Ch. 7— The Use of Animals in Testing • 151 
ing demands by the public for safe and effective 
products. 
The most appropriate animals are ideally those 
that, for the substance being tested, predict the 
human response most accurately . There is no other 
animal wholly identical to humans in terms of toxic 
effects. The choice of animal is influenced by 
known similarity to humans for the organ system 
or mechanism of interest, as well as convenience 
of breeding or purchasing, familiarity with the spe- 
cies, existing data, lifespan, ease of handling under 
experimental conditions, cost of obtaining and 
maintaining, litter size, and gestation period. Ro- 
dents have been used extensively, as have rabbits, 
primates, and dogs. 
Rodents have been used in almost all carcino- 
genicity testing despite the fact that such tests are 
the most difficult to extrapolate to humans. Mice 
and rats have been used because their lifespan is 
short, they are small and easily handled, and they 
have a number of metabolic pathways and patho- 
logical responses similar to those of humans . Some 
specially developed strains are sufficiently suscep- 
tible to cancer that test groups can be small. These 
factors contribute greatly to the economic feasi- 
bility of conducting carcinogenicity testing with 
rodents . Extensive experience in using them, and 
in using particular strains, is often an important 
reason for continuing their use (55). A large amount 
of data are already available on spontaneous tumors 
at specific organ sites (1). 
Although rodents are routinely used for many 
kinds of tests, other animals may be used for spe- 
cific reasons. For example, the rabbit is used for 
eye irritation tests because it has large, easily 
manipulated eyes and because its eyes have many 
characteristics found in human eyes (19). Hens 
have been shown to be a good model for delayed 
neurotoxic effects of organophosphorous com- 
pounds (12). 
Dose Levels and Route and 
Duration of Exposure 
The way in which exposure to a substance oc- 
curs can affect the kind and severity of toxic ef- 
fects. For example, if a chemical does not present 
a hazard when applied to skin because it is not 
absorbed, it may nonetheless be very toxic if taken 
orally. When the route of exposure does not affect 
the portion of the dose taken up or its distribution 
in the body, testing might be done in the manner 
most easily controlled. For other than the most 
preliminary tests to characterize toxicity, most 
would administer the substance by the same route 
as would occur in the course of accidental exposure 
or use by humans. Sometimes the palatability, solu- 
bility, stability, or volatility of a substance will de- 
termine which routes are feasible. 
Certain tests, such as the acute toxicity for a sin- 
gle exposure, are used as inexpensive screening 
tools for estimating the relative hazard presented 
by a substance. As discussed later in this chapter, 
the acute toxicity test known as the LD 50 is used 
in classification schemes for the transportation or 
disposal of chemicals. Acute toxicity testing might 
also be used to determine the risks of one-time 
exposure, as might occur in an accident. Ordinar- 
ily, the duration of exposure in an animal study 
is greater (at least in proportion to the lifespan) 
than the exposure period for which data will be 
used in extrapolating the risks to humans. 
The dose levels administered depend on a vari- 
ety of factors. On the one hand, it is not possible 
to detect long-term effects if the dose is so large 
that many animals die before the end of the test. 
On the other hand, administered doses represent- 
ative of human exposure levels may not produce 
detectable effects with what may be considered 
a reasonable number of test animals. Generally, 
three dose levels are used; they are chosen so as 
to span the range of responses from a “no-observed- 
effect level” to fully observable toxic effects. 
For carcinogenicity and other long-term testing, 
the highest dose should be one that will produce 
measurable toxicity without significantly altering 
lifespan. Other levels may depend on whether the 
carcinogenicity is being looked for in combination 
with chronic toxicity (55). The lowest dose could 
be one for which there are no observed effects 
or it might be related to the level of estimated hu- 
man exposure (38). 
Another approach is to choose doses that will 
yield levels in the blood similar to those expected 
for humans. Although this is perhaps a more real- 
istic test, effects may be more difficult to detect. 
In addition, the criterion of similarity may require 
