APPENDIX 22
Memorandum from Europeans for Medical
Advancement
We endorse the aims of the new chemicals policy
(ensuring a safer environment for both people and wildlife) and
agree that rapid action is needed to restrict the use of harmful
substances but our concern is that animal testing is not an appropriate
way to achieve that end. In fact, there can be no prospect
of fulfilling those aims if animal tests are employed to provide
the data. Apart from the timescale and expense involved being
utterly prohibitive, the results would be entirely meaningless
and even counter-productive: allowing unsafe chemicals to be deemed
safe. There are countless substances marketed today that are known
to be harmful to people and the environment. Animal testing has
not helped to prevent their use.
In contrast, using techniques based on molecular
toxicology and employing human tissues, chemicals' toxic risk
to humans (or the animal or plant species of interest) can be
defined with scientific confidence, accuracy and speed. The adoption
of such science based toxicology would have massive benefits for
consumer safety. For example, the reliable identification of carcinogens
and their subsequent removal from the environment would significantly
reduce the incidence of cancer.
Animal based toxicity testing is scientifically
flawed and dangerous for human health. Animal test results are
unreliable when applied to human beings and misleading when compared
to real life exposure. Animal toxicity tests have also not been
validated to modern standards. The proposed Chemicals Policy offers
the EU an opportunity to lead the world in the use of modern non-animal
tests. By using fast and efficient non-animal tests, it is possible
to gather enough relevant information on large numbers of chemicals
to classify and regulate harmful substances.
Science-based toxicity testing and proper use
of pre-existing data (especially from human exposure) is the only
scientifically acceptable way to restrict the use of harmful substances.
We will go on to explain, with examples, how
and why testing on animals is inappropriate for humans and even
for other animals.
The populations of three species of vultures
have decreased dramatically since the early 1990s because of a
veterinary drug, Diclofenac, used to reduce fever and treat lameness
in farm animals. The birds are dying of kidney failure, thanks
to drug residues in the carcasses of cattle, which constitute
the bulk of their diet. This is a prime example of how chemicals
affect different species in different ways.
Countless drugs, which have been safety-tested
in animals, go on to cause serious side-effects, including death,
in humans. Adverse drug reactions are the fourth leading cause
of death in the Western world, killing over 100,000 people every
year in the US alone. Clearly, the animal tests are not an effective
safety net. The fact that 80% of drugs fail in Phase I clinical
trials after passing animals tests is enough in and of itself
to indict the animal testing system.
Only by comparing the results from testing each
given substance or procedure in an animal species with data from
humans can we determine whether the animal is sufficiently similar
to humans to allow extrapolation. Not only that, but similar short-term
results could turn out to be completely different in the long
run—as is often the case with carcenogenicity. We can only
know which animals mimic humans after we study the human data.
Clearly, the predictive value of such models is nil.
Many environmental poisons have been permitted
continued sale and damaging exposure to humans because the manufacturers
were able to cite safety data from animals. The most obvious examples
include cigarettes and asbestos, both of which were claimed to
be safe long after human epidemiological data had shown otherwise.
The price of such unwarranted faith in animal experiments has
been many millions of human lives.
The following examples illustrate the failure
of the animal model in medical research, but the principle—that
inter-species extrapolation is hazardous—is clearly the
same.
The National Cancer Institute (NCI) tested 12
anti-cancer drugs on mice that are currently being used successfully
in humans. The scientists took mice that were growing 48 different
kinds of human cancers and treated them with the 12 drugs. They
found that 30/48 times, the drugs were ineffective in the mice.
In other words, 63% of the time, the mouse models with human tumors
inaccurately predicted human response.[53]
In a study that spanned over 10 years and has
not yet been repeated, the Food and Drug Administration (FDA)
began in 1976 to follow all the new medications it released for
side effects. In that study, the FDA found that out of 198 new
medications, 102 (52%) were either recalled or relabeled secondary
to side effects not predicted in animal tests.[54]
A similar study examined six drugs, the side effects of which
were already known in humans. The study found that animals correctly
predicted 22 side effects, but incorrectly identified 48 side
effects that did not occur in humans, while missing 20 side effects
that did occur in humans. This means that the animal models were
incorrect 68/90 times, or 76% of the time.[55]
More recent research indicated that a new drug has a one in five
chance of being relabeled or recalled due to serious adverse reactions.[56]
This, despite the fact that all new medications have undergone
extensive animal testing prior to being released to the public.
In August 2001, Mark Levin, Ph.D. and CEO of
Millenium Pharmaceuticals, presented data at the Drug Discovery
Technology Conference in Boston, MA regarding the inadequacy of
current animal models in drug testing. In the study he presented,
28 potential new drugs were tested in rats for liver toxicity.
Eleven of these drugs were shown to be toxic, while 17 were shown
to be safe. Twenty-two of the 28 potential drugs advanced into
human clinical trials, and the results revealed that of the 11
drugs that had been shown to be toxic in rats only two were toxic
in humans, while six were safe. Of the 17 drugs that were safe
in rats, eight were found to be safe in humans, while six were
found to be toxic to humans. Levin concluded that this basically
means the animals were about as accurate as "a coin toss."
