The acceptable daily intake (ADI) for food additives is commonly derived from the NOAEL (no-observed-adverse-effect level) in long-term animal in vivo studies. To derive an ADI a safety or uncertainty factor (commonly 100) is applied to the NOAEL in the most sensitive test species. The 100-fold safety factor is considered to be the product of both species and inter-individual differences in toxicokinetics and toxicodynamics. Although in vitro data have previously been considered during the risk assessment of food additives, they have generally had no direct influence on the calculation of ADI values. In this review 18 food additives are evaluated for the availability of in vitro toxicity data which might be used for the derivation of a specific data-derived uncertainty factor. For the majority of the food additives reviewed, additional in vitro tests have been conducted which supplement and support the short- and long-term in vivo toxicity studies. However, it was recognized that these in vitro studies could not be used in isolation to derive an ADI; only when sufficient in vivo mechanistic data are available can such information be used in a regulatory context. Additional short-term studies are proposed for the food additives which, if conducted, would provide data that could then be used for the calculation of data-derived uncertainty factors. Copyright (C) 2000 Elsevier Science B.V.