Food Technology Magazine | Applied Science
GRAS 31 Flavoring Substances
The 31st publication by the Expert Panel of the Flavor and Extract Manufacturers Association provides an update on recent progress in the consideration of flavoring ingredients generally recognized as safe under the Food Additives Amendment.
© Mizina/iStock/Getty Images Plus
[PLEASE DOWNLOAD THE ARTICLE PDF IN ORDER TO READ THE ACCOMPANYING GRAS TABLES.]
The FEMA GRAS Program has operated to assess the safety of flavor ingredients for their intended use in human food for more than 60 years. The GRAS provision of the 1958 Food Additives Amendment to the Federal Food, Drug, and Cosmetic Act defines a food additive as: “… any substance … which … may … [become] a component or… [affect] the characteristics of any food … if such substance is not generally recognized, among experts qualified by scientific training and experience to evaluate its safety, as having been adequately shown through scientific procedures … to be safe under the conditions of its intended use…” The FEMA GRAS program operates within the confines of the 1958 Food Additives Amendment using defined scientific procedures as published in Smith et al. 2005a and b and Cohen et al. 2018a to assess the safety of flavor ingredients under their conditions of intended use.
The FEMA Expert Panel has long operated under well-defined and thorough procedures to protect against potential conflicts of interest and bias in their assessments of GRAS status for flavor ingredients. The Expert Panel’s procedures are published (Marnett et al. 2013) and are publicly available on the FEMA website (www.femaflavor.org/gras#conflict).
This publication includes the results of the Expert Panel’s review of 49 new flavorings under their conditions of intended use (Tables 1 and 2). In addition, the Expert Panel determined that new use levels and/or use in new food categories for 25 flavor ingredients are consistent with their current FEMA GRAS status (Table 3). The Expert Panel removed the FEMA GRAS status for the uses of two substances.
Progress in the Reevaluation of Natural Flavor Complexes
Flavoring substances are often described as chemically defined substances (CDS) (e.g., isoamyl acetate) or natural flavor complexes (NFCs) (e.g., orange oil). Chemically defined substances are typically single chemical substances. In contrast, the NFCs are essential oils, absolutes, concretes, oleoresins, and/or extracts typically derived from the seeds, fruit, fruit peels, leaves, flowers, exudates, bark, twigs, and/or roots of plants and are complex mixtures. During the 63 years of the FEMA GRAS program, the FEMA Expert Panel has completed two reevaluations of the chemically defined flavor ingredients, and in 2015 the Expert Panel expanded its reevaluation program to include FEMA GRAS NFCs, focusing on NFCs listed in GRAS 3 (Hall and Oser 1965).
Since 2015, the FEMA Expert Panel has conducted updated safety evaluations of more than 200 NFCs and has published their results in a series of articles in Food and Chemical Toxicology (FCT). These publications are listed below, and two additional publications are in preparation. Each of these publications* focuses on a group of NFCs related by their similar composition and/or taxonomical properties and describes the safety evaluation of each NFC as well as information on their history of use, current usage, and manufacturing method(s). Since the constituents of many NFCs are secondary metabolite products of common plant biochemical pathways, they can be organized into a limited number of congeneric groups that share similar structural, metabolic, and toxicological properties. For the safety evaluation, the Panel applies the constituent-based, stepwise procedure it developed in the early 2000s, which was then updated prior to the beginning of the reevaluation program (Cohen et al. 2018a, Smith et al. 2005a).
Evaluation of Natural Flavor Complexes Containing Allylalkoxybenzene Constituents
An important issue considered by the FEMA Expert Panel in the safety evaluation of NFCs is the assessment of the risk posed by a relatively small number of common plant metabolites, such as safrole, methyl eugenol, estragole, myristicin, and parsley apiole. All these have the allylalkoxybenzene structural motif that exhibits genotoxic and/or carcinogenic properties and thus raises safety concerns. These common secondary metabolites often function within plants to attract pollinators or provide a defense against pathogens and/or insect herbivores predators and are present in over 450 different plant species including culinary herbs and spices such as nutmeg, mace, basil, parsley, and tarragon (Tan and Nishida 2012).
