The Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food is asked to advise the Commission on the implications for human health of chemically defined flavouring substances used in or on foodstuffs in the Member States. In particular, the Panel is asked to evaluate 18 flavouring substances in the Flavouring Group Evaluation FGE.13, using the procedure as referred to in the Commission Regulation EC No 1565/2000. These 18 flavouring substances belong to chemical group 14, Annex I of the Commission Regulation EC No 1565/2000. The present Flavouring Group Evaluation deals with 18 flavouring substances related to furfuryl alcohol and related substances and to sulphur substituted furan derivatives. They can be divided into two subgroups, consisting of nine flavouring substances each, depending on the absence or the presence of sulphur substituents in their molecule. One of the 18 flavouring substances possesses a chiral centre. In this case, the substance has been presented without any indication that the commercial flavouring substance has dominance of one or the other enantiomer. Two of the substances can exist as geometrical isomers and in both cases, no indication has been given that one of the possible isomers has preponderance in the commercial flavouring material. Sixteen of the flavouring substances are classified into structural class II, two are classified into structural class III. Sixteen of the flavouring substances in the present group have been reported to occur naturally in a wide range of food items. In its evaluation, the Panel as a default used the Maximised Survey-derived Daily Intakes (MSDIs) approach to estimate the per capita intakes of the flavouring substances in Europe. However, when the Panel examined the information provided by the European flavouring industry on the use levels in various foods, it appeared obvious that the MSDI approach in a number of cases would grossly underestimate the intake by regular consumers of products flavoured at the use level reported by the industry, especially in those cases where the annual production values were reported to be small. In consequence, the Panel had reservations about the data on use and use levels provided and the intake estimates obtained by the MSDI approach. In the absence of more precise information that would enable the Panel to make a more realistic estimate of the intakes of the flavouring substances, the Panel has decided also to perform an estimate of the daily intakes per person using a modified Theoretical Added Maximum Daily Intake (mTAMDI) approach based on the normal use levels reported by industry. In those cases where the mTAMDI approach indicated that the intake of a flavouring substance might exceed its corresponding threshold of concern, the Panel decided not to carry out a formal safety assessment using the Procedure. In these cases the Panel requires more precise data on use and use levels. According to the default MSDI approach, the 18 flavouring substances in this group have intakes in Europe from 0.0012 to 37 microgram/capita/day which are below the threshold of concern value for both structural class II (540 microgram/person/day) and structural class III (90 microgram/person /day) substances. The biotransformation of the four esters of furfuryl alcohol in the present Flavouring Group Evaluation leads to the formation of furfural, a reactive hepatotoxic aldehyde. Furfural is then oxidised to furoic acid, which can be conjugated with glycine yielding innocuous and readily excreted products. Also, the candidate substance ethyl furfuracrylate can be biotransformed to furoic acid. However, the furan ring of the candidate substance furoic acid and the furan moieties of the two candidate furoate esters may be completely oxidised to CO2, with the opening of the furan ring and production of reactive intermediates. Therefore it cannot be predicted that these eight flavouring substances included in subgroup 1 are metabolised to innocous products. In addition to the above mentioned pathways, 5-hydroxymethylfurfuraldehyde can be bioactivated to 5-[(sulfoxy)methyl] furfural, through sulfonation of its allylic hydroxyl functional group, catalyzed by sulfotransferases. The resulting ester has been demonstrated to induce genotoxic effects. Based on the general knowledge on the metabolism of sulphur-containing compounds, the flavouring substances bearing a free thiol group can be considered reactive per se interacting with endogenous sulphur-containing substances, e.g. glutathione and proteins, thus triggering adverse effects. The candidate furfuryl and furan monosulfides are expected to undergo oxidation mainly to the corresponding sulfoxides and sulfones. Alternatively they can be conjugated with glutathione, giving rise to mixed disulfides, which can be oxidised to thiosulfinates or thiosulfones or reduced to free thiols. Similar metabolic pathways may be predicted for the candidate disulfides and very likely for the trisulfide. Given the reactivity of thiol groups, whether free or resulting from di(tri)sulfide, and their importance in cell physiology, it cannot be excluded that all the nine flavouring substances included in subgroup 2 of the present Flavouring Group Evaluation interfere with normal cell function and therefore, they cannot be predicted to be metabolised to innocuous substances. Short-term and long-term toxicity studies are available for two flavouring substances included in subgroup 1, and for three related supporting substances, including furfural. They indicate that the liver is the critical target for their toxicity. Recently EFSA has established an ADI value of 0.5 mg/kg bw for furfural and the furfural component of furfural diethyl acetal. No toxicity data are available on flavouring substances included in subgroup 2; however results from toxicity studies on 16 supporting substances have been reported. Many of the available studies were performed either with a single dose level or multiple dose levels that produced no effects; the dose producing no adverse effects ranged between 0.45 and 10 mg/kg/day. Data on genotoxicity were available on two flavouring substances and on five structurally related substances. Overall, except for the flavouring substance 5-hydroxymethylfurfuraldehyde, the in vitro and in vivo data available do not give rise to concern with respect to genotoxicity of the remaining eight flavouring substances included in subgroup 1. Based on in vitro data on the mutagenic activity of a sulphate conjugate of 5-hydroxymethylfurfuraldehyde, there is sufficient evidence to raise concern about a genotoxic potential. Accordingly, the Procedure cannot be applied for this substance pending submission of in vivo genotoxicity data. The lack of data on the sulphur-containing flavouring substances included in subgroup 2, or on the structurally related substances, does not allow to conclude on their genotoxicity. It was considered that on the basis of the default MSDI approach the 17 flavouring substances which could be taken through the Procedure would not give rise to safety concerns at the estimated levels of intake arising from their use as flavouring substances. When the estimated intakes were based on the mTAMDI, they ranged from 75 to 3700 microgram/person/day for the 16 flavouring substances from structural class II. Thus, the intakes for nine of the flavouring substances were above the threshold of concern for structural class II of 540 microgram/person/day. The estimated intakes of two flavouring substances assigned to structural class III, based on the mTAMDI are 150 microgram/person/day, which is above the threshold of concern for structural class III of 90 microgram/person/day. Thus for 10 of the 17 flavouring substances taken through the Procedure, the intakes, estimated on the basis of the mTAMDI, exceed the relevant threshold for their structural class, to which the flavouring substance has been assigned. Therefore, for these 10 substances more reliable exposure data are required. On the basis of such additional data, these flavouring substances should be reconsidered along the steps of the Procedure. Following this procedure additional toxicological data might become necessary. In order to determine whether this evaluation could be applied to the materials of commerce, it is necessary to consider the available specifications: Adequate specifications including complete purity criteria and identity tests for the materials of commerce have been provided for the 17 flavouring substances which have been evaluated according to the Procedure, except that information on stereoisomerism is missing for three of the substances. Thus, the final evaluation of the materials of commerce cannot be performed for these three flavouring substances, pending further information.
