Knight A. 2007. Systematic reviews demonstrate poor clinical utility of animal experimentation. [online rapid response to Hackam DG. Translating animal research into clinical benefit. BMJ 2007;334:163-164.] http://www.bmj.com/cgi/eletters/334/7586/163#159752, accessed 18th Apr. 2007.
Since animal experimentation first began, advocates have claimed they are crucially important to the development of cures for human diseases. No matter how eminent such advocates and how earnest such claims, such opinions do not constitute an adequate form of evidence, however, given the contrary opinions of other scientists and the widespread social concern about animal experimentation. The need for firm evidence of the clinical utility or lack thereof of relatively expensive animal experimentation is also becoming ever more urgent given increasing competition for scarce research funding. Accordingly, systematic reviews of the clinical utility of animal experiments are warranted.
One of the first such reviews was published in 2005 by Lindl et al. . After examining citations of all published animal experiments conducted at three German universities between 1991 and 1993, Lindl et al. located only four publications indicating a direct connection with human outcomes. In each of these cases, however, the human outcomes were opposite to those obtained in animals.
In 2005 my colleagues and I published an examination of the concordance between various animal species and humans for 1,396 putative teratogens . We found a huge variability in positive predictability for known human teratogens, the mean of which was only 51%--hardly better than tossing a coin--with some commonly utilised species showing a high rate of false negative results.
In 2006 we demonstrated the limited utility of animal carcinogenicity data in deriving human risk assessments [3-4]. We found that for environmental contaminants of greatest US concern, the Environmental Protection Agency (EPA)--the US federal agency most responsible for protecting public health from environmental contaminants--considers animal data useful for predicting human carcinogenicity less than half (41.9%) the time. Furthermore, highly significant (p<0.0001) differences of human carcinogenicity classifications of identical chemicals between different agencies demonstrated that the EPA nevertheless remains over-reliant on animal carcinogenicity data, that it consequently tends to over-predict carcinogenic risk, and that the true human predictivity of animal carcinogenicity data is even poorer than indicated by EPA figures alone. EPA policy erroneously assuming that tumours in animals are necessarily indicative of human carcinogenicity was implicated as a primary cause of these errors.
Chimpanzees are the species most closely related to humans, and consequently, most likely to be predictive of human outcomes when used in biomedical research. However, in 2006 I demonstrated that of 95 chimpanzee experiments randomly selected from a population of 749 published during a recent decade, only 14.7% (14/95) were cited by 27 papers that appeared to describe well developed methods for combating human diseases . However, detailed examination of these medical papers failed to yield any chimpanzee study able to demonstrate an essential contribution, or, in most cases, a significant contribution of any kind, towards the development of the medical method described.
In 2006 Hackam & Redelmeier published a systematic review examining the frequency with which highly cited animal studies translated into successful human research . Of 76 animal studies with a median citation count of 889 (range: 639-2,233), only 36.8% (28/76) were replicated in human randomized trials. 18.4% (14/76) were contradicted by randomized trials, and 44.7% (34/76) had not been tested.
These systematic reviews clearly do not support the opinions of advocates that animal experiments are beneficial in the development of human therapeutic interventions and the assessment of human toxicity. On the contrary, they consistently demonstrate that animal experiments are of low clinical utility. In fact, no systematic review has been published in a peer-reviewed biomedical journal demonstrating that animal experiments are as beneficial in combating human diseases as advocates frequently claim. To persist with such claims contrary to this consistent and growing body of evidence is unscientific.
Differences between laboratory animals and humans that may contribute to their poor clinical utility include differing susceptibility to, aetiology and progression of diseases; differing absorption, tissue distribution, metabolism, and excretion of chemotherapeutic agents; and differences in the toxicity and efficacy of pharmaceuticals . The results are that laboratory animals are unreliably predictive of human outcomes, and that experiments on them constitute an inefficient means of combating human diseases.
The poor methodological quality of many animal experiments also contributes to their lack of clinical utility. This has also been demonstrated in several systematic reviews. In 2001 Horn et al. reviewed 20 animal studies examining the efficacy of Nimodipine in animal models of focal cerebral ischemia . Although animal studies are intended to be conducted prior to human trials to allow detection of potential toxicity and assessment of efficacy, in this case clinical trials proceeded concurrently with animal studies, which demonstrated equivocal evidence of efficacy.
In 2001 Roberts et al. systematically reviewed 44 randomised controlled animal studies examining the efficacy of fluid resuscitation in uncontrolled haemorrhage . They found that poor sample size choices left most studies underpowered, limiting the conclusions that could be derived, and that little information was provided on possible biases in the selection of treatment groups.
In 2004 Pound et al. reviewed six systematic reviews examining the extent to which animal experiments had informed human clinical research, and found that in two cases clinical trials were conducted concurrently with animal studies, in three cases clinical trials were conducted despite evidence of harm from prior animal studies, and in the remaining case the outcome of the animal study contradicted the findings of previous investigators, who appeared to have cited only studies that supported their prior views .
In 2006 O'Collins et al. examined 1,026 experimental treatments in acute stroke, finding that 114 drugs used clinically were no more effective in animal experiments than 912 drugs tested only in animals . Accordingly, they questioned whether the most efficacious drugs are, in fact, being selected for clinical trials. They called for greater rigor in the conduct, reporting, and analysis of animal data.