In the Handbook of Laboratory Animal Science
Volume II: Animal Models, the authors state:
The case of the huge 25-year screening program,
undertaken by the prestigious U.S. National Cancer Institute,
illustrates the kind of dilemma possible: in this program, 40,000
plant species were tested for anti-tumor activity. Several of
the plants proved effective and safe enough in the chosen animal
model to justify clinical trials in humans. In the end, none of
these drugs was found useful for therapy because of too high toxicity
or ineffectivity in humans. This means that despite 25 years of
intensive research and positive results in animal models, not
a single antitumor drug emerged from this work. As a consequence,
the NCI now uses human cancer cell lines for the screening of
cytotoxins.
Of 20 compounds known not to cause cancer in
humans, 19 did cause cancer in animals[57]
while of 19 compounds known to cause oral cancer in humans, only
seven caused cancer in mice and rats using standard NCI protocol.[58]
Of 22 drugs tested on animals and shown to be therapeutic in spinal
cord injury, none were effective in humans.[59]
The American Heart Association (AHA), the American
College of Emergency Physicians, the American College of Cardiology
(ACC), the European Resuscitation Council, the Heart and Stroke
Foundation of Canada, the Institute of Critical Care Medicine,
the Safar Center for Resuscitation Research, and the Society for
Academic Emergency Medicine stated the following in the journal
Circulation:
— concerning animal experiments into
cardio-pulmonary resuscitation, "Unfortunately, the results
of one lab may not be reproducible in another lab or in human
trials."
— for cardiac arrest, "high doses
of epinephrine therapy significantly improved survival in most
animal models but does not improve survival in humans."
— "species differ in response
to anesthesia and drugs, and may require different doses to produce
the same physiological response."
— "differences in metabolism,
physiological function, response to ischemia, hypoxia, hypercarbia
. . . return to spontaneous circulation . . .[are seen] in rats,
dogs and pigs."
— rats, dogs and pigs show "anatomical
differences [in] myocardial blood supply, pre-existing collateral
circulation, sensitization to arrhythmia . . . shape of chest."[60]
Although aptiganel (a n-methyl-D-aspartate (NMDA)
receptor blocker, manufactured under the brand name Cerestat),
was effective at providing brain protection against stroke in
animal models, large clinical studies revealed no positive effects
and possibly some harm when it was given to humans.[61]
More patients who received the drug died than those who did not,
and more side effects were observed in the group receiving aptiganel
than in the control group. No benefits were seen in patients treated
with aptiganel. In contrast to humans, rats given aptiganel showed
a decrease in brain damage by up to 70%. According to the Associated
Press, "Yet another experimental stroke drug that showed
great promise in animals has failed in humans, with the study
cut short because patients were dying or showing no improvement."[62]
Today we have many scientifically reliable research
and testing methods that we should be using instead of animal
studies. The following are examples of techniques that could be
used in toxicity testing:
1. In silico evaluation of the toxic
effects of a molecule, derived from its chemical structure (structure-activity
relationship).
2. Cellular studies performed on primary
cultures, as well as in established cell lines.
3. In vitro testing of enzyme activity.
4. Toxicogenomics, using DNA chips which
allow study of the expression of many genes simultaneously.
5. Toxicoproteomics, using protein chips
and other rapidly-developing new tools.
6. Identification of metabolites by mass
spectrometry, Raman spectroscopy, etc.
7. Molecular methods exist to assess membrane
toxicity (modification of polarity, size and structure of lipid
rafts, etc), and epigenetic toxicity (resulting in methylation,
acetylation or phosphorylation of chromatin, which can seriously
affect gene expression), etc.
8. Similarly, molecular methods exist to
assess immunotoxicity, neurotoxicity, reproductive and developmental
toxicity, signal transduction and other toxicities.
9. Assessment of cellular toxicity for classes
of the human population sharing common genetic polymorphisms could
be made, using class-specific DNA chips, which allow one to list
chemicals which are especially harmful, or conversely, safe, for
members of the class (polymorphism-specific toxicology).
For example, a large human polymorphism is found in the P450 family
of metabolising enzymes.
10. For certain classes of chemicals, effects
of micro-doses on functioning organs in situ could be monitored
in human volunteers under strict clinical test conditions, with
fully informed consent. Of particular value are imaging techniques
(MRI, PET scan, etc), which allow one to identify the organ targeted
by the xenobiotic, as well as its metabolism and elimination.
The only prospect of fulfilling the objectives
of REACH is through rapid implementation of such science-based
toxicology in place of traditional animal tests, which have no
scientific merit and actually endanger consumers, as well as being
totally impractical in terms of timescale and expense.
March 2004
53 Science vol 278, 7 Nov, 1997 p 1041. Back
54
GAO/PEMD-90-15 FDA Drug Review: Postapproval Risks 1976-85. Back
55
Lumley and Walker (Eds). Animal Toxicity Studies: Their Relevance
for Man Quay 1990 p 73. Back
56
JAMA 2002;287:2215-2220. Back
57
Mutagenesis 1987;2:73-78. Back
58
Fund Appl Toxicolo 1983;3:63-67. Back
59
J Am Parapl Soc 11;23-25, 1988. Back
60
Circulation 1996 pp 2326-2336. Back
61
JAMA 2001;286:2673 and Nature Medicine 2002;8:5. Back
62
New Stroke Drug Fails in Humans 5 December, 2001 CHICAGO (AP). Back
| 