While standard genotoxicity assays for allylalkoxybenzenes generally return negative results (Rietjens et al. 2014), incorporation of appropriate metabolic activating systems produce positive genotoxicity results (Herrmann et al. 2012, 2014, 2016). Evidence of carcinogenicity in the rodent liver has been reported for methyl eugenol, safrole, and estragole, and these compounds as well as those with a similar structural motif have been shown to undergo bioactivation to a reactive metabolite which may form DNA adducts (Miller et al. 1983, NTP 2000, Rietjens et al. 2014). Because the essential oils, oleoresins, and extracts derived from these plants unavoidably contain allylalkoxybenzene constituents, the FEMA Expert Panel carefully evaluated potential safety concerns and detailed exposure analyses, as described in a recent publication (Davidsen et al. 2023b).
These FEMA GRAS NFCs reevaluated in Davidsen et al. included Basil Oil (FEMA 2119), Basil Oleoresin (FEMA 2120), Estragon Oil (FEMA 2412), Mace Oil (FEMA 2653), Mace Oleoresin (FEMA 2654), Nutmeg Oil (FEMA 2793), Nutmeg Oleoresin (FEMA 5028), Parsley Oil (FEMA 2836), Parsley Oleoresin (FEMA 2837), and Snakeroot Canadian Oil (FEMA 3023). As part of their safety evaluation, the estimated current intake of each allylalkoxybenzene constituent from the use of the NFC as a flavor ingredient was compared to the Threshold of Toxicological Concern (TTC) of 0.15 µg/person/day for potential DNA-reactive mutagens and/or carcinogens (Boobis et al. 2017, EFSA Scientific Committee et al. 2019, Kroes et al. 2004). For many of the NFCs evaluated, the estimated daily intake of the allylalkoxybenzene constituent present in the NFC was below 0.15 µg/person/day and therefore not considered to raise a safety concern. When the estimated daily intake was greater than the TTC, the Panel conducted a further evaluation, as described in GRAS 29 (Cohen et al. 2020), by comparing the estimated daily intake of the allylalkoxybenzene constituent from the NFC to the BMDL10 values, defined as the lower confidence limit of the benchmark dose resulting in a 10% extra incidence in the number of animals developing liver adenomas and/or carcinomas compared to untreated animals.
For safrole, estragole, and methyl eugenol, rodent carcinogenicity studies were available. The adverse effects observed in the liver were analyzed using Benchmark Dose modeling with Bayesian model averaging using the U.S. Environmental Protection Agency BMDS software version 3.2 to determine the BMDL10. For myristicin, elemicin, and parsley apiole, for which carcinogenicity data were not available for benchmark dose modeling, BMDL10 values for these substances were estimated by read-across, using relative potency factors to methyl eugenol calculated by comparing in vivo and in vitro DNA adduct formation and/or physiologically based kinetic modeling studies of myristicin, elemicin, parsley apiole to methyl eugenol (Davidsen et al. 2023b). The FEMA Expert Panel used an MOE (margin of exposure) value of 10,000 to determine if consumption of the allylalkoxybenzene constituent from the use of the NFC raised a safety concern; this threshold for evaluating a risk by the MOE approach is similar to the value used by other scientific and regulatory bodies (EFSA 2005, EFSA Scientific Committee 2012, Health Canada 2021, JECFA 2005, JECFA 2016).
With the exception of Estragon Oil (FEMA 2412), for all the NFCs evaluated within Davidsen et al., the MOEs for the estimated intake of the respective allylalkoxybenzenes exceeded 10,000, resulting in the Panel’s conclusion that these estimated intakes of the allylalkoxybenzene constituents from the consumption of these NFCs as flavor ingredients do not present a safety concern (Davidsen et al. 2023b, EFSA 2005). The FEMA Expert Panel therefore reaffirmed their FEMA GRAS status. Because the MOE for the estimated intake of estragole from the consumption of Estragon Oil (FEMA 2412) was less than 10,000, the Panel required a refined estimate of intake based on the pattern of use. In response, industry-sponsored probabilistic modeling studies were conducted using a previously published approach (McNamara et al. 2003) and data from the U.S. National Health and Nutrition Examination Survey (NHANES). As the MOE calculations for estragole intake based on the mean and 90th percentile estimated intakes of Estragon Oil (FEMA 2412) from these refined estimates of intake were greater than 10,000, the FEMA Expert Panel also reaffirmed the GRAS status of Estragon Oil. Finally, the conditions of intended use for Basil Oil (FEMA 2119), Basil Oleoresin (FEMA 2120), Estragon Oil (FEMA 2412), Mace Oil (FEMA 2653), Mace Oleoresin (FEMA 2654), Nutmeg Oil (FEMA 2793), Parsley Oil (FEMA 2836), Parsley Oleoresin (FEMA 2837), and Snakeroot Canadian Oil (FEMA 3023) were revised to ensure consistency with the intake estimates used by the FEMA Expert Panel in their safety evaluations and are reported in Table 3.