The 2006 systematic review of highly cited animal studies by Hackam & Redelmeier found that only 37 (49%) were of good methodological quality. Few studies included random allocation of animals, adjustment for multiple hypothesis testing, or blinded assessment of outcomes .
The 2007 review by Perel et al. recently published in BMJ provides the latest addition to this growing evidence base . Upon examining interventions with unambiguous evidence of a treatment effect (benefit or harm) in clinical trials of corticosteroidal treatment for head injury, antifibrinolytics in haemorrhage, thrombolysis in acute ischaemic stroke, tirilazad in acute ischaemic stroke, antenatal corticosteroids to prevent neonatal respiratory distress syndrome, and bisphosphonates to treat osteoporosis, Perel et al. found that three interventions had similar outcomes in animal models, while three did not. Perel et al. reported that the animal studies varied in methodological quality and sample sizes, that randomisation and blinding were rarely reported and that publication bias was evident.
Perel et al. consequently called for the application to animal studies of standards for evidence based reporting similar to those applied to clinical trials. They also called for systematic reviews of animal experimental outcomes prior to progression to clinical trials, facilitating better detection of toxicity and efficacy, thereby potentially improving the safety and efficiency of the drug development process.
However, as Perel et al. stated, the discordance between human and animal outcomes is unlikely to result from the poor methodological quality of animal experiments alone. The failure of animal models to adequately represent human disease is another fundamental cause, which in contrast, may be impossible to correct, even in theory.
The decisions about whether or not to conduct research on animals must always involve weighing the animal and financial costs against the likely benefits of the research. The animal costs of invasive or harmful procedures are readily apparent. However, in 2004 a review of 80 studies by Balcombe et al. demonstrated that virtually all common laboratory species experience rapid, pronounced, and statistically significant elevations in a range of stress-related physiological responses to common laboratory procedures as well, including handling, blood collection, and gavaging . In 2006 Balcombe further demonstrated in a review of over 100 studies that the small, relatively barren cages in which the vast majority of laboratory rodents spend most of their lives also result in deleterious neuroanatomical, psychological and physiological effects . Such effects not only have profound ethical implications, but are also likely to distort scientific outcomes, through the stressful alteration of a range of physiological parameters, and through increased susceptibility to diseases secondary to stress-mediated immunosuppression.
Additionally, the human costs must also be considered, when other research fields or preventative medical strategies conceivably more likely to benefit human health may be deprived of funding consumed by relatively expensive animal research.
Given these profound bioethical and financial costs, and the widespread and legitimate public concern about animal experimentation, the onus is on those who would spend society's animal and financial resources on these experiments to clearly demonstrate that they are an efficient means of combating human diseases. To date, they have not done so.
Andrew Knight BSc., BVMS, CertAW, MRCVS Director, Animal Consultants International: an internationally-based organisation providing policy expertise on animal issues. Related works: http://www.animalconsultants.org/portfolio/portfolio.htm , http://www.animalexperimentfacts.info/ .
1. Lindl T., Vlkel M. & Kolar R. [Animal experiments in biomedical research. An evaluation of the clinical relevance of approved animal experimental projects: no evident implementation in human medicine within 10 years]. [German]. ALTEX 2005;22(3):143-51.
2. Bailey J, Knight A, Balcombe J. The future of teratology research is in vitro. Biogenic Amines 2005;19(2):97--145.
3. Knight A, Bailey J, Balcombe J. Which drugs cause cancer? Animal tests yield misleading results. British Medical Journal USA Oct. 2005;331:E389-91.
4. Knight A, Bailey J, Balcombe J. Animal carcinogenicity studies: 1 poor human predictivity. Alternatives to Laboratory Animals 2006; 34(1):19 -27.
5. Knight A, Bailey J, Balcombe J. Chimpanzee research: 2. lack of efficacy in combating human disease. Altex: Alternatives to Animal Experimentation 2006; 23(2):108.
6. Hackam & Redelmeier. Translation of research evidence from animals to humans. JAMA 2006;296(14):1731-2.
7. Bailey J. Non-human primates in medical research and drug development: a critical review. Biogenic Amines 2005;19(4-6):235--55.
8. Horn J., de Haan R.J., Vermeulen M., Luiten P.G. & Limburg M. Nimodipine in animal model experiments of focal cerebral ischemia: a systematic review. Stroke 2001;32(10):2433-8.
9. Roberts I, Kwan I, Evans P & Haig S. Does animal experimentation inform human healthcare? Observations from a systematic review of international animal experiments on fluid resuscitation. BMJ 2002;324:474-6.
10. Pound P., Ebrahim S., Sandercock P., Bracken M. & Roberts I. Where is the evidence that animal research benefits humans? British Medical Journal 2004;328:514-7.
11. O'Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp Bart H & Howells DW. 1,026 experimental treatments in acute stroke. Ann Neurol 2006;59:467--77.
12. Perel P, Roberts I, Sena E, Wheble P, Briscoe C, Sandercock P, Macleod M, Mignini LE, Jayaram P, Khan KS. Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ 2007;334:197-
13. Balcombe J, Barnard N, Sandusky C. Laboratory routines cause animal stress. Contemporary Topics in Laboratory Animal Science 2004;43(6):42-51.
14. Balcombe J. Laboratory environments and rodents' behavioural needs: a review. Laboratory Animals 2006;40(3):217-35.