Making regulatory decisions across the food ingredient spectrum.

Today, safety evaluation of food ingredients must be performed for a wide spectrum of substances and over a wide range of potential human exposures. Some of the traditional approaches to toxicological safety evaluation are not appropriate for certain of the more extreme examples across this spectrum. Yet the public and the regulated industry continue, rightfully, to expect that the system will function efficiently and effectively in their behalf. FDA must expand the range of its available tools to address the unique questions presented by nontraditional food chemicals, while maintaining the scientific credibility and integrity of the regulatory decision process that has protected public health for many decades. The safety standard that these materials must meet has not changed. Yet the types of questions pertinent to safety decisions may indeed be different than those traditionally used.

Safety assurance margins for food additives currently in use.

Toxicity data for a subset of food additives are combined (by means of a "Monte Carlo" computer analysis) with exposure information to indicate the characteristics of an overall profile of "safety assurance margins" for additives currently in use in the United States. It is demonstrated that safety concerns for (noncarcinogenic) toxic effects for these substances can seldom be expected a priori if they disappear into the U.S. food supply at levels of less than about 100,000 pounds annually.

FDA's priority-based assessment of food additives. III. Specific toxicity parameters.

In this third paper, we focus on a subset of 58 additives in the Priority-Based Assessment of Food Additives (PAFA) Project that have been associated in toxicity studies with specific toxic effects of inherently high concern. Additives producing these effects are assigned to categories based on chemical structure. A probability distribution of "R values" is calculated. We examine whether any apparent correlation exists between chemical structure category assignment and a tendency for compounds in a category to produce certain types of toxic response. Many chemical structure categories include additives having little or no toxicological data. Although this circumstance limits the scope of conclusions that may be drawn about structure-activity relationships, certain "trends" in the data appear to be worth additional future investigation.

FDA's priority-based assessment of food additives. II. General toxicity parameters.

We examine in greater depth some general toxicity parameters that have been compiled for food additives in the Priority-Based Assessment of Food Additives project. Correlations between chemical structure categories and basic parameters such as lowest effect level, highest no-effect level, and study duration indicate (among other things) trends in these parameters that are consistent in magnitude and direction with the Concern Level break points proposed in the FDA's Redbook. Certain data on the nature and extent of observed toxic effects of additives, when related to corresponding estimated levels of human intake, help to provide a profile of toxicological characteristics of a number of the regulated additives. The resulting breadth of such a profile reinforces the need for systematic priority setting in the allocation of resources to potential food additive safety concerns.

FDA's priority-based assessment of food additives. I. Preliminary results.

The principles outlined in the Bureau of Foods' recently published Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in Food (the Redbook) have been systematically applied to a Bureau of Foods food additive data base comprising nearly 1600 food and color additives and synthetic flavoring substances. The present publication is the first of several papers reviewing and analyzing the results of this work. This paper describes the content and organization of the scientific data base that the Bureau has compiled, and presents some preliminary general conclusions regarding the processing of these data according to Redbook principles.