WHO FOOD ADDITIVES SERIES: 52

First draft prepared by

Dr I. C. Munro
CANTOX Health Sciences International, Mississauga, Ontario, Canada

and

Dr A. Mattia
Division of Biotechnology and GRAS Notice Review, Office of Food Additive Safety, Center for Food Safety and Applied Nutrition, United States Food & Drug Administration, Maryland, USA

For nearly 40 years, the Joint FAO/WHO Expert Committee on Food Additives (JECFA) has played a leading role in the development of principles and procedures for the safety evaluation of flavouring agents. In 1996, the Committee began an ongoing programme to conduct safety evaluations and establish specifications for individual flavouring agents. On the basis of this experience, the Committee recognizes that the safety evaluation of flavouring agents presents unique challenges, principally owing to the fact that there are more than 2000 such agents used in commerce, the vast majority of which are added at extremely low levels in food. In view of the fact that the majority of individual flavouring agents occur naturally in food, the Committee has concluded that to evaluate each of these substances by traditional toxicological testing is simply not warranted in most cases (WHO, 1987).

In 1987, WHO classified flavouring agents into four groups:

1. (a) Artificial substances that are unlikely to occur naturally in food;
2. (b) Natural substances that are not normally consumed as food, their derived products, and the equivalent nature-identical flavourings;
3. (c) Herbs and spices, their derived products, and the equivalent nature-identical flavourings, and
4. (d) Natural flavouring substances obtained from vegetable and animal products and normally consumed as food whether processed or not, and their synthetic equivalents.

Since 1996, the Committee has evaluated the safety of approximately 1150 individual flavouring agents using the Procedure for the Safety Evaluation of Flavouring Agents (Annex 1, references 116 and 117 ). The Committe organized these agents into a number of well-defined chemical groups (i.e. congeneric groups) in order to effectively evaluate such a large number of substances. The evaluations conducted to date have demonstrated that flavouring agents within a congeneric group have a similar biochemical fate and toxicological potential. Most individual flavouring agents have been shown to be efficiently detoxicated to yield innocuous metabolites. Knowledge of their metabolic fate, coupled with low intake levels that are typically below thresholds of toxicological concern, has formed the basis for the majority of the Committee’s evaluations of individual flavouring agents. In instances where the metabolic fate of an individual flavouring agent was not well known or readily predictable, or where intake exceeded the threshold of toxicological concern for the relevant structural class, the Committee has relied on data on the toxicity of the agent itself or of a structurally related substance in order to perform a rigorous safety evaluation. The results of these evaluations comprise an extensive database, which has been published since 1996 in the WHO Food Additive Series as a series of monographs on congeneric groups.

A key development in the process by which the Committee evaluates the safety of flavouring agents has been the establishment of specifications, consistent with JECFA practice, for all flavouring agents. The criteria underpinning these specifications require that the chemical assay for individual flavouring agents be specified. The Committee has adopted the criterion of a minimum assay of 95% purity of the named flavouring agent. In some cases, flavouring agents have been reported to be of less than 95% purity. In such instances, the Committee has required the identification of the secondary components such that at least 95% of the chemical composition of the named flavouring agent can be accounted for. Key features of secondary components are that many are themselves flavouring agents that are likely have been evaluated previously by the Committee, as a member of a different congeneric group, or they are structurally related substances that belong to the same congeneric group as the named flavouring agent.

In order to provide for a rigorous safety evaluation when there are many secondary components, the Committee has collected and evaluated data on each of these components. This process has been facilitated by the fact that, since 1996, the Committee has reviewed and published monographs on most of the chemical groups of flavouring agents. Using these monographs on congeneric groups, the Committee now has the capacity to efficiently evaluate additional flavouring agents and numerous secondary components. An obvious extension of this process is that the Committee can begin to evaluate mixtures of flavouring agents, as found in natural flavouring complexes.

Natural flavouring complexes fall into the following categories:

1. Essential oils: the volatile flavouring constituents of plant sources (leaves, fruits, buds, bark, etc.) usually obtained by steam distillation, by expression or extraction, or some combination of these processes.
2. Extracts: the volatile and non-volatile flavouring constituents of plant sources, as described in (1) above, obtained by extraction with a permitted polar or non-polar solvent, the choice of which depends on the botanical source and the constituents desired.
3. Oleoresins, often called solid extracts: obtained by extraction as in (2) above, followed by removal of the solvent.

