick and passionate reply. Animal studies are expensive, long and inexact and bioinformatic models are making long and heavy strides in supplanting animal studies for efficacy, toxicity and quickness. However, like all new (and make no mistake about it, these are very new) technologies that are based not on seeing an actual physiological system react but on the results of an algorithm and a computer printout, this will take some time for general acceptance. This is my work and I would not want to take any medicine that I did not test on some type of living complex organism. The new technologies will prevail, being driven not only by compassion but also by good business sense. Everyone is working on this but the models are far from being validated and, in the meantime, science cannot drop everything to wait for that to happen. Be patient my friend because the path is drawn and, specifics aside, I assure you that the
scientific community is already heading down that road.

Animal Aid:

Your statements are encouraging, but there is no excusing the scientific community, the pharmaceutical industry and the regulatory authorities for dragging their heels for the past 50 years:

* animal experiments have never undergone validation studies, and yet government and industry continue to accept animal data for use in human medicine

* on average, a single non-animal replacement method requires 10 years in order to be validated - at this rate, animal experiments will persist for a few more hundred years at least
* the few systematic reviews already carried out comparing the results obtained from animal studies with their clinical application have revealed serious flaws in the methodology of the animal tests (e.g. Horn J, de Haan RJ, Vermeulen M, Luiten PG, L imburg M. Nimodipine in animal model experiments of focal cerebral ischemia: a systematic review. Stroke. 2001 Oct;32(10):2433-8. Review.)

You still have not responded to the specific points I raised. I must conclude that you either do not have the time or the expertise to do so. This is not meant as a criticism Ë just an observation.

Research Scientist:

My original reply to you was meant only to inform the site of the reason why

anti-emetics were tested in animals since I took the web site statement to be either ignorant or sarcastic or both. I hoped I was wrong and merely meant to inform. I work in the field of biomarker development and validation and we in this field are intimately familiar with both animal, in vitro, molecular and in silico models. We work every day to replace inefficient animal models with either more efficient models that maximize the information from any one animal or with evidence-based models that utilize anything from Phase 0 through Phase 4 clinical data. You will not read journal publications from me about this since I am not interested in publicity but doing the best job that I can. I have no need to worry about my standing in any community. I am also not dedicated, as you are, to immediate elimination of animal models but do applaud you for your diligence. My work, now and in the future, is with a different battle in the same war.