© fcafotodigital/iStock/Getty Images Plus
© fcafotodigital/iStock/Getty Images Plus
Impact of Microwave-Assisted Extraction on Composition of Flavorings
Microwave-assisted extraction is a process in which microwave energy is utilized to heat a vessel containing a mixture of solvent and a sample, such as botanical materials. As the solvent is heated, compounds from the botanical material are extracted into the solvent (Ahmad et al. 2021, Destandau et al. 2013). Due to localized heating of the botanical sample and subsequent release of target compounds into the solvent, this technique is considered by many in the industry to be a more efficient extraction process when compared to conventional methods that involve maceration and/or conductive heating (Ahmad et al. 2021, Destandau et al. 2013). Additional benefits include reduced solvent and energy usage, lower production costs, and the ability to adjust extraction conditions and the production process to preserve the integrity of botanical constituents (Ahmad et al. 2021, Belwal et al. 2018, Destandau et al. 2013, de Castro and Castillo-Peinado 2016).
Extraction of natural products from botanical materials using microwave energy assistance dates to the mid-1990s (Destandau et al. 2013). Ongoing research aims to enhance extraction efficiency and quality (Mandel and Tandey 2016, de Castro and Castillo-Peinado 2016, Belwal et al. 2018, Kala et al. 2016), and current efforts focus on:
- further reducing solvent use or adopting solvent-free methods;
- assessing the impacts of variables that could improve extraction results (such as the type of solvent used, the extraction time and microwave power, the vessel environment, and pre- and post-treatment of the botanical sample);
- removing undesirable compounds from the botanical sample;
- comparing products obtained from microwave-assisted extraction with those from traditional extraction methods;
- efficiently scaling up from the lab to an industrial setting, including studying the heat and mass transfer kinetics to optimize extraction;
- assessing climate and other environmental impacts.
Companies within the flavor industry are currently exploring the value of this technique to create botanical extracts that may be more faithful to the flavor and taste of the source botanical.
The FEMA Expert Panel has considered examples of microwave-assisted production processes and the related extraction products and concluded that it does not have concerns regarding the use of this technology for flavoring production. The Expert Panel noted that the use of microwave-assisted extraction may in some instances result in changes in the composition of the resulting extracts compared to extracts from more traditional methods (which may already have uses considered to be FEMA GRAS). The Expert Panel relies on a congeneric grouping approach for its safety evaluation of NFCs (Cohen et al. 2018a). It advises companies to consider whether microwave-assisted extraction could result in significant changes in the relative percentages of congeneric groups present in the extract compared to those initially evaluated for FEMA GRAS status (which were produced using more traditional methods). When congeneric group changes are likely to be substantial, companies should consider these as significant manufacturing changes that could require evaluation by the FEMA Expert Panel. Companies are encouraged to consult with the Panel’s Scientific Secretary.
Change in GRAS Status of estragole
The FEMA GRAS status of estragole (CAS 140-670-0, formerly FEMA No. 2411) under its conditions of intended use as a flavor ingredient was reviewed by the FEMA Expert Panel. The FEMA Expert Panel concluded that additional data were needed to support the continuation of its GRAS status, including data from studies that further probe the relevance of recently reported DNA adduct formation in in vitro studies conducted in human cell lines and their possible accumulation. Until such data are available for review by the Expert Panel, the flavor ingredient estragole, added as such, has been removed from the FEMA GRAS list.
Change in GRAS Status of 3-acetyl-2,5-dimethylfuran
The FEMA GRAS status of 3-acetyl-2,5-dimethylfuran (CAS 10599-70-9; formerly FEMA 3391) under conditions of intended use as a flavor ingredient was reviewed. The Panel concluded that additional data are required. Such data would include comprehensive metabolism and toxicity data as well as data that would support an in-depth evaluation of the mechanism of action for potential effects observed in toxicity and genotoxicity studies. Until such data are available and reviewed, the flavor ingredient 3-acetyl-2,5-dimethylfuran has been removed from the FEMA GRAS list.ft