These categories fall into groups (b), (c), and (d) listed above (WHO, 1987). The agents that comprise these categories range from those that consist almost entirely of a single chemical entity, such as bitter almond oil (benzaldehyde), to those with a highly complex composition (e.g. rosemary oil). In some instances, dozens of constituents are essential to the technical flavour characteristics of the natural flavouring complex. Although the chemical composition of these complexes is variable, their constituents can be assigned to relevant congeneric groups.

Table 1 gives examples of natural flavouring complexes and individual flavouring agents and provides an indication of the number of these substances that are used in commerce in the United States. Tables 2 and 3 list the constituents of two natural flavouring complexes, bois de rose oil and lemongrass oil, organized by congeneric group. While each of these complexes has numerous constituents, these constituents fit readily into congeneric groups, most of which have already been evaluated by the Committee.

Table 1. Examples of individual flavouring agents and natural flavouring complexes

^a Group as classified by WHO (1987) (see Introduction)

FCC, Food Chemical Codex

This monograph proposes that the existing Procedure for the Safety Evaluation of Flavouring Agents (Annex 1, references 116 and 117), as refined by the Committee (Annex 1, references 122, 131), be further modified to accommodate the safety evaluation of natural flavouring complexes. After consideration of an approach to determining specifications that could be used to specify a chemical assay for a natural flavouring complex, details of the suggested modifications to the existing Procedure in order to provide for the evaluation of natural flavouring complexes are outlined herein.

The Committee has always insisted upon specifications that define the substance being considered and, as previously noted, has typically required that a chemical assay for individual flavouring agents be at least 95%. In cases in which the named substance could not practically be freed from secondary components, the Committee has evaluated each secondary component individually. The emphasis throughout has been on assuring the identity and safety of the substance being evaluated.

In dealing with natural flavouring complexes, it may be useful to invoke compendia of specifications published by organizations that have previously considered these agents. The Food Chemicals Codex is one such organization, recognized by regulation in the United States and elsewhere. Food Chemicals Codex specifications typically list physiochemical properties, contaminants (including heavy metals and polynuclear aromatics) and microbiological parameters, to ensure that food ingredients, including flavouring agents, are safe for human consumption. For many natural flavouring complexes, Food Chemicals Codex also specifies a minimum chemical assay for key chemical constituents. For example, the chemical assay for caraway oil specifies that it should contain not less than 50%, by volume, of ketones calculated as carvone. The assay for peppermint oil is not less than 5% esters, calculated as menthyl acetate, and not less than 50% menthol. Clearly, flexibility is maintained in a chemical assay specified by Food Chemicals Codex in that a minimum level is specified for a chemical group that may be measured or calculated as a key chemical constituent in the natural flavouring complex. This approach could be used effectively to specify a chemical assay for natural flavouring complexes evaluated by the Committee.

To perform an evaluation, the Committee would organize the constituents of the natural flavouring complex into congeneric groups (e.g. phenols, tertiary alcohols, etc.) and perform a series of evaluations for the congeneric group. Within key congeneric groups (e.g. alicyclic secondary alcohols and ketones) selected constituents (e.g. menthol) will be associated with the technical flavouring characteristics of the natural flavouring complex (e.g. peppermint oil). In a manner entirely consistent with Food Chemicals Codex, the evaluation of congeneric groups within the complex could be linked by chemical assay to constituents that are fundamental to the flavouring characteristics of the natural flavouring complex.

An essential general principle is that the specification should be no more complex than is essential to assure the most critical aspects of safety, identity, and technical function. Determination of these key characteristics requires the use of methods shown to be reliable through collaborative studies. Existing relevant specifications from Food Chemicals Codex for bois de rose oil and lemongrass oil are provided in Tables 4 and 5 (Food Chemicals Codex, 1996).

The principles and procedures used by the Committee to estimate intake in the safety evaluation of individual flavouring agents are equally valid and appropriate for natural flavouring complexes, provided that the intended conditions of use are similar to those for individual flavouring agents. One obvious difference between individual flavouring agents and natural flavouring complexes is that some of the latter have much broader patterns of use in the food supply and much higher volumes of disappearance into the marketplace than do individual agents. Natural flavouring complexes such as vanilla extract and lemon oil are used in a wide variety of food categories that have high rates of consumption from products such as baked goods, beverages, soft and hard candy, and dairy products. Annual volumes of production for twelve natural flavouring complexes in use in the United States exceed 1 000 000 kg.

