An evaluation of the decision tree approach for assessing priorities for safety testing of food additives.

In this publication we report an evaluation of the decision tree scheme of Cramer, Ford and Hall (1978) for assigning priorities for toxicity testing of chemicals. The original scheme has been modified to allow more chemical structures to be considered and to take into account recent advances in toxicology. The majority of the food additives permitted in either the UK, USA or Canada have been processed through the modified decision tree questionnaire and their classification compared with currently available chronic toxicity data. A large proportion of the additives (53/73) assigned to the lowest toxicity (I) class have a low order of chronic oral toxicity as do many of the compounds assigned to the moderate toxicity (II) class. Although the majority of the additives assigned to the highest toxicity (III) class are substantially more toxic than those in the lower toxicity classes, some relatively innocuous compounds reached this classification. In addition, a few toxic compounds were assigned to the lowest toxicity class. The reasons for these incorrect assignments are discussed. It was concluded that the decision tree approach, although less discriminating than originally suggested, remains a useful method for classifying compounds in terms of their probable toxicity and that further modifications to the tree could be made.