The no-effect level and optimal use of toxicity data.

The term "no-observable-effect level" is preferable to "no-effect level" to indicate the impossibility of demonstrating no effect. The term "no-adverse-effect level" is sometimes used, indicating that not all changes are of toxicological relevance. The size of the safety factor can be justified scientifically only when good toxicokinetic and metabolic information exists. Actually, extrapolation and risk evaluation can be carried out in a proper way only when the mechanism underlying a particular effect is known. The use of recovery groups in long-term studies is recommended.