Chapter 7 
The Use of Animals in Testing 
Testing for the safety or efficacy of a substance 
or product accounts for a major use of animals 
as defined in this assessment, most of which are 
rats and mice (see ch. 3). Of these, probably the 
largest portion are used in developing drugs. A 
significant portion are also used to test other sub- 
stances— pesticides, industrial chemicals, and con- 
sumer products— to assess possible toxicity and 
to establish conditions under which they can be 
used safely. 
Research and testing have been differentiated 
for purposes of this assessment, but the bound- 
ary between them is not sharp. From the stand- 
point of developing alternatives, a key difference 
is that a particular test may be performed for hun- 
dreds, perhaps thousands of substances and use 
hundreds or thousands of animals, whereas a given 
research method will be used on far fewer. As a 
corollary there are far more research procedures 
than testing procedures from which to choose. Fur- 
thermore, individual researchers are much more 
likely to develop their own methods than are those 
conducting testing. These differences make the 
task of developing alternatives more manageable 
for testing than for research. 
Testing for efficacy has some attributes of re- 
search and some of toxicity testing. A particular 
protocol may be used on a small number of sub- 
stances and is likely to be tailored either to the 
application or to the family of substances being 
tested. Experimenters testing for efficacy need to 
have a better understanding of the mechanisms 
by which a particular effect occurs than those test- 
ing for toxicity, primarily because efficacy test- 
ing is closely related to the physiological mecha- 
nisms that the new drug may affect, whereas toxic 
effects may be quite independent. Finally, an im- 
portant distinction of efficacy testing is that the 
animals used would ordinarily be diseased. 
Other kinds of tests include those for safety 
other than for toxicity, as in testing of diagnostic 
techniques or quality control tests in the manu- 
facture of medical devices. These have end points 
even more specific than those for toxicity, and are 
thus good candidates for the development of alter- 
natives (see ch. 8). 
Toxicity testing is the focus of this and the fol- 
lowing chapter for three reasons. First, this is an 
area of animal use in testing in which the govern- 
ment has great influence on nongovernmental ac- 
tivities. Second, these tests are used in a more rou- 
tine fashion than are tests for efficacy or general 
safety and therefore have a greater tendency to 
lag in the application of state-of-the-art technol- 
ogy. Third, toxicity tests include methods that have 
attracted the largest political attack. 
All substances can be toxic at some exposure 
level, even water. Conversely, even substances 
known to be highly toxic may be harmless at low 
doses or under certain circumstances. Determin- 
ing the hazard to humans requires information 
about the potential hazard and the expected level 
of exposure, resulting in an estimate of the prob- 
ability that a substance will produce harm under 
certain conditions (8). This assessment of risk is 
a scientific endeavor, whereas the management 
of risk is a sociopolitical one (31,36). 
Although toxicity data on humans are invaluable 
in conducting risk assessments, they are usually 
unavailable. Some information comes from epi- 
demiologic studies or episodes of accidental hu- 
man exposure. Most often, however, testing on 
animals is used. An appropriate weight is given 
to the following factors on a case-bv-case basis, 
considering the seriousness of the hazard and the 
kind of assumptions needed to estimate risks to 
humans: 
• the relationship between dose and response; 
• the effects at the molecular, cellular, organ, 
organ system, and whole-organism levels; 
• conflicting results between studies and pos- 
sible explanations for the conflicts; 
• the effects of structurally similar substances 
on humans or animals; 
• any known metabolic differences between hu- 
mans and the test species that could affect the 
toxic response; and 
• statistical uncertainties and difficulties in ex- 
trapolating to a low dose (55). 
149 
I 
