190 • Alternatives to Animal Use in Research, Testing, and Education 
ing subjective measurements and reducing 
measurement errors. 
• Make greater use of depositories in standard- 
izing cell lines or strains of micro-organisms 
used for testing. 
• Study the relationship between physicochem- 
ical properties and pharmacokinetic proper- 
ties, as well as between physicochemical and 
toxicologic properties. 
• Develop techniques for detecting nonmuta- 
genic carcinogens. 
• Develop systematic methods for objectively 
evaluating new techniques. 
• Conduct postmarketing surveillance for ad- 
verse effects, noting any discrepancies with 
test results from animals. 
• Substitute very specific tests for the LD S0 and 
other general toxicity tests, particularly for 
substances having specialized uses, such as 
drugs. 
• Use skin irritation testing as a rough screen- 
ing tool for eye irritation. 
• Attempt to describe specific effects in eye ir- 
ritation studies, rather than reporting only 
the magnitude of the response. 
• Investigate specific effects such as neurotox- 
icity to the extent possible when conducting 
general toxicity tests. 
• Search for cell lines that retain their special 
functions upon replication and develop tech- 
niques for culturing them. 
• Evaluate the statistical precision needed in 
various circumstances with a view toward 
using the smallest number of animals likely 
to be adequate. 
• Use statistics to maximize the utility of results . 
Techniques such as blocking, covariance anal- 
ysis, and factorial design should be used rou- 
tinely. 
• Improve standards of care and diet to reduce 
background effects. 
• T ake care that those conducting tests are qual- 
ified to do so, including having been trained 
in humane handling of animals. 
• Combine tests wherever possible and keep 
them as short as possible, compatible with the 
nature of the test. 
• Place greater emphasis on "no observed effect 
levels” than on lethal doses when they have 
greater predictive value. 
• Use more than one species only to answer spe- 
cific questions, and not for general safety as- 
sessments. 
SUMMARY AND CONCLUSIONS 
There has been a small but significant shift away 
from whole-animal testing to in vitro and non- 
animal techniques in recent years, partly as a re- 
sult of advances in biological techniques and partly 
in response to political and economic pressures. 
Many new methods are being developed for com- 
monly used tests. Most of these are not yet vali- 
dated, but they already have potential uses for 
screening substances for the animal testing they 
may eventually replace. 
There are several kinds of alternatives . The first 
entails the continued, but modified, use of ani- 
mals— changes in experimental design or data anal- 
ysis so that fewer animals are needed or changes 
in protocols to reduce pain or distress. Living tis- 
sues, organs, and cells derived from humans or 
animals can sometimes be used instead of whole 
animals. These systems require a larger investment 
of time and money to develop than do modifica- 
tions of whole -animal techniques, but their advan- 
tages may also be greater. They are usually faster 
and often cheaper than the corresponding whole - 
animal test, and they have scientific advantages 
as well. However, they almost always are less 
predictive than whole-animal tests and often fail 
to provide reliable dose-response data, informa- 
tion that is critical in estimating potential toxicity 
to humans. 
Data, both anecdotal and epidemiologic, on toxic 
effects in inadvertently exposed humans are some- 
times useful. However, these data are often con- 
founded by lifestyle and exposure to other toxic 