Table 2. Constituents of bois de rose oil, organized by congeneric group^a

Table 3. Constituents of lemongrass oil organized by congeneric group^a

Notes to Table 3

Table 4. Food Chemicals Codex specifications for bois de rose oil

COE, Council of Europe; FCC, Food chemicals Codex; FEMA, Flavour and Extracts Manufacturers Association; ISO, International Organization for Standardization

Table 5. Food Chemicals Codex specifications for lemongrass oil

COE, Council of Europe; FCC, Food Chemical Codex; FDA, Food and Drug Administration; ISO, International Organization for Standardization

As has been common practice by the Committee since 1996, intake of individual flavouring agents is based on a conservative estimate, with the assumption that only 10% of the population consumes all of the substance as a flavouring agent, Clearly the use of natural flavouring complexes, such as vanilla extract and lemon oil, that are used at high volume is not limited to 10% of the population. It is recommended that in the case of natural flavouring complexes that are used at high volume (e.g. >100 000 kg/year) that the Committee use the actual percentage of users as estimated from databases on food intake. This would provide more realistic estimates of daily intake.

The data on poundage and intake for bois de rose oil and lemongrass oil are show in Table 6. For these natural flavouring complexes, it is valid to estimate daily intake using the principles and procedures that are used for individual flavouring agents.

Exposure to constituents of the natural flavouring complex in the various con-generic groups does not solely occur via intake of the complex. These constituents, mainly terpenes, are also present as common components of many traditional foods. In fact, the majority of terpenes that are constituents of a natural flavouring complex are also consumed in the normal diet. For example, exposure to linalool, limonene, linalyl acetate and other terpenes in bois de rose oil also occurs by daily consumption of spices (e.g. coriander), fruits (e.g. oranges), wine, and certain vegetables. Exposure to the constituents found in bois de rose oil occurs mainly from consumption of a normal diet. This is to be expected, given the ubiquitous presence of simple monoterpenes in all plants.

Table 6. Poundage and intake data for bois de rose oil and lemongrass oil

It is recommended that the current Procedure for the Safety Evaluation of Flavouring Agents be expanded to include the option to evaluate both individual flavouring agents and natural flavouring complexes. Like individual flavouring agents, flavouring agents present in a natural flavouring complex will be assigned to the appropriate congeneric groups. Subsequently, each congeneric group will be evaluated according to the current steps in the Procedure (see Figure 1). Once all congeneric groups contained within the natural flavouring complex have been evaluated, a conclusion can be made concerning the safety of the combination of congeneric groups that constitute the complex. This modification to the existing Procedure to allow the inclusion of assessments for congeneric groups will permit the evaluation of all types of flavouring agents, including relatively pure individual flavouring agents (chemical assay >95%), agents containing appreciable amounts of secondary components minimum chemical assay, <95%), and natural flavouring complexes containing different congeneric groups of individual flavouring agents. The proposed modification to the Procedure constitutes an efficient use of resources because it builds upon the evaluations previously conducted and relies on the extensive series of monographs on well-defined congeneric groups of flavouring agents already published by the Committee.

Figure 1. Procedure for the Safety Evaluation of Flavouring Agents and Natural Flavouring Complexes

As a test, two natural flavouring complexes, bois de rose oil and lemongrass oil, were evaluated using the modified Procedure. The results of these evaluations (shown in Table 7 and Table 8, respectively) indicate that at current levels of intake these natural flavouring complexes would not be expected to present a safety concern.

Table 7. Safety evaluation of bois de rose oil^a,b

Table 8. Safety evaluation of lemongrass oil^a,b

Notes to Table 8

Procedures that allow the Committee to evaluate individual flavouring agents as well as natural flavouring complexes are already in place. The examples provided indicate that the modified Procedure for the Safety Evaluation of Flavouring Agents also works for natural flavouring complexes. It is recommended that the Committee should evaluate several natural flavouring complexes using the modified Procedure at a future meeting.

Food Chemicals Codex (1996) Food Chemicals Codex, 4th Ed., Washington: National Academy Press.

WHO (1987) Principles for the Safety Assessment of Food Additives and Contaminants in Food (Environmental Health Criteria, No. 70), Geneva.

Acknowledgement

The authors offer special thanks to Dr Timothy Adams and Ms Christie Gavin of the Flavor and Extract Manufacturers of the United States for providing analytical data on flavour composition and for many useful ideas in the formulation of the modified Procedure for the Safety Evaluation Of Flavouring Agents.

    See Also:
       Toxicological Abbreviations