Contents Search publications                                         Previous  Next

The effects of chronic exposure are also tested in animals. In these tests, the animals are fed, inhale, or have the chemical painted on their skins throughout their lives. The type and amount of disease they develop are then compared with the effects in a control group. The controls should be the same as the exposed group in every respect except the chemical exposure. They should, for example, be the same strain of the same species, and be fed the same diet. The differences in rates of various diseases are then tested by various statistical means to see if they are significant. Having an adequate control group is just one element of good experimental design, as is the number of animals used. If a study is badly designed, its results will not be reliable.

For example, when testing carcinogens on animals the US National Cancer Institute says that the chemical should be tested at two doses in both sexes of two species of rodent, and each group should contain at least 50 animals. One dose is often the highest dose that will not kill, or acutely poison, the animal. Although this means that positive results will not be missed just because a high enough dose was not used, it means that predicting the effects of very low levels of the chemical is difficult. Trying to extrapolate (predict) what will happen at low doses, or what will happen in humans, are two of the major problems of animal testing.

Contents Search publications                                         Previous  Next